THE BIOGEOCHEMISTRY OF THE AMAZON BASIN
This page intentionally left blank
The Biogeochemistry of the Amazon Basin
Edited by MICHAEL E. MCCLAIN Florida International University, USA REYNALDO L. VICTORIA Universidade de Sao Paulo, Brazil JEFFREY E. RICHEY University of Washington, USA
OXFORD UNIVERSITY PRESS
2001
OXFORD UNIVERSITY PRESS
Oxford Athens Chennai Kolkata Nairobi
New York Auckland Bangkok Bogotá Buenos Aires Cape Town Dar es Salaam Delhi Florence Hong Kong Istanbul Karachi Kuala Lumpur Madrid Melbourne Mexico City Mumbai Paris Sao Paolo Shanghai Singapore Taipei Tokyo Toronto Warsaw
and associated companies in Berlin Ibadan
Copyright © 2001 by Oxford University Press, Inc. Published by Oxford University Press, Inc. 198 Madison Avenue, New York, New York 10016 Oxford is a registered trademark of Oxford University Press. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission of Oxford University Press. Library of Congress Cataloging-in-Publication Data The biogeochemistry of the Amazon Basin / edited by Michael E. McClain, Reynaldo L. Victoria, Jeffrey E. Richey. p. cm. Includes bibliographical references. ISBN 0-19-511431-0 1. Biogeochemistry—Amazon River Wathershed. I. McClain, Michael E. II. Victoria, Reynaldo L. III. Richey, Jeffrey Edward, 1946QH112 .B56 2000 577'.14'09811—dc21 00-041642
9 8 7 6 5 4 3 2 1 Printed in the United States of America on acid-free paper
Acknowledgments This book represents the efforts of many individuals and many projects. We collectively thank the individuals and funding agencies that made this work possible. The Editors' own contributions were made possible with support from the National Science Foundation Division of Environmental Biology and International Programs, the NASA Earth Observing System, the FundaCão de Amparo a Pesquisa do Estado de São Paulo, the Inter American Institute for Global Change Research, and the Andrew W. Mellon Foundation. We would particularly like to thank the reviewers of individual chapters: Dominique Arrouays, Robert Berner, Vladimir Elias, Christian Feller, Augusto Cesar Franco, Sana Gardescu, Jean Loup Guyot, Steven Hamilton, Robert Harrison, Dominique Irvine, Rod Keenan, George Klinge, Kirsten Laarkamp, Ariel Lugo, Flavio Luizao, Luiz Martinelli, Ernesto Medina, Michel Meybeck, Jose Natalino Silva, Jason Neff, Christopher Neill, Jorge Paolini, Walter de Paula Lima, Kathleen Ruttenberg, Miguel Simao Neto, Morgan Smidt, Boris Volkoff, Michael Williams, and Summer Wilson We would also like to thank Nancy Blanton, Cristina Dumlao, and Madelyn Mateo for final preparation of the manuscript.
This page intentionally left blank
Contents 1
2
3
4
5
6
7
8
9
The Relevance of Biogeochemistry to Amazon Development and Conservation Michael McClain
3
General Characteristics and Variability of Climate in the Amazon Basin and its Links to the Global Climate System Jose A. Marengo & Carlos A. Nobre
17
The Atmospheric Component of Biogeochemical Cycles in the Amazon Basin Paulo Artaxo
42
Soil versus Biological Controls on Nutrient Cycling in Terra Firme Forests Elvira Cuevas
53
Nutrient Cycling as a Function of Landscape and Biotic Characteristics in the Cerrados of Central Brazil M. Haridasan
68
Linking Biogeochemical Cycles to Cattle Pasture Management and Sustainability in The Amazon Basin Moacyr B. Dias-Filho, Eric A. Davidson & Claudio J. Reis de Carvalho
84
Nutrient Considerations in the Use of Silviculture for Land Development and Rehabilitation in the Amazon Florencia Montagnini
106
Extractive Reserves and Participatory Research as Factors in the Biogeochemistry of the Amazon Basin Foster Brown , Karen Kainer & Eufran do Amaral
122
The Recovery of Biomass, Nutrient Stocks, and Deep-Soil Functions in Secondary Forests Daniel Nepstad, Paulo R. S. Moutinho & Daniel Markewitz
139
Contents
viii
10
The Interface Between Economics and Nutrient Cycling in Amazon Land Development Carl F. Jordan
156
11
Carbon Storage in Biomass and Soils Martial Bernoux, Paulo M. A. GraCa, Carlos C. Cerri, Philip M. Fearnside, Brigitte J. Feigl & Marisa C. Piccolo
12
Terrestrial Inputs to Amazon Streams and Internal Biogeochemical Processing Michael E. McClain & Helmut Elsenbeer
185
Geo-ecological Controls on Elemental Fluxes in Communities of Higher Plants in Amazonian floodplains Maria Tereza F. Piedade, Martin Worbes & Wolfgang J. Junk
209
Biogeochemistry of Amazon Floodplain Lakes and Associated Wetlands John M. Melack & Bruce R. Forsberg
235
13
14
165
15
Organic Matter and Nutrients in the Mainstem Amazon River Allan H. Devol & John I. Hedges
275
16
Trace Elements in the Mainstem Amazon River Patrick T. Seyler & Gerald R. Boaventura
307
17
Biogeochemical Processes on the Amazon Shelf: Changes in Dissolved and Paniculate Fluxes During River/Ocean Mixing David J. DeMaster & Robert C. Aller
328
Contributors Andrea S. Alechandre, Department of Agrarian Sciences, Federal University of Acre, Rio Branco, AC 69.915, Brazil;
[email protected] Robert C. Aller, Marine Sciences Research Center, SUNY at Stony Brook, Stony Brook, NY 11794-5000, USA;
[email protected] Eufran do Amaral, Center for Agroforestry Research, Brazilian Agricultural Research Agency (CPAF/EMBRAPA), Rio Branco, AC, Brazil;
[email protected] Paulo Artaxo, Instituto de Fisica, Universidade de São Paulo, Rua do matao, Travessa R, 187, CEP 05508-900, Sao Paulo, Brazil;
[email protected] Martial Bernoux, Institut de Recherche pour le Développement (IRD), Centro de energia Nuclear na Agricultura (CENA), University of Sao Paulo (USP), Laboratorio Biogeoquimica do Solo, Caixa Postal 96, 13400-970 Piracicaba, SP, Brazil;
[email protected] Geraldo R. Boaventura, Universidade de Brasilia, Instituto de Geociências Campus Universitário Darcy Ribeiro, ICC-Central CEP: 70910-900 Brasilia, Brazil;
[email protected] I. Foster Brown, Woods Hole Research Center, Federal Fluminense University, Federal University of Acre, Parque Zoobotanico, Rio Branco, AC 69.915, Brazil;
[email protected] Cláudio Jose Reis de Carvalho, Laboratorio Ecofisiologia Vegetal, Embrapa Amazonia Oriental, C. Postal 48, 66017-970 Belém, PA, Brazil;
[email protected] Carlos C. Cerri, Centro de energia Nuclear na Agricultura (CENA), University of Sao Paulo - USP, Lab. Biogeoquimica do Solo, Caixa Postal 96, 13400-970 Piracicaba, SP, Brazil;
[email protected] Elvira Cuevas, Centro de Ecologia, Instituto Venezolano de Investigaciones Cientificas, Apartado 21827, Caracas 1020-A, Venezuela;
[email protected] Eric A. Davidson, The Woods Hole Research Center, P.O. Box 296, Woods Hole, MA 02543-0296, USA; edavidson® whrc. org David J. DeMaster, Deptment of Marine, Earth, and Atmospheric Sciences, North Carolina State University, Raleigh, NC 27695, USA;
[email protected]
x
Contributors
Allan H. Devol, School of Oceanography, Box 357940, University of Washington, Seattle, WA 98195, USA; devol@u. washington.edu Moacyr B. Dias-Filho, Laboratorio Ecofisiologia Vegetal, Embrapa Amazônia Oriental, C. Postal 48, 66017-970 Belém, PA, Brazil;
[email protected] Helmut Elsenbeer, Department of Civil and Environmental Engineering, University of Cincinnati, P.O. Box 210071, Cincinnati, OH 45221, USA;
[email protected] Philip M. Fearnside, Institute Nacional de Pesquisas da Amazonia (INPA), Coordenadoria de Pesquisas em Ecologia, Caixa Postal 478, 69011-970 Manaus, Amazonas, Brazil;
[email protected] Brigitte J. Feigl, Centro de energia Nuclear na Agricultura - CENA, University of Sao Paulo (USP), Laboratorio Biogeoquimica do Solo, Caixa Postal 96, 13400-970 Piracicaba, SP, Brazil;
[email protected] Bruce R. Forsberg, Instituto Nacional de Pesquisas da Amazonia (INPA), Coordenadoria de Pesquisas em Ecologia, Caixa Postal 478, 69011-970 Manaus, Amazonas, Brazil;
[email protected] Paulo M. A. Graca, Instituto Nacional de Pesquisas da Amazonia (INPA), Coordenadoria de Pesquisas em Ecologia, Caixa Postal 478, 69011-970 Manaus, Amazonas, Brazil;
[email protected] M. Haridasan, Departmento de Ecologia, Instituto de Ciências Biológicas, Universidade de Brasilia, 70919970 Brasilia, DF, Brazil;
[email protected] John 1. Hedges, School of Oceanography, Box 357940, University of Washington, Seattle, WA 98195, USA;
[email protected] Carl F. Jordan, Institute of Ecology, University of Georgia, Athens, GA 30602, USA; cfjordan® arches, uga. edu Wolfgang J. Junk, Max-Planck-Institut fur Limnologie, Arbeitsgruppe, Tropenokologie, Postfach 165, 24302 Plon, Germany;
[email protected] Karen A. Kainer, Tropical Conservation and Development Program, Center for Latin American Studies, University of Florida, P.O. Box 115531, Gainesville, FL 32611, USA;
[email protected] Daniel Markewitz, University of Georgia, Daniel B. Warnell School of Forest Resources, Athens, GA, 30602, USA;
[email protected] Jose A. Marengo, Centro de Previsao de Tempo e Estudos de Clima (CPTEC), Institute Nacional de Pesquisas Espaciais (INPE), Cacheoira Paulista, Sao Paulo, Brazil;
[email protected] Michael E. McClain, Department of Environmental Studies, Florida International University, Miami, FL 33199, USA;
[email protected]
Contributors
xi
John M. Melack, Donald Bren School of Environmental Science and Management, University of California, Santa Barbara, CA 93106-5131, USA;
[email protected] Florencia Montagnini, Area of Management and Conservation of Forests and Biodiversity, Centre Agronómico Tropical de Investigacion y Enseñanza (CATIE), 7170 Turrialba, Costa Rica;
[email protected] Paulo R. S. Moutinho, Instituto de Pesquisa Ambiental da Amazonia (IPAM), Eneas Pinheiro 1424, 66087-430 Belém, Para, Brazil; Daniel Nepstad, The Woods Hole Research Center, P.O. Box 296, Woods Hole, MA 02543-0296, USA;
[email protected] Carlos A. Nobre, Centre de Previsão de Tempo e Estudos de Clima (CPTEC), Instituto Nacional de Pesquisas Espaciais (INPE), Cacheoira Paulista, Sao Paulo, Brazil;
[email protected] Marisa de Cassia Piccolo, Centro de energia Nuclear na Agricultura (CENA), University of Sao Paulo USP, Lab. Biogeoquimica do Solo, Caixa Postal 96, 13400-970 Piracicaba, SP Brazil;
[email protected] Maria Tereza F. Piedade, Instituto Nacional de Pesquisas da Amazonia (INPA), C.P. 478, 69011-970 Manaus, AM, Brazil;
[email protected] Jeffrey E. Richey, School of Oceanography, Box 357940, University of Washington, Seattle, WA 98195, USA;
[email protected] Patrick T. Seyler, Département Milieux et Environnement, Institut de Recherche pour le Developpement, 213, Rue Lafayette, 75480 PARIS Cedex 10, France;
[email protected] Reynaldo L. Victoria, Centro de energia Nuclear na Agricultura (CENA), University of Sao Paulo (USP), Caixa Postal 96, 13400-970 Piracicaba, SP, Brazil;
[email protected] Martin Worbes, Forstbotanisches Institut, AuJSenstelle: Busgenweg 2, 37077 Gottingen, Germany ;
[email protected]
This page intentionally left blank
THE BIOGEOCHEMISTRY OF THE AMAZON BASIN
This page intentionally left blank
1 The Relevance of Biogeochemistry to Amazon Development and Conservation Michael E. McClain
To read the press of recent years, one might imagine that the fate of the world rests in the hands of those who would develop the Amazon basin. Waves of incoming colonists are blamed for the bulk of the deforestation and development (Schomberg 1998), but Asian logging firms, multinational oil companies, and gold miners are also portrayed as destructive agents hacking down the forest, systematically undermining its biodiversity, and severely contaminating its myriad ecosystems (Althaus 1996, Ferreira 1996, James 1998). The effects of these varied threats are regularly broadcast in alarming tones. Rueters News Service warned in January 1998 that "Brazil's Amazon rain forest, the world's richest trove of biological diversity and source of much of the Earth's oxygen, continues to be ravaged" (Craig 1998). And, in April 1999, a writer for the Associated Press communicated the "fear" of unspecified scientists that "damage to the rain forest... could throw the Earth's climate out of balance" (Donn 1999). Clearly, the fate of the Amazon and the implications of its fate to the overall Earth system are topics of enormous scientific and popular interest. While there is little disagreement that the complete destruction of Amazon forests would be catastrophic, what about partial deforestation of the region? How much, and which parts, of the Amazon can be converted to sustainable human land
uses without compromising the ecological integrity of the conserved areas? How might this development impact regional climate, adjoining coastal systems, and overall global processes? Answers to these volatile questions remain elusive and seemingly endless strands of controversy swirl about them. At the heart of the matter, yet largely beyond the public discussion, are biogeochemical cycles that support and regulate the functioning of the Amazonia's biological systems. Moreover, it is the incomplete understanding of these cycles that promotes uncertainty and feeds the controversy. The purpose of this book is to present a coherent assessment of our current understanding of the biogeochemical functioning of the Amazon basin. Although it is surely presumptuous to assume that this presentation will shed sufficient light on the uncertainties to eliminate the current controversies, we hope that it will provide a basis for lifting the discussion to a higher level. Cycles of carbon, nitrogen, phosphorus, oxygen, and many other bioactive elements link the region's flora and fauna to associated cycles of water, soil, and seasonal energy fluxes. Research into the biogeochemistry of Amazonia's natural ecosystems has revealed an unprecedented level of sophistication of adaptive mechanisms within the flora and fauna for conserving and accessing nutrients. Replacement of native vegetation with poorly
4
Michael E. McClain
adapted crops and pasture grasses often of global greenhouse gases. Recent articles leads to problems of nutrient availability and have highlighted the Amazon's role as a carultimately to crop failure and pasture degra- bon dioxide sink (Prentice and Lloyd 1998), dation. Sustainable agrarian development in and modeling results suggest that Amazon Amazonia is thus intimately related to funda- vegetation may account for between 5% and mental biogeochemical processes and to our 35% of the terrestrial sink for excess CO2 in understanding of those processes. the atmosphere (Tian et al. 1998). Flora and fauna composing the expansive Conversely, during El Nino years when the aquatic ecosystems exhibit a similar level of Amazon is drier than normal, these same adaptive sophistication, which enables a rich modeling results suggest that Amazon diversity of organisms to flourish in spite of vegetation may serve as a net source of CO2 modest nutrient inputs from lowland terrestri- to the atmosphere. The Amazon is also a al systems. The bounty of aquatic resources significant source of other greenhouse gases and water itself provides many essential serv- such as methane and nitrous oxide. When ices to the people of the Amazon. The these gaseous emissions are combined with region's natural fisheries provide a large pro- strong energy transfers linked to evaporation portion, and sometimes the majority, of ani- over the basin, the role of the Amazon in mal protein consumed by inhabitants. Fish are regional and global climate grows well also a valuable source of income to fishermen. beyond the areal coverage of the basin. By River and lake water satisfies nearly all of the considering the quantitative impact that water supply needs of Amazon peoples, Amazon development may have on these including drinking, cooking, bathing, and emissions and energy transfers, one begins to waste removal. While little water is used for understand the "fear" voiced in the irrigation, river channels and lakes are impor- opening paragraph. tant avenues of transportation and shipping. Each year, the Amazon river discharges Water bodies also provide abundant opportu- nearly 20% of the total volume of water delivnities for recreation and are enjoyed in this ered to the oceans by rivers. In fact, the way by most of the region's inhabitants. annual discharge of the Amazon is greater With each of the above-mentioned uses of than the combined annual discharge of water, quality (or chemical purity) is crucial. Earth's next eight largest rivers! AccompanyThis is especially true because water is ing this huge volume of water are generally not treated prior to human use in correspondingly large loads of salts, metals, the Amazon. Threats to water quality include organic compounds, and sediments. These excess erosion and nutrient runoff from inputs to the Atlantic Ocean take on both poorly managed fields and pastures, metal regional and global significance by fueling pollution from mining and other industrial coastal productivity and significantly impactactivities, and organic pollution from the ing the global cycles of certain elements. discharge of untreated human and animal wastes. The capacity of the river system to Biogeochemical Cycles in the assimilate these threats varies as a function of Atmosphere - Global Linkages discharge and biogeochemical processes operating to eliminate harmful conditions Biogeochemical cycles in the Amazon (McClain 2001). basin are linked to global cycles through On a global scale, and as Earth's largest exchange with the atmosphere and river conterminous area of wet tropical forest, the discharge into the Atlantic Ocean. Of these Amazon is viewed as a significant regulator two, exchange with the atmosphere is by far
The Relevance of Biogeochemistry to Amazon Development and Conservation
the most interactive and variable on seasonal and inter annual time scales. Furthermore, it is through atmospheric exchanges that the Amazon interacts directly with the global climate system. In chapter 2, Jose Marengo and Carlos Nobre examine the nature of climatic variability over the Amazon and the influence of the basin within the global climate system. Over most of the Amazon temperatures are remarkably uniform and do not vary appreciably with season. The Andean Amazon is of course an important exception, as temperatures there vary over large ranges. Precipitation, however, varies significantly with space and time throughout the basin and stands as the most important climatic variable in explaining the nature, distribution, and magnitude of certain biogeochemical processes. Variability in rainfall is tied to large-scale circulation patterns of air masses in the region and the interactions of these air masses with ocean surface waters. The tight coupling of these processes is evidenced by the profound impacts that El Nino Southern Oscillation (ENSO) events have upon rainfall in the basin. Overall, the Amazon experiences reduced rainfall during ENSO events, although certain areas experience elevated rains. The impact of such anomalies on land and aquatic biogeochemical cycles is likely to be significant but has not been adequately assessed as yet. The general reduction in rainfall across the Amazon during ENSO events raises the specter that global warming will produce similar reductions. This threat is amplified by the results of a growing number of largescale modeling efforts, which consistently show a decrease in precipitation due to deforestation. The combined effects of these disturbances could significantly alter the cycling of water and associated elements in the Amazon. In addition to climatic influences on land and aquatic biogeochemical cycles, there are also exchanges of bioactive elements
5
between the Amazon surface and atmosphere. These exchanges take the form of gases and aerosols, which have direct impacts on regional and global climate. In chapter 3, Paulo Artaxo discusses the forms and dynamics of these exchanges. Each year, Amazon forests and savannas emit large quantities of aerosols and associated bioactive elements to the atmosphere. In general, higher concentrations of bioactive elements are found in the coarser fraction of aerosols, which are recycled within the region itself. The finer fraction, however, may be transported for thousands of kilometers and thus exit the Amazon region. Atmospheric transport out of the region, although still poorly quantified, probably represents a significant loss of nutrients from the basin. During the dry season, when slash burning is most practiced, concentrations of aerosols in the less than 10 urn fraction increase by an order of magnitude to values ranging from 400-500 pg rrr3. Under certain circumstances of enhanced convection, these aerosols may rise to altitudes approaching 10 kilometers, where transport velocities are greatly accelerated and their distribution may be global. The biogeochemical roles of rainfall, aerosol, and gas emissions converge in the form of wet deposition, which is well documented as a significant source of nutrients and organic matter to Amazon surface systems (Andreae et al. 1990, Lesack and Melack 1996, Williams et al. 1997a). Important components of wet deposition are organic acids, which control rain pH values, and essential nutrients like phosphorus (P) and potassium (K).
Biogeochemical Cycles on Land Coping with Nutrient Scarcity According to an assessment by Ozorio de Almeida and Campari (1995), the bulk of deforestation in the Brazilian Amazon is carried out by people moving within the basin, as opposed to newly arriving colonists.
6
The internal movement is prompted by a loss in soil fertility, which drives families to abandon their pastures and fields and to clear new land elsewhere. An essential element in reducing deforestation is thus improving management of cultivated land, thereby extending its use and reducing the rate of land abandonment. Widespread use of fertilizers and highly mechanized agriculture is impractical for most people in the Amazon. So, improved land management for nutrient conservation must make maximal use of natural mechanisms of nutrient recycling and conservation. Management practices imported to the region regularly fail because they are adapted to conditions different than those present in the Amazon. New practices must be developed that are best suited to Amazon nutrient and soil/climate conditions. The luxurious growth which typifies Amazon forest is a clear indication that natural vegetation is well adapted to nutrient/soil/climate conditions of the region. Thus, important insights to sustainable land use are likely to be found by carefully examining the functioning of these systems.
Forests and savannas In chapter 4, Elvira Cuevas examines the mechanisms through which terra firme (or nonflooded) forests of the Amazon satisfy their nutrient needs on the generally nutrientpoor soils that cover greater than 80% of the basin. On these soils, forest growth is limited mainly by P, calcium (Ca), and magnesium (Mg), as evidenced by significant correlations between forest productivity measures and abundances of these elements (Cuevas and Medina 1986, Cuevas and Medina 1988). Strong and persistent leaching processes have severely depleted many Amazon soils in Ca and Mg and have produced conditions under which mineral forms of P are bound to iron and aluminum oxides and generally
Michael E. McClain
unavailable to plants. Under these conditions, the key to sustained nutrient flows seems to lie in organic pools of these elements and the vegetation's adaptations to access these pools. Along a toposequence in the upper Negro basin of Venezuela, Tiessen et al. (1994 a, b) documented that greater than 70% of available P occurred at the surface and in the upper organic horizons of the soil column. Of this P, 50% occurred in paniculate organic matter with a mean residence time of 4 years. The remaining mineral-associated P cycled over much longer time periods or was permanently sequestered in lateritic nodules. Amazon forest plants access organic pools of P, Ca, and Mg through surface mats of fine roots and symbiotic interactions with mycorrhizae fungi. Root mats in Amazon forests on average account for 30% of the total fine root mass, and root mat thickness is inversely correlated with nutrient availability. Mycorrhizae fungi stimulate phosphatase production in the vicinity of fine roots and thus enhance the availability of PO4 for uptake by fine roots. In fact, the combined action of surficial fine roots and mycorrhizae fungi has been shown to retain nearly 100% of added 32P in uptake experiments in the upper Negro basin (Cuevas and Medina 1988). It is important to note here that this emphasis on recycling means that the productivity of Amazon forests is linked to the rate of nutrient remineralization rather than the standing stock of nutrients at anytime. Furthermore, these recycling mechanisms are a direct adaptation to the humid conditions of the Amazon, where extreme leaching potential necessitates the concentration of fine roots at the surface and abundant water makes this root allocation scheme functionally possible. In the more arid savanna regions of the Amazon basin and Brazilian Cerrado, the problem of nutrient depleted soils is compounded by seasonal water shortages. As discussed by M. Haridasan in chapter 5, savanna vegetation is thus of greatly reduced
The Relevance of Biogeochemistry to Amazon Development and Conservation
biomass, with correspondingly reduced rates of nutrient cycling. Very little research has been conducted into the biogeochemical processes which link vegetation productivity to nutrient pools, but one especially important consideration is the role of fire in injecting periodic pulses of available nutrients into the system. Trees of the Cerrado are well adapted to the frequent fires and are seldom killed. Total vegetation cover responds quickly after fires and generally recovers with a few months of the fire event (Haridasan, this volume).
Pastures, plantations, and extractive reserves The conversion of Amazon forest to other land uses sends a shock through the ecosystem which significantly impacts underlying biogeochemical cycles. Deforestation accompanied by slash removal and burning initially releases large quantities of carbon and nitrogen to the atmosphere and decreases the standing stocks of these elements. Mineral nutrients such as P, Ca, Mg, and K are concentrated in ashes and are returned to the soil column as a pulse accompanying infiltrating rainwater. These transformations are most dramatic and enduring in the conversion of forest to pasture, where grasses replace trees and cattle replace forest animals. The \videspread creation of pastures in the Amazon has met with limited success because pasture productivity generally declines within three to five years of establishment, leading to expensive rehabilitation or, more likely, pasture abandonment. The key to extended pasture productivity lies in better management of nutrient cycles and therefore pasture fertility. In chapter 6, Moacyr Dias-Filho, Eric Davidson, and Claudio Reis de Carvalho discuss the role of biogeochemical cycles in regulating pasture productivity. The biogeochemistry of P is shown to be of particular
7
importance, as this element is most limiting. Of the approximately 200 kg ha'1 of P measured in the soils of one pasture in the eastern Amazon, less than 1% is thought to be available to plants. Because only a very small pool of P is available to maintain pasture grasses, and fertilization is economically difficult, management techniques which are normally ignored become important. For example, the animals themselves become a key link in the cycle which replenishes available P. Instead of returning to the soil surface as litterfall, as would be the case in natural Amazon forests, P taken up by plants is returned to the soil surface mainly as animal feces. This, in turn, leads to problems of P distribution, as cattle tend to deposit feces in concentrated areas rather than uniformly across the pasture (Buschbacher 1987). Cattle also alter nutrient cycles by compressing the soil surface and promoting overland hydrological flows and greater losses of P. Pasture creation impacts carbon (C) and nitrogen (N) cycles by reducing standing stocks and cycling rates of these elements. The main loss of C and N of course occurs in the aboveground biomass. In fact, C and N stocks may actually increase in pasture soils, but such increases are unpredictable and dependent on factors such as native soil fertility, fertilization, climate, fire frequency, and grazing intensity (Dias-Filho et al., this volume). Although standing stocks may rise or fall, research from the eastern and western Amazon shows that cycling rates consistently decrease. Net N mineralization and nitrification rates in pastures were found to range from less than 15% to 40% of rates measured in adjacent primary forest (unpublished data from Verchot and Davidson, presented in Dias-Filho et al., this volume). Similar patterns were observed for emissions of NO and N2O gases. These sharply reduced rates of N cycling raise the question of whether N might become limiting under certain conditions in Amazon pastures or secondary forests.
Michael E. McClain
8
As described by Florencia Montagnini in chapter 7, large-scale conversion of forest to silvicultural plantations has been carried out with generally disappointing results in the Amazon. Expansive stands of commercial species are highly susceptible to pests and disease, and in the Amazon they are dependent on additions of fertilizers to sustain multiple rotations. Thus it appears that such large-scale silviculture is unlikely to become a widespread use of Amazon land. Smaller scale mixed plantations, however, have a far greater potential as a sustainable land use and moreover as a tool in the rehabilitation of degraded landscapes. Such small-scale plantations place lesser demands on nutrient supply mechanisms and conserve a greater percentage of the system's natural nutrient conservation processes. Recognizing the special importance of organic nutrient pools in Amazon ecosystems, plantations can be used to produce mulch for effective nutrient application to crops or other managed lands. Depending on the nutrient content of different mulches and their biodegradability, specific needs can be satisfied. For example, quickly decomposing mulch may be added to short rotation crops to provide a pulse of nutrients to stimulate growth, or more slowly decomposing mulch may be applied to facilitate a more continuous source of nutrients. An especially valuable role for small plantations is the rehabilitation of degraded sites (Montagnini, this volume). Planting species with high nutrient use efficiencies will more quickly replenish organic matter levels in degraded soils. The form of Amazon land use which least impacts natural biogeochemical cycles is extractive reserves. In chapter 8, Foster Brown, Karen Kainer, and Eufran do Amaral describe how a typical extractivist household "stores" more than 50,000 Mg of carbon in their managed forest. A large portion of this carbon would have otherwise been released to the atmosphere as CO2 if the forest had
been cut. Because the forest remains intact, so do other components of the natural biogeochemical cycles of C, N, P, Ca, Mg, and other bioactive elements. Although extractive reserves are suitable for only 25% of the Brazilian Amazon, they could serve an important role in reducing emissions of C and N to the atmosphere.
Secondary forest As a result of the high rates of land abandonment in the Amazon, a great deal of land is undergoing forest regeneration. The rate of recovery of secondary forest is partially controlled by biogeochemical controls linked to soil fertility, and the forest itself plays a role in larger biogeochemical cycles by sequestering C and N within its increasing biomass, thus partially offsetting emissions from deforested areas. In chapter 9, Daniel Nepstad, Paulo Moutinho, Daniel Markewitz and Elizabeth Cheng calculate that carbon sequestration in second growth forests amounts to approximately 10% of that released through deforestation. These authors also raise the very intriguing issue of deep roots and their role in accessing nutrients for secondary forest growth. The significant investment made in superficial roots by trees of Amazon forests is a clear indication of the importance of nutrient recycling from organic pools at the soil surface. However, research from the central and eastern Amazon has shown that trees in seasonally dry forests also have roots extending to at least 18 m depth (Nepstad et al. 1994). While the main function of these roots appears to be the uptake of deep soil water and groundwater, there is also potential for these roots to access deeper nutrient pools in the soil column. Nepstad et al. (this volume) elaborate on this issue by demonstrating that secondary forests growing in the eastern Amazon have P and K nutrient needs that cannot be satisfied by available stocks in the
The Relevance of Biogeochemistry to Amazon Development and Conservation
upper soil levels. They also show that secondary forests may re-establish deep root systems (6 m deep) within 16 years. They further show that abundant fine roots infected with mycorrhizal fungi characterize the deep roots of secondary forest. These circumstantial data suggest that secondary forests are mining nutrient reserves in deeper soils, but conclusive data are still unavailable.
Economic consequences of nutrient scarcity on land Sustainable development in the Amazon hinges upon many factors which are ecological, physical, cultural, and economical. Among the most practical aspects of biogeochemical cycles in the Amazon are those that exhibit direct economical ties. In chapter 10, Carl Jordan discusses the interacting forces that bridge the gap between biogeochemical science and development economics. He recounts the dismal performance of such grand projects as the Jari plantation and the reported failure of as much as 85% of the ranches in a section of the eastern Amazon. Although intensive agriculture, characterized by application of fertilizers and pesticides, has been shown to be profitable in select parts of the Amazon, in general the inhabitants of the region must look to alternative means of economic development. Jordan asserts that production systems in the Amazon must be designed in such a way to mimic the structure of the original forest and its natural nutrient conserving mechanisms.
Carbon storage The consequences of deforestation, land management, and secondary growth extend beyond regional issues of conservation and sustainable land use to global issues of climate change. With CO2 and CH4 levels continuing to rise in the atmosphere, many global change scientists view the Amazon as
9
a key variable in the global carbon budget. Will the Amazon's forests act as a sink for atmospheric CO2, or will deforestation and burning of the region's forests only exacerbate the problem? Answers to these questions depend on the course of development (or conservation) in the region, but they also depend on exactly how much carbon the region's biota and soils can effectively hold. In chapter 11, Martial Bernoux, Paulo Graca, Carlos Cerri, Philip Fearnside, Brigitte Feigl, and Marisa Piccolo provide a detailed assessment of the quantity of carbon stored in soils and biota of the Brazilian Amazon basin. In total, the Brazilian Amazon is estimated to store more than 120 Pg (1015g) of reduced carbon, of which 34% is held in soils and 66% is held in aboveground vegetation. Of the several natural and managed vegetation types in the Amazon, natural forest is the most important C repository, accounting for 97% of the aboveground C. These estimates are likely to improve as more data are collected and better techniques of extrapolation are developed. However, it is already clear that the Amazon is a major reservoir of reduced carbon on the globe.
Biogeochemical Cycles in Aquatic Systems - The Balance of Water Quality Water is abundant in the Amazon, and nearly all living things in the basin are somehow dependent on this abundance. From the standpoint of development and conservation, the most crucial issue in water management is water quality. Clean water is fundamental to the maintenance of the Amazon's unique aquatic ecosystems and to the health of its people, who rely on surface water to satisfy their household water needs and to provide fish and other aquatic plants and organisms for their nutritional needs. Water quality is intricately linked to biogeochemical reactions such as the dissolution and decomposition of
10
organic wastes, the transport and scavenging of heavy metals, and the regulation of proper nutrient levels. These processes are similarly linked to hydrological variability in the system, which determines transport conditions and dilution through mixing. The continued development of the Amazon basin will increase demands placed on the region's surface water resources. In light of the limited economic resources of the nations composing the basin, it is likely that development will proceed without the implementation of adequate pollution control measures. For most activities, pollution "control" will be left to the natural assimilation capacity of the surface water system (McClain 2001). This is especially true in connection with the diffuse development activities in rural areas and small towns. Consequently, there is an acute need to better understand how the river system processes and transports organic matter, nutrients, and metals.
Michael E. McClain
this region, thus the influence of upland forests is reduced (McClain and Elsenbeer this volume). The generally greater relief of the western Amazon Andean foothills forms narrower stream valleys with less pronounced riparian wetland zones and therefore greater connectivity between streams and upland forests. The strong heterotrophic nature of Amazon streams makes them efficient at retaining both nutrients and organic matter input from adjoining lands. However, the capacity of these streams to assimilate contaminant levels of these same compounds is likely to be small. The greatest buffering capacity of Amazon stream systems likely lies in the riparian forest/wetland systems that adjoin them. Research by McClain et al. (1994) and Williams et al. (1997b) has demonstrated the ability of processes in the riparian zones of central Amazon watersheds to strip nitrate from upland groundwaters in forested and deforested settings. In the undisturbed setting Streams of the Barro Branco stream, nitrate levels were reduced by 80% in comparison with Streams are perhaps the most sensitive levels in upland groundwater entering the components of the surface water system riparian zone (McClain et al. 1994). Williams to processes on land due to their small et al. (1997b) found that groundwater nitrate size and limited dilution capabilities. In concentrations were elevated in groundwater chapter 12, Helmut Elsenbeer and I examine following cutting and burning of the forest, the role of streams in processing material and but upon passing through the riparian zone solute inputs from the Amazon landscape. levels decreased significantly. The same Detailed studies of biogeochemical processes authors reported that intact riparian forest in Amazon streams have been limited to only effectively buffered the stream from erosive a few sites, but results from these sites clear- inputs and their associated particulate nutrily indicate that inputs to streams from forests ent and organic matter inputs. vary as a function of soils, geomorphology, and rainfall characteristics. In the nutrient Lakes and floodplains depleted soils of the central Amazon, inputs of ions to the stream systems are reduced More than 1 million square kilometers and a significant proportion of ions, includ- of the Amazon basin may be classified as ing Ca, Mg, and K, may derive from vegeta- floodplain, wetland, or lake. These environtion sources via throughfall and overland ments range from narrow, sporadically flow of storm waters. Organic matter and flooded lowlands bordering streams to inorganic nitrogen also appear to derive massive, regularly flooded plains bordering largely from riparian forests and wetlands in the basin's largest rivers. Nearly 9000 lakes,
The Relevance of Biogeochemistry to Amazon Development and Conservation
11
covering an average maximum area of 67,900 control on species distributions is the km2, occur as integrated components of the biogeochemical nature of the river water and massive floodplains bordering the mainstem sediments. Distinct types of floodplain forest Amazon and its major tributaries (Sippel et al. develop adjacent to Whitewater rivers 1992). These varied aquatic habitats play a (varzea) and blackwater rivers (igapo). pivotal role in many ecological, as well as Piedade et al. show that geoecological biogeochemical, processes in the region. variables impact both the magnitude and From a practical standpoint, floodplains and timing of elemental fluxes in floodplains. their associated lakes provide essential Leaves of flooded forest in varzea are signifbreeding and feeding areas for many of the icantly enriched in nutrients in comparison to Amazon's most economically important fish leaves of flooded forest in igapo, and when species (Goulding 1980). Along Whitewater combined with productivity data these foliar rivers rich with Andean sediment, floodplains nutrient levels indicate more efficient nutrient also provide fertile soils for crops and use in igapo forest. By taking a more inpasture. Finally, these environments emit depth look at N pools and dynamics of the quantities of methane that are significant in varzea forest, Piedade et al. conclude that the global atmospheric methane budget and approximately 200 kg ha'1 of N (and associrelevant to studies of climate change (Bartlett ated nutrients) are input to the floodplain in the form of fresh leaf matter annually. The and Harriss 1993). In chapter 13, Maria Teresa Piedade, Martin annual pulse of nutrient input is even more Worbes, and Wolfgang Junk examine the dramatic in floodplain herbaceous communistrong links between ecological aspects of ties, which turnover completely each year. A the higher plants and elemental fluxes on stand of the grass Echinochloa polystachya Amazon floodplains. Distribution, species was found to take up 377 kg ha'1 of N from composition, biomass, and primary produc- river water and to return it to the soil surface tivity constitute fundamental ecological as fresh organic litter. Corresponding inputs parameters that exert marked control over of P and K were 51 and 1136 kg ha'1, respecelemental fluxes. Piedade et al. show that tively. Transformations such as these in nutrithese parameters are in turn linked to inter- ent forms are essential in maintaining the acting geomorphological and hydrological overall fertility of the environment and also factors. The most influential combined effect feeding the food chains which support of geomorphology and hydrology is period commercially important fish species. of inundation, which corresponds to the In chapter 14, John Melack and Bruce number of days per year that a given point is Forsberg provide a quantitative assessment of submerged by the oscillating flood waters. the role of floodplain lakes in regional cycles Species of the flooded forest are organized of C, N, and P. Floodplain lakes were found into zones which relate directly to period of to be important centers of organic carbon inundation. Species distributions throughout production and delivery to the river system. these zones vary as a function of both the The combined primary production of macrodegree of stress imposed and the adaptive phytes, forests, periphyton, and plankton strategies employed by the trees. Within the associated with floodplain lakes is estimated herbaceous community, distribution patterns at 117 Tg C yr1, of which only 24% is also vary according to the balance between remineralized in the lakes. As a result, an floodpulse-induced stresses and adaptive estimated 90 Tg C yr1 are delivered to the strategies, but additional factors linked to river system by continual and seasonal physical stability are also important. A final exchanges. This input alone amounts to
12
more than twice the annual export of organic C in the river system. Given that the river also receives organic C inputs from vast upland forests and other parts of the floodplain, these new data suggest that consumption of organic C is likely to be far higher in the river's mainstem than had been previously thought (see next section). Cycling of N and P in these lakes is closely correlated with that of organic C. According to Melack and Forsberg, productivity in floodplain lakes may be limited by either N or P in conjunction with spatial and temporal differences in the supplies of these nutrients across the spectrum of lakes. Melack and Forsberg discuss an array of anthropogenic impacts on Amazon floodplain lakes, including mining, hydroelectric dams, mercury pollution, and deforestation of adjoining uplands. Each of these activities upsets normal cycles of important elements, either by poisoning organisms which carryout the cycling (mining wastes and mercury pollution), altering the cycle of flooding to which organisms are adapted (dams), or otherwise changing the natural rates and forms of element inputs (dams, mining, and deforestation). Specific impacts of these activities are still mostly unstudied in the Amazon, but in light of the crucial role floodplain lakes play in larger elemental cycles and associated ecosystem products (for example, fisheries), negative impacts there are likely to propagate to much wider areas.
The mainstem river and its estuary The flow of water, solutes, and particulate matter in the mainstem Amazon river is the product of integrated hydrological, biological, physical, and biogeochemical processes across the entire basin. Of the many natural features of the Amazon, none more completely integrates these processes, because there is no corner of the basin that
Michael E. McClain
does not contribute some proportion of its water and associated load to the river system. The condition of the river may thus be viewed as a kind of proxy for the condition of the larger basin. As with aquatic resources throughout the Amazon, water quantity is not a problem, with the exception of occasional extreme flooding events. The main issue is water quality and the maintenance of water quality within ranges appropriate to support natural aquatic ecosystems and the water needs of the human population. Water quality in the mainstem Amazon remains excellent, with the exception of isolated sites of relatively small-scale contamination near the river's big cities. Thus biogeochemical investigations of the river enjoy the luxury of focusing mainly on natural, or background, compositional patterns and dynamics. In chapter 15, Alan Devol and John Hedges examine the relative role of upstream sources and in-channel processing in determining the dynamics of C, N, P, and oxygen (O) in the mainstem river. The major proportion of organic matter transported in the mainstem is refractory material of upstream origin, which appears to pass through the mainstem rather conservatively. Heterotrophic metabolism is fueled by a smaller, more labile pool of organic matter derived from local sources such as the floodplain (see Melack and Forsberg, above). Because the Amazon mainstem is almost entirely heterotrophic, the foodweb of the river's myriad aquatic organisms is based on the input of energy and nutrients from this more labile organic matter pool. Nearly all fine particulate forms of C, N, and P in the mainstem derive from Andean environments many thousands of kilometers upstream. This is a striking example of the interconnectivity and interdependency of processes across the Amazon. In an especially interesting application of biogeochemical investigative techniques, Devol and Hedges use rather sophisticated
The Relevance of Biogeochemistry to Amazon Development and Conservation
molecular and isotopic biomarkers to illuminate the finer details of organic matter dynamics in the mainstem. Sources of the main organic matter pools (dissolved, fine particulate, and coarse particulate) are identified using concordant data on elemental ratios, carbon stable isotope ratios, and relative abundances of specific compounds derived from lignin molecules. These data suggest that terrestrial vegetation is the predominant source of organic matter (OM) in the river, as opposed to plankton or floodplain grasses. Furthermore, subcomponents of the lignin molecular fraction point specifically to leaves as the source instead of the woody portions of trees. The course and degree of degradation within the OM fractions is shown to be quite distinct based on abundances of carbohydrates, amino acids, and lignin fractions. Dissolved OM is the most highly degraded fraction, while coarse particulate OM retains molecular signatures that are closest to fresh leaf material and fine particulate OM lies in between. When analyzed more closely, the consistent patterns in elemental and molecular variability between the fractions suggest that differential partitioning (or sorption) may account for final separation of refractory molecules into the fine particulate or dissolved phases. The transport dynamics implied by the foregoing data are confirmed by abundances of bomb carbon-14 in the different fractions. Devol and Hedges synthesize the above results into an elegant conceptual model explaining the processing of organic matter carried by the Amazon mainstem. The model, which they refer to as "Regional Chromatography", follows the processing from production of leaf litter in the basin's forests until the arrival of the final refractory OM in the mainstem river. Within the model, coarse particulate OM is transported to the river system via quick overland hydrological flowpaths or direct litterfall, whereupon it undergoes rapid decomposition but still
13
retains an appreciable proportion of its original biogeochemical signature upon reaching the mainstem. Conversely, fine particulate OM and dissolved OM are assumed to emanate from the soils after partitioning into their respective fractions, where movement to the river system occurs either as soil erosion (fine particulate OM) or subsurface runoff (dissolved OM). Although field investigations undertaken in upland portions of the basin to test the predictions of the Regional Chromatography model have not always yielded supporting data (McClain et al. 1997), the model continues to provide an internally consistent and conceptually appealing explanation for mainstem OM compositional patterns. An additional set of elements which play important roles in reactions ranging from biogeochemical to purely geochemical are trace metals. In chapter 16, Patrick Seyler and Geraldo Boaventura present a comprehensive survey of trace metal forms and fluxes in the mainstem Amazon and its major tributaries. Trace metal abundances between the mainstem and its tributaries vary as a function of the source rocks occurring in the basins and in function with the reactivity of the metals. Concentrations of most metals in the Amazon are comparable to those in other major rivers of the world, and the Andes stand out as the dominant source area for most metals transported by the river. In the Amazon mainstem, trace metals are transported primarily in particulate form. Transport in the dissolved phase predominated only in the Negro River, where many metals were complexed with dissolved organic matter. Seyler and Boaventura present new data illustrating the temporal variability of trace metal fluxes in the main river system. They identified four discrete patterns of variability which reflect the source, process of mobilization, and in-channel reactivity of the metals measured.
14
Currently, trace metal pollution is not a problem in the Amazon River, with the exception perhaps of localized small-scale occurrences. However, the basin includes several sites of elevated metal releases linked to gold and manganese ore mining and industrial discharges from the large cities. Seyler and Boaventura discuss the likelihood of contamination from these sources. Moreover, they come to the provocative conclusion that elevated fluxes of manganese, copper, vanadium, arsenic, and nickel may already be detectable at the river's mouth due to anthropogenic activities in the basin. Possible metal contamination should be taken seriously in the Amazon, as even small amounts of the more toxic metals may lead to widespread adverse effects in the river's aquatic systems. Whereas the Amazon River is the recipient of energy and nutrients provided by its upland basin and flood plain, it is an important provider of these things to the estuarine and coastal zone of the western Atlantic. In chapter 17, David DeMaster and Robert Aller examine the interactive physical and biogeochemical processes which regulate biological productivity on the Amazon shelf. In budgets calculated from recent data, they demonstrate that the river is the most important source of silica (Si) to the shelf area and therefore strongly controls production within the diatom community. Further-more, the bulk of this Si is not retained on the Amazon shelf but is instead transported offshore where it contributes to the nutrient needs of open-water plankton. The river is also an important source of P and N, but oceanic sources, and especially internal recycling, account for the largest inputs of these elements. Approximately 60% of riverine paniculate organic C is remineralized on the shelf, mostly on the seafloor. This is a small fraction of heterotrophic activity, however, as more labile marine particulate organic C is greatly preferred by heterotrophic organisms.
Michael E. McClain
By analyzing the relationship between C and O budgets, DeMaster and Aller conclude that biological production is approximately balanced by consumption, such that the shelf area exhibits no obvious autotrophic or heterotrophic character.
The Challenge for Today and the Future As the preceding paragraphs highlight, and as the following chapters will detail, biogeochemical cycles are at the heart of many of the key issues facing the Amazon basin. Both sustainable development and preservation of the region's unique ecosystems depend on our understanding of fundamental processes such as the balance of nutrients in land and aquatic ecosystems, the assimilation capacity of contaminants in these same ecosystems, and ultimately the tolerance of Amazon ecosystems to specific development pressures. The basin's singular importance in the larger Earth system means that the consequences of poor decisions in the Amazon will be felt in adjoining regions and perhaps even on the global scale. Amazon biogeochemists face the challenge of advancing understanding at a rate greater than the advancing development. In other words, if sound scientific understanding is to guide sustainable develop in the Amazon, then research must move ahead of development. This is a formidable challenge due to the large size of the basin and the relatively small research budget devoted to Amazon research by countries composing the basin and the international scientific funding agencies. The challenge is made more difficult by the unique environmental circumstances found in the Amazon, and thus the limitations against applying models borrowed from more intensively studied systems outside the tropics. Many of the most notorious failures to date in the development of the Amazon have come from applying models of land and
The Relevance of Biogeochemistry to Amazon Development and Conservation
water management borrowed from temperate zones. Whether these models assume an unrealistic level of soil fertility, unachievable financial resources, or other conditions taken for granted in temperate zones, they are predisposed to failure. Most biogeochemists working in the Amazon today understand that the key to sustainable development is to understand the integration of physical and biogeochemical factors responsible for the luxurious natural ecosystems of the basin and then to emulate
15
or attempt to preserve these factors in development programs. There are currently several ambitious research programs that aim to do just that, and the leaders of many of these are the authors of the following chapters. I invite you the reader to delve into the volume of good research presented in this book and to make use of it in the manner that best suits your own professional position and that best enables you to join in our efforts to guide Amazon conservation and development into the future along a sustainable path.
Literature Cited Althaus, D. 1996. "Amazon's empty legacy/Big oil responds to environment/Toll on rain forest, culture still unacceptable to critics." Houston Chronicle (TX), December 15, p. 27. Andreae, M. O., R. W. Talbot, H. Berresheim, and K. M. Beecher. 1990. "Precipitation chemistry in Central Amazonia." Journal of Geophysical Research 95(D10): 16987-16999. Bartlett, K. B., and R. C. Harriss. 1993- Review and assessment of methane emissions from wetlands. Chemosphere 26: 261-320. Burns, M. K. 1998. "Biological diversity being eroded in Amazon: Even in an Ecuadorean national park, where rain forest species are supposed to be protected, colonization and poaching threaten the delicate ecosystem." The Sun (Baltimore, MD), April 23, p. 2A. Buschbacher, R. 1987. "Cattle productivity and nutrient fluxes on an Amazon pasture." Biotropica 19: 200-207. Craig, J. 1998. Ravaging of Brazil's Amazon continues - survey, Reuters, January 26. Cuevas, E., E. and Medina. 1988. "Nutrient dynamics within Amazonian forests. II. Fine root growth, nutrient availability and leaf litter decomposition." Oecologia 76: 222235. Donn, J. 1999- Report: Amazon Rain Forest Fading, The Associated Press News Service, April 8, 1999. Ferreira, L. 1996. "Asian logging companies turning to Amazon rain forest," Kyodo News International, Inc., June 4. Gomez, R. 1994. "Contaminacion Ambiental en la Amazonia Peruana." Informe Tecnico de Avance of the Institute de Investigaciones de la Amazonia, Peruana, Iquitos. Goulding, M. 1980. The Fishes and the Forest. University of California Press, Los Angeles. James, I. 1998. The Endangered People of Peru, St. Petersburg Times (FL), July 5, 1998 p. ID .
Lesack, L. F. W., and J. M. Melack. 1996. "Mass balance of major solutes in a rainforest catchment in the Central Amazon: Implications for nutrient budgets in tropical rainforests." Biogeochemistry 32: 115-142. McClain, M.E. 2001. "Ecohydrology as a tool in the sustainable development of large tropical rivers". In: Ecohydrology From Concepts to Application. Zalewski, M. (Cambridge University Press, New York). McClain, M. E., J. E. Richey, J. A. Brandes, and T. P. Pimentel. 1997. "Dissolved organic matter and terrestriallotic linkages in the central Amazon basin of Brazil. "Global Biogeochemical Cycles 11: 295-311. McClain, M. E., J. E. Richey, and T. P. Pimentel. 1994. "Groundwater nitrogen dynamics at the terrestrial-lotic interface of a small catchment in the Central Amazon basin." Biogeochemistry 27: 113-127. Nepstad, D. C., C. R. de Carvalho, E. A. Davidson, P. Jipp, P. Lefebvre, G. H. Negreiros, E. D. da Silva, T. Stone, S. Trumbore, and S. Vieira. 1994. "The role of deep roots in the hydrological and carbon cycles of Amazonian forests and pastures." Nature 372: 666-669. Ozorio de Almeida, A. L., and J. S. Campari. 1995. Sustainable Settlement in the Brazilian Amazon. Oxford University Press, New York. p. 189. Prentice, I. C., and J. Lloyd. 1998. "C-quest in the Amazon basin." Nature 396: 619-620. Schomberg, W. 1998. Brazil says farmers are destroying the Amazon, Reuters, March 3. Sippel, S. J, S. K. Hamilton, and J. M. Melack. 1992. "Inundation area and morphometry of lakes on the Amazon River floodplain, Brazil." Archiv fur Hydrobiologie 123: 385-400. Tian, H., J. M. Melillo, D. W Kicklighter, A. D McGuire, J. V. K. Helfrich, B. Moore, and C. J. Vorosmarty. 1998. "Effect of interannual climate variability on carbon storage in Amazonian ecosystems." Nature 396: 664-667.
16 Tiessen, H., E. Cuevas, and P. Chacon. 1994a. "The role of soil organic matter in sustaining soil fertility." Nature 371: 783-785. Tiessen, H., P. Chacon, and E. Cuevas. 1994b. "Phosphorus and nitrogen status in soils along a toposequence of dystrophic rainforests on the upper Rio Negro." Oecologia 99: 145-150. Williams, M. R., T. R. Fisher, and J. M. Melack. 1997a.
Michael E. McClain "Chemical composition and deposition of rain in the central Amazon, Brazil," Atmospheric Environment 31, No. 2: 207-217. Williams, M. R., T. R. Fisher, and J. M. Melack. 1997b. "Solute dynamics in soil water and groundwater in a central Amazon catchment undergoing deforestation." Biogeochemistry 38: 303-335.
2
General Characteristics and Variability of Climate in the Amazon Basin and its Links to the Global Climate System Jose A. Marengo and Carlos A. Nobre
The Amazon region is of particular interest because it represents a large source of heat in the tropics and has been shown to have a significant impact on extratropical circulation, and it is Earth's largest and most intense land-based convective center. During the Southern Hemisphere summer when convection is best developed, the Amazon basin is one of the wettest regions on Earth. Amazonia is of course not isolated from the rest of the world, and a global perspective is needed to understand the nature and causes of climatological anomalies in Amazonia and how they feed back to influence the global climate system. The Amazon River system is the single, largest source of freshwater on Earth. The flow regime of this river system is relatively unimpacted by humans (Vorosmarty et al. 1997 a, b) and is subject to interannual variability in tropical precipitation that ultimately is translated into large variations in downstream hydrographs (Marengo et al. 1998a, Vorosmarty et al. 1996, Richey et al. 1989a, b). The recycling of local evaporation and precipitation by the forest accounts for a sizable portion of the regional water budget (Nobre et al. 1991, Eltahir 1996), and as large areas of the basin are subject to active deforestation there is grave concern about how such land surface disruptions may affect the water cycle in the tropics (see reviews in Lean et al. 1996).
Previous studies have emphasized either how large-scale atmospheric circulation or land surface conditions can directly control the seasonal changes in rainfall producing mechanisms. Studies invoking controls of convection and rainfall by large-scale circulation emphasize the relationship between the establishment of upper-tropospheric circulation over Bolivia and moisture transport from the Atlantic ocean for initiation of the wet season and its intensity (see reviews in Marengo et al. 1999). On the other hand, Eltahir and Pal (1996) have shown that Amazon convection is closely related to land surface humidity and temperature, while Fu et al. (1999) indicate that the wet season in the Amazon basin is controlled by both changes in land surface temperature and the sea surface temperature (SST) in the adjacent oceans, depending if the region is north-equatorial or southern Amazonia. A better understanding of rainfall variability and its mechanisms in Amazonia requires a clear documentation of the major elements of the warm season precipitation regime, within the context of the annual cycle, and longer time scale variations such as interannual and interdecadal. Rainfall variability in Amazonia has been the subject of several studies regarding physical causes, seasonal variations, and links to the Southern Oscillation (SO), and to SST conditions in the tropical Atlantic (see reviews in Ropelewski
18
Jose A. Marengo and Carlos A. Nobre
and Halpert 1987, 1989; Marengo 1992; Rao (Fig. 2.1c) show a persistent thermal lowet al. 1993, Nobre and Shukla 1996, Guyot et pressure system between 202 and 30s S over al. 1997, Marengo et al. 1998a, b). The low SO the Gran Chaco and Pampean Sierras, related phase, which is associated with the El Nino to the maximum of cloudiness over central phenomenon, is related to negative rainfall Amazonian and the Peruvian-Bolivian anomalies in northern and central Amazonia Altiplano. At this time of the year, cold fronts and anomalously low river levels in the that enter northern Argentina and southern Amazon River, while the high SO phase Brazil are frequently accompanied by (related to the La Nina phenomenon) features enhanced deep convection over western and anomalously wet seasons in northern and southern Amazonia (Garreaud and Wallace central Amazonia. Tendencies toward drier 1998) and by increased southward moisture conditions were observed during the El Nino flux from lower latitudes. It has been sugevents of poor rainy seasons of 1925-26, gested that a strong low-level jet east of the 1982-83 and more recently during 1997-98, Andes (Nogues-Paegle and Mo 1997) while wetter conditions were observed enhances this moisture flux. Another feature during the La Nina years of 1988-89 and 1995- is the strong and persistent northeasterly flow 96. The drought of 1998 in north and central from the North Atlantic into the Amazon Amazonia is being considered as the most basin, associated with the increased pressure over the tropical North Atlantic. In this region intense of the last 118 years. On longer time scales, studies have docu- there is an indication of the presence of a mented long-term variations in rainfall in the low-level northerly jet, with a maximal wind basin, associated with possible trends or cycles speed on the order of 15 m/s at 850 mb, at in both rainfall and river levels (Rocha et al. approximately 17s S and 62Q W, that is 1989, Chu et al. 1994, Dias de Paiva and Clarke responsible for the transport of water vapor 1995, Marengo et al. 1998a, Marengo et al. and heat to the region of Paraguay and 2001a, Zhou and Lau 1999), as well as climat- northern Argentina from the Amazon. At upper levels (Fig. 2.la), the elevated teric tendencies in the water balance and moisture transport in the basin (Costa and Foley rain of the Andes and the latent heat released 1999). However, no significant trends have through deep cumulus convection during the been detected in the rainfall regime of the warm season give rise to this upper-troposAmazon region, or on the discharges of the pheric high. The causes of the Bolivian high are therefore strongly related to latent heat Amazon river and its tributaries. release over the regions of precipitation (Figueroa et al. 1995, Lenters and Cook 1995, Seasonal Variation of Climate Seluchi et al. 1998). To the east of the Bolivian high (downstream), there is an in Amazonia upper-level cold trough sitting off the east cost of Brazil. That trough is associated with Atmospheric circulation and convection descending motion and is part of the mechaThe near surface and upper air circulation nisms that lead to the very low rainfall over the Americas, as deduced by the National observed in that region (Moura and Shukla Centers for Environmental Prediction (NCEP) 1981). Southern Amazonia and the Bolivian reanalysis (Kalnay et al. 1996), are shown in Altiplano are strongly heated during the Fig. 2.1a-d for the austral summer (December- austral warm season resulting in enhanced February) and winter (June-August)]. The near tropospheric zonal temperature gradients surface circulation patterns in austral summer and enhanced upper-tropospheric meridion-
General Characteristics and Variability of Climate in the Amazon Basin and its Links to the Global Climate System
Figure 2.1 a-b Upper-level circulation (200 hPa) over South America. (a) Austral summer (December-February), (b) austral winter (June-August).
Figure 2.1 c-d Near surface (850 hPa) over South America. (c) Austral summer (December-February), (d) Austral winter (June-August).
19
Jose A. Marengo and Carlos A. Nobre
20
al flows in the vicinity of both coasts of South America. The austral wintertime upper-level circulation (Fig. 2.1b) over South America is characterized by weak winds over the tropics, while the subtropical westerly jet is stronger and located more equatorward than in summer, in agreement with the descending branch of the Hadley-type circulation over that area. At lower levels, a northward-displaced near-equatorial lowpressure trough and SST maximum characterize the circulation (Fig. 2. Id). A northward cross-equatorial flow turning clockwise is found over the tropical eastern Pacific and Atlantic Oceans (Fig. 2.1b). Near the surface, during wintertime, surges of cold high-latitude air are known locally as "friagens." These events move across southeastern Brazil and the Amazon region from the south, greatly modifying the atmospheric structure and climatic conditions. Friagens can produce severe frosts in the coffee growing areas of southern Brazil and substantial cooling in the Amazon basin (Seluchi and Nery 1992, Marengo et al. 1997 a,b, Marengo 1999). The cold air heading these thrusts of high-latitude air may reach as far north as the equator. The events are relatively common in the Amazon region during the winter
Figure 2.2 Annual rainfall distribution in Amazonia [meters] (Marengo 1995b).
(May-September). Gulf et al. (1996) note that there were 6 such events in 1992 and 9 in 1993 while 14 events were reported during the winter of 1994. Summertime convection shows a maximum over southern Amazonia which extends to the southeast. This represents the South Atlantic Convergence Zone (SACZ) and is consistent with the low-level convergence shown in Fig. 2.1c. More information on the SACZ can be found in Nogues-Paegle and Mo (1997). During winter, the convection bands move northward towards the Northern Hemisphere, determining the rainy season in northern South America.
Rainfall The Amazon region is characterized by a strong annual cycle of rainfall, whose maximum follows the meridional displacement of solar heating to a first approximation (Rao and Hada 1990, Figueroa and Nobre 1990, Marengo 1995b). Southern Amazonia has distinct dry and rainy seasons, with a maximum of rainfall occurring in summer. In the northern and central regions there is almost no dry season, but approaching the equator there are distinct maximum in spring and "suppressed" maximum in fall. During the austral spring precipitation increases over the Amazon basin, and bands of precipitation develop linking tropical convection and precipitation activity in the lower extratropics of this region. Together with the OLR fields, rainfall displays a similar seasonal cycle. The heaviest rainfall and lowest OLR occurs over southern Amazonia. Annual variation of rainfall is linked to annual changes in large-scale upper-air circulation, and surface oceanic conditions have implicated the Atlantic, at least indirectly, as being responsible for interannual rainfall variability in the Amazon. For example, Marengo (1992) found more than average rainfall throughout the northern side of the
General Characteristics and Variability of Climate in the Amazon Basin and its Links to the Global Climate System
basin in southern summer in times of enhanced northeast trades due to increased moisture flux from the Atlantic. Strong Atlantic trades bringing anomalous moisture into the Amazon are associated with a southward displaced ITCZ, which is in turn related to an anomalous distribution of Atlantic SST (Moura and Shukla 1981, Nobre and Shukla 1996). Although the timing of the annual cycle of rainfall is largely controlled by the sun, rainfall in different parts of the basin is triggered by different rain producing mechanisms (Garreaud and Wallace 1997, 1998, Fisch et al. 1998). These mechanisms can be grouped into the following categories: * Diurnal deep convection resulting from surface warming, which is most prevalent in central Amazonia.
21
* Deep convection in northern Amazonia related to the ITCZ-like convection and moisture transported from the Atlantic. This depends on the meridional SST gradient of the tropical North Atlantic. * Instability lines originating near the mouth of the Amazon River that move inland and may even reach the Andes (Cavalcand 1982, Cohen et al. 1995) can initiate rainfall in northern Amazonia. * Convective activity at the meso scale and large scale, associated with the penetration of frontal systems in south/ southeast Brazil that can reach western southern Amazonia. As a consequence of recent international field campaigns if the Amazon basin (TRMM-
Figure 23 Annual cycle of rainfall in Amazonia (Marengo and Hastenrath 1993). Scale is indicated on the lower left side.
22
LBA and LBA-WET AMC, Marengo 2001 b) during the summer of 1999, a unique data set and knowledge was provided to allow, among other things, the description and analysis the diurnal and day-to-day variations of rainfall in southwestern Amazonia. By means of a high resolution time and space rainfall data, it was possible to get a better documentation of the diurnal cycle of rainfall in this part of the Amazon, and together with other observations such as radar and lighting and soundings there was possible to identify several episodes of rainfall and relate them to different rainfall mechanisms. Rainfall in the region shows clearly a preference for
Jose A. Marengo and Carlos A. Nobre
maximum rainfall intensity between 1200 and 1600 LST, and sometimes 1800 LST, while in some periods, a second maximum shows up between 0000 and 0400 LST. However, the more frequent maximum rainfall is detected at late afternoon. Annual rainfall in northern South America varies greatly, from less than 400 mm in northeast Brazil and the Caribbean coast of South America to more than 3000 mm in the upper watershed of the Rio Negro. Fig. 2.2 shows three centers of abundant rainfall in the Amazon basin. One is located in northwest Amazonia, with more than 3600 mm per year. Another region with abundant rainfall is
Figure 2.4 Daily Precipitation at Manaus, Brazil (03Q 50'S, 602 Ol'W) during two extremes of the Southern Oscillation
General Characteristics and Variability of Climate in the Amazon Basin and its Links to the Global Climate System
the central part of Amazonia around 50Q S, with 2400 mm per year. A third center is found close to the mouth of the Amazon river near Belem, with more than 2800 mm per year. This is in agreement with the circulation patterns over the Amazon basin and their variations as shown in Fig. 2.1c and 2.Id. In the upper Rio Negro basin, in northwestern Amazonia, rainfall is abundant throughout the year, reaching its maximum in April-June (Fig. 2.3), while southern Amazonia's rain peaks earlier (JanuaryMarch). The extreme high and localized values of precipitation in narrow strips along the eastern side of the Andean slopes are thought to be due to upglide condensation and a rain shadow effect on the lee side, so the localized maximum is due to the easterly winds being lifted when they flow over the Andes. The coastal maximum is caused by nocturnal convergence between the tradewind and the land breeze. The transition between the wet and dry seasons is short in Amazonia. The onset of the wet season is a relatively fast one that typically occurs within the within the period of a single month and may sometimes occupy only a single 5-day period. The transition to the dry season is on average, longer than a month. There have been relatively few efforts devoted to understanding wet season onset in the Amazon basin. Kousky (1988) determined the climatological onset, where onset progresses from northwest to southeast, although his criteria were never met in the extreme northwest. Withdrawal of the rainy season occurs toward the northwest. Marengo et al. (2001b) identified the onset and demise of the rainy season, with the onset dates progressing southeastward from near mid-September in the north and west to the beginning of October in the southeast. Near the mouth of the Amazon, onset occurs almost at the end of December, while north of the equator rainfall is apparently tied to Northern Hemisphere summer and does not
23
occur until mid-April. The withdrawal of the rainy season progresses toward the north, but moves slower and more systematically than does onset.
Figure 2.5 Rainfall and river variability in Amazonia, (a) Rainfall in the northern Amazonia, (expressed by the NAR index, Marengo [1992]), (b) discharges of the Amazon River at Obidos, (c) Solimoes river at Manacapuru, (d) Tapajoz River at Santarem, (e) Tapajos River at Itauituba (Marengo and Hastenrath 1993).
24
A very interesting aspect of the intraseasonal variability in the Amazon region and other tropical regions is the so-called veranico, or midsummer droughts, which occur during the austral summertime rainy season. The dry pells last between 5 days and 3 weeks and have a strong effect on agriculture if they occur near the beginning of the rainy season. Figure 2.4ab shows the daily rainfall from July 1988 to June 1989 (La Nina year) and from July 1982 to June 1983 (El Nino year), respectively, for Manaus. On average, the rainy season in Manaus starts in September, reaches its maximum in February-April and minimum in JuneAugust, when the end of it is detected. During the El Nino 1982-83, deficient rain during
Jose A. Marengo and Carlos A. Nobre
November 1982, January, February, and beginning of March 1983 was related to 3 extended veranicos of more than 15 days each.
Impacts of the Southern Oscillation in the Hydrometeorology of Amazonia Ropelewski and Halpert (1987, 1989), Kiladis and Diaz (1989) and Rogers (1988) have studied the anomalous patterns of rainfall all around the world during both extremes of the SO. As a result of very sparse data coverage in Amazonia, no clear signal of rain has been identified for this region, showing only negative rainfall anomalies over the mouth of the Amazon River. This was
Figure 2. 6 Rainfall anomalies during summer (DJF) and autumn (MAM) for El Nino 1972-73 (a),(b) and El Nino 1982-83 (c),(d). [Source: CPTEC].
General Characteristics and Variability of Climate in the Amazon Basin and its Links to the Global Climate System
observed for the El Nino 1982-83, while for the El Nino 1972-73 positive rainfall anomalies for the same region were reported. ENSO processes appear dominantly as interannual fluctuations in most climate records, at least in part because historical records are short and decadal fluctuations in them are few. However, at least some suggestions of long-term variations in the character of ENSO are evident. During this century, some decades have yielded frequent and more vigorous ENSO episodes while others have had relatively weak and infrequent episodes. Most notably, the time interval from the 1930s through the 1960s had weak ENSO variations relative to the first decades of the century and the decades beginning with the 1970s. Some decades (1950s and 1970s) seem to favor cold tropical episodes whereas others favor warm episodes [e.g. 1980s and 1990s] (Dettinger et al. 2000). On multiscale hydrologic variability associated with ENSO, Dettinger et al. (2000) indicated that influences are significant in several regions. Most notably, the Amazon basin (especially in the northern drainage basin) is dry in terms of both precipitation and flows during El Nino and La Nina years (Richey et al. 1989a). El Nino influences of Amazon streamflows are replaced by approximate mirror images during years of La Nina, although precipitation differences are not strictly mirror images (See their Fig. 1). Streamflow is not exactly symmetric in its response to the positive and negative SO conditions. Analyses of hydrometeorological series in Amazonia revealed the coherence of rainfall anomalies across northern part of the basin. For the period 1967-87, Marengo and Hastenrath (1993) compared the variability of rainfall in the northern portion (expressed by the NAR index) of the basin and data from the Solimoes, Tapajos, and Santarem rivers (Fig. 2.5a-e). This work showed that rainfall and river discharge anomalies in extreme years, such us the anomalously dry years 1979-80 and
25
1982-83 (El Nino) are mutually consistent, while the rest of the years do not show consistency. In fact, the river data show a period of relatively large discharges between 1972 and 1979 at the Solimoes basin, also observed at the Tapajos River and the main Amazon stream. For the whole basin, rainfall anomalies from the El Nino events 1972-73 and 1982-83 are depicted on Fig. 2.6a-d. The rainfall anomalies in the Amazon basin show a different behavior. At the end of 1972 (Fig. 2.6a), negative rainfall anomalies were observed over western Amazonia and near the mouth of the Amazon River, where anomalies reached -25 mm. Later, at the beginning and during summer and fall (Fig. 2.6b), positive rainfall anomalies (+100 mm) are shown in western and southern Amazonia extending towards the mouth of the Amazon. During the El Nino 1982-83, the anomalies in rainfall were distributed differently than on the 1972-73 event, in terms of intensity and location. At the end of 1982 and the beginning of summer 1983 (Fig. 2.6c) large negative rainfall anomalies occurred over central and eastern Amazonia, while small positive anomalies occurred over western Amazonia. During the fall (Fig. 2.6d), large negative rainfall anomalies are observed over western and central Amazonia, and especially the mouth of the Amazon river. At the end of the rainy season large negative rainfall anomalies are observed over "western Amazonia. Regarding streamflow data, the records of the Amazon, Negro, Xingu, and Tocantins rivers (Fig. 2.7a-d) are displayed for 2 El Nino years [1982-83 and 1986-87] and 2 La Nina years [1975-76 and 1988-891- Extreme years were chosen since they may show better the associations between rainfall and rain in their basins. The year before the peak of El Nino, the river discharge at Obidos are anomalously high, while on the year of El Nino the discharges are lower than average (Fig. 2.7a). During La Nina years, the discharges at Obidos are more than 7000 mVs above normal. At Manaus (Fig. 2.7b), the main
26
Jose A. Marengo and Carlos A. Nobre
Figure 2.7 Variations of river discharge/level in Amazonia during 2 El Nino (EN) and La Nina (NN) events (Marengo et al. 1998b).
difference with Obidos is the lack of anomalously large values on the year before the peak of the event. The El Nino in 1983 showed river levels below average, while during the peak of the Nino 1987 and the two La Nina year values larger than average were observed. At Belo Monte and Tucurui (Fig. 2.7c, d) the signal is mixed. The records for the El Nino year 1982-83 are larger than average, while for the other El Nino year 1986-87 the discharges are similar or just a bit larger than average. The discharges during the La Nina years show values that are near average. In principle, the records at Obidos should reflect the rainfall regimes from the Negro River, together with the Solimoes, Madeira and Purus. Chu (1982) found that two thirds of water level variability of the Amazon River at Manaus can be explained by the volume of
the Solimoes, while the rest comes from the Rio Negro basin. The large volumes of the Solimoes may upset the gauging values in lower courses of tributaries, producing what is known as the backwater effect (Meade et al. 1991, Guyot el al. 1997, Richey, personal communication 1998). The lower slope of the Solimoes and the large extension of the basin determine this effect that certainly would affect the discharges along the main stream, as well as the water levels measured in Manaus and in Iquitos (Marengo 1999)- This indicates that for some basins, the assumption that the discharges measured on large basins represent the hydrometeorological conditions at those sites is no longer valid. As seen in the previous section, with regard to El Nino/La Nina, if the main discharge is reduced during El Nino years, then so would tributary discharge be reduced. Conversely,
General Characteristics and Variability of Climate in the Amazon Basin and its Links to the Global Climate System
the opposite effect would prevail during La Nina. During the recent El Nino 1997-98, relatively dry conditions were observed during most of the summer of 1998 in central, western, and northern Amazonia. Rainfall over northern and central Amazonia from September to December 1997 (the beginning of the rainy season) showed negative anomalies up to 200 mm or more. Rainfall anomalies up to 200 mm below the normal were observed in Brazilian Amazonia especially to the north and western sections at the beginning of summer, while in the middle of summer negative rainfall anomalies were concentrated in the extreme west and northern sections of the basin. The unusually long, dry season in northern Amazonia during 1997 and the drought during summer 1998 produced an increase in burnings and forest fires, and the main cities in Amazonia were affected by the smoke of those fires.
Recent studies by Marengo (1999) for the Peruvian Amazon show negative water level anomalies for the Amazon river at Iquitos during very strong El Nino events, although the river flow measured at Iquitos is derived from such a large upstream area that it appears relatively insensitive to the El Nino/La Nina cycle in comparison with rainrelated anomalies at Iquitos. For the Colombian Amazon, observations indicate that ENSO's effect on river discharges occurs progressively later for rivers toward the east of the country and northern South America (Poveda and Mesa 1997). The impacts of La Nina are more pronounced than those of El Nino; however, there are instances in which rainfall anomalies are not associated with extreme phases of the ENSO, and vice versa. The 1982-83 and 1997-98 El Nino events are the strongest on record but did not produce intense dry anomalies. During 1957-60 Colombia experienced one of the longest dry
Table 2.1 Various estimates of the annual water budget of the Amazon basin. Precipitation Evapotranspiration Study (mm/yr) (mm/yr) Baumgartner and Reichel (1975) Villa Nova et al. (1976) Marques et al. (1977) Marques et al. (1980) Jordan and Heuveldop (1981) Leopoldo et al. (1982) Franken and Leopoldo (1984) Shuttleworth (1988a, b) Vorosmarty et al. (1989) Russell and Miller (1990) Nizhizawa and Koike (1992) Matsuyama (1992) Marengo et al. (1994) Costa and Foley (1999) Oki et al. (1999) Zeng (1999) Marengo et al. (2001 a)
27
2170 2000 2083 2328 3664 2076 2510 2636 2260 2010 2300 2153 2888 2166 2076 2044 2146
Source: Matsumyama 1992, Marengo et al. 1994, Costa and Foley 1999.
1185 1080 1000 1261 1905 1676 1641 1319 1250 1620 1451 1139 1616 1366 1023 1879 1581
Streamflow (mm/yr) 985 920 1083 1067 1759 400 869 1317 1010 380 849 849 1272 1106 1053 365 931
28
Jose A. Marengo and Carlos A. Nobre
Foley 1999). Several of these results represent the characteristics of the hydrological cycle near Manaus. Nishizawa and Koike (1992) slightly overestimated the annual evapotranspiration ratio, since both evapotranspiration and drainage are overestimated. Differences in results are due to different areas considered for the basin, different precipitation networks and methods of assessment (isolated stations, Manaus- only rain, the EOS-DNAEE network) and the methods used to determine the annual water balance, either using the ET = P - R approximation, or using the water vapor flux convergence equation, for the entire Amazon basin, or near Manaus only. For more discussion of these vales, see Matsuyama (1992). Hydrology has emphasized the movement Water Balance of Amazonia of water in the terrestrial system, considering forcings such as precipitation and potential The water balance of the Amazon basin is evapotranspiration and responses such as of great importance, due to the presence of streamflow and storage. Regional water balthe world's largest hydrographic system. ances are calculated based on streamflow There is a concern that large-scale land use data, observed precipitation, and evapotranchanges may significantly change the flow spiration measurements. In this context, regimes of rivers within the region and the water that evaporates from land surface is land-atmosphere exchange of moisture. The lost to the system if it is advected out of the lack of continuous precipitation and evapo- prescribed region by atmospheric motion, ration data, and of measurements of river but recycled in the system if it falls again as discharge along the Amazon and its main precipitation (Brubaker et al. 1993). Studies tributaries has forced many scientists to use on recycling of water in the hydrological indirect methods for determining the water cycle of the Amazon basin have been perbalance for the region. Early studies by Salati formed on the last two decades by Molion and Marques (1984) attempted to quantify (1975), Lettau et al. (1979), Salati (1987), the components of the water balance: Salati et al. (1979), Salati and Voce (1984), average precipitation was estimated as and Eltahir and Bras (1993). 11.9 x 1012 m3 yr1 (Villa Nova et al. 1976); The moisture balance in the Amazon discharge of the Amazon at Obidos was region indicates a substantial role of evapoestimated as 5.5 x 1012 m3 yr1, from Oltman (1967); and evapotranspiration was estimated by the Penman method as 6.4 x 1012 m3 yr1. Table 2.2 Estimates of water vapor in The latter is in good agreement with the dif- Amazonia during the ABLE-2B experiment.
seasons on record, but the 1957-58 El Nino was not particularly marked either in intensity or duration. One of the rainiest years on record was 1971, which was accompanied by only a moderate La Nina effect. Clearly, other factors affect Colombian hydrology besides ENSO, and their dependence is probably nonlinear. During the recent E Nino 1997-98, deficient rainfall in Northern Amazonia, including Iquitos, Peru and the State of Roraima, Brazil, was observed during most of the rainy season. Droughts that lead to forest fires were detected in 1997/98, as it were observed during other very strong ENSO events of: 1911/12, 1925/26, 1982/83.
ference between precipitation and river discharge. The results of other studies of the annual water budget in Amazonia are listed in Table 2.1 (compiled from Matsuyama 1992, Marengo et al. 1994, and Costa and
Precipitation Water vapor divergence Evapotranspiration Source: Rocha 1992
295.8 mm/month 187.7 mm/month 108.1 mm/month
General Characteristics and Variability of Climate in the Amazon Basin and its Links to the Global Climate System
transpiration (Lettau et al. 1979, Salati et al. 1979, Marques et al. 1980). The aerological estimates of the evapotranspiration over the eastern and central parts of the basin averaged 3 to 3-5 mm/day. More recent estimates based on ABLE-2B and FLUAMAZON estimated the evapotranspiration as high as 5 mm/day with significant transient variations (Nobre et al. 1991). During highly active precipitation episodes moisture convergence can account for 70 to 80% of precipitation. However, in monthly or longer-term averages evapotranspiration is responsible for approximately 55% of precipitation. The high variability of the moisture balance components has been identified in the aerological method (Rocha 1992) and satellite derived estimates (Cutrim et al. 1995). During the rainy season in Amazonia, water vapor flux transported across the Andes is small compared with the incoming flux, and the water vapor convergence area is connected with the SACZ. Conversely, during the dry season the water vapor transported across the Andes is as large as the incoming flux near the equator. Estimates of the water balance for the ABLE-2B period (April 13-May 13, 1987) using the divergence equation are shown in Table 2.2. The approach of the water balance by meteorology has focused more on the dynamics and thermodynamics of the atmosphere at larger scales. Moisture evaporated from the land surface can act not only as a source of water for later precipitation but can also change the thermodynamic structure of the atmosphere, altering circulation and convection. For the Amazon basin, meteorological studies of the water balance have ranged from large-scale atmospheric water budgets based on the ECMWF reanalysis or FGGE gridded data (Matsuyama 1992, Oki et al. 1995, Rao et al. 1996), and previously based on climatic observations of precipitation, a few
29
radiosonde stations, and river discharges (Marques et al. 1977, 1979a, b, 1980a, b, Chu, 1982, Marengo et al. 1994). Rao et al. (1996) used the ECMWF data from 1985-89 and their results suggested that (a) the Amazon basin is the principal source of moisture for central Brazil for the period September-February, and (b) the water vapor flux from the equatorial Atlantic associated with tradewinds is the main moisture source for the Amazon basin. However, the interannual variations of these fluxes and their associations with El Nino/La Nina have not been studied yet. It is also clear that a variety of human activities can act to modify various aspects of surface hydrologic systems. For example, changes in land cover can significantly affect the surface water and energy balance through changes in net radiation, evapotranspiration, and runoff (Costa and Foley 1999). However, because of the intricate relationships between the atmosphere, terrestrial ecosystems, and surface hydrological systems, it is still difficult to gauge the importance of human activities in the Amazonian hydrologic cycle. The availability of worldwide gridded reanalysis (NCEP, ECMWF, NASA-DAO) has allowed assessments of the water vapor transport into Amazonia, and in fact some recent works have shown contradictory results in terms of trends in input moisture into the Amazon basin. For instance, Costa and Foley (1999) identified a statistically significant decreasing trend in the atmospheric transport of water both into and out of the Amazon basin, based on 20 years (1976-96) of the NCEP reanalysis. On the other hand, Curtis and Hastenrath (1999) have identified statistically significant upward trends of lower tropospheric convergence, upward flow, convergence of atmospheric water vapor transport, and precipitable water over the Amazon basin, based on the analysis of 40 years (1958-97) of the NCEP reanalysis.
30
More recently, Marengo et al. (2001a) assesses the components of the water budget in the Amazon basin and their time and space variations, by using a combination of hydrometeorological observations and moisture fluxes derived from the NCEP/ NCAR reanalyses, for the period 1970-98. Their results show that there is a seasonality and interannual variability of the water balance that varies across the basin. Because of its larger size, southern Amazonia dominates the seasonal cycle of the water balance of the entire region, while at interannual time scales rainfall anomalies in northern Amazonia modulate the water budget of the entire basin. This is due to a relatively stronger links between the hydrometeorology of northern Amazonia and the tropical Pacific interannual variability. In the entire Amazonia, precipitation exceeds evaporation representing a sink of moisture (P>E). Our estimates of the Amazon region's water balance do not show a closure of the budget, with an average imbalance of almost 44%, meaning that some of the moisture that converges in the Amazon region is not unaccounted for. The imbalance is larger in southern Amazonia as compared to northern Amazonia. Large uncertainties are detected in the moisture convergence fields derived from the reanalyses (sometimes on the order of 40% around the mean values), and in the precipitation based from observations due to sampling problems. However, variability in the moisture convergence and observed rainfall and runoff matches quite well. The cycling ratio exhibits a slight and non-significant negative trend, which is steeper in northern Amazonia, showing that for the period of the study there are no signals of an intensification of the hydrological cycle (and higher cycling rates) in northern Amazonia, while southern Amazonia exhibits a systematic increase in cycling ratio.
Jose A. Marengo and Carlos A. Nobre
Long-Term Climate Variability and Change in Amazonia The Amazon humid forest plays an important role in the water cycle and water balance of much of South American. Several model studies and field experiments show that large part of the rainfall in the region originates as water recycled in the forest. However, under the current rate of deforestation of no more than 10% per year, rainfall and streamflow observations across the basin do not exhibit any significant trends yet. The Amazon basin has experienced increased deforestation in time, and the resulting possible changes in regional and global climate have motivated a host of climate model experiments (Zeng 1999). Possible scenarios include a reduction in precipitation and evapotranspiration rates and an increase in regional air temperatures. However, the southern portions of the basin (which have the highest rates of deforestation) actually do not show reduced precipitation. For a review on the results of several experiments using different models, please refer to Fisch et al. (1998) and Gash and Nobre (1997). It is also clear that a variety of human activities can act to modify various aspects of surface hydrologic systems. For example, changes in land cover can significantly affect the surface water and energy balance through changes in net radiation, evapotranspiration, and runoff. However, because of the intricate relationships between the atmosphere, terrestrial ecosystems, and surface hydrological systems, it is still difficult to gauge the importance of human activities in the Amazonian hydrologic cycle. Previous studies have considered the long-term variability of precipitation in Amazonia, either based on records at individual stations, several stations with short time series (Rocha et al. 1989, Marengo 1992, Dias de Paiva and Clarke 1995, Chu et al.
General Characteristics and Variability of Climate in the Amazon Basin and its Links to the Global Climate System
31
Table 2.3 Deforestation rates in Brazilian Amazonia (km2/ yr). State Acre Amapa Amazonas Maranhao Mato Grosso Para Rondonia Roraima Tocantins Amazonia % /yr
78-88* 620 60 1510 2450 5140 6990 2340 290 1650 21130 0.54
88-89 540 130 1180 1420 5960 5750 1430 630 730 17860 0.48
89-90 550 250 520 100 4020 4890 1670 150 580 13810 0.37
90-91 380 410 980 670 2840 3780 1110 420 440 11130 0.30
91-92 400 36 799 135 4674 3787 2265 281 409 13786 0.37
92-94** 482 0 370 372 6220 4284 2595 240 330 14896 0.40
94-95 1208 9 2114 1745 10391 7485 4730 220 797 29059 0.81
95-96 433 0 1023 1061 6543 6135 2432 214 320 18161 0.51
Source: PRODES/INPE. * Average for the period 1978-88. ** Average for the period 1992-94.
1994), or based on discharges of the Amazon rivers and its tributaries (Gentry and LopezParodi 1980, Marengo et al. 1998a, Marengo 1995b, Guyot et al. 1997). These studies identified short periods of low-frequency variation, meaning upward or downward trends in natural climate variability rather than an effect of global climate change. Zhou and Lau (1999) have investigated the interannual and decadal variation of summer rainfall in Amazonia and found that the ENSO influence is dominant in the interannual scale, confirming earlier results by Marengo (1992). On the decadal and longer time scale the footprints of SST variation are found in the rainfall observation. The decadal mode shows a meridional displacement of the precipitation belt over the equatorial Atlantic, which may be attributed to the dipole pattern of anomalous SST across the equator. This SST dipole and its influence on circulation and interdecadal rainfall variations in Northeast Brazil have been discussed by Nobre and Shukla (1996). Regarding interdecadal circulation changes in the tropical South Atlantic, studies by Wagner (1996) have identified a substantial
warming trend in the surface waters during the austral summer months of February and March. As a consequence of the strengthened interhemispheric SST gradient in this season, the position of the ITCZ was displaced southward, and rainfall in northeast Brazil increased over 1951-90 with a statistically significant tendency. Since the North Atlantic is a source of moisture for northern Amazonia, this southward displaced ITCZ would also imply a systematically increased moisture flux from the Atlantic into Amazonia. Interdecadal and long-term variability of rainfall in Amazonia have been studied by Marengo et al. (2001 d). They based the analysis on the Indices of rainfall for Northern Amazonia NAR and southern Amazonia SAR, defined in Marengo (1992), and updated and upgraded. The NAR and SAR time series available since the late 1920's have shown a non-significant decreasing trend, and a positive trend significant at 5% level, respectively (Fig. 2.8a, b). However, the most remarkable feature is the presence of alternating phases or cycles of relatively wetter and drier periods rather than any unidirectional trends.
32
The maximum values of NAR were during the anomalously wet period of the late 1960's and the minimum was on the second half of the 1990's . SAR shows an opposite pattern. For NAR, the periods around 1945-50 and 1975-80 were transitional periods leading to relatively wetter and drier conditions, respectively. The relatively wet period from the late 1960's and middle 1970's reported in earlier publications (Marengo et al. 2001 c) is consistent with a period with positive NAR (Fig. 2.8 a), and Rio Negro water levels at Manaus. A period with negative NAR indices has been identified from the 1980's and extending into the present. On the other hand SAR shows such periods of transition a bit earlier (194045) for the first one and almost coinciding for
Jose A. Marengo and Carlos A. Nobre
the second one on the second half of 1970s. Since approximately 1992 both NAR and SAR series have been decreasing. There is a period between the end of the 1960's to the beginning of the 1970's where NAR shows opposite tendencies. For the rest of the period, NAR, together with the discharges of the Amazon River at Obidos and the rio Negro in Manaus (not shown) exhibit similar variability in time. The 10-year centered moving average curves in Fig. 2.8a, b are superimposed on the interannual variability of these hydrometeorological and circulation indices. The curves show that on decadal time scales the most notable variation identified in the
Figure 2.8 Trends and cycles in Northern Amazonian Rainfall as represented by the NAR index (Marengo 1999), for the period 1929-96.
General Characteristics and Variability of Climate in the Amazon Basin and its Links to the Global Climate System
33
Table 2.4 Comparisons of GCM simulations of Amazon deforestation experiments. Results show the differences between deforested minus control run. Experiment
AE*(mm/day)
AT*(K)
AP*(mm/day)
AC*(mm/day)
CCM, BATS Dickinson, Henderson-Sellers (1988) Dickinson and Kennedy (1992) Henderson-Sellers et al. (1993) Hahman and Dickinson (1995) Zeng et al. (1996) Hahmann and Dickinson (1997)
-0.5 -0.7 -0.6 -0.4 -2.0 -0.4
+3.0 +0.6 +0.5 +0.8 +1.0
0.0 -1.4 -1.6 -0.8 -3.1 -1.0
+0.5 -0.7 -1.0 -0.4 -1.1 -0.6
GENESIS Costa and Foley (2000)**
-0.4
+3.5
-0.4
0.0
UKMOGCM Lean, Warrilow (1989) Lean and Rowntree (1993) Lean, Rowntree (1997) Lean et al. (1996)
-0.6 -0.6 -0.8 -0.8
+2.0 1.9 +2.3 +2.3
-1.3 -0.8 -0.3 -0.4
-0.7 -0.3 +0.5 +0.4
COLA GCM, SiB, SsiB Shukla et al. (1990), Nobre et al. (199D -1.4
+2.5
-1.8
-0.4
Dirmeyer, Shukla (1994) Sud et al. (1990)
+2.0
-0.7 -1.5
-0.3 -0.3
+3.0
-0.3 -0.7 -1.5
-0.1 0.3 -0.3
+3.8 +0.1
+1.0 -0.5
+3.7 -0.1
-0.4 -1.2
GLA GCM, SiB
Sud et al. (1996a) Sud et al. (1996b) Walker et al. (1995)
-0.2 -1.0 -1.2
LMD GCM, SECHIBA Polcher, Laval (1994a) Polcher, Laval (1994b)
-2.7
-0.4
* AE is change in evapotranspiration (mm/day), AT is the change in surface air temperature (aK), AP is the change in precipitation (mm/day), AC is the moisture convergence, calculated as the difference of AP and AE (AC=AP-AE). ** Combined effects of deforestation and double atmospheric CC^-
northern Amazonia is a 2 decade-long contrasting variability, with changes in sign around the 1940's and the 1970's. The change appears to have been associated with weakening of ENSO and, possibly, a weakening of connections between its atmospheric and oceanic components
(Dettinger et al. 2000, Marengo et al. 2001d). These associations are more apparent during the period since the middle 1970s when there were more moderate to strong El Nino events. The variability of NAR and SAR reflects the ENSO effects at interannual scales, and on decadal scales with less rain in
34
Jose A. Marengo and Carlos A. Nobre
northern Amazonia in a time when more Oki et al. 1995, Costa and Foley 1999). To frequent and strong ENSO events occurring relate GCM source runoff to the available hydrographic data, runoff routing models since mid 1970's. Guyot et al. (1997) show that the Solimoes were developed and applied to regional or river in Manacapuru exhibit positive long- continental scale. The scheme of Vorosmarty term trends, together with the Madeira river at et al. (1989) was applied with water balance Porto Velho and Xingu river in Altamira. models for several worldwide large basins. These results are consistent with the positive Marengo et al. (1994) applied the river trends of the Rio Negro river in Manaus routing scheme developed by Miller et al. (Marengo 1995a), but none of those trends (1994) to the Amazon basin and its main subbasins and identified the impacts of the reach statistical significance. Regarding tendencies in air temperature, parameterization of land surface interactions Sansigolo et al. (1992) assessed tendencies on the model generated runoff, as well as the in several major cities in Brazil and identified size of the basin. positive trends significant at 95% in Belem, Many areas of Amazonia and the Pantanal while nonsignificant trends were identified for region, when viewed at GCM grid resoluManaus and Cuiaba. More recently, Victoria et tions, have greater than 10% open-water al. (1998) used the records of 17 stations coverage. The low resolution of global GCMs spread across the Amazon region for the cannot resolve even large features such as period 1918-95 to try to identify signals of the Amazon floodplain or the Pantanal, global warming in Amazonia. They found an which may have significant impacts on increasing trend (+0.63SO per century that regional energy and water fluxes between is slightly higher than the trend found for land and the atmosphere. Patchiness in the the Southern Hemisphere (+0.562C per pattern of vegetation and deforestation, century). The results of 2 nonparametrical sta- evidenced in satellite images over Rondonia, tistical tests showed that the observed trend is also impossible for low-resolution GCMs to started to be significant in the late 1960s to the represent. Higher resolution will allow better early 1970s. However, a conclusive relation- representation of these areas in the models. ship between warming trends and deforesta- Coupling to river flux models over regions tion in Amazonia is difficult to establish with and subbasins also requires higher resolutions than traditional GCMs can supply. the data available. Higher-resolution regional modeling will Climate and Hydrological Modeling greatly facilitate our hydrologic studies.
in Amazonia
Overview of experiments on Climate models have been used to simulate Amazon deforestation
the behavior of large-scale hydrologic systems such as the Amazon basin in response either to regional deforestation or doubling CO2 (see reviews in Marengo et al. 1994 and Costa and Foley 1999). In addition, several authors have simulated hydrological processes and river discharge within atmospheric general circulation models (GCM) (Vorosmarty et al. 1989, Miller et al. 1994, Marengo et al. 1994, Sausen et al. 1994,
Regarding Amazon deforestation, the largest deforestation rates are observed in southwestern and eastern Amazonia, the so-called Deforestation Arc. The Institute Nacional de Pesquisas Espaciais (INPE) of Brazil developed the Deforestation Project (in Portuguese, PRODES Programa de Monitoramento de Desflorestamento na Amazonia) based on LANDSAT images for
General Characteristics and Variability of Climate in the Amazon Basin and its Links to the Global Climate System
monitoring deforestation in the Brazilian Amazon. The estimates of deforestation rates indicated in Table 2.3 show an increase of 33% in the annual deforestation rate, varying from 0.30% in 1991 to 0.40% in 1994. In 1997 the rate increased to 0.81%. Deforestation was greater on the Brazilian states of Para, Mato Grosso, Rondonia and Acre, while the deforestation rates decreased in Amazonas, Amapa, Roraima, Maranhao, and Tocantins. Maps derived from LANDSAT products indicate that deforestation is increasing in the savanna region while smaller in the tropical rain forest. Atmospheric GCMs simulations have suggested a possible change in the regional and global climate as the results of tropical deforestation (see reviews in Zeng 1998). Under a hypothesized Amazon basin deforestation scenario, almost all models show a significant reduction in precipitation and evapotranspiration , and most found a decrease in moisture convergence. The picture becomes more complicated when one examines the sensitivities to changes in individual land surface properties such as albedo and roughness length (Dirmeyer and Shukla 1994, Hahmann and Dickinson 1997). Different experiments emphasize the importance of different aspects of the land atmospheric systems. Many more issues arise, such as the roles of moist convection (Polcher and Laval 1994a, b), the effects of a warmer ground (Eltahir and Bras 1993), the effects of a more realistic surface representation (Lean et al. 1996), and the combined effect of deforestation and increased CO2 in the atmosphere (Costa and Foley, 2000). Previous modeling studies with GCMs have suggested that large-scale changes in Amazonia land surface cover may affect the regional climate. Sensitivity studies with these models have shown that realistic descriptions of the land surface are critical to these results. For example, the predicted climatic effects of large-scale deforestation
35
can be highly dependent on albedo, surface roughness, or soil hydraulic properties. The spatial variation of these and other properties are not well established for Amazonia. Experiments by Lean and Rowntree (1993) have shown similar results, but differing in the rate of change. Eltahir and Bras (1993) have studied the sensitivity of regional climate deforestation in the Amazon basin, finding decreases up to 10% in precipitation for deforestation in patches of 250 km on a side in a mesoscale model. Dirmeyer and Shukla (1994) used the SiB model coupled to the COLA GCM and found that the change in climate, particularly in precipitation, is strongly dependent on the change of surface albedo that accompanies deforestation. Polcher and Laval (1994a, b) and Polcher (1995) used the French LMD model coupled to the SECHIBA land surface parameterization in an experiment on Amazonian and African deforestation. They found that the model produces a reduction in evaporation and an increase of soil temperature and, in contrast to similar experiments, a significant increase of moisture convergence was obtained. This increase is attributed to an enhancement of the convective activity of the ITCZ. Walker et al. (1995) and Sud et al. (1996a, b) studied the impact of in situ deforestation in Amazonia on regional climate. They have found that evapotranspiration decreases and land surface outgoing longwave radiation and sensible heat flux increase, thereby warming and drying the planetary boundary layer. A decrease in precipitation is also found, consistent with the reduction in atmospheric flux of moisture convergence. Their results are consistent with earlier simulation studies of deforestation. Rahman and Dickinson (1995) and Zeng et al. (1996) used the GCM with BATS and a primitive equation model, respectively. Zeng and collaborators showed a much-weakened Atlantic Hadley-Walker circulation as a result of a strong positive
36
Jose A. Marengo and Carlos A. Nobre
feedback loop in the atmospheric circulation However, both processes work to warm the and the hydrological cycle. Continuous Amazon basin. Deforestation results in an deforestation suggests that the replacement increased surface temperature, largely of forest by grassland may be able to sustain because of decreases in evapotranspiration. dry climate. Table 2.4 summarizes the results The combined effect of deforestation and of simulations of Amazon deforestation double CO2 is an increase in temperature on experiments. In some of the model experi- the order of 3-5QC. ences (Lean and Rowntree 1997) relative changes in the magnitude of rainfall and Conclusions evapotranspiration indicate that increased moisture convergence partially compensates Regional studies discussed in this chapter for the reduced evaporation, in contrast to clearly demonstrate the temporal and spatial many previous deforestation experiments. dependence of Amazon rainfall on variations The wide disparity of results indicated in in the meridional positions of the ITCZ and Table 2.4 may be partially explained by the regional features such as the sea-breeze specification of vegetation and soil character- interactions near the mouth of the Amazon, istics between experiments and the differing the convection forced by the Andes in model resolutions. The availability of western Amazonia, and the effects of cold extensive measurements taken during the fronts organizing summertime convection in ABRACOS field campaigns (Lean et al. 1996, southern Amazonia. These also determine Gash et al. 1996, Gulf et al. 1996, Gash and different diurnal cycles of precipitation Nobre 1997) for paired sites of forest and across the basin. Equatorial Amazonia does cleared areas provides a major opportunity to not experience a dry season, as southern or introduce some consensus across the breadth eastern Amazonia does. Southern Amazonia of experiments. Gash and Nobre (1997) using rainfall peaks in early summer, while central the ABRACOS measurements of surface Amazonia rainfall peaks in late summer. climate demonstrated the local-scale, meso- Consequently, the season of largest river scale and large-scale climatic impacts of discharges occurs in early winter, a few Amazonian deforestation. The difference in months after the peak of the rainy season in radiation and energy balance between northern and western Amazonia. Daily forest and clearing produces higher air variability of rainfall is characterized by the temperatures in the clearings, particularly in presence of dry spells (veranicos) during the dry season. In areas of substantial defor- the rainy summer season extending between estation, higher sensible heat fluxes from the November and February. Northern Amazonia cleared forest produce deeper convective exhibits longer and more intense veranicos boundary layers, with differences in cloud during El Nino years, as compared to cover being observed and meso-scale circu- short and less intense veranicos during La lations being predicted. Nina years. A recent study by Costa and Foley (2000) The tropical Atlantic is the main source of examined the combined effects of deforesta- moisture for the Amazon via the tradewinds, tion and doubled atmospheric CO2 concentra- and the intensity and timing of this transport tions on the climate of Amazonia. He found is dependent on the interannual variability of that the effects of deforestation and increasing tropical Atlantic circulation and SST meridCO2 concentrations on evapotranspiration and ional gradients. Furthermore, signals of the precipitation tend to counteract one another. tropical Pacific in Amazonia are observed
General Characteristics and Variability of Climate in the Amazon Basin and its Links to the Global Climate System
especially during very strong El Nino events, such as 1982-83 and in 1997-98. These two events featured large negative rainfall anomalies in the northern and central parts of the basin, due to a suppressed convection and relatively strong subsidence in the region. The strongest signal of the SO is detected in northern Amazonia, while rain in southern Amazonia is not affected by the extremes of the SO. Anomalously low rainfall favors the propagation of forest fires in northern Amazonia, and also causes problems with the generation of hydroelectricity and transportation due to very low levels in the main Amazonian rivers. Recent studies have shown a decadal variability in rainfall and riverflows in Amazonia, which is consistent with similar time scale behavior in Northeast Brazil, southern Brazil and Northern Argentina, and which is coherent with decadal scale variability in SST in the tropical and the southern Atlantic. On longer time scales, no unidirectional tendencies in rainfall and rivers have been found in Amazonia, as one would expect from increased deforestation in the basin and as is predicted by model
37
experiments on regional deforestation. Most of the climate models show similar tendencies even though magnitudes on change differ. Analysis of climatic data from forest and recently deforested areas in Amazonia has corroborated changes in air temperature, evaporation, albedo, and humidity due to the removal of the forest. A positive trend in air temperature has been identified from observations, even though these records are sometimes not long enough or do not cover enough of the region to affirm that a signal of global warming or deforestation is detected in Amazonia. The motivation for large-scale studies, such as the Large Scale Biosphere Atmosphere Experiment in Amazonia (LBA), is the need to understand the different terms in the atmospheric and surface water, carbon and energy budget on a basin-wide scale. These must be understood both as they are now and as they may evolve in the future in response to vegetation change and human-induced climate change. This new understanding will be developed using a host of new observational evidence gathered in Amazonia and hydrological and GCM models.
Literature Cited Baumgartner, A., and R. Riechel. 1975. The World Water Balance. Elsevier, New York, pp. 464. Brubaker, K., D. Entekhabi, and P. Eagleson. 1993. "Estimation of continental precipitation recycling." Journal of Climate 6: 1077-1089. Cavalcanti, I. F. A. 1982. "Urn estudo sobre interacoes entre sistemas de circulacao de escala sinotica e circulacoes locais." Relatorio Tecnico INPE 2492—TDL/097, p. 113 Chu, P. S. 1982. Diagnostics of climate anomalies in tropical Brazil. Ph.D. Dissertation. Department of Meteorology, University of Wisconsin, Madison. USA. Chu, P. S., -P. Yu, and S. Hastenrath. 1994. "Detecting climate change concurrent with deforestation in the Amazon basin: which way has it gone?" Bulletin American Meteorological Society. 75: 579-583. Cohen, J. C. P., M. A. F. Silva Dias, and C. A. Nobre. 1995. "Environmental conditions associated with Amazonian Squall Lines: a case study." Monthly Weather Review. 123(11): 3163-3174.
Costa, M. H., and J. Foley. 1999. "Trends in the Hydrological Cycle of the Amazon basin." Journal of Geophysical Research. 104: 14189-14198. Costa, M., Foley, J. 2000. "Combined effects of deforestation and double atmospheric CO2 concentrations on the climate of Amazonia." Journal of Climate. 13:18-34. Gulf, A. J., Esteves, A. Marques Filho, and H. da Rocha. 1996. "Radiation, temperature and humidity over forest and pasture in Amazonia." In: Amazonian Deforestation and Climate, eds. J. Gash, C. Nobre, J. Roberts, and R. Victoria, pp. 175-192. Curtis, S., and S. Hastenrath, 1999. Trends of upper-air circulation and water vapour over equatorial South America and adjacent oceans. International Journal of Climatology 19: 863-876. Cutrim, E., D. Martin, and R. Rabin. 1995. "Enhancement of cumulus clouds over deforested lands in Amazonia." Bulletin American Meteorological Society 76: 1801-1805.
38
Dettinger, M., D. Cayan, G. McCabe, andj. Marengo. 2000. "Multiscale ENSO Variability and Hydrological Systems." In: El Nino: Multiscale Variability and its Impacts on Natural Ecosystems and Society, eds. Henry Diaz, Vera Markgraf (Cambridge University Press), pp. 113-148 Dias de Paiva, E. M. C, and R. Clarke. 1995. "Time trends in rainfall records in Amazonia." Bulletin American Meteorological Society 75: 579-583. Dickinson, R., and A. Henderson-Sellers. 1988. "Modeling tropical deforestation: A study of GCM land-surface parameterization." Quarterly Journal Royal Meteorological Society 114: 439-462. Dickinson, R., and P. Kennedy. 1992. "Impacts on regional climate of Amazon deforestation." Geophysical Research Letters. 19: 1947-1950. Dirmeyer, P., and J. Shukla. 1994. "Albedo as a modulator of climate response to tropical deforestation." Journal of Geophysical Research 99: 20863-20877. Eltahir, E., and R. Bras. 1993. "On the response of the tropical atmosphere to large-scale deforestation." Quarterly Journal Royal Meteorological Society 119: 779-783. Eltahir, E. 1996. "Role of vegetation in sustaining large-scale atmospheric circulations in the tropics." Journal of Geophysical Research 101: 4255-4268. Eltahir, E., and J. Pal. 1996. "Relationships between surface conditions and subsequent rainfall in convective storms." Journal of Geophysical Research 101: 26237-26245. Figueroa, N., and C. Nobre. 1990. "Precipitation distribution over central and western tropical South America." Climanalise. 5: 36-40. Figueroa, N., P. Satyamurty, and P. L. Silva Dias. 1995. "Simulations of the summer circulation over the South American region with an eta coordinate model." Journal of Atmospheric Science 52: 1573-1584. Fisch, G., J. Marengo, and C. Nobre. 1998. "Uma revisao geral sobre o Clima da Amazonia." Acta Amazonica. 28: 101-126. Franken, W., and P. Leopoldo. 1984. "Hydrology of catchment areas in Central-Amazonian forest streams." In: The Amazon, Limnology and Landscape Ecology of a Mighty Tropical River and its Basin, ed. H. Soili (W. Junk, Dordrecht, The Netherlands), pp. 501-519. Fu, R., B. Zhu, and R. Dickinson. 1999. "How do atmosphere and land surface influence seasonal changes of convection in the tropical Amazon?" Journal of Climate 12: 1306-1321. Garreaud, R., and J. M. Wallace. 1997. "The diurnal march of convective cloudiness over the Americas." Monthly Weather Review 125: 3157-3171. Garreaud, R., and J. M. Wallace. 1998. "Summertime incursions of midlatitude air into subtropical and tropical South America." Monthly Weather Review 126: 27143-2733. Gash, J., and C. Nobre. 1997. "Climatic effects of Amazonian deforestation: Some results from ABRACOS." Bulletin American Meteorological Society 78: 823-830. Gash, J., C. Nobre. J. Roberts, and R. Victoria. 1996. "An overview of ABRACOS." In: Amazonian Deforestation and Climate, eds. J. Gash, C. Nobre, J. Roberts, and R. Victoria (John Wiley & Sons, New York), pp. 1-14.
Jose A. Marengo and Carlos A. Nobre
Gentry, A., and J. Lopez-Parodi. 1980. "Deforestation and increased flooding in the upper-Amazon." Science 210: 1354-1356. Guyot, J.-L., J. Callede, M. Molinier, W. Guimaraes, and E. de Oliveira. 1997. "La variabilite hydrologique actuelle dans le bassin de 1'Amazone. Seminario Internacional Consequencias climaticas e Hidrologicas del Evento El Nino a Escala Regional y local." Memorias Tecnicas. 26-29 November Quito, Ecuador, pp. 285-293. Hahmann, A., and R. Dickinson. 1995. "Performance and sensitivity of the RCCM2/BATS model to tropical deforestation over the Amazon Basin." Conf. proc. XXI IUGG general Assembly, Boulder, CO. Hahmann, A., and R. Dickinson. 1997. "RCCM2 BATS model over tropical South America: application to tropical deforestation." Journal of Climate 10: 1944-1964. Hastenrath, S. 1995. "Climate and Dynamics of the Tropics." Kluwer. p. 256. Henderson-Sellers, A., R. Dickinson, T. Durbidge, P. Kennedy, K. McGuffie, and A. Pitman. 1993. "Tropical deforestation: modeling local to regional scale climate change." Journal of Geophysical Research 98: 7289-7315. Jordan, C., and J. Heuveldop. 1981. "The water balance of an Amazonian rain forest." Acta Amazonica 11: 87-92. Kalnay, E., and coauthors. 1996. "The NCEP/NCAR Reanalysis Project." Bulletin American Meteorological Society 77: 437-471. Kiladis, G. N., and H. F. Diaz. 1989- "Global climatic anomalies associated with extremes in the Southern Oscillation." Journal of Climate. 2: 1069-1090. Kousky, V. E. 1988."Pentad outgoing longwave radiation climatology for the South American sector." Revista Brasileira Meteorologia. 3: 217-231. Lean, J., and D. Warrilow. 1989. "Climatic impact of Amazon deforestation." Nature 342: 311-313. Lean, J., and P. Rowntree. 1993. "A GCM simulation of the impact of Amazon deforestation on climate using an improved canopy representation." Quarterly Journal Royal Meteorological Society. 119: 509-530. Lean J., and P. Rowntree. 1997. "Understanding the sensitivity of a GCM simulation of Amazonian deforestation to specification of vegetation and soil characteristics." Journal of Climate 6: 1216-1235. Lean, J., C. Bunton, C. Nobre, and P. Rowntree. 1996. "The simulated impact of Amazonian deforestation on climate using measured ABRACOS vegetation characteristics." In: Amazonian Deforestation and Climate, eds. J. Gash, C. Nobre, J. Roberts, and R. Victoria (John Wiley & Sons, New York), pp. 549-576. Lenters, J., and K. H. Cook. 1995. "Simulation and diagnosis of the regional summertime precipitation climatology of South America." Journal of Climate 8: 2988-3005. Leopold, L. 1962. "Rivers," American Scientist 50: 511-537. Leopoldo, P., W. Franken, E. Matsui, and M. Roibeiro. 1982. "Estimativa da evapotranspiracao da floresta Amazonica de terra firme." Acta Amazonica 12: 23-28. Lettau, H., K. Lettau, and L. Molion. 1979. "Amazonia's hydrologic cycle and the role of atmospheric recycling in assessing deforestation effects." Monthly Weather Review. 107: 227-238.
General Characteristics and Variability of Climate in the Amazon Basin and its Links to the Global Climate System Marengo, J. 1992 "Interannual variability of surface climate in the Amazon Basin." International Journal of Climatology. 12: 853-863. Marengo, J. 1995a. "Variations and Change in South American Streamflow." Climate Change 31: 99-117. Marengo, J. 1995b. "Interannual variability of deep convection in the tropical South American sector as deduced from ISCCP C2 data." International Journal of Climatology 15: 995-1010. Marengo, J. 1999- "Factores medio-ambientales del area de estudio: Climatologia en la Zona de Iquitos. Parte III." In: Geoecologia y Desarrollo de la Zona de Iquitos, Peru, eds. R. Kalliola, and S. Flores University of Turku, Finland, pp.35-55 Marengo, J., and S. Hastenrath. 1993. "Case studies of extreme climatic events in the Amazon basin." Journal of Climate 6: 617-627. Marengo, J., J. Miller, G. Russell, C. Rosenzweig, and F. Abramopoulos. 1994. "Calculations of river-runoff in the GISS GCM: impact of a new land surface and runoff routing model in the hydrology of the Amazon River." Climate Dynamics 10: 349-361. Marengo, J., A. Cornejo, P. Satyamurty, C. A. Nobre, and W. Sea. 1997a. "Cold waves in the South American continent. The strong event of June 1994." Monthly Weather Review 125: 2759-2786. Marengo, J., C. Nobre, and A. Gulf. 1997b. "Climatic impacts of Friagens in forested and deforested areas of the Amazon Basin." Journal of Applied Meteorology 36: 1553-1566. Marengo, J., Liebmann, B. Kousky, V., Filizola, N., Wainer I, 2001a. "On the onset and end of the rainy season in the Brazilian Amazon Basin." Journal of Climate, 14: 833-852. Marengo, J., Nobre, C. A., Sanches, M. B. 2001b. "On the atmospheric water balance and moisture cycling in the Amazon basin. Characteristics and space-time variability." Submitted to Journal of Hydrometeorology. Marengo, J., Fisch, G., Vendrame, I., Cervantes, P., 2001c. "On the diurnal and day-to-day variability of rainfall in Southwest Amazonia during the TRMM-BA and LBAWET AMC campaign of the summer of 1999." Submitted Journal of Geophysical Research. Marengo, J, Bhatt, U., Cunningham, C., 2001d. "Decadal and multi-decadal climate variability in the Amazon basin." Submitted International Journal of Climatology. Marengo, J., 1999. "Factores medio-ambientales del area de estudio: Climatologia en la Zona de Iquitos. Parte HI." In Geoecologia y Desarrollo de la Zona de Iquitos, Peru. R. Kalliola, S. Flores Eds. University of Turku, Finland, pp. 35-55. Marengo, J., J. Tomasella, Uvo. 1998a. "Long-term streamflow and rainfall fluctuations in tropical South America: Amazonia, Eastern Brazil and Northwest Peru." Journal of Geophysical Research 103: 1775-1783. Marengo, J., C. A. Nobre, and G. Sampaio. 1998b. "On the associations between hydrometeorological conditions in Amazonia and the extremes of the Southern Oscillation." In: extended abstracts of Memorias Tecnicas, Seminario internacional Consecuencias climaticas e hidrologicas
39
del evento El Nino a escala Regional y Local. 26-29 November 1997, Quito, Ecuador, pp. 257-266. Marengo, J., B. Liebmann, V. Kouksy, N. Filizola, and I. Wainer. 1999. "On the onset and end of the rainy season in the Brazilian Amazon Basin." Journal of Climate. Submitted. Marques, J., M. dos Santos, N. Villa Nova, and E. Salati. 1977. "Precipitable water and water wapor flux between Belem and Manaus." Acta Amazonica 7: 355-362. Marques, J., M. dos Santos, and E. Salati. 1979a. "O armazenamento atmosferico de vapor dagua sobre a regiao Amazonica." Acta Amazonica 9: 715—721. Marques, J., M. dos Santos, and E. Salati. 1979b. "O Campo de fluxo de vapor de agua sobre a regiao Amazonica." Acta Amazonica 9: 701-713. Marques, J., M. dos Santos, and E. Salati. 1980a. "Calculo de evapotranspiracao na bacia Amazonica a traves do metodo aerologico." Acta Amazonica 10: 357-361. Marques, J., M. dos Santos, and E. Salati. 1980b. "A divergencia do campo do fluxo de vapor dagua e as chuvas na regiao Amazonica." Acta Amazonica 10: 133-140. Matsuyama, H. 1992. "The water budget in the Amazon River Basin during the FGGE Period." Journal of Meteorological Society Japan 70: 1071-1083Meade, R., J. Rayol, S. Conceicao, and J. Natividade. 1991. "Backwater effects in the Amazon river basin of Brazil." Environmental Geological Science 18: 105-114. Miller, J., and G. Russell. 1992. "The impact of global warming on river runoff." Journal of Geophysical Research 97: 2757-2764. Miller, J., G. Russell, and G. Caliri. 1994. "Continental scale river flow in climate models." Journal of Climate7: 914-928. Milliman, J., and R. Meade. 1983. "World-wide delivery of river sediment to the oceans." Journal of Geology 91: 1-21. Molion, L. 1975. A climatonomic study of the energy and moisture fluxes of the Amazon basin with consideration on deforestation. Ph.D. Dissertation, Department of Meteorology, University of Wisconsin, Madison, p. 190. Moura, A. D., and J. Shukla. 1981. "On the dynamics of the droughts in Northeast Brazil: observations, theory and numerical experiments with a general circulation model." Journal of Atmospheric Science 38 (12): 2653-2673. Nace, R. I. 1972. "World hydrology: status and prospects," in IAHS-UNESCO-WMO-ed., World Water Balance 1: 1-10. Nizhizawa, T., and Y. Koike. 1992. Amazon-Ecology and Development, Iwanami, Tokyo, p. 221. Nizhizawa, T, and M. Tanaka. 1983. "The annual change in the tropospheric circulation and the rainfall in South America." Archives of Meteorological Geophysical Bioklimatologie Series 33: 107-116. Nobre, C., P. Sellers, and J. Shukla. 1991. "Amazonian deforestation and regional climate change." Journal of Climate 4: 957-988. Nobre, C. A., J. C. Gash, R. Hutjes, D. Jacob, A. Janetos, P. Kabat, M. Keller, J. Marengo, J. R. McNeal, P. Sellers, D.
40
Wickland, and S. Wofsy. 1996. The Large Scale Biosphere-Atmosphere Experiment in Amazonia (LBA). Concise Experimental Plan. Compiled by the LBA Science Planning Group. Staring Center-DLO, Wageningen, The Netherlands. Nobre, P., and J. Shukla. 1996. "Variations of sea surface temperature, wind stress, and rainfall over the tropical Atlantic and South America." Journal of Climate 9: 2464-2479. Nogues-Paegle, J., and K. M Mo. 1997. "Alternating wet and dry conditions over South America during Summer." Monthly Weather Review 125: 279-291. Oki, T., K. Musiake, H. Matsuyama, and K. Masuda. 1995. "Global atmospheric water balance and runoff from large river basins." Hydrological Processes 9: 655-678. Oki, T., 1999. The global water cycle. Chapter 1.2, Global Energy and Water Cycles, Eds. K. Browning and R. Gurney, Cambridge University Press, pp. 10-27. Oilman, R. 1967. "Reconnaissance investigations of the discharge and water quality of the Amazon," In: Atas do Simp, sobre Biota Amazonica, Rio de Janeiro, CNPq, vol. 3, pp. 163-185. Polcher, J. 1995. "Sensitivity of tropical convection to land surface process." Journal of Atmospheric Science 52: 3143-3161. Polcher, J., and K. Laval. 1994a. "The impact of African and Amazonian deforestation on tropical climate." Journal of Hydrology 155: 389-405Polcher, J., and K. Laval. 1994b. "A statistical study of the regional impact of deforestation on climate in the LMD GCM." Climate Dynamics 10: 205-219. Poveda, G., and O. Mesa. 1997. "Feedbacks between hydrological processes in tropical South America and large-scale ocean atmospheric phenomenon." Journal of Climate 10: 2690-2702. Rao, V. B., M. C. Lima, and S. H. Franchito. 1993. "Seasonal and interannual variations of rainfall over eastern Northeast Brazil." Journal of Climate 6: 1754-1763. Rao, V. B., and K. Hada. 1990. "Characteristics of rainfall over Brazil: Annual variations and connection with the Southern Oscillation." Theoretical and Applied Climatology 42: 81-91. Rao, V. B., I. Cavalcanti, and K. Hada. 1996. "Annual variation of rainfall over Brazil and water vapor characteristics over South America." Journal of Geophysical Research 101: 26539-26551. Richey, J., C. Nobre, and C. Deser. 1989a. "Amazon River discharge and climate variability: 1903 to 1985." Science 246: 101-103. Richey, J. E., L. A. K. Mertes, T. Dunne, R. L. Victoria, B. L. Forsberg, A. C. N. S. Tancredi, and E. Oliveira. 1989b. "Sources and routing of the Amazon River flood wave. Global Biogeochem." Cycles 3: 191-204. Rocha, R. P. 1992. "Simulacao numerica de sistema de mesoescala sobre a America do Sul." Tese de Mestrado. IAG/USP. Rocha, H., C. Nobre, and M. Barros. 1989. "Variabilidade natural de longo plazo no ciclo hidrologico da Amazonia." Climanalise 4(12): 36-42. Rogers, J. C. 1988. "Precipitation variability over the
Jose A. Marengo and Carlos A. Nobre Caribbean and tropical Americas associated with the Southern Oscillation." Journal of Climate 1: 172-182. Ropelewski, C., and M. Halpert. 1987. "Global and regional scale precipitation patterns associated with the El Nino-Southern Oscillation." Monthly Weather Review 115: 1606-1626. Ropelewski, C., and M. Halpert. 1989. "Precipitation patterns associated with the High Index of the Southern Oscillation." Journal of Climate 2: 268-284. Russell, G., and J. Miller. 1990. "Global river runoff calculated from a global atmospheric general circulation model." Journal of Hydrology 117: 241-254. Salati, E. 1987. "The forest and the hydrological cycle." In: The Geophysiology of Amazonia, ed. R. E. Dickinson (Wiley, New York), pp. 273-296. Salati, E., and P. Voce. 1984. "Amazon basin: A system in equilibrium," Science 225: 128-138. Salati, E., and J. Marques. 1984. "Climatology of the Amazon region." In: The Amazon: Limnology and Landscape Ecology of a Mighty Tropical River and its Basin., ed. H. Sioli (W. Junk, Dordrecht, The Netherlands). Salati, E., and P. B. Vose. 1984. "Amazon Basin: A system in equilibrium." Science 225: 129. Salati, E., A. DaH'Ollio, E. Matsui, and J. Gat. 1979. "Recycling of water in the Amazon basin: An isotopic study." Water Resource Research 15: 1250-1258. Sansigolo, C., R. Rodrigues, and Etchichury. 1992. "Tendencias nas temperaturas medias do Brasil." Anais do VII Congresso Brasileiro de Meteorologia. Vol. 1, 367-371. Sausen, R., S. Schubert, and L. Dumenil. 1994. "A model of river runoff for use in coupled atmosphere-cean models." Journal of Hydrology 155: 337-352. Seluchi, M., and J. Nery. 1992. "Condiciones meteorologicas asociadas a la ocurrencia de heladas en la region de Maringa." (in Spanish). Revista Brasileira de Meteorologia 7: 523-534. Seluchi, M., Y. V. Serafini, and H. Le Treut. 1998. "The impact of the Andes on transient atmopsheric systems: A comparison between observations and GCM results." Monthly Weather Review 126: 8905-912. Shukla, J., C. Nobre, and P. Sellers. 1990. "Amazonia deforestation and climate change," Science 247: 1322-1325Shuttleworth, W. J. 1988a. "Evaporation from Amazonian rainforests." Philosophical Transactions of Royal Society of London B233. 321-346. Shuttleworth, W. J. 1988b. "Macrohydrology—the new challenge for process hydrology." Journal of Hydrology 100: 31-56. Sud, Y., Y. Sellers, Y. Mintz, G. Chou, G. Walker, and W. Smith. 1990. "Influence of the biosphere on global circulation and hydrologic cycle- a GCM simulation experiment." Agricultural and Forest Meteorology 52: 133-180. Sud, Y, R. Yang, and G. Walker. 1996a. "Impact of in situ deforestation in Amazonia on the regional climate: General circulation model simulation study." Journal of Geophysical Research 101: 7095-7109. Sud, Y., G. Walker, J.-H. Kim, G. Listen, P. Sellers, and W. Lau. 1996b. "Biogeophysical consequences
General Characteristics and Variability of Climate in the Amazon Basin and its Links to the Global Climate System of a Tropical Deforestation Scenario: A GCM simulation study." Journal of Climate 9: 3226-3247. Victoria, R., L. Matinelli, J. Moraes, M. Ballester, A. Krusche, G. Pellegrino, R. Almeida, and J. Richey. 1998. "Surface air temperature variations in the Amazon region and its border during this century." Journal of Climate 1: 1105-1110. Villa Nova, N., E. Salati, and E. Matsui. 1976. "Estimativa da evapotranspiragao na bacia Amazonica." Acta Amazonica 6: 215-228. Vorosmarty, C, C. Willmott, B. Choudhury, A. Schloss, T. Stearns, S. Roberson, and T. Dorman. 1996. "Analyzing the discharge regime of a large tropical river through remote sensing, ground-based climatic data, and modeling." Water Resource Research 32: 3137-3150. Vorosmarty, C. J., K. Sharma, B. Fekete, A. H. Copeland, J. Holden, J. Marble, and J. A. Lough. 1997a. "The storage and aging of continental runoff in large reservoir systems of the world." Ambio 26: 210-19Vorosmarty, C. J., M. Meybeck, B. Fekete, and K. Sharma. 1997b. "The potential impact of neo-Castorization on sediment transport by the global network of rivers." In: Human Impact on Erosion and Sedimentation, eds. D. Walling, and J.-L. Probst (IAHS Press, Wallingford U.K.), pp. 261-72.
41
Vorosmarty, C., B. Moore, A. Grace, L. Gildea, J. Melillo, B. Peterson, E. Rastetter, and P. Steudler. 1989. "Continental scale models for water balance and fluvial transport: An application to South America." Global Biogeochemistry 3: 241-265. Wagner, R. 1996. "Decadal-scale trends in mechanisms controlling meridional sea surface temperature gradients in the tropical Atlantic." Journal of Geophysical Research 101: 16683-16694. Walker, G., Y. Sud, and R. Atlas. 1995. "Impact of the ongoing Amazonian deforestation on local precipitation: A GCM simulation study." Bulletin American Meteorological Society 76: 346-361. Zeng, N., 1999"Seasonal cycle and interannual variability in the Amazon hydrologic cycle." Journal of Geophysical Research. 104, 9097-9106. Zeng, N. 1998: "Understanding climate sensitivity to tropical deforestation in a mechanistic model." Journal of Climate 11: 1969-1975. Zeng, N., R. Dickinson, and X. Zeng. 1996. "Climatic impact of Amazon deforestation—A mechanistic study." Journal of Climate 9: 859-883. Zhou, J., and K.M. Lau. 1999. "Interannual and decadal variability of summer rainfall over South America." Journal of Climate (accepted).
3 The Atmospheric Component of Biogeochemical Cycles in the Amazon Basin Paulo Artaxo
Tropical forests, with their high biological activity, have the potential to emit large amounts of trace gases and aerosol particles to the atmosphere. The accelerated development and land clearing that is occurring in large areas of the Amazon basin suggest that anthropogenic effects on natural biogeochemical cycles are already occurring (Gash et al. 1996). The atmosphere plays a key role in this process. The tropics are the part of the globe with the most rapidly growing population, the most dramatic industrial expansion and the most rapid and pervasive change in land use and land cover. Also the tropics contain the largest standing stocks of terrestrial vegetation and have the highest rates of photosynthesis and respiration. It is likely that changes in tropical land use will have a profound impact on the global atmosphere (Andreae 1998, Andreae and Crutzen 1997). A significant fraction of nutrients are transported or dislocated through the atmosphere in the form of trace gases, aerosol particles, and rainwater (Keller et al. 1991). Also the global effects of carbon dioxide, methane, nitrous oxide, and other trace gases have in the forest ecosystems a key partner. The large emissions of isoprene, terpenes, and many other volatile organic compounds could impact carbon cycling and the production of secondary aerosol particles over the Amazon region. Vegetation is a natural source of many types of aerosol particles that
play an important role in the radiation budget over large areas (Artaxo et al. 1998). There are 5 major reservoirs in the Earth system: atmosphere, biosphere (vegetation, animals), soils, hydrosphere (oceans, lakes, rivers, groundwater), and the lithosphere (Earth crust). Elemental cycles of carbon, oxygen, nitrogen, sulfur, phosphorus, and other elements interact with the different reservoirs of the Earth system. The carbon cycle has important aspects in tropical forests due to the large amount of carbon stored in the tropical forests and the high rate of tropical deforestation (Jacob 1999). In Amazonia there are two very different atmospheric conditions: the wet season (mostly from November to June) and the dry season (July-October) (see Marengo and Nobre, this volume). Biomass burning emissions dominate completely the atmospheric concentrations over large areas of the Amazon basin during the dry season (Artaxo et al. 1988). In the wet season, a very clean atmosphere shows very low concentrations for most of trace gases and aerosol concentration (Artaxo et al. 1990).
Trace Gases Relevant to the Biogeochemical Cycles in Amazonia Nitrogen and oxygen account for 99% of the Earth's atmosphere. Their concentrations have stayed nearly constant over the
The Atmospheric Component of Biogeochemical Cycles in the Amazon Basin
43
last several hundred million years. Anthro- tant source of CO2, and the average soil pogenic influences are altering the concen- respiration rate depends on soil moisture tration of most of the other relevant gases, and temperature. Biomass burning influences which occur in only trace amounts. Carbon significantly soil respiration rates for the dioxide (CO2) is chemically inert in cerrado and savanna ecosystems (Matson the troposphere, with a lifetime of about and Harriss 1995). 105 years. Methane (CH4) has an atmospherMethane (CH4) has important soil, water, ic lifetime of about 9-6 years and is chemi- and vegetation emissions, and it is the second cally active and emitted in large amounts most important greenhouse gas after CO2. from the flooded forest, from swamps, and Biomass burning of cerrado and tropical forest through biomass burning. Carbon monoxide is an important global source of methane. The (CO) is emitted through incomplete combus- CH4 emission rates for soils depend strongly tion and in soil processes, whereas oxides of on soil moisture. Large positive and negative nitrogen (NO, NO 2 , N2O) are also emitted as fluxes were observed for savannas in Africa, a consequence of biogenic processes in soils depending on flooding condition and ecosysand through biomass burning (Warneck and tem type. Termite and enteric fermentation in Zellner 1999). The tropics present the highest large herbivores are also a significant source of concentrations of the radical OH, because of methane (Jacob 1999). the high water vapor and UV radiation Globally, the oxides of nitrogen, NO (nitric (Andrea and Crutzen 1997). As OH is respon- oxide), NO2 (nitrogen oxide), and N2O sible for most of the removal mechanism of (nitrous oxide), are key species involved in the O3, CH4, CO, and other trace gases from the chemistry of the troposphere and stratosphere. atmosphere, it influences significantly tropi- NO and N2O are produced mostly by microcal trace gas concentrations (Jacob 1999, bial soil activity, whereas biomass burning is Brasseur et al. 1999). Tropical savannah soils also an important source of NO. Nitric oxide is are important sources of many trace gases a species involved in the photochemical prosuch as CO2, CH4, N2O, and NO (Sanhueza duction of ozone in the troposphere, is et al. 1994). After emissions some of these involved in the chemical production of nitric gases are chemically transformed under the acid, and is an important component of acid precipitation. Nitrous oxide plays a key role in very active tropical atmosphere. Carbon dioxide (CO2) is the most impor- stratospheric ozone depletion and is an important trace gas in the forest ecosystem, and it tant greenhouse gas, with a global warming is subject to strong and complex biogeo- potential more than 200 times that of CO2. Carbon monoxide (CO) strongly influences chemical interactions (Houghton et al. 1998). CO2 participates in respiration and photo- the concentration of the radical OH in the synthesis and it is the driving force behind tropical atmosphere. CO oxidation can lead to most of the biological processes in the soil either production or destruction of ozone, and vegetation. Several short-term recent depending on the NOX mixing ratio. Tropical experiments show an accumulation of car- soils are either a sink or a weak source of CO, bon in the forest ecosystem. More long-term where photochemical oxidation of methane measurements and better integration of car- and other hydrocarbons and biomass burning bon flux measurements with other ecosystem emissions are the predominant CO sources. Tropospheric ozone (O3) is a gas that has and climatic measurements are needed to better address the issue of whether tropical no direct emission sources and is the third forest is a sink or a source of CO2 to the most important greenhouse gas after CO2 atmosphere. Soil respiration is also an impor- and CH4. The lifetime of O3 in the
44
Paulo Artaxo
atmosphere is on the order of days to weeks, observed in regions very far from the emisleading to a highly variable temporal and sions, such as tropical Pacific and southern spatial distribution. The chemical precursors Atlantic oceans. of O3 are hydrocarbons (including CH4 and nonmethane hydrocarbons-NMHCs), CO, The Role of Aerosols in Biogeochemical and the oxides of nitrogen (NOX). NOX play Cycles in Amazonia a particularly important role in the ozone budget, since their abundance determines Aerosol particles are important to the if the photochemical oxidation of hydrocar- Amazon ecosystem because of several bons and CO results in net O3 production or roles they play in the atmosphere and in destruction. During the biomass burning biogeochemical cycles. Aerosol particles are period in Amazonia, ozone concentrations responsible for the airborne transport of reach 60-90 ppb and may actually become phosphorus, calcium, sulfur, nitrogen comtoxic to vegetation. Concentrations of ozone pounds, and other essential nutrients. in Amazonia during the wet season are very Aerosols also act as cloud condensation low, in the range of 8-12 ppb at midday. nuclei affecting the cloud formation mechaNMHCs are an important emission nisms (Rogers et al. 1992). Aerosols affect category for tropical forests. Guenther et al. regional and global climate through their (1995) estimate that vegetation is the source radiative properties, reducing the amount of of over 90% of all NMHC in the global atmos- solar radiation that reaches the ecosystem. phere and Amazonia is the largest contribu- The forests emit large amounts of aerosol tor. The reason for this large natural particles to the atmosphere as part of their biogenic emission is still unknown. This natural metabolism. During the wet season, category includes isoprene, terpenes, and when no biomass burning occurs, aerosol hundreds of volatile organic compounds. concentrations for particles less than 10 pm There are large variations in emission rates are about 10-20 pg/m3. Aerosol particles in for different plant species and different land- forest regions are also formed as secondary scapes. Some recent studies suggest that a products from natural emissions of biogenic plant emits isoprene as part of the leaf tem- hydrocarbons, such as terpenes. A large perature regulation mechanism. Terpenes are fraction of the aerosol mass (70-85%) is in the know to form aerosol particles via efficient form of organic matter, transporting mechanisms and are thus responsible for a significant amounts of carbon in the form of significant part of the total organic aerosol many different organic compounds. mass. Several NMHCs have mass fractions in Vegetation has long been recognized as an the aerosol and gas phases (Hewitt 1998, important source of both primary and Andreae and Crutzen 1997). secondary aerosol particles. Forest vegetation Biomass burning is an important global is the principal global source of atmospheric source of hydrogenated species (e.g. CH4, organic particles, and in a tropical forest NMHCs, CH3C1, and CH3Br), as well as other natural vegetation plays a major role in trace gases such as CO (Crutzen and Andreae airborne particle concentrations. Only a few 1990). These species are produced predomi- studies of natural biogenic aerosols from nantly in smoldering conditions characterized vegetation in tropical rain forests have been by insufficient O2 supply. Methyl bromide undertaken (Artaxo et al. 1988, 1990, 1994, and methyl chloride are important in stratos- 1998, Echalar et al. 1998). Natural biogenic pheric ozone depletion. Ozone secondarily aerosols consist of many different types of produced by biomass burning emissions is particles, including pollen, spores, bacteria,
The Atmospheric Component of Biogeochemical Cycles in the Amazon Basin
algae, protozoa, fungi, fragments of leaves, excrement, and fragments of insects. This aerosol component is mainly in the coarsesize fraction (dp > 2 pm). The mechanisms of particle emission are still not well
45
understood, but probably include mechanical abrasion by wind, biological activity of microorganisms on plant surfaces and forest litter, and plant physiological processes such as transpiration and guttation. These process-
Table 3.1 Elemental composition of natural biogenic particles in the Amazon basin*. Cuiaba Wet Season Coarse Mode Fine Mode
Alta Floresta Wet Season Coarse Mode Fine Mode Mean
Standard
Mean
3.5
16.4
-
9.4 105
3.9
4.6 170 2.3 94 8.0 2.1 5.5 0.53 24
0.41 110 23 57 7.4 130 1.3 74 7.2 1.7 3.2 0.46 23
0.72 33 96 2.3 170 1.03 80 12.0 3.4 0.49 1.2 0.57
1.2
1.7
0.69 1.19
0.99 0.72 0.12 0.10 0.08 0.16 0.42 0.43 0.14
Variable
Mean
Standard
Mass** BC** Mg Al Si P S Cl K Ca Ti V Cr Mn Fe Ni Cu Zn Ga Ge As Se Br Rb Sr Zr Au Pb
5.5 0.66
103 27 74
0.10 0.12 0.19 0.12 1.37 0.53 0.21 0.81 0.18 0.52
.68
0.13 0.30
103 230 270 86 140 41 270 44 17 1.05 2.21 3.0 190 0.49 1.2 2.3 0.59 0.233 0.24 0.21 0.68 0.80 0.67 0.90 0.29 1.0
240 230 56 103 24 180 28 15 0.76 -
3.1 190
Mean
Standard
1.8
11.7
0.33
-
5.7 270 590 13 39 9.1 100 120 26 1.4 0.88 2.5 470 0.40 3.5 8.3 -
Standard
30 67 1.4 110 0.93 44 6.5
2.6 0.43
2.3
1.16
3.4 790
82
60
0.19 0.73
1.6
2.0
0.76 0.015 0.17 0.30 0.46 0.54 0.54 0.29 0.18 1.4
-
-
0.10 0.72 1.1 0.062 0.30 0.21 0.14 0.055 -
1.24
0.86
0.104 0.64 0.32 0.16 0.104
36 1.6
0.37
0.48 1.7
-
407 970 18 65 14.3 200 180
0.55 2.4 4.5 0.68 0.91 0.96 1.44 3.8
0.51 0.68 0.68 0.92 -
3.0
* Averages and standard deviations are presented for the atmospheric concentration (in ng m~3) of the participate mass, black carbon, and aerosol trace elements in samples collected in Alta Floresta from August 92 to March 95, and in Cuiaba from July 90 to August 95. ** Aerosol mass and black carbon (BC) are expressed in ug m~3.
46
es may generate particles containing biogenic-related elements such as Na, Mg, P, S, K, Ca, Zn, and Rb. Also, a significant fraction of aerosols is comprised of secondary aerosols, formed by gas-to-particle conversion of organic and sulfur-related natural biogenie gases. These biogenic particles are mostly submicrometer in size. Bacteria in forested areas were found in the size range of 0.5 to 2.5 pm. Biological activity of microorganisms on leaf surfaces and forest litter results in airborne particles. Windblown pollens certainly contribute to coarse particles in forested areas. The transpiration of plants can lead to migration of Ca, SO4, Cl, K, Mg and Na to the atmosphere. The biogenic-related elements (K, P, S, Zn, Rb, and others) are essential to superior plants. They are present in the fluids circulating in the plant and are released from the leaves to the atmosphere. Natural biogenic aerosol particles emitted by plants play an important role in nutrient cycling in tropical ecosystems. Tropical ecosystems maintain a delicate nutrient balance characterized by intense internal recycling and depend on atmospheric input of certain mineral nutrients to fulfill certain nutrient requirements. It has been shown that biogenic particles may influence cloud properties (Schnell 1982) and can act as cloud condensation nuclei, potentially affecting cloud formation mechanisms and cloud dynamics. These particles can travel for long distances (Talbot et al. 1990, Artaxo et al. 1998). Table 3.1 shows the average elemental composition of natural biogenic aerosol particles collected in two background sampling stations in Amazonia (Artaxo et al. 1998, Echalar et al. 1998). The first site (Alta Floresta) is a forest site, and the second (Cuiaba) is a cerrado vegetation site. The average aerosol mass and black carbon concentration is also presented in Table 3.1. It is possible to observe that most of the elements occur predominantly in the coarse
Paulo Artaxo
mode, with sulfur being the only exception. The aerosol mass for particles less than 10 pm is 15-20 pg/m3, a very significant aerosol mass. Phosphorus is strongly present in the coarse mode with more than 90% of P concentrations in particles larger than 2 pm. This means that the cycling of phosphorus is mostly local or regional. Figure 3.1 shows the time series of aerosol mass concentrations for Alta Floresta from 1991 to 1998. Alta Floresta is in the northern part of Mato Grosso state and is a place where heavy biomass burning occurs from July to October. The large increase in aerosol concentrations during the biomass burning period is remarkable. Background aerosol concentrations of about 10-20 pg/m3 of particles less than 10 pm in size during the wet season can increase to 400-500 pg/m3 during the dry season. Even if nutrients comprise only a small fraction of this aerosol fraction, it can represent a significant mass fraction of the elements that are lost from the ecosystem. The natural biogenic aerosol component is comprised of particles that can travel for thousands of kilometers and has an atmospheric residence time of 2-10 days. The composition of these natural biogenic aerosol particles can be observed in Table 3.1, for particles collected in Alta Floresta and Cuiaba and for the fine and coarse mode fractions. We observe that phosphorus, for example, shows concentrations of about 90 ng/m3 in Alta Floresta, with 80% of this concentration in the coarse mode fraction. The coarse fraction has mostly local or regional impact. In contrast, Table 3.2 shows the average elemental concentration for aerosols during the biomass burning period. Concentrations presented in Table 3-2 come from samples collected in the INPE Bandeirante aircraft during the SCAR-B experiment and also ground-based concentrations for Alta Floresta and Cuiaba. Sulfur can reach an average of 1000 ng/m3 during
The Atmospheric Component of Biogeochemical Cycles in the Amazon Basin
47
Alta Floresta Aerosol Mass Concentration
Fig. 3.1 Time series of PM10 aerosol mass concentrations in pg/m for Alta Floresta from 1992 to 1998. Aerosol mass concentration are for particles less than lOum. the dry season, compared with 300 ng/m3 et al. 1994, Echalar et al. 1995, Andreae et al. during the wet season. For the past 20 years 1998). Trace gas emissions in the burning of most of the tropical forest areas have been African savannah and the Brazilian cerrado under strong pressure by rapid change of are also similar (Scholes et al. 1996, Scholes land cover, with forest and adjacent savannas and Andreae 2000). The pyrogenic particles being cleared, most of the time through the are efficient cloud condensation nuclei use of fire, and converted to pasture and (CCN) (Rogers et al 1992). The abundance of agricultural fields at a substantial rate (Artaxo CCN in the dry season over Amazonia and Africa certainly has a strong effect on cloud et al. 1998). Pyrogenic and natural biogenic emissions formation and precipitation patterns, even to the atmosphere in the Amazon basin may for regions far from biomass burning, but this have an impact on the global tropospheric issue has not yet been properly addressed. The aerosols emitted during biomass chemistry, because this region exhibits intense convective activity that injects gases burning can travel long distances. Figure 3.2 and aerosols to high altitudes where they can shows a three-dimensional trajectory analysis be transported over long distances (Andreae for aerosol emissions during the dry season et al. 1996, 1998, Echalar et al. 1998). (Freitas et al. 1997). From each biomass Elemental concentrations in aerosols from burning spot, the air masses that transport Amazonia and Africa are very similar (Artaxo the aerosol particles and trace gases stay at
48
Paulo Artaxo
Table 3.2 Elemental composition of biomass burning aerosols in the Amazon basin measured during the SCAR-B experiment and comparison with long-term ground-based measurements at Alta Floresta and Cuiaba. SCAR-B Mean
Standard Deviation
SCAR-B Minimum
SCAR-B Maximum
Alta Floresta Dry Season
Cuiaba Dry Season
Na
95.4
59.9
30.0
659
385
_ -
2292 3126
1808 2104
196 350 713
270 2356 7297 8673
.
Mg AI Si P S Cl K Ca Ti V Cr Mn Fe Cu Zn Rb Sr Br Ga As Zr Mass* BC*
140
136 900
559
122 1130
Element
(ng nr3)
1198 222
1581 1251 121 3.67 5.14 67.7 1132 2.90 10.7 8.82 16.4 13.1 0.49 0.21 6.10 107 5.49
229 1076
1134
123 3.43 1.86 64.4
985 1.71
6.3 5-32 14.2 8.7 0.34 0.16 3.69 60.7 3.92
(ng nr3)
3.50 59.0 27.6 137.3 36.2 9.29 0.54 2.29 10.8 76.9 0.64 1.92 1.55 1.44 2.40 0.01 0.041 2.08 8.05 0.23
(ng nr3)
4333 1291 5581 5722
533 14.2 8.16 322 4327 8.77 27.7 25.9 59.5 36.9 1.46 0.85 15.7 297 17.5
(ng nr3)
1130 2540 2620
125
1590 270 162 7.1 41.6 18.8 1623
7.9 13.1
(ng m-3)
1342 2930 21.7 560 36 1020 489 130.8 6.37
4.6 16.5 1990 5.91 11.85
10 5.4
-
19.0 2.36
6.1 -
2.7 8.2 81.2
5.73
3.34
5.86 49-1
2.6
3
Aerosol mass and black carbon (BC) are expressed in pg m' .
low altitude (below 3000 meters) in the Amazon region. When an air mass reaches the Andes, it will rise to higher altitudes because of orographical effects and will reach the Pacific Ocean in only a few days. In the south of Brazil, when Amazon-derived air masses encounter cold fronts they can rise to very high altitudes of about 10 kilometers where atmospheric transport is fast and very efficient. High ozone and aerosols in the South Atlantic were observed during August-
September as a result of biomass burning in Amazonia and Africa (Andreae et al. 1996). Talbot et al. (1990) and Swap et al. (1995) have shown the chemistry changes associated with long-range transport of aerosols in the Amazon basin. Also, the Amazon region could receive a large influx of trace elements from Sahara dust and tropical Africa (Swap et al. 1995). The magnitude of this flux that can carry essential nutrients to the Amazon basin is still unknown.
The Atmospheric Component of Biogeochemical Cycles in the Amazon Basin
The Role of Wet Deposition in Amazonia The Amazon basin is characterized by high rainfall rates, and wet deposition could be significant in the biogeochemical cycles of several nutrients (Andreae et al. 1990). The biogeochemistry of trace elements in the atmosphere is influenced by different sources of gases and aerosols. During a rain event, these gases and aerosols are incorporated into raindrops, and a significant amount of these aerosols are cloud processed, and incorporated into cloud droplets from the very beginning. During the Amazonian wet season, rainwater chemistry is dominated by natural biogenic emissions, consisting mostly of
49
organic acids, potassium, phosphorus, and other components (Lesack and Melack 1996). In the wet season, average pH is about 5.0 (Williams et al. 1997, Andreae et al. 1990). During the biomass burning period, rainwater pH decreases to about 4.5, mostly because of enhanced concentrations of organic acids, ammonia, and nitrates. Williams et al. also measured DOC (Dissolved Organic Carbon) at average values of 138 and 238 pM for the dry and wet seasons, respectively. The annual volume weighted mean of 159 pM implies an annual deposition of about 0.4 moles C nr2. The acidity was basically determined by the concentrations of organic acids. The concentration of acetate was 9-3 peq H and formate was 2.9 peq H. The average conductivity was
Fig. 3.2 Three dimensional forward air mass trajectories starting from biomass burning spots on August 22, 1985, during the SCAR-B experiment. Shades of gray indicate elevation in meters.
50
Paulo Artaxo
Table 3.3 Rainwater composition at Lake Calado, West of Manaus. Volume Weighted Mean ((eqH) Solute
H pH NH4 Na K Ca Mg Cl SO4 NO3 P04 TDP (/tM)*
Annual
Dry Season
17.0
4.7 3.0 2.4 0.8 2.4 0.9
Wet Season
Annual
Dry Season
Wet Season
31.7
11.2
46.8
24.7
22.1
4.5 7.4 3-4
5.0 1.2 2.1 0.7 2.4 1.0 4.6
8.2 6.7 2.3 6.6 2.5
-
1.6 3.3
5.4
5.8 2.6 0.9 1.9 0.5 3.5 2.2 5.0 0.05
0.04
0.4
0.3
2.0 4.2
1.2 2.4 0.6 4.5 2.8 6.4
0.03 0.25
0.06 0.47
4.6
Deposition (meq nr2)
0.02 0.15
12.6 11.5 0.09
0.7
2.4 4.1 1.4 4.7 2.0 9.1 3.2
6.5
Source: Adapted from Williams et al. 1997. * TDP is the total dissolved phosphorus.
6.5 pScrrr1, with larger conductivity during the dry season. Williams et al. (1997) observed that solute deposition occurred in proportion to total rainfall, suggesting that there was a rather constant source of aerosols in the atmosphere throughout the year. Andreae et al. (1990), during two campaigns in the 1987 wet season and the 1985 dry season during the ABLE experiments, measured similar rainwater concentrations. Significant marine influences were observed in the rainwater collected near Manaus, indicating the long-range transport of marine aerosols within the Amazon basin. A large fraction of the sulfate content of wet season rain appears to be of marine origin. The emissions of biogenic sulfur and nitrogen species from the rain forest ecosystem and their subsequent oxidation in the atmosphere appear to make only a small contribution to the deposition flux of sulfate and nitrate in central Amazonia. Forti et al. (1991) analyzed samples of rainwater and throughfall during ABLE-2B for a site near Manaus (Reserva Ducke). The throughfall was enriched 40-90% compared with the rainwater for Mg, K, Na, Ca, Cl, and SO4. Absorption of NH4+ by the
vegetation was also observed during the dry period. These values are similar to rainwater composition in Africa (Freydier et al. 1998), where a pH value of 5.0 was observed. In the tropical forest of Costa Rica, Eklund et al. (1997) also identified organic acids as the major determinant of precipitation pH, but they observed a higher marine aerosol component than that in Central Amazonia. Measurement of total nutrient deposition requires studies integrating rainfall, throughfall, and stemflow in the wet component, as well as aerosol dry deposition. No studies in Amazonia have yet integrated all these components. Throughfall studies indicate a large enrichment of potassium, on the order of 2.7 to 10.2, leached from plant leaves (Holscher et al. 1998). This large variation in potassium enrichment could be caused by different floristic composition and precipitation patterns.
Conclusions The atmosphere in tropical forests is a very active part of biogeochemical cycles that
The Atmospheric Component of Biogeochemical Cycles in the Amazon Basin
include the soil, vegetation, and hydrological pathways. Trace gases, aerosol particles, and rainwater interact with the system in a variety of ways. Atmospheric transport extends local or regional processes to far larger scales, reaching even to global scales. Integrated studies involving wet and dry deposition fluxes for key nutrients in Amazonia and other tropical areas are needed. Also the issue of how CCN affects cloud formation mechanisms in the dry season must be more thoroughly investigated to assess the impact of anthropogenic influences in tropical forests. A flux of trace elements and nutrients from Africa to Amazonia and from Africa to the Indian Ocean and Australia has been well
51
documented qualitatively for many years. Efforts in quantifying these fluxes are very important for improving our understanding of how tropical ecosystems use and exchange key nutrients. Acknowledgements: I thank Andi Andreae for providing very useful unpublished manuscripts, and for very useful discussions over the last two years in the chapter's theme. Thanks are due to Alcides C. Ribeiro, Ana L. Loureiro, Tarsis Germano, and LAMFI staff for assistance during aerosol sampling and PIXE analysis. This work was financed through grant 97/11358-9 from FAPESP-Fundacao de Amparo a Pesquisa do Estado de Sao Paulo.
Literature Cited Andreae, M. O., R. W. Talbot, H. Berresheim, and K. M. Beecher. 1990. "Precipitation chemistry in Central Amazonia." Journal of Geophysical Research 95, D10: 16987-16999. Andreae, M. O., J. Fishman, and J. Lindesay. 1996. "The Southern Tropical Atlantic Region Experiment (STARE): TRansport and Atmospheric Chemistry near the EquatorAtlantic (TRACE-A) and Southern African Fire/ Atmosphere Research Initiative (SAFARI): An introduction." Journal of Geophysical Research 101: 23519-23520. Andreae, M. O., and P. J. Crutzen. 1997. "Atmospheric aerosols: biogeochemical sources and role in atmospheric chemistry." Science 276: 1052-1058. Andreae, M. O. 1998. "Feedbacks and interactions between Global change, atmospheric chemistry and the biosphere." Paper presented at the Workshop on GeosphereBiosphere Interactions and Climate. Pontifical Academy of Sciences, Vatican City, 9-13 November 1998. Andreae, M. O., T. W. Andreae, H. Annegarn, F. Beer, H. Cachier, W. Elbert, G. W. Harris, W. Maenhaut, I. Salma, R. Swap, F. G. Wienhold, and T. Zenker. 1998. "Airborne studies of aerosol emissions from savanna fires in southern Africa: 2. Aerosol chemical composition." Journal of Geophysical Research 103: 32119-32128. Artaxo, P., H. Storms, F. Bruynseels, R. Van Grieken, and W. Maenhaut. 1988. "Composition and sources of aerosols from the Amazon Basin." Journal of Geophysical Research 93: 1605-1615. Artaxo, P., W. Maenhaut, H. Storms, and R. Van Grieken. 1990. "Aerosol characteristics and sources for the Amazon Basin during the wet season." Journal of Geophysical Research 95: 16971-16985.
Artaxo, P., F. Gerab, M. A. Yamasoe, andj. V. Martins. 1994. "Fine Mode Aerosol Composition in Three Long Term Atmospheric Monitoring Sampling Stations in the Amazon Basin." Journal of Geophysical Research 99, Dll: 22857-22868. Artaxo, P., E. T. Fernandes, J. V. Martins, M. A. Yamasoe, P. V. Hobbs, W. Maenhaut, K. M. Longo, A. Castanho. 1998. "Large Scale Aerosol Source Apportionment in Amazonia." Journal of Geophysical Research 103, D24: 31837-31848. Brasseur, G., J. J. Orlando, and G. S. Tyndall. 1999. Atmospheric Chemistry and Global Change-Topics in Environmental Chemistry, Oxford University Press, New York. Crutzen, P., and M. O. Andreae. 1990. "Biomass burning in the tropics: Impact on atmospheric chemistry and biogeochemical cycles." Science 250: 1669-1678. Echalar, F., A. Gaudichet, H. Cachier, and P. Artaxo. 1995. "Aerosol emissions by tropical forest and savanna biomass burning: characteristic trace elements and fluxes." Geophysical Research Letters 22: 3039-3042. Echalar, F., P. Artaxo, F. Gerab, M. A. Yamasoe, J. V. Martins, K. M. Longo, W. Maenhaut, and B. N. Holben. 1998. "Aerosol composition and variability in the Amazon basin." Journal of Geophysical Research 103, D24: 31849-31866. Eklund, T. J., W. H. McDowell, and C. M. Pringle. 1997. "Seasonal variation of tropical precipitation chemistry: La Selva, Costa Rica." Atmospheric Environment 31, 23: 3903-3910. Freitas, S. R., K. M. Longo, M. A. F. Silva Dias, and P. Artaxo. 1997. "Numerical modeling of air mass trajectories from the
52
biomass burning areas of the Amazon basin." Annais da Academia Brasileira de Ciencias 68: 193-206. Forti, M. C., and L. M. Moreira-Nordemann. 1991. "Rainwater and through/all chemistry in a Terra Firme' rain forest: Central Amazonia," Journal of Geophysical Research 96: 7415-7421. Freydier, F., B. Dupret, andj. P. Lacaux. 1998. "Precipitation chemistry in intertropical Africa." Atmospheric Environment 32: 749-765. Gash, J. H. C., C. A. Nobre, J. M. Roberts, and R. L. Victoria. 1996. "Amazonian Deforestation and Climate, John Wiley & Sons, England. Guenther, A., C. N. Hewitt, D. Erickson, R. Fall, C. Geron, T. Graedel, P. Harley, L. Klinger, M. Lerdau, W. A. McKay, T. Pierce, B. Scholes, R. Steinbrecher, R. Tallamraju, J. Taylor, and P. A. Zimmerman. 1995. "Global model of natural volatile organic compound emissions." Journal of Geophysical Research 100: 8873-8892. Holscher, D., T. D. de A. Sa, F. F. Moller, M. Denich, and H. Folster. 1998. "Rainfall partitioning and related hydrochemical fluxes in a diverse and in a mono specific secondary vegetation stand in Eastern Amazonia." Oecologia 114: 251-257. Hewitt, C. N. 1998. Reactive Hydrocarbons in the Atmosphere, Academic Press, San Diego. Houghton, R. A., E. A. Davidson, and G. M. Woodwell. 1998. "Missing sinks, feedbacks, and understanding the role of terrestrial ecosystems in the global carbon balance." Global Biogeochemical Cycles 12: 25-34. Jacob, D. 1999. Introduction to Atmospheric Chemistry, Princeton University Press, New Jersey. Keller, M., D. J. Jacob, S. C. Wofsy, and R. C. Harriss. 1991. "Effects of tropical deforestation on global and regional atmospheric chemistry." Climatic Change 19: 139-158. Lesack, L. F. W., and J. M. Melack. 1996. "Mass balance of major solutes in a rainforest catchment in the Central
Paulo Artaxo Amazon: Implications for nutrient budgets in tropical rainforests." Biogeochemistry 32: 115-142. Matson, P. A., and R. C. Harriss. 1995. Biogenic Trace Gas Emissions, Blackwell, Oxford, p. 384. Rogers, C. E, J. G. Hudson, B. Zielinska, R. J. Tanner, J. Hallett, and J. G. Watson. 1992. "Cloud condensation nuclei from biomass burning." In: Global Biomass Burning: atmospheric, climatic, and biospheric implications, ed. J. Levine (MIT press, Cambridge, MA). Sanhueza, E., L. Cardenas, L. Donoso, and M. Santana. 1994. "Effect of plowing on CO2, CO, CH4, N2O, and NO fluxes from tropical savannah soils." Journal of Geophysical Research 99: 16429-16434. Schnell, R. C. 1982. "Kenyan lea litter: A source of ice nuclei." Tellus 34: 92-95. Scholes, M., and M. O. Andreae. 2000. "Biogenic and pyrogenic emissions from Africa and their impact on the global atmosphere." Ambio 29:23-29. Scholes, R. J., D. Ward, and C. O. Justice. 1996. "Emissions of trace gases and aerosol particles due to vegetation burning in southern-hemisphere Africa," Journal of Geophysical Research 101: 23677-23682. Swap, R., M. Garstang, S. Macko, P. Tyson, W. Maenhaut, P. Artaxo, P. Kallberg, and R. Talbot. 1996. "The long-range transport of southern African aerosols to the tropical South Atlantic." Journal of Geophysical Research 101: 23777-23792. Talbot, R. W., M. O. Andreae, H. Berresheim, P. Artaxo, M. Garstang, R. C. Harriss, K. M. Beecher, and S. M. Li. 1990. "Aerosol chemistry during the wet season in central Amazonia: The influence of long-range transport." Journal of Geophysical Research 95: 16955-16970. Warneck, P., and R. Zellner. 1999. Global Aspects of Atmospheric Chemistry, Dietrich Steinkopf, Germany. Williams, M. R., T. Fisher, and J. M. Melack. 1997. "Chemical composition and deposition of rain in the central Amazon, Brazil." Atmospheric Environment 31, 2: 207-217.
4 Soil versus Biological Controls on Nutrient Cycling in Terra Firme Forests Elvira Cuevas
Terra firme forests are those that by definition are not permanently or seasonally flooded (terra firme meaning "firm terrain"). This type of forest encompasses the Amazon and Orinoco basins, stretching from the lower slopes of the Andes, east to the Guianas, and south to about 15°S in western Brazil and northern Bolivia (Richards 1996). Structural and compositional variability in these forests in the Amazon basin is very wide as a result of climate differences and geomorphological position. The region is not climatically uniform; the central and much of the southern parts have less and more seasonal rainfall than the eastern and western parts (Walsh 1996). These differences have direct and indirect ecological significance, as phenology and biological processes related to nutrient availability will be strongly influenced by both factors (Cuevas and Medina 1986, 1988, 1990, Medina and Cuevas 1989). Periods of two or more consecutive dry days are ecologically significant in a humid area such as San Carlos de Rio Negro, in the northern part of the Amazon, because of low water retention capacity in the widespread sandy soils. In lower geomorphological positions, dry spells of 5-10 days may result in fluctuations of the water table from 0.4-1.0 m (Herrera 1977, Bongers et al. 1985). In areas with a more strongly seasonal climate, roots have been found extending to 18 m depth (Nepstad et al. 1995). This may explain the
presence of evergreen forest in the seasonally dry eastern Amazon. Structure and physiognomy of terra firme forests is very similar throughout Amazonia, but floristically it is quite variable due to different compositions in the subbasins of the Amazon's major tributaries. These subbasins are located within geochemical regions that can be differentiated based on the physicochemical properties of drainage waters (Sioli 1975, Fittkau 1971, Fittkau et al. 1975). Blackwater rivers, such as the Rio Negro, drain mostly sandy podsolized soils low in most essential nutrients for plant growth. They are characterized by a high content of humic acids, which remain dissolved because of the predominant low concentrations of polyvalent cations, mainly Ca2+ and Mg2+. Whitewater rivers, on the other hand, have a higher sediment content and are relatively rich in cations because they drain geologically younger Andean areas. Clearwater rivers, which are common in the central Amazon region, also drain nutrient-poor soils. On a small-scale basis, the composition of mixed forests is related to environmental factors such as topography, drainage, and soil characteristics (Medina and Cuevas 1989, Steege 1993, Duivenvoorden and Lips 1995, Brouwer 1996, Richards 1996). The history of erosion and redeposition in the Amazon basin has led to an extensive leaching of nutrients, particularly the common
54
cations K+, Ca2+, and Mg2+ and inorganic phosphate reserves (Fittkau et al. 1975, Furch and Klinge 1989). This impoverishment is particularly intensive in the central and northern areas of the basin characterized by annual rainfall around or above 2000 mm (Salati et al. 1978, Matsui et al. 1982). Terra firme forests growing on Oxisols/ Ultisols seem to be primarily limited by phosphorus (Vitousek 1984, Vitousek and Sanford 1986, Cuevas and Medina 1986, Cuevas and Medina 1988), while at the same time being tolerant of high soil concentrations of mobile aluminum (Sprick 1979, Sobrado and Medina 1980). Cuevas and Medina (1988) showed that terra firme forests can also be limited by calcium and magnesium, which corroborated the hypothesis of Furch and Klinge (1989) who measured low Ca and Mg contents in biotic and abiotic compartments in terra firme forests of Amazonia. Plant adaptation to these conditions requires highly efficient uptake and/or utilization of nutrients, especially phosphorus, calcium and magnesium (Marschner 1995). Biomass development (Jordan 1985) and regeneration capacity (Uhl 1987) are not always clearly associated with differences in soil fertility, but they are certainly related to rates and patterns of nutrient cycling (Vitousek and Sanford 1986, Medina and Cuevas 1989, Tiessen et al. 1994b). On poor soils, nutrients may cycle without substantial losses from the system (Baillie 1989, Burnham 1989). In such dystrophic systems, organic matter and particularly the forest litter mat may play an essential role in conserving nutrients for sustaining forest production (Stark and Jordan 1978). Nutrient recycling and supply is tightly regulated by largely biotic processes, such as retranslocation or resorption prior to leaf abscission (Medina and Cuevas 1989, Cuevas and Medina 1990). Nutrients are also reutilized from decomposing residues through rapid mineralization and uptake by a dense
Elvira Cuevas
fine-root mat, or ectorganic horizon, associated with the litter layer (Cuevas and Medina 1986, Cuevas and Medina 1988). In addition, nutrient cycling and turnover seem to be tied to organic matter dynamics (Tiessen et al. 1994a), illustrating again the importance of biological processes as the main regulators of nutrient availability. Understanding the changes to soil biogeochemical cycles that influence soil fertility after clearing is a key condition for predicting ecological consequences of deforestation and for implementing effective management schemes (see Dias Filho et al., this volume). A central question to be addressed then is "what role does organic matter dynamics play in biogeochemical cycling in terra firme forests?"
Soils under Terra Firme Forests The Amazon region is dominated by old, highly weathered, leached soils due to large areas of tectonically and geomorphologically stable land surfaces (Baillie 1996). The main soils developed under these conditions, which account for around 70% of the Amazon basin (Richter and Babbar 199D, can be grouped under modal kaolisols (Baillie 1996), which include the ferralitic soils or ferralsols (FAO-UNESCO 1974), or Oxisols and Ultisols (Soil Survey Staff 1975). The classification is based on the importance of iron and aluminum in the elemental composition or of sesquioxides and kanditic aluminosilicates in the clay minerals (Bleeker 1983). Because of the advanced weathering state of these soils, quartz is the only primary mineral remaining which forms a significant sand fraction and gives a coarse texture to the soils. All other primary minerals have been weathered to low-activity, secondary clay minerals with few permanent negative charges (Buol 1985). Labile aluminum occurs either in the soil solution or adsorbed onto cation exchange complexes. These soils are acidic (pH 3-5.5), with a considerable
Soil versus Biological Controls on Nutrient Cycling in Terra Firme Forests
depletion of basic cations and very low phosphorus availability due to the high P fixing capability of the clays and co-precipitation with labile forms of aluminum and iron. Because of the weathering conditions and the predominance of low-activity clays, these soils have very low cation availability ( K+, Ca2+, and Mg2+). Clay kaolisols can also be present in some areas. pH, exchangeable base status, and available phosphate contents are still low by global standards but are slightly higher than in modal kaolisols. The ranges, however, overlap considerably and aluminum is still the dominant exchangeable cation in the soil. Total contents of potassium, magnesium, and to a lesser extent, phosphorus, are considerably higher, mostly held in non-exchangeable forms by the nonkandite aluminosilicates (Baillie 1996). The remaining, nonkaolisol, soils underlying terra firme forests are mostly podzols and white sands. These developed from very sandy and quartzose parent materials on gently sloping tops of old alluvial terraces, and on plateaux and gentle dip slopes. These soils have limited available moisture, drainage, and aeration. They also have limited supplies of all nutrients, especially nitrogen, and can have strong to moderate acidity. For a detailed description of these soils and the others mentioned above please refer to Baillie (1996). The differences in the soils presented above are related mainly to variations in their parent materials and thus could reflect the geological structure of the region. A good example is the study by Duivenvoorden and Lips (1995) in the middle Caqueta region of the Colombian Amazon. These authors distinguished between two groups of welldrained upland soils: soils of the Ali-Acrisol group occurring on clayey sediments of Andean provenance (clay kaolisols) and soils of the Acri-Ferrarsol group occurring on Palaeozoic sandstone plateaus and on deposits derived from the Precambrian Shield
55
(modal kaolisols). The parent material of the first group is less weathered and still contains appreciable amounts of clay minerals other than kaolinite, such as vermiculite (18%) and illite (10%). This is evidenced by high total Mg, K, and Na concentrations, low exchangeable bases, and higher exchangeable acidity. The second group, on the other hand, derives from highly weathered parent material with low levels of nutrient reserves, low exchangeable bases and exchangeable acidity, and an absolute dominance of kaolinite (over 95%) in the clay fraction. The conditions expressed above, which impose substrates with low to very low nutrient availability, especially the lithogenic ones, will determine nutrient limitations to plant growth. Vegetation modifies these constraints through its capacity to accumulate and concentrate carbon and nutrients.
Species Richness and Composition in Terra Firme Forests Species richness and composition can be related to soil properties (Richards 1996). Steege (1993) found that species distribution patterns and population structure in Guyana were related to soil type; however, soil drainage played an important role in species occurrence. Eperua rubiginosa, which can be found in higher numbers growing on soils with impeded drainage conditions, had a high presence on poorly drained sandy Ferralsols. The study by Duivenvoorden and Lips (1995) in the Colombian Amazon showed that patterns of tree species composition in terra firme forests depended significantly on soil properties even though the edaphic component explained only a small fraction of the variance. There were, however, no significant differences with respect to tree species density (Table 4.1). On the other hand, when density of tree species was analyzed in proportion to stem density, there was a higher species density in the relatively
56
Elvira Cuevas
richer soils. This corroborates what Baillie et al. (1987) showed for a mixed Diptercarp forest wherein properties determined by lithology of the parent material were more important for species distribution than properties such as organic matter and exchangeable cations. In fact, organic matter and exchangeable nutrient patterns are likely the products of species distributions through differing rates and amounts of nutrient return to the forest floor via litterfall and root residues. These plant inputs control the formation of soil organic matter and the magnitude of variation of the superficial soil contents of nutrients (Lugo et al. 1990, Cuevas and Lugo 1998, Tiessen et al. 1994b).
Soil Nutrients and Biomass Aboveground biomass of terra firme forests can differ by as much as 250 Mg ha"1 (Table 4.2). Medina and Cuevas (1996) proposed that differences in the range of aboveground biomass calculated by Brown and Lugo (1992) could result from the variability of landscapes, but also that they could be a result of different stages of succession after partial exploitation in the past. When taking into consideration the broad spectrum of soils and climate variation present in the locations in Table 4.2 and in the sites conidered by Brown and Lugo (1992), it is
likely that biomass differences could be explained by differences in water and nutrient availability, although successional stage can not be ruled out. The differences in amounts and distribution of total biomass reinforce the hypothesis of nutrient availability, especially P, being the overriding determinant for biomass development. Forests with similar amounts of aboveground biomass can differ up to 100 Mg ha'1 in total biomass when the belowground component is also measured (Table 4.2). The biomass contribution of roots can vary between 9 to 20% in those forests. Root/shoot ratios can range between 6 and 33, suggesting that belowground biomass may be allocated to increase nutrient and/or water uptake. Regulation of the root/shoot ratio according to growing conditions is well known (Russell 1977), as there is a functional balance between the two. Growth and maintenance of roots can exert a strong drain of carbon from the aboveground part (Marschner 1995), thus along a water and/or nutrient availability gradient, aboveground biomass will be limited towards the extreme of availability, resulting in higher carbon allocation belowground. This certainly seems to be the case presented in Table 4.2 as the terra firme forests from Venezuela tend towards the extreme of limited soil nutrient availability (Medina and Cuevas 1989).
Table 4.1 Tree species density and cumulative amount of tree species (DBH > 10 cm) in 0.1 ha nlots in well-drained uplands from the middle Caaueta areas. Soil type Ali-Acrisol Acri-Ferralsol Tree species/0.1 ha-1 Tree species/stem density Cumulative No. Tree species/19 plots of 0.1 ha Cumulative No. stems/ 19 plots of 0.1 ha Source: Duivenvoorden and Lips 1995 * n = 21 plots in the Ali-Acrisol soil, 19 in the Acri-Ferralsol soil. " Denotes significant differences between soil types.
39 ± 8.3 0.58 ± 0.09** 431 ± 9** 1307 ± 18**
37.9 ± 9.2 0.50 ± 0.10 379 1430
Soil versus Biological Controls on Nutrient Cycling in Terra Firme Forests
57
Table 4.2 Belowground and aboveground biomass, percent contribution of belowground and root/shoot ratios in terra firme forests. Location
Belowground Mg ha'*
Aboveground Mg ha"1
Total Mg ha"1
% Roots
66
407
473 541
14
6
Klinge 1976
108
20
25
Russell 1983
542
20
25 19 24
Poels 1987 Uhl and Jordan 1984
Brazil Brazil Surinam Venezuela Venezuela
108
433 434
49 57
261
310
16
234
292
20
Venezuela
61 61
185
246
25
335
396
15
423 30
465 28
9 28
Venezuela Venezuela CV (%)
42 37
Here, not only is total biomass less than in other forests, but root/shoot ratios are also considerably higher. In contrast to shoot growth, root growth is much less inhibited under P deficiency (Marschner 1995). This can be construed as a mechanism to maximize soil P uptake, especially when the element is strongly limiting for growth or when ratios among specific soil resources are unusually skewed (Lajtha and Harrison 1995). In the toposequence of San Carlos de Rio Negro, Tiessen et al. (1994b) indeed found that distributions of carbon, nitrogen, and phosphorus of total plant biomass (above- plus belowground, Medina and Cuevas 1989) follow those of soil C, N,
Root/Shoot
Source
Saldarriaga 1985 Sanford 1989 Sanford 1989 Buschbacher 1984
33 18 10 44
and available P (Table 4.3). The terra firme forests at the top and midslope had considerably lower biomass P than the Amazon Caatinga forests in the lower slopes, which related to the very low P availability of these soils (shown both in absolute amounts and in the extreme skewness of the N/P ratios of the soil). Both biomass and soil N were higher in these two sites. Available P from leaf litter and the ectorganic horizon in the terra firme forests and in leaf litter and humic horizon in the Tall Amazon Caatinga accounted for 71, 85, and 72%, respectively, of the total available P. This illustrates that available P may be closely related to the amount of organic residues present, and to
Table 4.3 Comparison of nitrogen and available phosphorus in the soil and total biomass (aboveplus belowground) for three slope positions in a toposequence in San Carlos de Rio Negro. Forest Type
Soil Type
Slope Position
SoilN (gm-2)
Biomass N (gm-2)
Soil Resin P Biomass P (gm-2) (g m-2)
Terra firme Terra firme Tall Amazon Caatinga
Plinthic hapludox Plinthic hapludox Tropaquod
Top Middle Lower
869 1333 414
1485 1817 1145
2.6 4.0 5.5
Sources: Soil data from Tiessen et al. 1994b; biomass data from Medina and Cuevas 1989.
48 55 101
58
their turnover. In situations such as this, where organic matter mineralization is the most important process providing available P, the involvement of the mineral soil is minimized. Tiessen et al. (1994a), working in the same terra firme forest, estimated a soil organic matter budget down to 60 cm depth which showed the following distribution of carbon, nitrogen, and phosphorus: 60% of C, 65% of N, and 50% of P occurred in particulate organic matter with a mean residence time of less than 4 years; 27% of C, 29% of N, and 33% of organic P were mineral-associated and had a mean residence time near 50 years. The remainder of the organic matter and associated nutrients were sequestered in lateritic nodules (Fig. 4.1). Thus, ecosystem stability is dependent on the maintenance and cycling of the faster pool of organic matter in the system.
Elvira Cuevas
Productivity and Nutrient Fluxes in Terra Forests Ecosystem productivity is an index integrating the cumulative effects of many processes and interactions which proceed simultaneously in the system. Whether or not a relation exists between soil fertility and forest productivity is still under discussion (Proctor 1992, Silver 1994, Thompson et al. 1992). There is, however, no doubt that soil fertility influences input/output budgets, and that soil fertility influences the rate of recovery of the forest ecosystem after disturbance (Bruijnzeel 1992). Measurement of litterfall and its associated fluxes of nutrients is a practical way to evaluate production capacity and availability of nutrients in tropical forests. Mass and nutrient fluxes of litterfall in terra firme forests can vary considerably according to specific loca-
Fig. 4.1 Organic matter stocks in soil and litter layer under a terra firme forest in San Carlos de Rio Negro, Venezuela (data from Tiessen et al. 1994a).
Soil versus Biological Controls on Nutrient Cycling in Terra Firme Forests
59
Table 4.4 Mass and nutrient fluxes in leaf litter from terra firme forests growing on Ultisols/Oxisols. Site
Mass (Mg ha-1 yr1)
N
P
K (kg ha-1 yr--1)
Ca
Mg
Source
5.4 6.1
73.2 97.6
1.1 1.0
22.7 13.4
8.9 12.2
12.4 7.3
Brower 1996 Duivenvoorden and Lips 1995
6.8
87.7
0.8
19.7
10.9
8.8
5-4
83-7
0.8
7.8
5.4
4.3
7.6
121.1
2.1
15.4
13.0
5.2
7.0
85
1.8
9
12
7
Guama, Para, Brazil Tucurui, Para, Brazil Manaus, Brazil
7.3 4.8 5.6
122.6 95.5 84.0
3.0 2.4 1.7
12.4 17.8 11.2
22.6 35.5 11.2
20.4 11.0 11.2
Central Amazonia, Brazil Maraca, Roraima, Brazil
5.4 6.3
97.2 81.9
1.1 3.7
8.1 29.6
20.5 46.6
9.7 17.
Average sd CV (%) Spodosols/Entisols*
6.2
93.6 15.8 17
1.8 1.0
15.2 6.8
53
45
18.1 12.6 70
10.4
0.9 15
Average sd CV (%)
4.8 0.6 10
49.8 13.4 30
1.8
13.1 8.2 60
24.8
9.6 2.3 20
Mabura Hill, Guyana Aracuara, Colombia
San Carlos de Rio Negro, Venezuela
0.9 50
15.9 60
Cuevas and Medina 1986 Medina and Cuevas 1989 Klinge 1977 Silva 1984 Klinge and Rodrigues 1968a,b Luizao 1989 Scott et al. 1992
4.9 47
* Data from Brower 1993, Luizao 1995, Duivenvoorden and Lips 1995, Cuevas and Medina 1986, Herrera 1979.
tions and sites. The range of mass values presented in Table 4.4, 4.8 - 7.6 Mg ha"1 yr1, are very similar to that reported by Vitousek and Sanford (1986) for forests growing on infertile Oxisols/Ultisols worldwide. No relationship has been found between annual rainfall and annual leaf litterfall, which indicates that other factors, such as nutrient availability, play an important role in the productivity of these forests. As with mass, nutrient inputs into the system from leaf litterfall vary considerably; however, the degree of variation is higher for lithogenic elements such as P, K, Ca, and Mg than for N. The coefficient of variation of total and N mass are similar (15 and 17%, respectively) indicating the biogenic input of this element and the relatively small variation among sites.
Potassium and Mg also have a similar degree of variation, around 45%, whereas P and especially Ca are the elements with the higher degree of variation in these forests (53 and 67%, respectively). The Maraca site in Brazil is cycling 7 times more Ca and 4 times more P via leaf fall than the poorest site in Colombia. Conversely, N is 1.3 times higher in the poorest Colombian site than in the Maraca site. In other words, nutrient fluxes other than N seem to be reflecting the variable lithology and degree of weathering of the parent material, while N varies according to the organic matter dynamics of the system. When tierra firme forests growing on Ultisols/Oxisols are compared with forests growing on Spodosols/Entisols, only one specific difference is noted (Table 4.4). The
60
range of values for leaf fall mass and nutrient fluxes overlap between the two groups, with the exception of nitrogen, which is considerably lower in Spodosols/Entisols. Nitrogen fluxes to the forest floor in terra firme forests are twice the amount in forests growing on white sands. Average P fluxes are similar despite the fact that the organic matter flux is higher in terra firme forests growing on Ultisol/Oxisols. Using an index of nutrient use efficiency (mass of litter produced/mass of nutrient in leaf litter, Vitousek 1984), there are contrasting patterns for the two groups according to soil type and nutrient availability (Table 4.5). Nitrogen use efficiency is higher in Spodosols/Entisols (96 v 64, respectively) while P use efficiency is higher in Ultisols/ Oxisols (4096 v 2067, respectively). This reflects the relative impoverishment in available P and N enrichment in the soil and vegetation of terra firme forests (Cuevas and
Fig. 4.2 Distribution of dry mass/P ratios and P/N ratios from terra firme forests growing on Ultisols/Oxisols (references in Table 4.4). MFS=Medium Fertility Soil
Elvira Cuevas
Medina 1986, 1988, 1990, Medina and Cuevas 1989, Tiessen et al. 1994a, Duivervoorden and Lips 1995). Phosphorus/nitrogen ratios of leaf litterfall are inversely related to P use efficiency as measured by the organic matter/P ratio (Fig. 4.2). There is no relationship between P/N (or N/P) ratios and the efficiency of N use in these forests. This indicates that P/N ratios are mostly regulated by the availability of P and not of N, as the latter is not limiting for these types of forests (Cuevas and Medina 1988, Medina and Cuevas 1989). The same general relationship was found by Medina and Cuevas (1994) when taking into consideration a wider set of data encompassing humid tropical forests growing on soils of contrasting fertility. The forests from Colombia, Guyana, and Venezuela have high P use efficiency and low P/N indicating strong P limitation; the Brazilian forests range from low to medium P use efficiency suggesting a higher variability in P availability in these systems. The Maraca forest, on the other hand, reflects the effect of seasonality in enhancing nutrient availability. Complex nonsclerphyllous forests can develop in apparent nutrientdeficient soils (Thompson et al. 1992), where fine litter production is among the highest reported for tropical forests (8.85-9-52 Mg ha-iyr1) (Scott et al. 1992). Phosphorus use efficiency and P/N ratio of leaf litterfall are in the medium range, indicating higher P availability in the soil. The lack of correspondence between leaf chemistry and soil nutrients may be attributed to the strong seasonality of rainfall, which causes pulses in decomposition of soil organic matter and subsequent nutrient release (Chapin and Bieleski 1982). Fungal and other microbial biomass fluctuate rapidly in response to wetting and drying cycles, even in nonseasonal tropical forests, leading to net nutrient immobilization followed by pulses of mineralization (Lodge 1993). Thus, the rate of nutrient circulation in
61
Soil versus Biological Controls on Nutrient Cycling in Terra Firme Forests
Table 4.5 Ratio of mass and nutrients circulated in litterfall as an index of nutrient use effciency (Vitousek 1984). Site Mabura Hill, Guyana Aracuara, Colombia
Mass/Nitrogen
Mass/Phosphorus
74
4909 6100 8500 6750 3619 3889 2433 2000 3294 4909 1703 4373 48
62
78 64
San Carlos de Rio Negro, Venezuela Guama, Para, Brazil Tucurui, Para, Brazil Manaus, Brazil Central Amazonia, Brazil Maraca, Roraima, Brazil
63 82 60 50 67 56 77
—Average —CV (%)
67 15
Sources: Sources of data are presented in Table 4.4.
the forest floor, rather than the absolute amounts, explains the relatively high production rates, and high fluxes of P in litterfall.
Fine Roots and Nutrient Cycling All terra firme forests develop an ectorganic horizon composed mostly of a root mat, of variable thickness, on top of mineral soil. The thickness of the root mat, made up of apogeotropically growing fine roots (Sanford 1987), can vary from a few centimeters to 80 cm depending on site characteristics and nearness to tree trunks and can make up an average of 30% of total fine-root biomass in these systems (Stark and Spratt 1977). The proportion of fine roots in the ectorganic horizon can vary from 12 to 54% of total fineroot biomass depending on the sites studied (Table 4.6). The absence of a superficial root mat at the Roraima site can be accounted for by increased nutrient availability due to seasonal pulses in the system. Sanford (1989) found that terra firme forests in San Carlos de Rio Negro have 95% of total root biomass in the upper 20 cm of mineral soil and in the root mat, of which 48% is in the surface root
mat alone. Ageotropism, or lack of growth toward the soil, can result from the very low Ca levels in the soil system, as Ca is responsible for the gravitropic response in plant roots (Bennet et al. 1990). Production of roots on top of the mineral soil has been explained as a consequence of the low nutrient availability in Amazon forests (Herrera et al. 1978, Cuevas and Medina 1983, Medina and Cuevas 1989). Vertical root distribution results from differential nutrient availability in the soil profile (Berish 1982, Berish and Ewel 1988). Shallow rooted systems may be a result of litter and soil organic matter production and decomposition rates in systems where nutrient input from litter exceeds that of nutrient release by soil weathering, as is the case of Ca, Mg, and P in terra firme forests (Medina and Cuevas 1989). In the Middle Caqueta region of Colombia, for example, Ca and Mg concentrations in the L and F layers are between 15 and 20 times higher than in the mineral soil (Duivenvoorden and Lips 1995). Soils under terra firme forests can vary in the amount and availability of Ca and Mg as previously indicated. Cuevas and Medina
62
Elvira Cuevas
Table 4.6 Total mass ± s.d. (live + dead) of fine roots (< 5 mm diameter) in ectorganic horizons and mineral soils at 0-40 cm depth in terra firme forests. Site Middle Caqueta area, Colombia!* Soils Ali-Acrisol group Soil Acri-Ferralsol group San Carlos de Rio Negro, Venezuela2** Dark red concretionary ferralsol Red-yellow concretionary ferralsol Manaus, Brazil3*** Roraima, Brazil4**** * ** *** *"*
Ectorganic horizon Mineral Soil (g/m2) (g/m2)
Total (g/m2)
Proportion in Ectorganic (%)
200 ±130 600 ± 380
1500 ± 450 1900 ± 730
1700 ± 530 2500 ± 730
12 ± 7 23 ± 14
2940 ± 2650
3070 ± 1580
6010
49
5750 ± 4680
4950 ± 2920
10700
54
500 0
900 700
1400 700
36 0
Duivenvoorden and Lips 1995, Sanford 1989, Klinge 1976, Thompson et al. 1992
(1986) proposed that terra firme forests were primarily limited by Ca, Mg, K, and P. In the same system, fine-root production in the root mat was significantly increased in growth cylinders with added Ca or P (Cuevas and Medina 1988). The role of Ca and Mg on fine root growth was confirmed by measuring root growth inside of decomposition bags, where root biomass was weakly and negatively correlated with N but positively and strongly correlated with Ca and Mg (Cuevas and Medina 1988). Duivenvoorden and Lips (1995) found that the amount of fine roots in the very superficial F horizon in the ectorganic layer was significantly higher in Ali-acrisols and Acri-ferralsols when compared to forests in relatively richer soils, such as those on well-drained floodplains or in white sand soils of the same area. There were no significant differences in N and P concentrations among the different areas; however, Ca and Mg had significantly lower concentrations in the acrisols and ferralsols. Fine roots in the F horizon of these
two soils are associated with proportional decreases of Ca and Mg concentrations from the L to F horizon. These results correspond to the those from litterbag studies in San Carlos de Rio Negro, Venezuela, where attachment of fine roots had a positive effect upon Ca and Mg mobility in surface root mats (Cuevas and Medina 1988). Magnesium deficiency can severely impair root growth (Marschner 1995), impacting not only the acquisition of Mg but also other mineral nutrients and of water. This may also reduce drought resistance and adaptation to nutrient poor sites. Calcium is essential for the strengthening of cell walls and plant tissues. It is also essential for root extension, as root growth is dependent upon exogenous Ca supply (Marschner 1995). An exploring root system with a large surface area (which may include the mycorrhizal component) is important in water- or P limited situations, whereas such properties are less important for mobile ions, such as nitrate. These conditions suggest that the nutrient saving mecha-
Soil versus Biological Controls on Nutrient Cycling in Terra Firme Forests
nism of fine-root accumulation in very superficial soil horizons is primarily directed at conservation of Ca and Mg. In areas with a more strongly seasonal climate, deeper root distribution up to 18 m can be found (Nepstad et al. 1995). This deeper distribution is not only important in accessing water but also nutrients resulting from weathering, such as Mg (Brouwer 1996).
Role of Symbiotic Interactions in Nutrient Cycling The presence and amount of active roots have a direct effect on microbial activities and nutrient dynamics via rhizosphere interactions (Newman 1985). Areas rich in organic matter have a higher proportion of labile nutrients which can be absorbed by fine roots either directly or via mycorrhizal symbiosis (Janos 1983, Hogberg 1986, Alexander 1989, St. John 1983, Lodge 1993, among others). The degree to which this mycorrhizal symbiosis is relevant to the cycling of nutrients within the ecosystem depends on the inherent fertility of the underlying mineral soil. Salcedo et al. (1991), working in an Atlantic coastal forest in Recife, Brazil, showed that phosphorus from the litter/fermentation layer is cycled back to the vegetation via mycorrhizae-mediated mechanisms. However, 6l% of the added 32P moved down to the mineral soil, where P in the soil solution is controlled by microbial biomass activity. In contrast, Stark and Jordan (1978), working on a P deficient upland terra firme forest in San Carlos de Rio Negro, Venezuela (Cuevas and Medina 1988), found that nearly 100% of the added $2P was retained in the root mat associated with the litter layer, with less than 0.1% moving down to the surface of the mineral soil. A superficial and dense root distribution is an important adaptation of plants on soils that show high P fixation, when water is
63
not a limiting resource. A root/litter mat allows direct uptake of mineralized P by the vegetation, thereby shunting nutrient cycling via mineral soil-microbial biomass mediated processes (Tiessen et al. 1994b). In terra firme forests with higher nutrient availability, the microbial biomass component in the soil plays a more important role in nutrient cycling (Salcedo et al. 1991, Luizao et al. 1992, Feigl et al. 1995). The presence of a root mat or ectorganic layer (Stark and Spratt 1977, Duivervoorden and Lips 1995, Luizao 1995) of varying thickness increases the surface area for absorbing nutrients mineralized from decomposing litter, the principal source of nutrients in these systems (Medina and Cuevas 1989, Tiessen et al 1994a). Haynes (personal communication) measured values of 20 and 2 me/100 g dry weight for the cation exchange capacity and anion exchange capacity of roots in a Terra firme forest in San Carlos de Rio Negro. Retention of cations and anions from decomposing litter could be a consequence of the ion exchange capacity of the root mat. This mechanism is similar to those operating in other forest ecosystems, even on sites richer in nutrients. However, the process of nutrient capture may be especially effective in terra firme forests as a result of very high root densities in the upper soil layers, or above the soil, and the occurrence of heavy mycorrhizal infection throughout (Went and Stark 1968, Janos 1983, St. John 1980, St. John and Uhl 1983, Caceres 1989, Moyersoen 1993). In addition, vesiculararbuscular mycorrhizal fungi (VAM) stimulate phosphatase production in the extended rhizosphere, thus enhancing P availability via organic P mineralization (Jayachandran et al. 1992). Medina and Cuevas (1994) analyzed available data for potential mineralizaton of nitrogen and nitrification rates from tropical forests. Nitrogen mineralization rates in
64
Elvira Cuevas
estimated the degree of change in topsoil C due to slash-burn cultivation on a red-yellow concretionary sandy ferralsol in San Carlos de Rio Negro, Venezuela (Table 4.7). After three years, which is the average period of crop production before site abandonment, 81% of the C in the organic layer was lost. This quantity represented only 29% of the C in the whole soil down to 15 cm depth. Under the strong leaching regime of San Carlos, mineralization of organic N provides a steady supply for the vegetation via the rapid turnover of litter and organic matter (Cuevas and Medina 1986, 1988, Tiessen et al. 1994a). Tiessen et al. (1994b) characterized the P status and P transformations in both coarse and fine soil fractions in the same toposequence under mature forests and showed that the litter and root mat layers above the tierra firme soils maintained higher total P levels than the underlying mineral soil. In the organic mats, 50-60% of the P is in inorganic forms, which maintain high levels of available (resin extractable) P. As a result, the litter mats account for over 70% of the total available P in the top 72 cm of soil. Of the total P, about 8% is resin extractable, reflecting in part the high contribution and P availability in the Mechanisms of Sustainability organic layer. A factor analysis showed that available P may be closely related to the The intricate mechanisms of forest adapta- amounts of organic residues present, and by tion and the importance of biotic recycling of implication to their turnover. nutrients raise the question, what role does The natural rates of nutrient cycling are the mineral soil play in the nutrient transfor- accelerated under agriculture, where over half mations of the ecosystem? In regions of the of the organic-bound nutrients will be minerAmazon basin with soils of higher clay and alized in around 2 years. The remainder will nutrient contents, crop production can be contribute small amounts of nutrients for a fursustained with relatively low-input technolo- ther 25 years (Tiessen et al. 1994a). Thus, agrigies (Sanchez and Benitez 1987). Agricultural culture is not sustainable without nutrient use of other more sandy, low-nutrient sites is inputs beyond 3 years, although the release of limited to small (< 1 ha) slash-burn sites with remaining nutrients can provide for the a 3-5 year cultivation cycle. The end of the reestablishment of secondary successions. cycle usually coincides with the disappear- Tiessen et al (1994b) proposed that secondary ance of the ectorganic horizon, indicating succession will be as much P as N-limited, that mineral soils contribute little to sustained since much of the natural P availability is crop production. Tiessen et al. (1994a) controlled by organic matter dynamics.
Amazon terra firme forests are lower than those measured on richer soils in Costa Rica and Barro Colorado, Panama. In general all the organic N mineralized is subsequently nitrified, leaving little or no free NH^ Under natural conditions, however, NH4 could be incorporated into biomass, immobilized in the soil, or taken up by the vegetation, resulting in a reduction of nitrification rates. In relatively infertile soils of the Amazon basin, nitrogen mineralization rates are much higher in the upper soil organic layer (including root mat if present) than in the mineral soil underneath (Vitousek and Matson 1988, Montagnini and Buschbacher 1989). While almost 100% of the N mineralized is also nitrified, the organic layer nitrification amounts to only 25-39% in Brazil and 44-51% in Venezuela. Mycorrhiza in general enhances ammonium uptake (Alexander 1989). The high amount of fine roots, coupled with the very high micorrhizal association in the ectorganic layer and the superficial organic layer in these forests, allows the incorporation of the mineralized nitrogen, thus preventing significant losses of this element.
Soil versus Biological Controls on Nutrient Cycling in Terra Firme Forests
65
Table 4.7 Changes in topsoil carbon on cultivation in a typical 3 yr slash and burn site on a Red-yellow concretionary sandy ferralsol in San Carlos de Rio Negro, Venezuela. C Content Native (g nv2) Organic layer Whole soil (-15 cm) Coarse C, root C Root C alone
7000 5100 2800 500
C Content after 3 Years 1300
3600 ND 500
C Loss
Mean Residence Time C Loss (yr)
(g m-2)
C Loss (%)
5700 1500
81 29
2 9
0
0
0.4
Source: Tiessen et al. 1994.
Secondary forest growth shows rates of productivity similar to or greater than those of intact mature forest and can reestablish relatively high soil organic matter levels (Uhl 1987), although basal area and total biomass similar to the mature forest are reached only after 190 years in areas like San Carlos de Rio Negro (Saldarriaga et al. 1988). Saldarriaga (1987) studied 23 sites near San Carlos, the oldest of which had been abandoned for 80 years. He found that K recovers most quickly, followed by P. The recovery of N stocks to the level of the original forest was estimated to be around a hundred years. On relatively richer soils, recovery of nutrient stocks will happen more quickly. In areas with high rainfall and limited stabilization of organic and other materials by silts and clays (10% of soil weight in the case of San Carlos), fertilization is not a viable option even for cash crops, unless a successful litter-organic matter cycle can be estab-
lished via alternative management techniques. In less leached areas fertilization might be an option; however, the potential for crop failure is still there due to the effect of droughts and the rapid degradation of the soil organic matter. Both scenarios make the management of organic matter an essential condition in crop sustainability. We can then consider that the quantification of rates of nutrient cycling associated with soil organic matter turnover under primary and secondary forests could be a predictive tool for evaluating the potential of soils for agriculture and for subsequent forest recovery. The development of easily measured indicators for nutrient cycling, and the evaluation of nutrient cycling and organic matter turnover in natural and derived systems, are essential for decisions concerning land use and management. Measurements of static nutrient pools are simply inadequate.
Literature Cited Alexander, I. 1989. "Mycorrhizas in tropical forests." In: Mineral Nutrients in Tropical Forests and Savanna Ecosystems, ed. J. Proctor (Blackwell Scientific Publications, Oxford), pp. 169-188. Baillie, I. C. 1989. "Soil characterisitcs and classification in relation to the mineral nutrition of tropical wooded ecosystems." In: Mineral Nutrients in Tropical Forests and Savanna Ecosystems, ed. J. Proctor (Blackwell Scientific Publications, Oxford), pp. 15-26.
Baillie, I. C. 1996. "Soils of the humid tropics." In: The Tropical Rain Forest. Second Edition, ed. P.W. Richards (Cambridge University Press, Cambridge), pp. 256-286. Baillie, I. C., P. S. Ashton, M. N. Court, J. A. R. Anderson, E. A. Fitzpatrick, and J. Tinsley. 1987. "Site characteristics and the distribution of tree species in Mixed Dipterocarp Forest on tertiary sediments in Central Sarawak, Malaysia." Journal of Tropical Ecology 3: 201-220.
66 Bennett, R. J., C. M. Bree, and V. H. Bandu. 1990. "A role for Ca2+ in the cellular differentiation of root cap cells: a re-examination of root growth control mechanisms." Environmental Experimental Botony 30: 515-523. Berish, C. 1982. "Root biomass and surface area in three successional tropical forests." Canadian Journal of Forest Research 12: 699-704. Berish, C., and J. J. Ewel. 1988. "Root development in simple and complex tropical successional systems." Plant Soil 106: 73-84. Bleeker, P. 1983- Soils of Papua New Guinea. Australian National University Press, Canberra. Bongers, F., D. Engelen, and H. Klinge. 1985- "Phytomass structure of natural plant communities on spodosols in southern Venezuela: the bana woodland." Vegetation 63: 13-34. Brouwer, L. C. 1996. Nutrient cycling in pristine and logged tropical rain forest. A study in Guyana. Tropenbos Guya Series 1. Thesis Universiteit Utrecht. The Netherlands. Brown, S., and E. Lugo. 1992. "Above-ground biomass estimates for tropical moist forests of the Brazilian Amazon." Interciencia 17: 8-16. Bruijnzeel, L. A. 1992. "Managing tropical forest watersheds for production: where contradictory theory and practice co-exist." In: Wise Management of Tropical Forests, eds. F. R. Miller, and K. L. Adam (Oxford Forestry Institute, Oxford), pp. 37-75. Buol, S. W. 1985- "Mineralogy classes in soil families with low activity clays." In: Mineral Classification of Soils, ed. J. A. Kittrick (Spec Publ 16. Soil Sci. Soc. Amer. and Amer. Soc. Agr. Madison, Wisconsin), pp. 169-178. Buschbacher, R. J. 1984. Changes in productivity and nutrient cycling following conversion of Amazon forest to pasture. PhD Thesis, University of Georgia, Athens, GA. Burnham, 1989. "Pedological processes in temperate and tropical soils." In: Mineral Nutrients in Tropical Forests and Savanna Ecosystems, ed. J. Proctor (Blackwell Scientific Publications, Oxford), pp. 27-41. Caceres, A. 1989- Las micorrizas vesiculo-arbusculares en un bosque humedo tropical y su evolucion luego de la perturbacion (conuco) y la sucesion por 60 anos en San Carlos de Rio Negro, T.F. Amazonas. Masters Thesis. Institute Venezolano de Investigaciones Cientificas. Caracas, Venezuela. Chapin, F. S., and R. L. Bieleski. 1982. "Mild phosphorus stress in barley and a related low-phosphorus-adapted barley grass: Phosphorus fractions and phosphate absorption in relation to growth." Physiology Plant. 54: 309-317. Cuevas, E., and E. Medina. 1983. "Rootproduction and organic matter decomposition in a terra firme forest of the upper Rio negro basin." In: International Symp. Root Ecol. and its Practical Applications, ed. L. Kutschera (Gumpenstein, Austria), pp. 653-666. Cuevas, E., and E. Medina. 1986. "Nutrient dynamics within Amazonian forest ecosystems. I. Nutrient flux in fine litter fall and efficiency of nutrient utilization." Oecologia 68: 466-472. Cuevas, E., and E. Medina. 1988. "Nutrient dynamics within Amazonian forests. II. Fine root growth, nutrient availability and leaf litter decomposition." Oecologia 76: 222-235.
Elvira Cuevas Cuevas, E., and E. Medina. 1990. "Phosphorus/nitrogen interactions in adjacent Amazon forests with contrasting soils and water availability." In: Phosphorus Cycles in Terrestrial and Aquatic Ecosystems, eds. H. Tiessen, D. Lopez-Hernandez, and I.H. Salcedo (Proc. SCOPE-UNEP Regional Workshop 3: South and Central America. Saskatchewan Institute of Pedology. Saskatoon, Canada), pp. 83-94. Cuevas, E., and A. E. Lugo. 1998. "Dynamics of organic matter and nutrient return from litterfall in stands of ten tropical plantation species." Forest Ecology and Management. 112: 263-279Duivenvoorden, J. F., and J. M. Lips. 1995. A land-ecological study of soils, vegetation, and plant diversity in Colombian Amazonia. The TROPENBOS Foundation. Wagenigen, The Netherlands. FAO/UNESCO. 1974. Soil Map of the World 1: 5 000 000. I. Legend. UNESCO, Paris. Feigl, B. J., G. P. Sparling, D. J. Ross, and C.C. Cent 1995. "Soil microbial biomass in Amazonian soils: evaluation of methods and estimates of pool sizes." Soil Biology and Biochemistry 27: 1467-1472. Fittkau, E. -J. 1971. Okologische Gliederung des AmazonasGebietes auf geochemischer Grundlage. Miinsterige Forschungen zur Geologic und Palaontologie 20/21: 35-50. Fittkau, E.-J., W. Junk, H. Klinge,. and H. Sioli. 1975"Substrate and vegetation in the Amazon." In: Berichte der Internationalen Vereinungung fur Vegetationskunde, ed. R. Tiixen (J. Cramer. Vaduz, Liechstenstein), pp. 73-90. Furch, K., and H. Klinge. 1989. "Chemical relationships between vegetation, soil and water in contrasting inundation areas of Amazonia." In: Mineral Nutrients in Tropical Forest and Savanna Ecosystems, ed. J. Proctor (Blackwell Scientific Publications, Oxford), pp. 189-204. Herrera, R. 1977. "Correlation with vegetation types." In: Soil and terrain conditions in the International Amazon Project at San carlosde Rio Negro, Venezuela, ed. E. F. Briinig (Trans. International MAB/IUFRO Workshop on Tropical Rainforest Ecosystems Research. Hamburg, West Germany), pp. 182-188. Herrera, F., C. F. Jordan, H. Klinge, and E. Medina. 1978. "Amazon ecosystems. Their structure and functioning with particular emphasis on nutrients." Interciencia 3: 223-232. Hogberg, P. 1986. "Soil nutrient availability, root symbioses and tree species composition in tropical Africa; a review." Journal of Tropical Ecology 2: 359-372. Janos, D. P. 1983. "Tropical mycorrhizas, nutrient cycles and plant growth." In: Tropical Rain Forest: Ecology and Management, eds. S. L. Sutton, T. C. Whitmore, and A. C. Chadwick (Blackwell Scientific Publications, Oxford), pp. 327-345. Jayachendran, K., A. P. Schwab, and B. A. D. Hetrick. 1992. "Mineralization of organic phosphorus by vesiculararbuscular mycorrhizal fungi." Soil Biology and Biochemistry 24: 897-903Jordan, C. F. 1985. Nutrient Cycling in Tropical Forest Ecosystems. (John Wiley & Sons, Chichester). Klinge, H. 1973. "Root mass estimation in lowland tropical rain forests of central Amazonia, Brazil. I. Fine root mass-
Soil versus Biological Controls on Nutrient Cycling in Terra Firme Forests es of a pale yellow latosol and a giant humus podzol." Tropical Ecology 14: 29-38. Lajtha, K., and A. F. Harrison. 1995. "Strategies of phosphorus acquisition and conservation by plant species and communities." In: Phosphorus in the Global Environment, ed. H. Tiessen (John Wiley & Sons, Chichester), pp. 139-147. Lodge, D. J. 1993. "Nutrient cycling by fungi in wet tropical forests." In: Aspects of tropical Mycology, eds. S. Isaac, J. C. Frankland, R. Watling, and A. J. Whalley (Cambridge University Press, Cambridge), pp. 37-58. Lugo, A. E., E. Cuevas, and M. J. Sanchez. 1990. "Nutrients and mass in litter and top soil of ten tropical tree plantations." Plant Soil 125: 263-280. Luizao, F. J. 1989- "Litter production and mineral element input to the forest floor in a Central Amazonian forest." Geojournal 19: 407-417. Luizao, F. J. 1995. Ecological studies in contrasting forest types in central Amazonia. Ph.D. Thesis. Stirling University, U. K. Luizao, R. C. C., T. A. Bonde, and T. Roswall. 1992. "Seasonal variation of soil microbial biomass-the effects of clearfelling a tropical rainforest and establishment of pasture in the central Amazon." Soil Biology and Biochemistry 24: 805-813. Marschner, H. 1995. Mineral Nutrition of Higher Plants. Academic Press. London. Matsui, E., E. Salati, M. N. G. Ribeiro, C. M. dos Reis, A. C. S. N. Tancredi, and J. R. Gat. 1982. "Precipitation in the Central Amazon Basin: the isotopic composition of rain and atmospheric moisture at Belem and Manaus." Acta Amazonica 13: 307-369. Medina, E., and E. Cuevas. 1989. "Patterns of nutrient accumulation and release in Amazonian forests of the Upper Rio Negro basin." In: Mineral Nutrients in Tropical Forest and Savanna Ecosystems, ed. J. Proctor (Blackwell Scientific Publications, Oxford), pp. 217-240. Medina, E., and E. Cuevas. 1994. "Mineral nutrition: humid tropical forests." Progress in Botany 55: 115-129Medina, E., and E. Cuevas. 1996. "Biomass production and accumulation in nutrient-limited rain forests: implications for responses to global change." In: Amazonian Deforestation and Climate, eds. J. H. C. Gash, C. A. Nobre, J. M. Roberts, and R.L. Victoria (John Wiley & Sons, Chichester), pp. 221-239Moyersoen, B. 1993- "Ectomicorrizas y micorrizas vesiculoarbusculares en Caatinga Amazonica del Sur de Venezuela." Scientia Guaianae 3. Caracas.
67
Montagnini, F., and R. Buschbacher. 1989. "Nitrification rates in two undisturbed tropical rain forest and three slash-and-burn sites of the Venezuelan Amazon." Biotropica 21: 9-14. Nepstad, D. C., C. R. De Carvalho, E. A. Davidson, P. H. Jipp, P. A. Lefebvre, G. H. Negreiros, E. D. da Silva, T. A. Stone, S. E. Trumbore, and S. Vieira. 1995. "The deep-soil link between water and carbon cycles of Amazonian forests and pastures." Nature 372: 666-667. Newman, E. I. 1985. "The rhizosphere: carbon sources and microbial populations." In: Ecological Interactions in the Soil, ed. A. H. Fitter (Blackwell Scientific Publications, Oxford). Poels, R. L. H. 1987. Soils, water and nutrients in a forest ecosystem in Suriname. Ph.D. Thesis, Agricultural University Wagenigen, The Netherlands. Proctor, J. 1992. "Soils and mineral nutrients: what do we know, and what do we need to know, for wise rain forest management?" In: Wise Management of Tropical Forests. Oxford Forestry Institute, eds. F. R. Miller, and K. L. Adam (Oxford), pp. 27-35. P. Chacon, and E. Cuevas 1994b. "Phosphorus and nitrogen status in soils along a toposequence of dystrophic rainforests on the upper Rio Negro." Oecologia 99: 145-150. Thompson, J., J. Proctor, V. Viana, W. Milliken, J. A. Ratter, and D. A. Scott. 1992. "Ecological studies on a lowland evergreen rainforest on Maraca Island, Roraima, Brazil. I. Physical environment, forest structure and leaf chemistry." Journal of Ecology 80: 689-703. Uhl, C. 1987. "Factors controlling succession following slash-and-burn agriculture in Amazonia." Journal of Ecology 75: 377-407. Uhl, C., and C. F. Jordan. 1984. "Succession and nutrient dynamics following forest cutting and burning in Amazonia." Ecology 65: 1476-1490. Vitousek, P. M. 1984. "Litterfall, nutrient cycling, and nutrient limitation in tropical forests." Ecology 65: 285-298. Vitousek, P. M., and P. A. Matson. 1988. "Nitrogen transformations in a range of tropical forest soils." Soil Biology and Biochemistry 20: 361-367. Vitousek, P. M., and R. L. Sanford, Jr. 1986. "Nutrient cycling in moist tropical forest." Annual Review of Ecological Systems 17: 137-167. Walsh, R. P. D. 1996. "Climate." In: The Tropical Rain Forest. Second Edition, ed. P. W. Richards (Cambridge University Press, Cambridge), pp. 159-205. Went, F., and N. Stark. 1968. "Mycorrhiza." BioScience 18: 1035-1039.
5 Nutrient Cycling as a Function of Landscape and Biotic Characteristics in the Cerrados of Central Brazil M. Haridasan
The cerrados of central Brazil have long been designated savannas without sufficient understanding of the structure and functioning of the different vegetation forms in the region. Excessive emphasis on identifying similarities with other savannas in Africa and Australia, and even within South America outside Brazil, prevented researchers from recognizing the cerrados' special features and interdependence among themselves in the landscape where they occur. The more extensive cerrado sensu stricto on dystrophic soils, and to a lesser extent the gallery forests known locally as matas de galeria or matas ciliares, dominated the attention of most botanists and other researchers (Ratter and Dargie 1992, Furley 1992, Furley and Ratter 1988, Furley et al. 1992). Even with increasing interest in biodiversity and ecosystem functioning, very little ongoing research is reported on nutrient cycling from the cerrado region (Solbrig et al. 1996). Information available in the literature is restricted to isolated attempts to describe and quantify specific processes like litterfall and decomposition (Peres et al. 1983), rainwater composition (Schiavini 1983, Delitti 1984), soil fertility gradients (Lopes and Cox 1977), leaf nutrient concentrations (Haridasan 1987, 1992, Araujo and Haridasan 1988), primary productivity of the ground layer (Batmanian and Haridasan 1985, Meirelles and Henrique 1992), effects of burning (Coutinho 1990,
Kauffman et al. 1994, Miranda et al. 1996c) and activities of soil fauna (Constantino 1988, Egler and Haridasan 1987, Oliveira Jr. 1985) at specific sites within a particular vegetation. Results of long duration experiments from permanent plots or watersheds are not yet reported in the literature. Very little information is available on the food webs or the role of fauna in nutrient cycling. Research on specific processes like CO2 emission on an ecosystem basis is quite recent (Miranda et al. 1996a, b, Mier et al. 1996). The following discussion is therefore restricted to the occurrence of different vegetation forms in the cerrado region and environmental factors affecting their distribution and functioning in relation to nutrient availability and nutrient cycling processes. One of the difficulties in getting information on research already carried out in Brazil is that the dissertations of graduate students in the universities are seldom published in indexed journals. There has been no attempt to compare the cerrados with other savannas in South America, other continents, or Amazon forests because the information available from the cerrados on quantitative aspects of nutrient cycling in natural ecosystems is very meager (Baruch et al. 1996). There are several important studies reported from the Amazon region on nutrient concentrations in plants, litter decomposition, and hydrological cycles over the last couple of
Nutrient Cycling as a Function of Landscape and Biotic Characteristics in the Cerrados of Central Brazil
decades (Vitousek and Sanford 1986, Proctor 1989, Jordan, 1987, Medina and Cuevas 1996, Klinge et al. 1995) but there are no studies of all pools and flows, especially those involving fauna and foodwebs.
69
periods. The density, height, canopy cover, and species composition of trees vary with soil depth, presence of concretions and soil fertility. As slope increases and deep Latosols give way to Entisols or Inceptisols, tree density and biomass diminish, giving rise to Geology and Geomorphology campo sujo (a low tree and shrub savanna) or campo limpo on nutrient-poor parent The modern Brazilian landscape began as material. Deciduous forests occur on lime a vast peneplain produced by denudation outcrops and well-drained calcium-rich soils. between the late Cretaceous and mid- Gallery forests occur on dissected terrain Tertiary. Uplifting and polycyclic stream along water courses on lower pedi-plains incisions during the later Tertiary dissected and young Pleistocene landforms (Emmerich the upland. Present-day plateaus (known 1990). When the water table is present near locally as chapadas) are the end products the surface, cerrado woodlands give way to of this old cycle of erosion (Cole 1986) and campo limpo (Fig. 5.1) and often to buritizais their surfaces are remnants of the original (Fig. 5.2). A few tree species of the cerrado South American land surface (King 1956, like Curatella americana L. and Byrsonima Braun 1971). The hilly lands represent a crassifolia (L.) Kunth, however, seem to tolnew cycle of dissection. Intensive weathering erate poor drainage conditions and occur in of the Tertiary/Quaternary sediments on so-called hydrologic (Ratter 1992) or hyperthese plateaus gave rise to different soils seasonal or semiseasonal (Sarmiento 1992) of the Oxisol order, depending on the local savannas, where lateral roots of trees are relief and drainage conditions (Curi and often restricted to upper soil layers not Franzmeier 1984, Macedo and Bryant 1987, reached by seasonal water tables (Dubs Rodrigues and Klamt 1978). Soils on 1992). Mesophytic forests also occur on hill slopes are generally shallow Entisols slopes on dissected surfaces, probably or Inceptisols. because of higher availability of moisture in Geomorphological evolution of the landscape played the most important role in the distribution of vegetation forms in the cerrado region (Cole 1986, 1992, Emmerich 1990). Topography, water table and parent material determine the boundaries among vegetation forms ranging from cerradao (a dense woodland) and cerrado sensu stricto (an open woodland) on well-drained soils to campo limpo (an open grassland), gallery forests and buritizais (swamps with the predominance of the burity palm, Mauritia viniferd) on poorly drained soils. The open woodland form of cerrado sensu stricto Figure 5.1 Campo /zmpo-cerrado sensu stricto occurs generally only on well-drained soils of transition on a Typic Quartzipsamment. Water peneplain or pediplain plateaus, because table near the soil surface is responsible for most of the woody species of the cerrado do the absence of cerrado tree species in the not tolerate waterlogging for extended campo limpo in the foreground.
70
the surface layers during the dry season due to subsurface seepage from the surrounding areas and release of nutrients through weathering of underlying parent material (Burnham 1989, Cole 1986). Thus, in any watershed of the cerrado region the landscape consists of a mosaic of soil and vegetation forms on different erosion surfaces. It is only reasonable to assume that past processes such as erosion and pediplanation played an important role in creating gradients in soil fertility and vegetation biomass within such watersheds. For example, organic matter accumulates in waterlogged and poorly drained soils but decomposes relatively rapidly in well-drained deep latosols. Besides retarding aerobic microbial decomposition processes, waterlogging also reduces the activities of soil fauna like termites and ants. Intense leaching of soluble nutrients like K and Mg from the plateaus benefited vegetation in the low-lying terrain. These gallery forests then act as sinks for leached nutrients, trapping them within their biomass and preventing their loss from the ecosystem as a whole (Ponomareva and Dokuchayev 1984).
M. Haridasan
The most important influence of geology on the vegetation of cerrado is the association of deciduous forests with limestone outcrops and the Ca-rich, mesotrophic soils derived from them (Ratter 1992, Ratter and Dargie 1992). Otherwise, cerrado vegetation is characterized by uniform edaphic conditions occurring over vast extensions of old planation surfaces with relict Latosols developed from underlying sandstone formations (Cole 1986). Sandy soils (Quartzipsamments) extend over 30 million hectares in the cerrado region, supporting mainly the cerrado sensu stricto vegetation when there is no waterlogging. Approximately 50% of the cerrado region belongs to the Precambrian, 30% to the Cretaceous, and 20% to other geological systems (Parada and Andrade 1976). The Precambrian system can further be subdivided into the Inferior Precambrian period, dominated by rocks of medium to high metamorphism (slate, phyllite, gneiss, and quartzite), and the Superior Precambrian, dominated by rocks of low metamorphism (metasedimentary rocks). The limestones which occur in the cerrado region (the Araxa, Arai, and Tocantins groups) belong to the Inferior Precambrian system. The superior Precambrian system (Bambui group) includes some of the phosphatic rocks encountered in the cerrado region of central Brazil, as in Patos de Minas, as well as limestones which are often dolomitic.
Climate
Figure 5.2 Gallery forest (to the left in the background) and the buritizal (to the right) in a cerrado landscape. The burity palms occur where drainage is impeded and in the gallery along the water courses.
Climate has a profound influence on the nature of native cerrado sensu stricto vegetation, though it is not the unique determinant of its origin and maintenance. Average annual temperatures range from 18 to 262C in the cerrado region (Eiten 1972). Average temperatures of the hottest and coldest months are only slightly different in the core region but daily variations,
Nutrient Cycling as a Function of Landscape and Biotic Characteristics in the Cerrados of Central Brazil
especially during the dry season, can be greater than the seasonal variation. The lowest minimum temperatures as well as the largest variations in temperature occur in the southern regions of the cerrado in Sao Paulo state. Day length varies very little throughout the year. Air humidity is high during the wet (rainy) season but can be very low during the dry season. Potential evapotranspiration is in the range of 1700-1800 mm, and water deficits affect plant growth only during the prolonged dry season in the core region. Thus climatic conditions are totally favorable for the maintenance of evergreen vegetation forms in the cerrado region. The main growth-limiting factor is the low availability of soil water in the upper layers of welldrained soils on upland plateaus during the dry season. The consequences of low availability of soil water are the deciduousness of many tree and shrub species and the low primary productivity of grasses in the dry season. Differential growth rates and leaf production during the dry and wet seasons among tree species are well documented in the literature. Litter accumulation and decomposition are
71
also affected by seasonal variations in humidity (Schiavini 1983). The lack of organic matter accumulation in well-drained soils is a direct consequence of the favorable conditions for decomposition throughout the year. Research on physiological aspects of plant metabolism in native plants as related to water and nutrient stress is meager.
Soils The different physiognomic forms of vegetation in the cerrado region are a direct consequence of the edaphic gradients associated with geomorphological variations. The distribution of soils in the cerrado region is shown in Table 5.1. Deep, well-drained latosols are most common on plateaus and in valleys when slope does not exceed 8%, except when the parent material is predominantly quartzitic in nature. Sandy soils, classified as areia quartzosa in the Brazilian system of soil classification (Typic Quartzipsamments in Soil Taxonomy), extend over 34 million ha or 20% of the cerrado region. Ultisols make up less than 5% of the total area. Fertile soils, defined as
Table 5.1 Distribution of soils in the cerrado region of central Brazil (Hoeflich et al. 1977)
Soil classification Soil Taxonomy
Brazilian system
FAO
Latossolos Vermelho-Amarelo Vermelho Escuro Latossolo Roxo
Ferralsols Acric Ferralsols Orthic Ferralsols Rhodic Ferralsols Arenosols Plinthic Acrisols Lithosols
Areia Quartzosa Lateritas hidromorficas Litossolo Podzolicos Vermelho Amarelo Distrofico Orthic Acrisols Eutrofico Ferric Luvisols Total
M ha
Area % Total
Oxisols Acrustox Haplustox Haplustox Psamments Plintaquults Lithic Dystropepts
94.5
69.7 17.9 6.9 34.3 17.0 15.1
56 41 11 4 20 10 9
Ustalts Ustalts
2.1 7.0
1 4
175.0
100
72
mesotrophic soils and identified on the basis of base saturation (greater than 50%) in the Brazilian system account for about 1% of the area. The nonexistence of a hierarchical scheme of soil classification in the Brazilian system makes it difficult to establish clear-cut soil-vegetation associations from surveys such as those of RADAMBRASIL. While such surveys describe variations in soils and vegetations separately and use natural vegetations to subdivide broad soil classes, they do not establish any correlations between soil properties and vegetation parameters (Haridasan 1993). For example, a soil survey report for the Federal District in central Brazil (EMBRAPA 1978) identifies four vegetation forms (forests, cerradao, cerrado, and campo sujo) associated with dark red latosols (Latossolo Vermelho Escuro) without discussing variations in soil properties which cause this gradient in vegetation. Botanists and ecologists, on the other hand, often discuss gradients in vegetation and soilavailable nutrients across different classes of soils (for example, Oxisols versus Entisols) without paying attention to important factors that place these soils in different categories at the highest level. Most comparisons of cerrados and other savannas in the literature focus on cerrado sensu stricto on deep latosols, but these comparisons have several limitations. These comparisons do not, for example, give proper attention to variations in geomorphological aspects, water table fluctuations or physical properties of soils. Attempts to correlate vegetation parameters with chemical properties of soils, especially analytical data on available nutrients in the surface layers are more common but should be viewed with caution for several reasons. First of all, such correlations by themselves do not establish a cause-and- effect relationship between soil properties and vegetation parameters. More often, the vegetation
M. Haridasan
should be affecting the available nutrient status of the surface soil through litterfall than the other way around. Secondly, attempts to establish a general deficiency parameter are inadequate unless specific conditions are defined under which individual nutrients become limiting. Another common error is to attribute undue significance to small, though consistent, quantitative variations in analytical parameters. Ratter and coworkers placed due emphasis on the influence of high calcium levels in the soil on the cerrado vegetation when they concluded that the floristic composition and phytosociological parameters of cerradoes and cerrados on more fertile soils can be quite different from those on dystrophic soils (Ratter 1971, 1992, Ratter et al. 1977, 1978). Many species occur only on dystrophic soils, others only on mesotrophic soils. Even when species indifferent to soil fertility occur in different communities their relative importance can vary to a great extent (Araujo and Haridasan 1988). Nutritional adaptations of such species and communities have received very little attention for any useful discussion at a community or ecosystem level (Haridasan 1992). What seems rather obvious at present is that calcareous soils generally have enough nutrients to permit transfer of large quantities of nutrients to plant biomass during the rainy season when the vegetation is green. We have discussed in an earlier publication the difficulties in comparing nutritional adaptations of native plant species in different communities (Araujo and Haridasan 1988). Occurrence of species exclusive to each locality or soil, differences in distribution of common species among sites, distinct characteristics of functional groups such as leguminous species, calcifugous or calcicole species, and aluminium accumulators make such comparisons on a community level very difficult. Without such details, it is difficult to define ecosystem
Nutrient Cycling as a Function of Landscape and Biotic Characteristics in the Cerrados of Central Brazil
processes which will determine nutrient pools and fluxes. Whatever data we have generally available in most publications are mean concentrations of litter or leaves of all species in an area. For example, it is not all species which contribute to fixation of atmospheric nitrogen and contribute to a general improvement of soil fertility. Calcicole species can probably survive well only on soils with neutral or alkaline pH. The ecological significance of the differences in relative importance of functional groups such as Al-accumulators or Leguminosae in native communities is rarely discussed in the literature (Haridasan and Araujo 1988).
Natural Vegetation Among the different physiognomic forms of vegetation occurring in the cerrado region, the most extensive is the cerrado sensu stricto. The floristic composition of different vegetation forms, geographic distribution of species and their relationship to Amazonian forests are well discussed in the literature. However, variations in phytosociology on regional and local levels have been reported on only more recently (Ratter and Dargie 1992). Detailed studies on soil-plant associa-
73
tions involving phytosociology of native species and environmental gradients in soil properties are also recent (Haridasan and Araujo 1988, Araujo 1992, Haridasan et al. 1997, Silva Junior 1995, Oliveira-Filho et al. 1989). The following discussion is restricted to only those aspects which are known to affect nutrient cycling directly or indirectly without going into details regarding the structure and functioning or determinants of the cerrado vegetation.
Tree biomass and nutrient pools Probably the most important and characteristic aspect of the cerrado sensu stricto vegetation, as far as nutrient cycling is concerned, is its low biomass. Very few measurements are available of the biomass of the different components—trees, shrubs and the ground layer—of the cerrado vegetation. Silva (1990), based on measurements of the weight of all individual trees with a minimum of 5 cm girth at a height of 30 cm in three 20 m x 30 m plots, determined the live aerial biomass of trees alone to be 21.38 Mg ha"1 (Table 5.2). These measurements were carried out by felling all trees in three 20 m x 30 m plots at the end of the dry season. Since some of the trees shed leaves during
Table 5.2 Live aerial biomass (dry weight) of the different components of the trees (girth > 5cm at 30 cm height) of a cerrado sensu stricto on a dark red latosol (Haplustox) at the Fazenda Agua Limpa in Brasilia.
Source: Silva, 1990
Tree Component
Dry weight (Mg ha-1)
Trunk Thick branches Fine branches Leaves Fruits Total
9.40 6.58 4.27 1.04 0.09 21.38
M. Haridasan
74
Table 5.3 Comparison of the aerial biomass (dry matter) and nutrient stocks of trees of a cerrado sensu stricto in Brasilia and an Amazonian seasonal varzea forest in Ilha de Marchantaria. Cerrado sensu stricto*
Vegetation
Amazonia 1**
Amazonia 2** 254.6
9.9
97.5 38.2
N
not reported
387.1
1107.2
P
4.9
K
30.5 24.9 12.0
31.5 665.6 805.8 163.0
67.1 1806.2 2988.2
Biomass, Mg ha'1 2
21.4 1
Basal area, m ha'
95.7
Proportion Amazonia 1 Cerrado
Proportion Amazonia 2 Cerrado
4.55 3.86
11.90 9.67
6.43 21.82 32.36 13.58
13.69 59-22 120.00 28.48
Nutrients, kg ha'1
Ca Mg
341.7
* Silva 1990. ** Klinge et al. 1995.
the dry season, the dry weight of leaves could be an underestimate. A comparison of these data with those of an Amazonian forest shows that the aerial biomass of the trees of a cerrado sensu stricto in central Brazil may be only 8 to 22% of that of an Amazonian forest, and the basal area only 10 to 26% (Table 5.3). This difference in biomass reflects directly on the nutrient pools in the biomass. A comparison of the data reported by Klinge et al. (1995) for the aboveground biomass and nutrient stock in two inundation forests in the Ilha de Marchantaria with the data for a cerrado sensu stricto from central Brazil (Silva 1990) illustrates how nutritionally poor the cerrado is in quantitative terms. The proportions of stock of essential nutrients in the tree biomass of cerrado are: 7 to 16% for P, 1.7 to 4.6% for K, 0.83 to 3.09% for Ca, and 3-5 to 7.4% for Mg. Thus Ca, K, and Mg seem to be much more deficient in the cerrados than P. We have no corresponding data for the stock of nutrients in the root biomass of trees for comparison among the two ecosystems. This comparison is only illustrative of two specific sites. Estimates of aboveground biomass for the Amazonian forests may vary
considerably depending the site and methodology (see Cuevas, this volume; Brown and Lugo 1992; Fearnside 1992). Similarly, any comparisons on a regional scale with cerrados are complicated by the large variety of the vegetation forms. Abdala et al. (1997) estimated the total aboveground live biomass of trees of another cerrado sensu stricto on a dark red latosol to be 22.9 Mg ha-1 (Table 5.4). This estimate was based on the measurements of basal areas and plant height and estimates of cylindrical volume of individual species from two 20 m x 50 m plots. The biomass of shrubs totaled 3-1 Mg ha"1; grasses contributed 4.13 Mg ha'1, or 74%, of the total biomass of the herbaceous layer (5.58 Mg ha"1). Litter was estimated at 5.19 Mg ha'1. Their estimate of belowground biomass of living and dead roots of this vegetation was 41.1 Mg ha'1 to a depth of 6.2 m. Thus the belowground biomass was slightly greater than that aboveground (39.8 Mg ha"1 including litter and dead trees). The belowground component should therefore exercise an enormous influence on nutrient cycling, but we have no estimates of the nutrient content of the root biomass or the rate of root
Nutrient Cycling as a Function of Landscape and Biotic Characteristics in the Cerrados of Central Brazil
decomposition. This is in addition to the organic matter component of the soil, which was estimated at 642 Mg ha'1 in the top 6.2 m of soil. The proportion of root biomass as compared to live aerial biomass in the cerrado is thus much higher than in tropical forests where root mass varies anywhere from about one-thirtieth to one-half of the aerial biomass (Cuevas this volume, Medina and Cuevas 1996). Castro (1995) established the variation in aerial and belowground biomass along a vegetation gradient from campo limpo to cerrado sensu stricto near Brasilia. The author's estimates of the aerial biomass of two cerrado sensu stricto sites show how variable such estimates can be. Her estimate of 12.9 Mg ha'1 for tree biomass amounted to only 58% of that measured by Silva (1990) and 54% of the estimate of Abdala et al. (1997). The author established this relationship by estimating aerial biomass from measurements of diameter and height using equations of biomass for tropical forests. However, her measurements of belowground biomass confirm that root biomass (46.6 to 52.9 Mg ha'1) exceeds the aerial biomass
75
(24.8 Mg ha'1) in a cerrado sensu stricto. In the campo limpo, which is devoid of trees, the aerial biomass (5.5 Mg ha"1) amounted to only 34% of the root biomass (16.3 Mg ha'1 ) in the top 2 m of soil. This proportion was 31% in the campo sujo with its sparse distribution of trees and 47.9 to 53-3 in the cerrado sensu stricto. Thus the different cerrado vegetation types showed a higher root/shoot ratio than many tropical forests.
Ground layer biomass and nutrient pools Batmanian and Haridasan (1985) estimated the primary production of the herbaceous layer of a cerrado sensu stricto to be 3.27 Mg ha'1 yr1 using the peak biomass method, with the grasses contributing 68 to 78% of the live biomass of the herbaceous layer. Thus, the transfer of nutrients from soil or litter to live biomass should be of the order of 20 kg ha'1 yr1 of N, 2 kg P, 20 kg K, 10 kg Ca, and 3 kg Mg. Even allowing for the higher estimates of primary production and live biomass of the herbaceous layer from other studies, these amounts are insignificant when
Table 5A Partitioning of biomass in a cerrado sensu stricto on a dark red latosol near Brasilia in central Brasil (Abdala et al. 1998) Component Trees (girth > 6cm at 30 cm) Stems and branches Leaves Dead trees Shrubs (girth < 6cm at 30 cm) Leaves Stems and branches Herbs Grasses Non-grasses Litter Total aboveground biomass Roots Total plant biomass
Dry weight, Mg ha'1
21.7 1.2 3.0 0.3 2.8 4.1 1.5 5.2 39.8 41.1 80.9
76
M. Haridasan
Table 5.5 Aerial biomass and nutrient stock of the ground layer vegetation of a cerrado sensu stricto in Brasilia.
Biomass, kg ha"1 Nutrient concentrations, mg g*1 P K Ca Mg Nutrient stock, kg ha"1 P K Ca Mg
Grasses
Nongrasses
778
663
0.53 5.79 1.85 1.09
0.53 5.38 6.35 1.72
0.412 4.505 1.439 0.848
0.351 3.567 4.210 1.140
Source: Based on data for unfertilized plots in Villela and Haridasan 1994
compared to the uptake by any field crop or the amounts of fertilizer application in agriculture (see for example, Fassbender et al. 1985). Thus it is evident that the native grasses and herbaceous species are adapted for survival under low soil fertility conditions. However, this does not imply that they are not capable of responding to higher availability of nutrients through fertilizer applications. The estimate of Villela and Haridasan (1994) for the biomass of the herbaceous layer of unirrigated plots of a cerrado sensu stricto in Brasilia was only 1441 kg ha"1. Grasses contributed 54% of the biomass of the herbaceous layer (Table 5.5). The nutrient concentrations and stock of Ca and Mg were higher in the nongrasses, while the stock of K was higher in the grasses. Monitoring any alteration of the total nutrient pool or fluxes due to fire or disturbance becomes a difficult task because of the very small quantities involved. Irrigation during the dry season could increase the live aerial biomass of the grasses in a cerrado sensu stricto, and liming could increase the uptake of Ca by the herbaceous layer. Haridasan et
al. (1997) reported that liming resulted in a linear increase in the concentration of Ca in the biomass of grasses and 5 herbaceous species of the ground layer of the same cerrado, proportional to the dosage of liming and the concentrations of N, P, and K increased generally with the application of fertilizers. These results indicate a general deficiency of nutrients in the soils of the cerrado and the capacity of the native species to respond to improvement in soil fertility. However, quick recovery of the native vegetation after fires should be in part due to the capacity for growth under low nutrient requirements.
Nutrient Fluxes The processes which contribute to transfers of nutrients among different pools in a natural ecosystem include uptake by plants, transfer to primary and secondary consumers, return to the organic and inorganic pool in the soil through litter of plant origin and animal excreta and dead bodies, decomposition of roots, rainfall, and fixation of
Nutrient Cycling as a Function of Landscape and Biotic Characteristics in the Cerrados of Central Brazil
77
Table 5.6 Production of litter and their nutrient concentrations in different physiognomic forms of vegetation in the cerrado region of central Brazil. Vegetation type Reference Cerrado Peres et al. (1983) Schiavini (1983) Delitti (1984) Pompeia (1989) Cerradao Peres et al. (1983)
Production (Mg ha-1 y-D
N (mg g-1)
(mg g'1)
2.1
8.5 (5.8-10.6) 13.4 14.6
0.6 (0.5-0.7) 0.3 0.6 0.5
1.6 (0.6-2.8)
8.2 (5.8-11.8)
0.6 (0.5-0.7)
2.4 3-2
7.8
atmospheric nitrogen (Fassbender 1985). The present author is not aware of any estimates of the transfer of nutrients to the animal foodweb from plant biomass or the size of the animal pools in natural cerrado ecosystems. Also lacking are studies on decomposition rates of roots. Whatever information is available on production and decomposition of litter and rainfall inputs are reviewed here. The role of fauna like termites and ants in recycling organic matter and altering soil properties creating gradients in fertility have been very little investigated (Egler and Haridasan 1987). The higher levels of available nutrients in abandoned termite mounds, for example, could facilitate the germination and early growth of tree seedlings during their disintegration. The termites could also play an important role in the disintegration and decomposition of dry litter in "welldrained soils.
Litter production and decomposition
P
K Ca Mg (mg g-1) (mg g-1) (mg g-1) 3.2
2.0 2.2 2.1
(2.1-4.4) 1.5 6.4 8.3
1.2 (0.7-1.9) 1.2 1.7 1.9
1.6 (0.8-2.8)
3.4 (2.1-4.4)
1.4 (0.7-1.9)
comparable in spite of the higher production in the cerradao (7.8 Mg ha^yr"1). Thus the flow of nutrients from the plant biomass to the soil surface could be about 3.7 times higher in the cerradao than the cerrado. Their results of similar nutrient concentrations in the litter in cerrado and cerradao were consistent with the observations of Ribeiro (1983) who compared the leaf concentrations of nutrients of individual tree species of a cerrado and cerradao on dystrophic soils. The N concentrations (8 mg g'1) reported by Peres et al. (1983) were considerably lower than the values (1.34 mg g"1) reported by Delitti (1984) for a cerrado in the Sao Paulo state of Brazil. Available reports indicate wide variations in the estimates of litter decomposition rates probably because of the differences in methodology and the ambients where the studies were conducted. Long-term studies are lacking for definite conclusions.
Rainfall inputs Estimates of litter production in the cerrado sensu stricto varies from 2.1 to 3.2 Mg ha'1 yr1 (Table 5.6). Peres et al. (1983) concluded that the concentrations of nutrients in the litter in a cerrado, and cerradao on dystrophic soils of similar fertility near Brasilia, were
Because cerrado soils and native vegetation are generally deficient in nutrients, even the small quantities of nutrients in the rainfall might be an important source for the recuperation of the native vegetation after
M. Haridasan
78
perturbations and fires (Coutinho 1979, 1990, Pivello-Pompeia and Coutinho 1992). However, data available so far are too few for any definite conclusions.
Mineral nutrition of native plants vis-a-vis nutrient cycling and ecosystem functioning Very few ecosystem studies emphasize interspecific variations in the uptake and translocation of nutrients. It has been shown that nutrient concentrations in leaves and other components do not reflect the nutrient levels in the soil because of species-specific uptake (Ernst and Tolsma 1989, Araujo and Haridasan 1988). While specific physiological adaptations of native species such as nitrogen fixation and mycorrhyzal associations have received research attention, the relative contributions of such functional groups in different native communities have seldom been quantified. For example, Haridasan (1987) hypothesized that Al-accumulating species could have a competitive advantage in dystrophic soils where Al saturation of soils could be a limiting factor for plant growth. In many situations of extreme stress in the cerrado region, Al-accumulating species show relatively high importance values. Among the 28 species occurring at 49 or more of the 96 sites of cerrado throughout Brazil with information available on phytosociological parameters of woody species, Ratter et al. (1996) cites 4 known Al-accumulators, Qualea grandiflora, Q. parviflora, Q. multiflora, and Salvertia convallariodora. Vochysia thyrsoidea Pohl, an Al-accumulating species very common on deep well-drained latosols, tolerates waterlogging on one extreme and resists drought well on shallow sands of quartzitic parent material at the other extreme. Another Al-accumulating species, Vochysia divergens, occurs as monodominant species in seasonally flooded forests of the Pantanal region (Nascimento and Cunha 1989).
Disturbances and nutrient cycling Besides deforestation and land clearing for agricultural crops, pasture, and silvicultural plantations, large areas of cerrados are destroyed by mining activities which include the complete removal of topsoil. Spontaneous revegetation or colonization by native species often depends on the depth of topsoil removal and the physical and chemical properties of the remaining subsoil. Even when there are no physical limitations and only a few centimeters of soil are removed, regeneration of secondary vegetation often takes several years. In many places vegetative propagation from deep roots is a source of reestablishment of shrubs and tree species in the absence of seed propagation and seedling establishment. More often soil compaction and depauperate subsoil are the main restrictions for natural recolonization, even by grass and herbaceous species from surrounding native vegetation. De-compaction and manuring results in spontaneous recolonization in such cases. In a recent field experiment we have been able to demonstrate that lack of nutrients is indeed the most limiting factor preventing the colonization of degraded areas in the latosol region of cerrados (Leite et al. 1994). Addition of organic matter supplements like castor bean cake and peat resulted in spontaneous invasion and establishment of herbaceous and shrub species providing hundred percent vegetative cover within a couple of years in an area which had not shown any signs of recovery even after 30 years of abandonment. Inputs through rainfall and other sources were not sufficient to produce any significant impact on the establishment of secondary vegetation though such inputs may be significant in natural ecosystems.
Fire in the cerrados Among the natural disturbances that the cerrado sensu stricto suffers periodically,
Nutrient Cycling as a Function of Landscape and Biotic Characteristics in the Cerrados of Central Brazil
Figure 5.3 A cerrado in Brasilia during the dry season showing the fuel load in the herbaceous layer.
79
a predominance of non-grasses but within a year the original proportion of grass biomass may be restored (Batmanian and Haridasan 1985). Several dormant species of the ground layer produce flowers and seeds only after a fire. Kauffman et al. (1994) estimated the fuel loads along a vegetation gradient from campo limpo to cerrado sensu stricto near Brasilia. In the cerrado only 27% of the fuel load of 10 Mg ha*1 was comprised of graminoids; the remainder was deadwood and leaf litter. They estimated the nutrient pools in combustible components in the cerrado sensu stricto to be 54.7 kg ha'1 N, 13.8 kg ha-1 K, 3.5 kg ha-1 P, and 30.5 kg ha'1 Ca. They concluded that the total biomass of the herbaceous layer of the cerrados was similar to that of other savanna ecosystems. The authors concluded that any loss of N due to fire was negligible compared to the N pool in the soil. Biological N fixation and precipitation inputs would compensate for such losses. Similarly, precipitation inputs would compensate for the loss of P, K and Ca (Schiavini 1983, Coutinho 1979, Pivello-
none is more important than fire (Eiten 1972, Coutinho 1982), which is more frequent during the long dry season when the herbaceous layer dries out. The intensity of the fire depends on the quantity of fuel material accumulated at the soil surface. Quantity of fuel material is, in turn, dependent on the period intervening since the last fire, the primary production of the ground layer, and the quantity of litter produced by the shrubs and trees (Fig. 5.3). These fires often totally destroy the aerial portion of the ground layer vegetation but seldom kill any tall trees, though the high temperatures of the flames may result in trees shedding most of their leaves a few days after the fire (Fig. 5.4). Cerrado trees and shrubs are able to survive the frequent fires because of several adaptive strategies such as thick, corky barks and the ability to sprout adventitious buds from deeper lying tissues such as wood cambium. What is impressive is the ability of the ground layer vegetation to sprout within a Figure 5.4 A cerrado sensus stricto in Brasilia few weeks following the fire, often providing immediately after a fire during the dry season. total vegetative cover within three to four The ground layer is completely burnt while months (Fig. 5.5). During the first few leaves of the trees have dried due to heat months the ground layer biomass may show from the fire and will shed in a few days.
80
M. Haridasan
Pompeia and Coutinho 1992) within one to three years.
Soil erosion and nutrient cycling The great susceptibility of latosols to erosion following removal of native vegetation and soil compaction is an important aspect to be considered in nutrient loss studies. Accelerated erosion and the formation of wide deep gullies alter the landscape and bring the unwelcome consequences of sedimentation and destruction of water courses and reservoirs. Ross (1992) concluded that rates of soil erosion after clearance of native vegetation were up to eight times higher than those after partial clearance and under forest in Roraima in the Amazon region. Soil erosion resulted in accelerated rates of litter removal from the soil surface and significantly higher losses of nitrogen, calcium and potassium from the soil nutrient pool. Similar consequences can be expected in the cerrados. Devoid of natural vegetation, the soils of the cerrado region are highly susceptible to erosion by water. High rainfall intensities favor severe erosion of all forms in unprotected soils. The consequences forn nutrient and water cycling are several. First, the removal of the weathered well-structured surface soil, which is relatively richer in nutrients than the underlying material, leaves behind compact impermeable nutrient-poor regolith. This compacted surface is incapable of supporting spontaneous recolonization by native species. The removal of porous top soil also implies a loss of water storage capacity, resulting in excessive runoff during the rainy season and drought during the dry season. Construction of roadways and mining activities often result in the formation of long, wide and deep gullies affecting the whole drainage pattern of entire watersheds because of the high intensity rains in the region.
Figure 5.5 Sprouting of the ground layer vegetation of campo sujo on sandy soil within a month after the fire during the dry season.
Conclusions Richter and Babbar (199D concluded their review on soil diversity in the tropics stating, "to speak carelessly about 'tropical soil' greatly oversimplifies the complexity and diversity of ecosystems in this 5-billion hectare region. As soil diversity at all spatial scales (from region to microsite) is better documented, understanding and use of tropical ecosystems can only improve." To speak carelessly about "savannas", similarly, oversimplifies the complexity and diversity of ecosystems in the tropical region. Documentation of spatial variability of determinants such as geomorphological features, soil properties, water regimes, flora and fauna and ecosystem processes is still too sparse to define and understand different facets of nutrient cycling in Amazon savannas and the cerrados of central Brazil. Generalizations based on information from isolated site-specific studies may not be totally valid on an ecosystem basis because of these variations. Future research on nutrient cycling in cerrados should emphasize the specific adaptations of native communities and individual species
Nutrient Cycling as a Function of Landscape and Biotic Characteristics in the Cerrados of Central Brazil
or functional groups as components of these natural ecosystems. These adaptations are crucial to overcoming often extreme stress conditions, including nutrient deficiency and alu-
81
minum toxicity in strongly acid soils and water deficits during the dry season. Any research on nutrient pools and fluxes involving primary and secondary consumers will be practically new.
Literature Cited Abdala, G. C., L. S. Caldas, M. Haridasan, and G. Eiten 1998. Below-ground organic matter and root-shoot ratio in a cerrado in central Brazil. Brazilian Journal of Ecology (1998) 2:11-23 Araujo, G. M. 1992 Comparacao da estrutura e do teor de nutrientes nos solos e nas folhas de especies arboreas de duas matas mesofilas semideciduas no Triangulo Mineiro. Ph. D. Thesis. Universidade Estadual de Campinas, Campinas, Brazil. Araujo, G. M., and M. Haridasan. 1988. A comparison of the nutrient status of two forests on dystrophic and mesotrophic soils in the cerrado region of central Brazil. Communications in Soil Science and Plant Analysis 19:1075-1089. Baruch, Z., A. J. Belsky, L. Bulla, A. C. Franco, I. Garay, M. Haridasan, P. Lavelle, E. Medina, and G. Sarmiento. 1996. Biodiversity as regulator of energy flow, water use and nutrient cycling, in O. T. Solbrig, E. Medina, and J. F. Silva, editors. Biodiversity and savanna ecosystem processes - a global perspective. Ecological Studies, Vol. 121. Springer-Verlag, Berlin, Heidelberg. Batmanian, G. J., and M. Haridasan. 1985. Primary production and accumulation of nutrients by the ground layer community of a cerrado vegetation of central Brazil. Plant and Soil 88: 437-440. Brown, S., and A. E. Lugo. 1992. Aboveground biomass estimates for tropical moist forests of the Brazilian Amazon. Interciencia 17(Jan-Feb):8-18. Braun, O. P. G. 1971. Contribuicao a geomorfologia do Brasil. R. bras. Geogr. 32(3):3-39. Burnham, C. P. 1989.Pedological processes and nutrient supply from parent material in tropical soils. Pages 2741. in J. Proctor. Mineral nutrients in tropical and savannah ecosystems. Blackwell, Oxford. Castro, E. A. 1995. Biomass, nutrient pools and response to fire in the Brazilian cerrado. Master's thesis. Oregon State University. Corvallis. Cole, M. M. 1986. The savannas: Biogeography and Geobotany. Academic Press. London. Cole, M. M. 1992. Influence of physical factors on the nature and dynamics of forest-savanna boundaries. Pages 63-75. in P. A. Furley, J Proctor, and J. A. Ratter, editors Nature and Dynamics of Forest-Savanna Boundaries. Chapman and Hall, London. Constantino, R. 1988. Influencia da macrofauna na dinamica de nutrientes do folhedo em decomposicao em cerr do sensu stricto. Master's thesis. Departamento de Biologia Vegetal, Universidade de Brasilia, Brasilia, Brazil. Coutinho, L. M. 1979. Aspectos ecologicos do fogo no cerrado. III. A precipitacao atmosferica de nutrientes minerais. Revista Brasileira de Botanica 2:97-101. Coutinho, L. M. 1982. Ecological effects of fire in Brazilian cerrados. Pages 273-291. in B. J. Huntley and B. H.
Walker, editors. Ecology of tropical savannas. Ecological Studies, Vol. 42. Springer-Verlag, Berlin. Coutinho, L. M. 1990. Fire in the Ecology of Brazilian Cerrado. Pages 82-105. inj. G. Goldammer, editor. Fire in the tropical biota: Ecological processes and global challenges. Ecological Studies, Vol. 84. Springer-Verlag, Berlin. Curi, N., and D. P. Franzmeier. 1984. Toposequence of oxisols from the Central Plateau of central Brazil. Soil Sci. Soc. Am. J. 48:1245-1248. Delitti, W. C. 1984. Aspectos comparatives de ciclagem de nutrientes na mata ciliar, no campo cerrado e na floresta implantada de Pinus elliotti Engelm. var. elliottii (Mogi-Guacu, SP). Ph. D. thesis, Universidade de Sao Paulo, Sao Paulo, Brazil. Dubs, B. 1992. Observations on the differentiation of woodland and wet savanna habitats in the Pantanal of Mato Grosso, Brazil. Pages 431-449. in P. A. Furley, J Proctor, andj. A. Ratter, editors. Nature and Dynamics of Forest-Savanna Boundaries. Chapman and Hall, London. Egler, I., and M. Haridasan. 1987. Alteration of soil properties by Procornitermes araujoii Emerson (Isoptera, Termitidae) in latosols of the cerrado region of central Brazil. Pages 280-308. in J. J. San Jose, and R Monies, editors. La capacidad bioprodutiva de sabanas. IVIC, Caracas, Venezuela. Eiten, G. 1972. The cerrado vegetation of central Brazil. Bot. Rev. 38:201-341. EMBRAPA, 1978. Levantamento de reconhecimento dos solos do Distrito Federal. Boletim Tecnico no. 53. Service Nacional de Levantamento e Conservacao de Solos. Rio de Janeiro, Brazil. Emmerich, K. H. 1990. Influence of landform, landscape development and soil moisture balance on forest and savanna ecosystem patterns in Brazil. Pedologie. XL(1):5-17. Ernst, W. H. O., and D. J. Tolsma. 1989. Mineral nutrients in some Botswana savanna types. Pages 97-119. in J. Proctor. Mineral nutrients in tropical and savannah ecosystems. Blackwell, Oxford. Fassbender, H. W. 1985. Ciclos da materia organica e dos nutrientes em ecossistemas florestais dos tropicos. Pages 202-230. in P. C. Rosand, editor. Reciclagem de nutrientes e agricultura de baixos insumos nos tropicos. CEPLAC/SBCS, Ilheus, Bahia, Brazil. Fassbender, H. W., L. Alpizar, J. Heuveldop, G. Enriquez, and H. Folster. 1985. Ciclos da materia organica e dos nutrientes em agrossistemas com cacaueiros. Pages 230257. in P. C. Rosand, editor. Reciclagem de nutrientes e agricultura de baixos insumos nos tropicos. CEPLAC/SBCS, Ilheus, Bahia, Brazil. Fearnside, P. M. 1992. Forest biomass in Brazilian Amazonia: comments on the estimate by Brown and Lugo. Interciencia 17 Qan-Feb):19-27. Furley, P. A. 1992. Edaphic changes at forest-savanna
82
boundary with particular reference to the neotropics. Pages 91-117. in P. A. Furley, J Proctor, and J. A. Ratter, editors. Nature and Dynamics of Forest-Savanna Boundaries. Chapman and Hall, London. Furley, P. A., and J. A. Ratter. 1988. Soil resources and plant communities of the central Brazilian cerrado and their development. Journal of Biogeography 15:97-108. Furley, P. A., J. Proctor, and J. A. Ratter. 1992. Nature and dynamics of forest-savanna boundaries. Chapman and Hall, London. Haridasan, M. 1987. Distribution and mineral nutrition of aluminium-accumulating species in different plant communities of the cerrado region of central Brazil. Pages 309-348. in J. J. San Jose, and R Montes. editors. La capacidad bioprodutiva de sabanas, IVIC. Caracas, Venezuela. Haridasan, M. 1992. Observations on soils, foliar nutrient concentrations and floristic composition of cerrado sensu stricto and cerradao communities in central Brazil. Pages 171-184. in P. A. Furley, J Proctor, and J. A. Ratter, editors. Nature and Dynamics of Forest-Savanna Boundaries. Chapman and Hall, London. Haridasan, M. 1993 Solos. Pages 309-328. in M. NovaesPinto, editor. Cerrado: caracterizacao, ocupacao e perspectivas. Second Edition. Editora Universidade de Brasilia, Brasilia. Haridasan, M., and Araujo, G. M. 1988. Aluminium-accumulating species in two forest communities in the cerrado region of central Brazil. Forest Ecology and Management 24: 15-26. Haridasan, M., J. M. Felfili, M. C. Silva Jr., A. V. Rezende, and P. E. N. Silva 1997. Gradient analysis of soil properties and phytosociological parameters of some gallery forests on the Chapada dos Veadeiros in the cerrado region of central Brazil. Pages 168-180. in Proceedings of the International Symposium on Assessment and Monitoring of Forests in Tropical Dry Regions with Special Reference to Gallery Forests. Editora da Universidade de Brasilia, Brasilia, Brazil. Haridasan, M., A. A. M. C. Pinheiro, and F. R. R. Torres. 1997. Resposta de algumas especies do estrato rasteiro de um cerrado a calagem e a adubacao. Pages 87-91. in Leite, L. L., and C. H. Saito, editors. Contribuicao ao conhecimento ecologico do cerrado. Universidade de Brasilia, Brasilia, Brazil. Hoeflich, E. R., J. C. Cruz, J. Pereira, F. F. Duque, and H. Tollini. 1977. Sistema de producao agricola no cerrado. Pages 37-58. in M. G. Ferri, editor. IV Simposio sobre o Cerrado: Bases para Utilizacao Agropecuaria. Editora Itatiaia and Editora da Universidade de Sao Paulo, Sao Paulo, Brazil. Jordan, C. F. 1987. Amazonian rainforests: Ecosystem disturbance and recovery. Springer, New York. King, L. C. 1956. Geomorfologia do Brasil central. R. Bras. Geog. 18(2):l47-265. Klinge, H., J. Adis, and M. Worbes. 1995. The vegetation of a seasonal varzea forest in the lower Solimoes river, Brazilian Amazonia. Acta Amazonica. 25:201-220. Kauffman, J. B., D. L. Cummings, and D. E. Ward. 1994. Relationships of fire, biomass and nutrient dynamics along a vegetation gradient in the Brazilian cerrado. Journal of Ecology. 82:519-531. Leite, L. L., C. R. Martins, and M. Haridasan. 1994. Efeito da
M. Haridasan descompactacao e adubacao do solo na revegetacao espontanea de uma cascalheira no Parque Nacional de Brasilia. Pages 527-534. in Anais do I Simposio Sul-americano e II Simposio Nacional sobre Recuperacao de Areas Degradadas. Foz de Iguacu, Parana, Brazil. Lopes, A. J., and F. R. Cox. 1977. A survey of the fertility status of surface soils under cerrado vegetation of Brazil. Soil. Sci. Soc. Am. Jour. 41:752-757. Macedo, J., and R. B. Bryant. 1987. Morphology, mineralogy, and genesis of a hydrosequence of oxisols in Brazil. Soil Sci. Soc. Am. J. 51:690-698. Medina, E., and E. Cuevas. 1996. Biomass production and accumulation in nutrient-limited rain forests: implications for responses to global change. Pages 221-239. in J. H. C. Gash, C. A. Nobre, J. M. Roberts, and R. L. Victoria, editors. Amazonian deforestation and climate. Institute of Hydrology, London. Meirelles, M. L, and R. P. B. Henriques. 1992. Producao primaria liquida em area queimada e nao queimada de campo sujo de cerrado, Planaltina, DF. Acta bot. bras 6:3-14. Meir, P., J. Grace, A. Miranda, and J. Lloyd. 1996. Soil respiration in a rainforest in Amazonia and in cerrado in central Brazil. Pages 319-329. in J. H. C. Gash, C. A. Nobre, J. M. Roberts, and R. L. Victoria, editors. Amazonian deforestation and climate. Institute of Hydrology, London. Miranda, A. C., H. S. Miranda, J. Lloyd, J. Grace, J. A. Mclntyre, P. Meir, P. Riggan, R. Lockwood, and J. Brass. 1996a. Carbon dioxide fluxes over a cerrado sensu stricto in central Brazil. Pages 353-363. in J. H. C. Gash, C. A. Nobre, J. M. Roberts, and R. L. Victoria, editors. Amazonian deforestation and climate. Institute of Hydrology, London. Miranda, A. C., H. S. Miranda, J. Lloyd, J. Grace, R. J. Francey, J. A. Mclntyre, P. Riggan, R. Lockwood, and J. Brass. 1996b. Fluxes of carbon, water and energy over Brazilian cerrado: An analysis of eddy covariation and stable isotope. Plant Cell and Environment. Miranda, H. S., C. H. Saito, and B. F. S. Dias. 1996c. Impactos de queimadas em areas de cerrado e restinga. Departamento de Ecologia, Universidade de Brasilia, Brasilia, Brazil. Nascimento, M. T, and C. N. Cunha. 1989. Estrutura e composicao floristica de um cambarazal no Pantanal de Pocone - MT. Act. Bot. Bras. 3:3-23Oliveira-Filho, A. T., G. J. Shepherd, F. R. Martins, and W. H. Stubblebine. 1989. Environmental factors affecting physiognomic and floristic variation in an area of cerrado in central Brazil. Journal of Tropical Ecology 5:413:431. Oliveira Jr., R. 1985. Efeito dos microartropodos sobre a decomposicao do folhedo em um cerrado. Master's thesis. Departamento de Biologia Vegetal, Universidade de Brasilia, Brasilia, Brazil. Parada, J. M., and S. M. Andrade. 1976. Cerrados - Recursos Minerais. Pages 195-209 in M. G. Ferri, editor. IV Simposio sobre o Cerrado: Bases para Utilizacao Agropecuaria. Editora Itatiaia and Editora da Universidade de Sao Paulo, Sao Paulo, Brazil. Peres, J. R. R., R. Suhet, M. A. T. Vargas, and A. Drosdowicz. 1983. Litter production in two cerrado vegetations in Brazil. Pesquisa Agropecuaria Brasileira 18:1037-43.
Nutrient Cycling as a Function of Landscape and Biotic Characteristics in the Cerrados of Central Brazil Pivello-Pompeia, V. R., and L. M. Coutinho 1992. Transfer of macro-nutrients to the atmosphere during experimental burnings in an open cerrado (Brazilian savanna). Journal of Tropical Ecology. 8:487-497. Pompeia, S. L. 1989. Aspectos da dinamica dos nutrientes minerals em solo sob vegetacao de campo cerrado (Mogi-Guafu, SP). Master's thesis, Universidade de Sao Paulo, Sao Paulo, Brazil. Ponomareva, V. V., and V. V. Dokuchayev.1984. Watermineral nutrition of plants as a major factor in phylogenesis and soil formation. Pochvovedeniye. 1984(4):29-38. Proctor, J. 1989- Mineral nutrients in tropical and savannah ecosystems. Blackwell, Oxford. Ratter, J. A. 1971. Some notes on two types of cerradao occurring in North Eastern Mato Grosso. Pages 100-102. in M. G. Ferri, editor. Ill Simposio sobre o cerrado. Editora da Universidade de Sao Paulo, Sao Paulo. Ratter J. A., P. W. Richards, G. Argent, and D. R. Gifford. 1977. Observances adicionais sobre o cerradao de solos mesotroficos no Brasil central. Pages 306-316. in M. G. Ferri, editor. IV Simposio sobre o cerrado. Editora da Universidade de Sao Paulo, Sao Paulo. Ratter J. A., P. W. Richards, G. Argent, and D. R. Gifford. 1978. Observations on the forests of some mesotrophic soils in central Brazil. Revta brasil Bot 1:47-58 Ratter, J. A., and T. C. D. Dargie. 1992. An analysis of the floristic composition of 26 cerrado areas in Brazil. Edinburgh Journal of Botany, 49:235-250. Ratter, J. A. 1992. Transitions between cerrado and forest vegetation in Brazil. Pages 417-419- in P. A. Furley, J Proctor, and J. A. Ratter, editors. Nature and Dynamics of Forest-Savanna Boundaries. Chapman and Hall, London. Ratter, J. A., S. Bridgewater, R. Atkinson, and J. F. Ribeiro. 1996. Analysis of the floristic composition of the Brazilian cerrado vegetation. II: Comparison of the wood) ".gelation of 98 areas. Edinburgh Journal of Botany. 53:153-180. Ribeiro, J. F. 1983. Comparacao da concentracao de nutrientes na vegetacao arborea e nos solos de um cerrado e um cerradao no Distrito Federal. Master's thesis.
83
Departamento de Ecologia, Universidade de Brasilia, Brasilia, Brazil. Richter, D. D., and L. I. Babbar. 1991. Soil diversity in the tropics. Advances in Ecological Research 21:315-389. Rodrigues, T. E., and E. Klamt. 1978. Mineralogia e genese de uma seqiiencia de solos no Distrito Federal. R. Bras. Cienc. Solo. 2:132-143. Ross, S. M. 1992. Soil and litter nutrient losses in forest clearings close to a forest-savanna boundary on Maraca island, Roraima, Brazil. Pages 119-143. in P. A. Furley, J Proctor, and J. A. Ratter, editors. Nature and Dynamics of Forest-Savanna Boundaries. Chapman and Hall, London. Sarmiento, G. 1992. A conceptual model relating environmental factors and vegetation formations in the lowlands of tropical south America. Pages 583-601. in P. A. Furley, J Proctor, and J. A. Ratter, editors. Nature and Dynamics of Forest-Savanna Boundaries. Chapman and Hall, London. Schiavini, I. 1983. Alguns aspectos da ciclagem de nutrientes em uma area de cerrado (Brasilia, DF): chuva, producao e decomposicao de liter. Master's thesis. Departamento de Ecologia, Universidade de Brasilia, Brasilia, Brazil. Silva, F. C. 1990. Compartilhamento de nutrientes em diferentes componentes da biomassa aerea em especies arboreas de um cerrado. Master's thesis. Departamento de Ecologia, Universidade de Brasilia, Brasilia, Brazil. Silva Junior, M. C. 1995. Tree communities of the gallery forests of IBGE Ecological Reserve, Federal District, Brazil. Ph. D. Thesis. University of Edinburgh, Edinburgh, UK. Solbrig, O. T., E. Medina, and J. F. Silva. 1996. Biodiversity and savanna ecosystem processes - a global perspective. Ecological Studies, Vol. 121. Springer-Verlag, Berlin, Heidelberg. Villela, D. M. V., and M. Haridasan. 1994. Response of the ground layer community of a cerrado vegetation in central Brazil to liming and irrigation. Plant and Soil 163:25-31. Vitousek, P. M., and R. L. Sanford. 1986. Nutrient cycling in moist tropical forest. Annual Review of Ecology and Systematics 17:137-167
6 Linking Biogeochemical Cycles to Cattle Pasture Management and Sustainability in the Amazon Basin Moacyr B. Dias-Filho, Eric A. Davidson, Claudio J. Reis de Carvalho
Pasture development has become the observed (reviewed by Serrao and Toledo largest anthropogenic disturbance of forest 1990). If left uncontrolled, these invader land in the Amazon basin (Skole et al. 1994, species slowly become dominant and lead to Serrao and Toledo 1990). The area of forests "pasture degradation," a condition characterconverted to cattle pasture in Amazonia is ized by a complete dominance of the weedy currently estimated at approximately 20 community. If left to secondary succession, million hectares. In the Brazilian Amazon forest vegetation usually becomes reestabbasin, most of the conversion of forest land lished on these degraded pasture lands in the to pasture began during the early 1960s to Amazon, although the species composition is the late 1980s, as a consequence of the usually different than that of the primary opening of Amazon highways and govern- forest (Nepstad 1989). The nutrient status of ment policies aimed at regional develop- the degraded pasture soils is among the ment (Hecht 1982, Nepstad et al. 1991, factors that affect the rate of regrowth of the secondary forests. Serrao et al. 1979). One of the first attempts to study soil Pasture productivity and longevity in the Amazon basin seem to be closely related nutrient dynamics under cultivated pastures to soil fertility and nutrient cycling (e.g., Dias in the Amazon basin was conducted in the Filho and Serrao 1987, Serrao et al. 1979). early 1970s by Falesi (1976). The results of Thus, understanding the major biogeo- that chronosequence study in different soil chemical cycles that influence soil fertility types suggested that soil nutrient cycling in under pasture is vital for predicting the pastures differed from that of the traditional consequences of continued conversion of slash-and-burn agriculture. The decline in the tropical forests to cattle pastures. This under- levels of some nutrients in the soil was found standing is also important for devising to be more gradual, and the decline in management technologies that enhance the productivity over time in planted pastures sustainability of these areas and thus slow could, at least in part, be associated with the behavior of available phosphorus in the soil. further deforestation. Although during the first three to five years Although the sampling methodology of after establishment, the productivity of pas- Falesi's chronosequence study has been tures is often good, after that period a rapid criticized (see Hecht 1982), other studies decline in productivity of the planted grasses considering the dynamics of soil fertility in associated with an increased presence of pasture areas in the Amazon basin found herbaceous and woody invaders is generally similar results (Diez et al. 1991, Hecht 1982,
Linking Biogeochemical Cycles to Cattle Pasture Management and Sustainability in the Amazon Basin
Buschbacher 1984, Moraes et al. 1996, Teixeira 1987). The connection between soil fertility (mainly available phosphorus) and pasture productivity and longevity in the Amazon basin 'was further established in many studies conducted from the late 1970s to the early 1990s (Azevedo et al. 1982a, 1982b; Dias Filho and Serrao 1980, 1981, 1982, 1987, Dias Filho et al. 1989, Dias Filho and Simao Neto 1992, Embrapa 1980, Italiano et al. 1982, Serrao et al. 1979). The promotion of efficient nutrient recycling has been acknowledged as a means of enhancing the stability of pasture systems in tropical America (e.g., Dias Filho 1986, Serrao and Toledo 1990, Spain and Salinas 1985). In the Amazon basin, efficient nutrient recycling is even more important, as most of the soils where pasture development is concentrated are nutrient-poor Oxisols and Ultisols. However, there has been very little progress in understanding major biogeochemical cycles in active and abandoned pastures in the Amazon basin. An attempt to directly address the role of grazing animals in nutrient recycling of an active pasture in Amazonia was made by Buschbacher (1984, 1987). More detailed studies have followed this, however, without directly addressing the role of cattle and grazing, and considering specifically only nitrogen and carbon dynamics in pasture soils (Dejardins et al. 1994, Feigl et al. 1995a, 1995b, Neill et al. 1995, 1997, Nepstad et al. 1994, Trumbore et. al. 1995). We review these studies here, and we attempt to identify the most important issues for our understanding of pasture biochemistry: How does nutrient availability in pastures change over time? What nutrients are most critical? What is the role of cattle and grazing? What are the effects of altered hydrology? How does weed invasion affect pasture biogeochemistry? Also, we attempt to identify what management practices are important to promote pasture longevity and sustainability, and to reclaim degraded pasture areas.
85
Land Clearing Methods and Pasture Establishment The method of land clearing can have important and long-lasting consequences for pasture productivity and longevity (Dias Filho 1986, Dias and Northcliff 1985, Seubert et al. 1977). Land clearing for pasture establishment on forested areas in the Amazon basin normally involves cutting the forest and then burning of the vegetation (Dias Filho 1986, Serrao and Toledo 1990, Toledo and Morales 1979). Clearing is normally done with chain saws, generally after the valuable wood species have been removed. Mechanical clearing for pasture establishment using heavy machinery is seldom practiced in the Amazon basin, mainly because of high operational costs, but also due to the detrimental effects mechanical clearing is known to have on soil properties, reducing pasture productivity (reviewed by Dias Filho 1986). Perhaps one of the earliest and most influential studies that compared the agricultural impact of manual (slash-and-burn) and mechanical (bulldozer) clearing for pasture establishment in the Amazon basin was that of Seubert et al. (1977) in Peru. In that study, pasture (Panicum maximum) biomass production, over a 2-year period, was 68 % higher in the area cleared manually. Other studies have followed, with the general conclusion that pastures manually cleared from forest were more productive and that mechanical clearing adversely affected soil physical and chemical properties (e.g., Alegre 1985, Toledo and Morales 1979, Dias and Northcliff 1985). The basic difference between the two land clearing methods is that in the "manual" process, fire is the major agent employed to remove vegetation cover and prepare the land for pasture establishment. From an agricultural perspective, burning can be regarded as a fast and economical process to
86
easily prepare the area for cultivation. From an ecological viewpoint, however, it causes immediate disruption of major biogeochemical cycles through the production and release into the atmosphere of particulates and environmentally significant gases, and the release into the soil of relatively large amounts of nutrients previously immobilized in the aboveground biomass. Forest biomass burning also affects the hydrological cycle by changing rates of land evaporation and water runoff. Losses from the system due to combustion of biomass can be quite high for carbon, nitrogen, and sulfur, and less for major nutrients like phosphorus, potassium, and calcium (Ewel et al. 1981, Frost and Robertson 1985, Kauffman et al. 1995, Sanchez 1976, Wright and Bailey 1982). However, immediate improvement in the soil chemical characteristics - higher pH, reduced aluminum toxicity, and higher nutrient content—is usually observed after forest cutting and burning in the Amazon basin (Buschbacher 1984, Dantas 1989, Dantas and Matos 1980, Diez et al. 1991, Falesi 1976, Seubert et al. 1977, Smyth and Bastos 1984). The general improvement in soil chemical properties after burning is mainly due to constituents in the ash that provide substantial inputs of bases and other elements to the soil. Also, heating caused by burning may contribute to the improvement of soil fertility by enhancing the mineralization of elements (N, Ca, Mg, and P) formerly linked and complexed with organic matter (Giovannini et al. 1990). During pasture establishment, available soil nutrients are generally at their highest levels due to ash deposition (after forest clearing) or chemical fertilization (in reclaimed pastures). In addition, the soil is left unprotected against high temperatures, which accelerates the decomposition rates of fresh organic matter, and against wind and raindrop impact, which contribute to erosion
Moacyr B. Dias-Filho et al.
and soil compaction. For these reasons, a fast and efficient soil cover by the forage plant is key in determining the future of pasture productive longevity. Any failure in pasture establishment will expose the area to an increased nutrient loss and weed invasion that ultimately may lead to premature pasture degradation. According to Dias Filho (1986), forage seed quality, seeding rates, time of planting and of beginning grazing, and the agronomic characteristics of the forage species are the most crucial aspects affecting successful pasture establishment.
Carbon Cycling The most important effect of forest-topasture conversion on the carbon cycle is the release of 100-200 tons C ha'1 from aboveground forest biomass to the atmosphere. However, soil C stocks are often as large or larger than aboveground biomass-C, and the changes in soil C stocks, although much less than changes in aboveground C, can also be significant (Nepstad et al. 1994). Inventories of soil C stocks following tropical pasture formation have shown increases (Bushbacher 1984, Cerri et al. 1992, Chone et al. 1991, Feigl et al. 1995, Fisher et al. 1994, Neill et al. 1996), decreases (Bushbacher et al. 1988, Desjardin et al. 1994, Detweiler 1986, Eden et al. 1991, Garcia-Oliva et al. 1994, Luizao et al. 1992, Street 1982, Veldkamp 1994), and mixed results (Brown and Lugo 1990, Trumbore et al. 1995). The key to understanding these conflicting responses is variation in the productivity of pasture grasses. A comparison of C budgets for forests and pastures in the eastern Amazon was made by Trumbore et al. (1995). In a reformed and fertilized pasture of Brachiaria brizantha, they estimated gains, relative to forest soil C stocks, of over 20 tons soil C ha~ 1 in the top 1 m of soil and a loss of about 0.5 tons C ha'1 in the 1-8 m soil depth interval during the first 5 years following pasture
Linking Biogeochemical Cycles to Cattle Pasture Management and Sustainability in the Amazon Basin
87
reformation. In contrast, they calculated a 10-100 cm depth interval and to 80-90% loss of about 9 tons soil C ha'1 in the top below 1 m depth. Hence, large changes in 1 m of an abandoned, degraded pasture soil soil C can be observed near the surface in a and no change below 1 m. As shown in matter of years, although modest changes in Table 6.1, total C inputs were greatest in the modern soil C at depth are also significant forest, due to high rates of litterfall, but when summed over the entire profile of 8 or 20-30% of the litter C probably decomposes more meters. to CO2 within the litter layer. Root inputs Losses of soil C were also found in pastures were greatest in the top 1 m of soil in the of the eastern Amazon by Desjardin reformed productive pasture, which resulted et al. (1994), whereas increases of soil C in in gains of soil C stocks there. However, the Amazonian cattle pastures have been reportdeeply rooted woody vegetation had been ed in the "western Brazilian Amazon (Cerri et removed from the reformed pasture, result- al. 1992, Chone et al. 1991, Feigl et al. 1995, ing in very low rates of root inputs to deep Neill et al. 1996). The studies in the western soils, causing losses of deep soil C in the Brazilian Amazon were of pasture soils that productive pasture. Low rates of litter and had not been fertilized. It is tempting to root inputs in the degraded pasture caused interpret these results as an indication that loss of soil C in the topsoil. Hence, both total the soils in the west tend to have greater productivity and the distribution of C inputs native soil fertility than those in the east, thus to the soil among litter and roots of varying characteristically supporting more productive depth are the key to determining whether C pastures without fertilization, but this broad will be lost or gained in topsoil and subsoil. generalization may be too simplistic. It is true These changes in soil C stocks are relative- that the probability of finding eutrophic soils ly rapid, due to rapid turnover of soil organ- increases in a westerly direction toward the ic matter (SOM) in these tropical soils. Using Andes, but both eastern and western regions radiocarbon derived from atmospheric test- have a wide variety of soil types. The studies ing of nuclear weapons in the 1960s as a in the western Brazilian Amazon included tracer, Trumbore et al. (1995) estimated that Oxisols, Ultisols, and Alfisols, but these the mean residence time of C in the top 10 taxonomic classifications are too broad for cm of soil is about 3 years for 30% of the reliable regional extrapolation. Both eutrophSOM and 10-30 years for 60% of the SOM. ic and dystrophic soils occur within the Only about 10% of the C in the top 10 cm of orders of Oxisols and Ultisols. soil cycles on a millennial time scale. This In addition to soil fertility, there are differvery old C fraction increases to 40-80% in the ences in climate and there may be differences
Table 6.1 Comparison of C inputs to soils at Fazenda Vitoria, Paragominas, Brazil. C Inputs
Forest (tons C ha-1)
Degraded Pasture (tons C ha-1)
Productive Pasture (tons C ha-1)
Litter Roots (0-1 m depth) Roots (1-5 m depth) Total
4.5 1.9 0.9 7.3
1.4 1.2 1.3 3.9
2.0 3.4 0.2 5.6
Source: Trumbore et al. 1995
Moacyr B. Dias-Filho et al.
88
in management practices between the eastern and western regions of the Brazilian Amazon basin. Most of the eastern Amazon, including the Paragominas study site of Trumbore et al. (1995), has a 3-5 month dry season that limits pasture productivity due to severe water limitation and that may increase the frequency of fire. Frequent fire would reduce nitrogen stocks that might, in turn, decrease grass productivity. Regional differences in pasture management are partly historical. Forest clearing along the Belem-Brasilia highway began in the 1960s and 1970s (Nepstad et al. 199D, when the use of less productive grass species was common and when less was known about optimal grazing intensity, which resulted in severe overgrazing in many areas. Examples of overgrazing probably exist in all regions, but they may be more common in the eastern Amazon. In summary, we know that pasture productivity is the key to predicting whether soil C stocks will increase or decrease following forest conversion to pasture, and we know that several factors, including native soil fertility, fertilization, climate, fire frequency, and grazing intensity, influence pasture productivity. We do not, however, know which of these factors has had the greatest influence in the past and which
will be most important in the future for determining productivity in Amazonian pastures. Nor do we know the areal extent of degraded v. productive pastures. Quantitative estimates of areas of productive and nonproductive pasture lands would be necessary to calculate basin-wide effects of forest-to-pasture conversion on the regional C budget. Finally, we wish to emphasize that sustainability of pasture management practices and maintenance of soil C stocks are practically synonymous.
Phosphorus Cycling Of all the essential plant nutrients, P is most commonly considered a limiting factor to primary productivity in tropical ecosystems on highly weathered soils. Due to its low mobility in the soil and high stability (no biologically important form of P is gaseous), and because inputs (e.g., wet and dry deposition) are roughly equivalent or even higher than outputs (e.g., erosion and runoff), P exists in relatively constant amounts or even increases over time scales of years to centuries in some natural terrestrial ecosystems. In managed ecosystems, however, net losses of P may be high because of harvest of products like grain,
Table 6.2 Forms of P in the top 10 cm of mineral soil at Fazenda Vitoria, Paragominas, Para. NaOH Extractable Site
Primary forest Degraded pasture Reformed pasture 20-yr Secondary forest
Total P (kgPha-i)
194 250 298 209
Organic-P (kg P ha-1)
Inorganic-P (kg P ha-1)
19 30 36 26
26 30 38 28
Mehlich 3 (kgPha- 1 )
1.6 1.2 1.1 0.8
Source: Unpublished data from D. Markewitz and E. A. Davidson. ' Higher values in the pastures are due more to higher bulk density in the compacted pasture soils than due to higher P concentrations.
Linking Biogeochemical Cycles to Cattle Pasture Management and Sustainability in the Amazon Basin
fiber, and roots. The generally higher vulnerability of these ecosystems to surface runoff and erosion can also contribute to P losses. This condition results from incomplete soil coverage by the crop plant and lower infiltration capacity of disturbed soils. Highly weathered tropical soils of oxic and kaolinitic mineralogies often contain large stocks of total P, but the vast majority of this is bound to iron and aluminum oxides in forms that are generally thought to be un-available to plants. Table 6.2 demonstrates that the stocks of total P in the top 10 cm of the mineral soil of a kaolinitic Oxisol at Paragominas, Para, are about 200 kg ha'1, but that less than 1% of the total P is extracted by the Mehlich 3 extraction method. Agronomists often use the Mehlich extraction as an index of P available to plants, but, clearly, this pool is insufficient to provide plant needs of P and must be replenished frequently from other pools (Mengel and Kirkby 1987). It is unclear whether the P that replenishes the plant-available Mehlich pool is derived from enzymatic decomposition of organic P or from gradual desorption of inorganic P, or both. Below 10 cm depth, the concentrations of total P remain similar to those in the top 10 cm, indicating that P is abundant in these soils. However, the concentrations of NaOH-extractable and
Mehlich-extractable P decline to barely detectable levels below 50 cm depth. Phosphorus is most likely a limiting nutrient for plant growth in those ecosystems. For the great majority of Amazon basin soils, available P in the topsoil is known to be naturally low (Leon and Hammond 1985). After forest clearing and burning, available P in the topsoil usually increases due to ash deposition and slash decomposition (e.g., Falesi 1976, Seubert et al. 1977). However, unlike other nutrients which are often maintained at more or less stable levels under pasture, available P in top soil often falls sharply with time after pasture establishment (Falesi 1976), suggesting that a mining effect might be occurring. This decrease in available soil P is normally associated with a decrease in forage grass biomass production, generally leading to pasture degradation (i.e., an increase in weed biomass) and abandonment. In these degraded pasture areas, forage grasses often respond to P fertilization (Table 6.3) but fail to respond to fertilization with other nutrients (Table 6.4), implying that P availability is the major impediment to pasture productivity. In view of its importance to pasture primary production, maintaining relatively high levels of available P is a major challenge to pasture managers in the Amazon basin. However, insufficient knowledge of the biogeochemistry of P in active Amazonian pastures is a
Table 6.3 Response of Andropogon gayanus to P and N fertilization. Fertilizer (kg ha'1)*
Dry Matter (kg ha-1)**
0N +0P
828 1456 405 2875
0 N + 50 P 75 N + 0 P 75 N + 50 P
89
Source: Dias Filho (1986) * Area near Paragominas, Brazil, on yellow latosol (Haplustox) originally under degraded pasture, reclaimed by burning, tilling, planting, and fertilizing in May 1980 (end of wet season). ** Dry matter production evaluated at 20 cm from the soil surface in February 1981.
Moacyr B. Dias-Filho et al.
90
Table 6.4 Dry matter production of a 12-year-old degraded Panicum maximum pasture after weeding and fertilizer application at Paragominas on a yellow latosol (Haplustox). Treatment Treatment
Dry Matter Matter (kg ha' ha-1)*
Control Legumes interplanted Complete** minus N Complete minus P Complete minus K Complete minus S Complete minus lime Complete minus micronutrients Complete
7428 6127 12508 4654 11120 15667 12183 11630 12249
Source: Bias Filho and Serrao (1987). * Tukey's HSD = 6230. ** Complete = 150 N +100 P + 100 K + 50 S + 1000 Lime + 30 micronutrients (in kg ha'1).
major limitation to achieving this goal. The biogeochemical P cycling of active pastures differs from forest and, to some extent, abandoned pasture and agricultural ecosystems in its complexity and unpredictability. Cattle grazing affects the rate of P movement and distribution within the system and increases its potential for loss. In active pasture systems for beef production, the harvested product is the animal, with a potential to export relatively low amounts of P. Considering an average carrying capacity of 0.8 animal units per hectare per year (1 animal unit = 450 kg of live weight) for a typical Brachiaria brizantha pasture with medium productivity, the amount of P that could potentially be exported from the system in one year, due to meat production, would be around 2.5 kg per ha (based on a body composition of 0.7 % P for beef cattle). Under the management conditions typical of most beef cattle pasture systems in the Amazon basin, animals would get their dietary P from two major sources: mineral supplementation and pasture plants. Involuntary soil intake could also be a potential, but difficult to quantify, input source of P.
Because of the low P content (usually < 0.2 %) of forage plants, P supplementation is commonly practiced in most ranches throughout the Amazon basin, particularly in Brazilian Amazonia. Assuming a stocking rate of 0.8 animal units per hectare per year, the annual input of P to the system through animal consumption of mineral supplementation would be around 2.0 kg, which is close to the amount expected to be exported annually by animal products (2.5 kg). The amount thus needed to balance the cycle would be only 0.5 kg. In the absence of phosphate fertilizer inputs, this P must come from the soil pool reserves, through forage consumption. Under the assumption of a forage dry matter P content of 0.12 %, and a stocking rate of 0.8 animal units per hectare, we can estimate that 3.9 kg of P would be ingested by the animal in one year through forage consumption. This gives a total amount of available P for recycling of 3.4 kg ha'1 yr1 (3-9-0.5). Therefore, most of the P taken from the soil by plant uptake and consumed in the forage by the grazing animal on a daily basis, would be returned to the system through animal excreta.
Linking Biogeochemical Cycles to Cattle Pasture Management and Sustainability in the Amazon Basin
Although it may seem straightforward that there is an efficient P recycling through the return of ingested P in animal excreta, in practical terms this efficiency is highly questionable, because feces and urine are not distributed uniformly over the pasture but deposited on specific areas. Usually, areas close to drinking troughs, mineral feeders, or shade receive greater amounts of feces and urine. For example, in a Brachiaria decumbens pasture in the Venezuelan Amazonia, it was found that more than one-half of the dung excreted by Zebu cattle was concentrated in only 30 percent of the pasture area (Buschbacher 1987). The proximity of cattle gathering areas to streams and ponds produces the potential for runoff of P to significantly alter nutrient cycling in aquatic ecosystems, but this topic has not been studied to our knowledge. In addition to greatly altering P distribution within the pasture area due to the patchy distribution of excretion, cattle can also indirectly enhance the potential of P movement in the landscape. This is because cattle transit within the pasture area can create trails which are channels for runoff, and overgrazing and trampling may result in open areas that facilitate losses through runoff of paniculate and dissolved P and erosion of surface soil and plant material. Since most of the excreted P is in the feces
91
and most of the feces are deposited in areas where cattle transit and concentration are more intense, the potential for P loss from the system into surface waters through erosion and runoff can be greatly enhanced during the heavy rains typical of most areas in the Amazon basin. Comparison of the water infiltration capacity of adjacent forest and reclaimed pasture soils in the Paragominas region, eastern Amazonia, has shown that water can infiltrate up to twenty times slower in the soil under pasture (M. B. Dias Filho, unpublished data). Moreover, P transferred in feces to bare trails and campsites will have little or no effect on forage production and P recycling. Cattle also have a great impact on the internal P (nutrient) cycling of forage species. Because grazing is selective and often the plant parts with the greatest amount of nutrients (i.e., young leaves) are preferentially consumed, there is a profound impact on the resorption efficiency of P, as well as other nutrients. This, in turn, stimulates soil P uptake, and if recycling through animal excretion is not efficient (which is often the case, mainly under low stocking rates) and there is no P fertilizer input, a slow depletion of soil P will usually occur. The amount of P removed from the soil by forage grass in a typical Amazonian pasture can be substantial (Table 6.5). Ongoing
Table 6.5 Annual dry matter production (mean of two years) and P uptake (kg of P in plant tops) of Brachiaria brizantha cv. Marandu. Fertilizer (kg P ha'1)*
Dry Matter (kg ha-1 year1)**
P Uptake (kg ha-1 year1)**
0
13900 (696)
13.7 (1.07)
22
23334 (834)
29.5 (1.75)
44
28159 (1651)
39.7 (2.26)
Source: Adapted from Dias Filho and Simao Neto (1992). * Area at Paragominas, Brazilian Amazonia on yellow latosol (Haplustox) originally under degraded pasture, reclaimed by burning, tilling, planting and fertilizing with simple superphosphate. ** Data are mean (± s.e.) n= 4.
92
Moacyr B. Dias-Filho et al.
Table 6.6 P and nitrogen concentration* (% of dry weight) of forage grass (Hyparrhenia rufa), herbaceous weedy vegetation and litter** of a 10-year-old degraded pasture in eastern Brazilian Amazonia. Source
P
N
Grass Weeds Litter
0.10 0.20 0.06
0.63 1.70 0.78
Source: Dantas (1989). * Mean of three sample periods during 2 consecutive years. ** Dead and dry material from many species (mainly weeds).
research (C. J. R. de Carvalho and M. B. Dias Filho, unpublished data) has shown that forage grasses like Brachiaria humidicola and B. brizantha have the ability to respond to low available P in the soil by secreting acid phosphatase at the rhizosphere. Acid phosphatase can solubilize organic phosphate in the soil, increasing the availability of P for plant uptake. If not grazed, or if the pasture is not burned, most of the P absorbed in plant tops will be immobilized by plant tissue and, thus, will be unavailable in the soil. Transfer of P and N by nutrient resorption in forage plants and in the weedy vegetation prior to leaf fall induces low nutrient concentration in litter (Table 6.6) and may in part explain the decreasing amount of available soil P often associated with an increase in the weed biomass of a degrading pasture or with pasture age. Without this biological immobilization, however, these nutrients would be more vulnerable to loss through runoff and erosion. Immobilization of P in the soil in the short or long term by geochemical fixation processes might, in some situations, play an important role in the cycling of P in pasture ecosystems. However, unlike in the acid savanna soils, where high P fixation is prevalent, only a relatively small proportion of Amazon soils (mainly Oxisols and Ultisols of
clayey topsoil texture) are known to have the capacity to fix large quantities of P into relatively insoluble forms (Sanchez 1987). For this reason, P fixation has received far less attention than the role of the biologically mediated organic P transformations in this ecosystem. Geo-chemically fixed P is probably not available for plant uptake and subsequent animal consumption, but could be vulnerable to loss from the system via soil erosion. In summary, we know that available P is often found in low amounts in the great majority of Amazon basin soils. After forest clearing and burning, available P levels increase in topsoil, but during pasture utilization these levels often fall sharply. This decrease in P availability is usually associated with a decline in pasture productivity and an increase in weed biomass. The accumulation of weed biomass sequesters soil P and, thus, temporarily decreases its availability in the soil. In these degraded pastures, forage grasses often respond to P fertilization but fail to respond to fertilization with other nutrients, strongly suggesting that P availability is the major limitation to pasture productivity. Cattle probably play a key role in the P dynamics of active pastures areas by affecting P distribution patterns within the pasture area and facilitating P losses through runoff and erosion.
Linking Biogeochemical Cycles to Cattle Pasture Management and Sustainability in the Amazon Basin
93
Calopogonium spp., Desmodium spp.) can be found in most active pasture areas, and their Land use changes, such as conversion contribution to the N budget is unknown. of tropical forests to cattle pastures, affect When degraded pastures are reformed and biological N fixation, mineralization, nitrifi- fertilized with P, grass biomass production cation, and denitrification (Davidson et al. can be substantial (Tables 6.4 and 6.5), and 1993, Keller and Reiners 1994, Matson et al. it is difficult to believe that all of the N in 1987, Montagnini and Buschbacher 1989, this new plant biomass ( > 100 kg N ha*1) Neill et al. 1995). Matson et al. (1987) esti- could be derived from nutrient-poor degradmated that N mobilized annually from ed pasture soil. Perhaps symbiotic N fixation deforestation was equivalent to more than by the spontaneous herbaceous legumes half of the industrial N fixed globally and or associative N fixation in the grass rhizosgreater than the total amount of N delivered phere may be necessary to supply sufficient to oceans by rivers. N to meet plant needs. If N fixation does Tropical ecosystems on highly weathered occur in active pastures, that may explain soils present an enigma to our current the absence of a response in grass growth to understanding of the N cycle. Based on N fertilization. ratios of elements found in foliage and litAssociative N fixation, the process in terfall of both leguminous and nonlegumi- which free-living N fixing bacteria utilize nous trees, N appears to be an abundant root exudates as energy sources and can nutrient relative to P in forest vegetation maintain active nitrogenase enzymes in growing on many soils dominated by the low O2 environment of the rhizosphere, kaolinitic and oxic mineralogies (Vitousek has been inferred for tropical grasses such and Sanford 1986). And yet, despite the as sugar cane (Dobereiner et al. 1972) apparent relative abundance of N in these and several species of pasture grasses ecosystems, leguminous trees are abundant (Boddey and Victoria 1986). Conditions in in these forests, sometimes being nodulated the grass rhizosphere may be conducive to and sometimes not nodulated (Salati et al free-living N fixing bacteria, including 1982). Do these legumes allocate energy to abundant root exudates, respiration that N fixation despite high availability of N in consumes O2, soil pH usually in the range these ecosystems? Is N abundant in these of 6 to 7, and sufficient P from fertilization. forest ecosystems because symbiotic N fixa- Given the high productivity of these tion exceeds plant demands? (See Cuevas, tropical grasses, it seems plausible that their this volume, for a thorough discussion of rhizosphere N fixation could be significant, nutrient cycling in Amazon forests.) but conclusive evidence for associative Biological N fixation may also be impor- N fixation in pastures is still lacking. tant in active and abandoned pastures, Chronosequence studies in Rondonia although the evidence for it is mostly showed similar or higher N stocks in the inferential. In abandoned pastures, nodulat- surface soil of pastures compared to ed leguminous species are common early forests, although spatial heterogeneity made successional plants, suggesting that symbiot- it difficult to conclude that significant ic N fixation is advantageous for growth in increases in N stocks due to associative N these degraded soils. The use of planted fixation in the pastures had occurred grass-legume pastures is not common in (Piccolo et al. 1994). the Amazon, but spontaneous growth of Comparisons of total N stocks may belie herbaceous legumes (e.g., Centrosema spp., more important differences in "plant-avail-
Nitrogen Cycling
Moacyr B. Dias-Filho et al.
94
able N" among forest and pasture soil. Total ed pasture (Table 6.7). These results provide N is a poor indicator of plant-available, further inferential evidence that biological N actively cycling N pools. Preliminary results fixation or some other N inputs result in from studies at Paragominas (L. V. Verchot partial recovery of the N cycle in secondary and E. A. Davidson, unpublished data) forests and improved pastures. Nutrients mineralized from soil organic show that net N mineralization and net nitrification rates are low in the pastures matter during forest clearing usually support compared to mature forest (Table 6.7). The shifting agriculture or productive pastures NO and N2O flux data follow the same trend for only a few years in many tropical as the net N mineralization data: primary regions, resulting in nutrient-poor, degraded forest > secondary forest > improved pas- soils in abandoned farms and pastures ture > degraded pasture (Table 6.7). (Keller and Reiners 1994, Tiessen et al. 1994, Because N gas emissions are strongly affect- Serrao and Toledo 1990). In the case of ed by N cycling rates and N availability in N, there are several mechanisms by which the soil (Firestone and Davidson 1989), N can be rendered unavailable, redistribthese data also imply decreased N availabil- uted, or lost from pasture soils. Root systems of tropical pasture grasses ity in the pastures. In Rondonia, a similar pattern has been observed, in which cattle and decomposing litter are known to pastures ranging from 3 to 80 years old had immobilize significant amounts of N lower rates of net N mineralization and net (Bushby et al. 1992, Robbins et al. 1989), nitrification than did adjacent forest soils perhaps into organic-N forms that are not readily remineralized. This immobiliza(Neill et al. 1995). The secondary forest at Paragominas has tion may induce N deficiency by reducing recovered much of the net N mineralization both pasture and animal production in some capacity that is characteristic of primary tropical pasture ecosystems (Robbins et al. forest, and the improved (reformed) pasture 1989). It is expected that an increase in stockhas higher net nitrification than the degrad- ing rate or pasture burning, by reducing the
Table 6.7 Mineralization indices and N trace gas emissions for soils from primary forest, secondary forest, reformed pasture, and degraded pasture at Fazenda Vitoria.* Land Use
NetN Mineralization* * (mg N g-i 7d-J)
Primary forest
2.57 (0.62) a
2.25 (0.22) a
2.5 (0.7) a
1.2 (0.3) a
Secondary forest
1.55 (0.39) ab
0.30 (0.29) be
0.9 (0.2) b
0.3 (0.1) b
Net Nitrification** (mg N g-1 Td-1)
N2O Emissions*** (kg N ha-1 yr1)
NO Emissions*** (mg N g-i 7d-i)
Improved pasture
0.97 (0.09) be
0.46 (0.09) b
0.3 (0.3) b
0.8 (0.3) b
Degraded pasture
0.33 (0.11) c
-0.22 (0.05) c
0.0 (0.1) b
0.7 (0.3) b
Source: L. V. Verchot and E. A. Davidson, unpublished data. * Means and (s.e.) within a column followed by the same letter are not statistically different at = 0.05 by Duncan's Multiple Range test. ** Net N mineralization and net nitrification values were obtained from laboratory incubations (7 days) at room temperature. *" Annual trace gas emissions are extrapolated from measurements on 9 dates during the wet season and 6 dates during the dry season.
Linking Biogeochemical Cycles to Cattle Pasture Management and Sustainability in the Amazon Basin
amount of litter, can reverse this trend, but this mobilization of N would also make the N in the system more prone to loss from the ecosystem. Nitrogen losses due to the direct effect of the grazing ruminant have usually not been accounted for in studies of N dynamics in active Amazonian pastures (but see Buschbacher 1987). Cattle interfere in the internal N cycle of pasture plants, alter nitrogen distribution in the pasture area, and change botanical composition, and these effects are important for the N cycle of active pastures. Much of the N loss that occurs in active pastures is accounted for by the heavy deposition of excreta around campsites, drinking and mineralization troughs, and cattle trails. The greatest potential for N loss are in urine patches which may receive very high amounts of N (> 300 kg N ha^1). Nitrogen returned in urine is rapidly hydrolyzed to ammonia, becoming available for plant uptake. However, because of the very high amount of N in urine patches, plant uptake and utilization is usually not efficient, and large amounts of N may be lost through ammonia volatilization, which is further enhanced by the high temperatures at ground level characteristic of tropical pastures. As pasture utilization increases it is expected that N losses from the system would also increase as a result of more N passing through the animal and being excreted. For a typical active B. brizantha beef cattle pasture in Brazilian Amazonia with an average carrying capacity of 0.8 animal units per hectare per, the amount of N that would be converted in live weight gain and potentially exported from the system in one year as animal product would be about 7 kg N ha*1 (based on a body composition of 2% N for beef cattle). Assuming an annual herbage consumption of about 3300 kg (9 kgd'1 x 365d) of B. brizantha dry matter per hectare, we can estimate that 33 kg of N are ingested by the animal in one year
95
(based on an average N content of 1% in B. brizantha dry matter). Considering that only about 10% of the N ingested by the animal is retained, 30 kg of the ingested N is excreted in urine and dung. If a 30% recovery of excreta N by pasture plants is assumed (Ball et al. 1979, Ledgard and Sanders 1982, Ledgard et al. 1982), then about 20 kg of N returned to each hectare of pasture via excreta could potentially be lost during the period of one year through volatilization of ammonia, leaching, erosion, and denitrification. Since N fertilization and protein or urea supplementation are not common management practices in beef cattle pasture systems in the Amazon basin, N inputs are mostly restricted to biological N fixation and additions in precipitation. Loss of N from the system must be balanced by atmospheric deposition, N fixation, or depletion of soil N through mineralization of soil organic matter. Increases in emissions of NO and N2O have been observed in young tropical cattle pastures relative to forests (Keller et al. 1993, Luizao et al. 1992), which is due to an initial pulse of net N mineralization following forest clearing that provides mineral N for nitrifying and denitrifying bacteria (Firestone and Davidson, 1989)- However, Keller and Reiners (1994) found that N gas emissions declined below levels of the primary forest in pastures that were older than 10 years in the Atlantic coastal plain of Costa Rica. There are indications that the period of enhanced N gas emissions in young pastures of the Amazon may be more brief (i.e., only 1—3 years) than "was observed in Costa Rica (personal communication of preliminary results from Rondonia by Paul Steudler and from Para by Louis Verchot). The reason for this difference is unknown, but the volcanic soils of Costa Rica may be more fertile with respect to N availability. If it is true that N2O emissions from soils of most Amazonian pastures
96
Moacyr B. Dias-Filho et al.
that are > 3 years old are lower than emissions from the primary forests (Table 6.7), then one effect of deforestation at the basin scale may be a net reduction in N2O emissions. Seasonality of precipitation is also an important factor controlling N trace gas emissions from Amazonian soils. At Paragominas in the highly seasonal eastern Amazon, N2O is the most important N gas during the wet season when diffusivity is restricted, thus favoring the more reduced N gas form (Table 6.8). During the dry season, the more oxidized NO form is the dominant N gas emitted. Differences in total N gas emissions among land uses is controlled by N availability as indicated by net N mineralization and net nitrification (Table 6.7), while the seasonal variation in soil moisture content provides the controls over the relative proportions of NO and N2O emissions (Table 6.8). In addition to affecting trace gas emissions, lower N availability in old pastures relative to primary forests may also have implications for stream water chemistry. Preliminary results from studies in Rondonia and Para indicate very low concentrations of NO3 in first-order streams draining pastures (C. Neill and D. Markewitz, personal communication). In summary, N cycling processes clearly change when forests are converted to cattle pastures. In Amazonia, soil N availability
generally declines (except perhaps during a brief period of high N availability immediately after clearing), which results in lower emissions of N trace gases and probably lower hydrologic export of NO3. Direct export of N via the cattle is estimated at about 7 kg N ha"1. Cattle are also important in redistributing N within the pasture by concentrating N in excreta around campsites, drinking troughs, and cattle trails, where the potential for further N loss may be as high as 20 kg N ha"1. It is not clear whether the N needs of pasture grasses are provided by associative N fixation or if the most successful introduced grass species are simply well adapted to obtaining N from nutrient-poor soils. Similarly, it is unclear whether reduced N availability in old pasture soils limits rates of regrowth of secondary forests upon pasture abandonment, or if N fixation rates in young secondary forests keep up with demand for N while some other factors limit forest regrowth. For much of the Amazon basin, N fertilization does not appear necessary, but nevertheless, the N nutrition of pasture grasses needs to be better understood for developing sustainable management practices. Moreover, altered N cycling processes following land use change clearly affect the exchange of N between the atmosphere and the biosphere and between terrestrial and aquatic ecosystems.
Table 6.8 Preliminary results for seasonal variation in N gas emissions at Fazenda Vitoria, Paragominas, Brazil, 1995. March(Wet Season) N2O NO (ng N cm-2 hr1) (ng N cnr2 hr1) Primary forest Secondary forest Active pasture Degraded pasture
2.6
1.6 1.0 0.7
1.0 0.4 0.9 0.5
Source: L. V. Verchot and E. A. Davidson, unpublished data.
September(DrySeason) N20 NO (ng N cnr2 hr1) (ng N cm'2 hr-1)
0.6 0.3 0.1 0.1
1.5 0.3 0.2 0.4
Linking Biogeochemical Cycles to Cattle Pasture Management and Sustainability in the Amazon Basin
Acidity and Base Cations The effective cation exchange capacity (ECEC) of the variable-charge kaolinitic and oxic soils of the Amazon is typically low, and what little ECEC exists is pH dependent. At pH 4 to 5 (in water), which is characteristic of most of the primary forest soils, the ECEC is often in the range of 2-3 cmol charge per 100 g topsoil. Depending on the soil type, 20-80% of the ECEC may be base saturated (with the lower end of this range being more common), and the remainder is exchangeable acidity (mostly aluminum). When the forest is cleared and the site is burned, the pH of the new pasture soil often increases to 6 or 7. The ECEC often doubles and the base saturation increases. In effect, a significant fraction of the base cations (Ca, Mg, K) that were present in the above ground biomass of the forest is retained in the soil as a result of the increased ECEC caused by the change in pH of these variable charge soils. While observations of increased base cation content in the soil following fire are common (Andreux and Cerri 1989, Ewel et al. 1981, Kaufman et al. 1995, Moraes et al. 1996, Sanchez et al. 1983) cation budgets that trace masses of nutrients from the primary forest biomass to the soil and then to pasture or secondary forest vegetation are lacking. Kaufman et al. (1995) calculate that < 10% of the base cations in the aboveground forest biomass are lost during forest clearing and burning in Para and Rondonia, but they caution that most of the cations were found in the ash immediately after the fire, which could be prone to subsequent loss by erosion. On the other hand, if the ash becomes incorporated into the soil, the increased pH-dependent ECEC of the soil may be sufficient to retain those nutrients. Studying chronosequences of pastures in Para and Mato Grosso, Falesi (1976) found that the pH, ECEC, and base saturation of the topsoil remained elevated in all pastures. The
97
oldest pasture he studied was 11 years. These results indicate that, provided serious erosion is avoided by good pasture management, base cations can be retained. A similar result was observed by Moraes et al. (1996) in two pasture chronosequences in Rondonia. The pH and the sum of base cations remained elevated in pasture soils relative to the forest in all cases except an 81-year-old pasture, where the pH was still slightly elevated and the sum of base cations was about the same as the primary forest. They also show a trend of gradual decreasing pH and sum of base cations in pastures older than 5 years. Whether the gradual loss of base cations is due to leaching, erosion, repeated fires, or harvest exports is not known. Hence, base cations are not retained by the soil indefinitely, but it appears that losses often do not become significant for a decade or more. Fertilization trials conducted throughout the Amazon basin have shown no significant responses of unproductive (i.e., degraded) pasture areas, ranging from 8 to 13 years old, to liming or potassium amendments (Dias Filho and Serrao 1987, Serrao et al. 1979), reflecting the adequate pH and base saturation status of these areas. As in the case for P and N, cations are obtained by foraging over a large area and are then deposited as excreta on a small area, thereby becoming concentrated, principally near waterers, mineral feeders, and campsites. Potassium is mostly excreted in the urine, while the major pathway of excretion for calcium and magnesium is in the feces (reviewed by Barrow 1987). About 90% (circa 70 kg ha"1 yr1 for a B. brizantha pasture •with a medium productivity) of the ingested potassium is returned to the soil via excreta in an immediately available form (K). Because it is excreted in a soluble form, highly concentrated in urine patches, small amounts of K can be lost from soil through leaching if it is not efficiently absorbed by roots or retained in the soil. Pasture over-
98
Moacyr B. Dias-Filho et al.
grazing and the premature grazing of recent- ary forest growth is limited by the availability of ly burned areas can greatly contribute to these nutrients. Intuitively, it seems that growth increase these losses by directly interfering in of pasture and forest vegetation on highly the uptake efficiency by pasture species and weathered, acidic soils with low cation by increasing the potential of erosion and exchange capacity should be limited by the leaching. Periodic pasture burning could aid availability of base cations, but Amazonian pasin the cycling of base cations, particularly the ture soils appear surprisingly capable of retainless mobile Ca that is immobilized in slash ing base cations, and the same may be true for (woody residue from forest), litter, and soils of secondary forests. standing weedy vegetation. On the other hand, repeated fires or fires too late in the Hydrological Cycle dry season, when pasture recovery would be slow, could make base cations (and other The modification of natural ecosystems, as when all the plant cover is substituted nutrients) more prone to loss by erosion. For beef cattle pasture systems in the during conversion of forests to cattle pasAmazon basin, only calcium, among the base tures, affects the transport of water within cations, is normally added to the system via these systems. These changes result from pasture management practices. This is modifications of turbulence, radiation balbecause animal supplementation with bone ances, temperature, and humidity. Pasture meal or bicalcium phosphate is commonly canopies are more uniform than forests, have practiced in most beef cattle systems a more smooth surface, and, therefore, pasthroughout Amazonia. For a typical active tures have a different microclimate (Grace pasture with a medium carrying capacity, the 1994). The difference in the conditions amount of calcium entering the system in between pastures and forests are greatest in one year through mineral supplementation the dry season. Measurements made above would be around 4.0 kg per hectare which is the canopy in active pastures of B. brizantha close to the amount expected to be exported in Ji-Parana, Rondonia, during the dry season in animal products. show that the specific humidity deficit (D) We are not aware of published results on the above the surface in an area of pasture is changes in soil acidity and base cation reten- higher (19.0 g kg"1), than that of a forest area tion in secondary forests of the Amazon. (12.5 g kg-1) (McWilliam et al. 1996). Close to Preliminary results of a study of a 20-year-old the soil surface, D is even higher. In a forest at Paragominas indicate that soil pH, degraded pasture in the region of ECEC, and base saturation remain elevated in Paragominas, mean maximum values of D 10 the top 20 cm of mineral soil (D. Markewitz cm from the soil surface were in the order of and E. A. Davidson, unpublished data). 24.7 g kg"1 in the dry season, while below However, this forest is one of the sites the canopy of an adjacent primary forest classified by Uhl et al. (1988) has having had these values were not above 6.0 g kg'1. The "light use" while in the pasture phase. It is pos- higher values of D in areas of degraded sible that more intensive use of pastures that pastures were a result of the high temperamight result in greater harvest export, greater tures and low air humidity typical of these erosion, or a longer period for leaching may areas during the dry season (Nepstad et al. result in depleted base cation stocks upon pas- 1996). During the rainy season, the ture abandonment. Woody forest vegetation maximum temperatures and D are similar requires large quantities of cations, particularly between forest areas and abandoned Ca, but we do not know if the rate of second- pastures (Nepstad 1989). Studies conducted
Linking Biogeochemical Cycles to Cattle Pasture Management and Sustainability in the Amazon Basin
during many years in Amazonia show that the substitution of forest by pastures has caused a reduction of 11% in net radiation, due mainly to differences of albedo and long-wave net radiation (Wright et al. 1992, Gulf et al. 1996). Mobilization of "water from the soil is closely related to root depth and root density in each layer of soil. Fine roots of active B. brizantha pastures, established in deeply weathered clayey soils in eastern Amazonia, reach depths of 8 m or more (Nepstad et al. 1994). In abandoned pastures (50% B. humidicola and P. maximum cover and 50% invading shrubs and small trees), fine roots ( < 1 mm in diameter) were found at depths of 12 m (Nepstad et al. 1994). Fine-root biomass in the superficial soil layers of an active pasture in Paragominas, eastern Amazonia, was 3 times higher than that found in an adjacent primary forest area. Fine root biomass in the active pasture decreased by a factor of 100 between the surface and 6 m depth. In an abandoned pasture area, the distribution pattern of fine-root biomass was similar to that observed in the deeper soil layers of the forest ecosystem. This pattern is associated with the fine roots of the existing dicotyledonous invading species. The monitoring of the soil plant available water (PAW) between 0 and 8 m depth in the soil has shown that during the dry season, in active B. brizantha pastures areas, a considerable use of water reserves occurs primarily in the top 2 m of soil. This is similar to observations made in primary and secondary forest areas. However, below 2 m, the depletion of the soil water reserves was greater in the forest ecosystem (Jipp et al. 1998). In general, active pasture ecosystems have a greater proportion of fine roots in soil layers between 0—2 meters, and the water in these layers is depleted more quickly, while a major part of the water reserve in the soil is stored in deeper layers. As the pasture
99
is invaded by dicotyledonous species, which often have deeper root systems, nutrient cycling may be intensified as water reserves (and nutrients) in deep soil layers are accessed. For example, the surface soil below Cordia multispicata plants in degraded pastures is richer in Ca, Mg, and K relative to surface soil under other species as a result of inputs from the nutrient rich Cordia litter (Vieira et al. 1994). It has also been shown that in Cecropia palmata, a pioneer species that maintains high stomatal conductance, P levels in leaf tissue are always high, even in dry periods (Denich 1989). These results suggest that the ability of these species to access soil water and maintain a transpirational flux may also be influencing their ability to absorb soil nutrients. The progression of plant succession through the formation of "vegetation islands" and later secondary forest greatly changes the hydrological cycle at a local scale. This can be demonstrated by the similarity between the water depletion profiles obtained in secondary and primary forests (Jipp et al. 1998). If the reduction in plant-available water below 4 m depth in pastures is small and the losses of rainwater through grass canopy interception are also low, it is possible that there is a greater infiltration to the water table from pasture soils, as well as nutrient losses through erosion and runoff during the rainy season (Dunin 1987, Nepstad et al. 1994). Most of the water lost to the atmosphere in the pasture system is subjected to stomatal regulation, which is influenced by environmental variables such as available water in the soil, and D. In Manaus, central Amazonia, at the beginning of the dry season, very high values for stomatal conductance (gs) at mid day were observed in the forage grass B. decumbens. These reached extremes of 1.0 mol nr2 s"1, without a correlation between gs and D. However, these values fell drastically in response to
Moacyr B. Dias-Filho et al.
100
reductions of soil available water (Roberts et al. 1996). On the other hand, the values for maximum gs obtained in pastures of P. maximum during the rainy season in Maraba, eastern Amazonia, rarely exceed 0.40 mol nr2 s"1, in the period between 10:00 am and 2:00 pm. The most frequently observed values were around 0.25 mol rrr2 s"1. Stomatal behavior of P. maximum showed a negative correlation with an increase of D in the air in the dry season. The lowest values of gs (0.10 mol rrr2 s'1) were observed in periods when D was greater than 15 g kg"1- In B. brizantha pastures, no clear relationship between gs and D was observed. However, the effect of the reduction of soil available water was drastic, causing the gs close to midday to be reduced from values of 0.62 mol rrr2 s"1 to only 0.17 mol rrr2 s'1 at the beginning of the dry season (MeWilliam et al. 1996). Under normal conditions of water availability, values for LAI in pastures of B. brizantha have been measured above 4.0. However with the establishment of a water deficit in the soil, these values decrease to below two or even lower in pastures of P. maximum (Roberts et al. 1996). A similar situation is found in abandoned pastures in eastern Amazonia, where a reduction of approximately 68% of green tissue has been observed in the dry season, while in an adjacent area of primary forest this reduction was only 16% (Nepstad et al. 1994). Primary forests, which have deep root systems and little seasonal variation in LAI, maintain stable subcanopy microclimatic conditions and transpirational flux, even during the dry season. Because of an evergreen forest canopy, the return of the rainy season has less impact on the microclimate near the soil in the forest than in the pastures, and the deep soil water stores are also more efficiently recharged in the forest. In the active pastures, the response to environmental water deficits can be influ-
enced by the grass species. In general, however, gs values are normally high when water supply is adequate, but decreases abruptly in situations when both high evapotranspiration demand and low rainfall occur. Despite stomatal control of water vapor loss, the drying of most aboveground pasture biomass is usually observed in areas with an intense dry season. This leads to a dramatic reduction in the flux to the atmosphere of water stored in soil layers below 4 m. This desiccation and reduction in LAI exposes pasture areas to more severe conditions of temperature, wind, and humidity, which consequently cause greater disposition to fire. After the beginning of the rainy season, the upper layers of the pasture soil become saturated quickly, creating a risk of surface runoff, erosion of sediments and nutrients to rivers, and increased water loss to groundwater seepage. In abandoned pastures invaded by deep-rooted dicotyledonous shrubs, some of the plant-available water below the rooting depth of the grasses can be transpired. As secondary forests grow, the hydrologic characteristics of the forest become more like those of the primary forest.
Conclusions The biogeochemistry of active pastures is perhaps one of the least studied processes within managed systems of the Amazon basin. Although the dynamics of the major nutrients in this ecosystem are fairly well known, mostly through agronomic trials conducted during the last two decades, the biogeochemical processes governing these dynamics are still poorly understood. One of the greatest gaps in understanding the biogeochemistry of active pasture systems concerns the role of the grazing cattle in affecting the nutrient cycling processes. Recent studies conducted in active Amazon basin pastures have emphasized the dynam-
Linking Biogeochemical Cycles to Cattle Pasture Management and Sustainability in the Amazon Basin
ics of N and C without paying sufficient attention to important pasture management issues like stocking rate or the frequency (and methods) of weed control. Phosphorus, probably the most important element for pasture primary production in the Amazon basin, has received far less attention than N and C. Most of what we know about phosphorus cycling in active Amazonian pastures is derived from agronomic trials conducted in the Brazilian Amazon and extrapolation of research results obtained in other regions. The increased vulnerability of bioactive elements to loss from the pasture ecosystem is often believed to be one of the greatest differences between this and the antecedent primary forest ecosystem. While this is true during the pasture establishment phase, when the soil is left unprotected from rain, wind, and excessive solar heat, it has been observed that well-managed pastures are usually capable of maintaining soil nutrients at rather constant levels through time, the exception being available phosphorus. In many circumstances, low-productivity pastures can be reclaimed with relatively small additions of phosphate fertilizer, strongly suggesting that all other major plant nutrients (and soil pH) are present in adequate levels in the soil. While phosphorus nutrition can be met relatively easily with fertilization, another challenge for developing sustainable management practices in pastures of the
101
Amazon basin is to balance the need for and means of weed control in order to avoid land degradation. Weed control for pasture reclamation often includes the use of herbicides and/or the removal of tree root biomass by bulldozers, which prevents resprouting of treelets. This practice, along with frequent use of fire, may very likely delay the rate of regrowth of secondary forests, if the pasture is eventually abandoned. Hence, while some management practices, like maintaining good ground cover and minimizing erosion, are salutary for both agroecosystems and successional forest ecosystems, other practices, like bulldozing and herbiciding, may slow ecosystem recovery and commit the site to long-term dependency on intensive inputs to maintain pasture productivity. Intensification of pasture management, if done soundly, may lead to sustainable and permanent pasture usage for some areas. If done poorly, it may lead to impoverished and degraded lands. As abandoned pasture areas become increasingly invaded by shrubs and small trees, which in time lead to secondary forest formation, the biogeochemical cycles of plant nutrients and the hydrological cycle are expected, eventually, to resemble the cycles originally found in the primary forest. The rate of recuperation of biogeochemical cycles in secondary forests and the factors that influence those rates, however, deserve further attention.
Literature Cited Alegre, J. C. 1985. Effects of land clearing and preparation methods on soil physical and chemical properties and crop performance on an Ultisol in the Amazon Basin. Ph.D. thesis, Raleigh, North Carolina State University. Andreux, R, and C. C. Cerri. 1989. "Current trends in the research on soil changes due to deforestation, burning and cultivation in the Brazilian tropics." Toxicology and Environmental Chemistry 20/21: 275-283. Azevedo, G. P. C. de, A. P Camarao, J. B. da Veiga, and E.A.S. Serrao. 1982a. "Introdu^ao e avaliagao de forrageiras no
municipio de Maraba-PA." Belem, EMBRAPA-CPATU. p. 12. (EMBRAPA-CPATU. Boletim de Pesquisa, 46). Azevedo, G. P. C. de, A. P Camarao, J. B. da Veiga, and E.A.S. Serrao. 1982a. 1982b. "Introducao e avaliacao de forrageiras no municipio de Sao Joao do Araguaia, Estado do Para." Belem, EMBRAPA-CPATU. p. 23. (EMBRAPA-CPATU. Boletim de Pesquisa, 47). Ball, P. R., D. R. Keeney, P. W. Theobald, and P. Nes. 1979. "Nitrogen balance in urine-affected areas of a N.Z. pasture." Journal of Agronomy 71: 309—314.
102
Boddey, R. M., and R. L. Victoria. 1986. "Estimation of biological nitrogen fixation associated with Brachiaria and Paspalum grasses using 15N labeled organic matter and fertilizer." Plant and Soil 90: 265-292. Borrow, N. J. 1987. "Return of nutrients by animals." In: Managed Grasslands: analytical studies, Ecosystems of the world; 17B, ed. R. W. Snaydon (Elsevier, Amesterdam), pp. 181-186. Brown, S., and A. E. Lugo. 1990. "Effects of forest clearing and succession on the carbon and nitrogen content of soils in Puerto Rico and U.S. Virgin Islands." Plant and Soil 124: 53-64. Buschbacher, R. 1984. Changes in productivity and nutrient cycling following conversion of Amazon rainforest to pasture. Ph.D. Dissertation, University of Georgia, Athens. Buschbacher, R. 1986. "Tropical deforestation and pasture development." Bioscience 36: 22-28. Buschbacher, R. 1987. "Cattle productivity and nutrient fluxes on an Amazon pasture." Biotropica 19: 200-207. Buschbacher, R., C. Uhl, and E. A. S. Serrao. 1988. "Abandoned pastures in eastern Amazonia II. Nutrient stocks in the soil and vegetation." Journal of Ecology 76: 682-699. Bushby, H. V. A., I. Vallis, and R. J. K. Myers. 1992. "Dynamic of C in a pasture grass (Panicum maximum var. trichoglume)-soi\ system." Soil Biology and Biochemistry 24: 381-387. Cerri, C. C., B. Volkoff, and F. Andreux. 1991. "Nature and behavior of organic matter in soils under natural forest, and after deforestation, burning and cultivation, near Manaus." Forest Ecology and Management 38: 247-257. Cerri, C. C., J. F. L. Moraes and B. Volkoff. 1992. "Dinamica do carbono organico em solos vinculados a pastagens da Amazonia Brasileira." Rev. INITA Inv. Agr. No. 1, Tomo I, pp. 95-102. Charley, J. L., and B. N. Richards. 1983. "Nutrient allocation in plant communities: mineral cycling in terrestrial ecosystems." In: Physiological plant ecology IV: ecosystem processes, mineral cycling, productivity and man's influence, ed. Vol. 12D, eds. O. L. Lange, P. S. Nobel, C. B. Osmond, and H Ziegler, (Encyclopedia of plant physiology; new ser.), pp. 5-45. Chone, T., F. Andreux, J. C. Correa, B. Volkoff, and C. C. Cerri. 1991. "Changes in organic matter in an oxisol from the central Amazonian forest during eight years as pasture, determined by 13C composition." In: Diversity of environmental biogeochemistry, ed. J. Berthelin (Elsevier, New York), pp. 307-^05. Culf, A. D., J. L. Esteves, A. de O. Marques Filho, and H. R. da Rocha. 1996. "Radiation, temperature and humidity over forest and pasture in Amazonia." In: Amazonian deforestation and climate, eds. J. H. C. Gash, C. A. Nobre, J. M. Roberts, and R. L. Victoria (John Wiley & Sons, London), pp. 175-191. Dantas, M. 1989. Studies on succession in cleared areas of the Amazon rain forest. Ph.D. dissertation. Oxford University, Oxford. Dantas, M., and A. de O. Matos. 1980. "Estudos fitoecologicos do tropico umido Brasileiro: III. Conteudo de nutrientes em cinzas de floresta e capoeira, Capitao Poco, PA." Belem, EMBRAPA-CPATU. p. 23. (EMBRAPACPATU. Boletim de Pesquisa, 24).
Moacyr B. Dias-Filho et al. Davidson, E. A., P. A. Matson, P. M. Vitousek, and R. Riley. 1993. "Processes regulating soil emissions of NO and N2O in a seasonally dry tropical forest." Ecology 74: 130-139. Denich, M. 1989. "Estudo da importancia de uma vegetacao secundaria nova para o incremento da produtividade do sistema de producao na Amazonia oriental brasileira." EMBRAPA/CPATU—GTZ, Belem, Brasil. Desjardins, T., F. Andreux, B. Volkoff, and C. C. Cerri. 1994. "Organic carbon and 13C contents in soils and soil size-fractions, and their changes due to deforestation and pasture installation in eastern Amazonia." Geoderma 61: 103-118. Detweiler, R. P. 1986. "Land use change and the global carbon cycle: the role of tropical soils." Biogeochemistry 2: 67-93Dias, A. C. C. P., and S. Nortcliff. 1985. "Effects of two land clearing methods on the physical properties of an Oxisol in the Brazilian Amazon." Tropical Agriculture. 62: 207-212. Dias Filho, M. B., 1986. "Especies forrageiras e estabelecimento de pastagens na Amazonia." In: Pastagens na Amazonia, eds. A. M. Peixoto, J. C. de Moura, and V. P. de Faria (Congresso Brasileiro de Pastagens. FEALQ, Piracicaba, Brazil), pp. 27-54. Dias Filho, M. B., and E. A. S. Serrao. 1980. "Recuperacao de pastagens de capim coloniao {Panicum maximum) atraves de niveis de fosforo e introducao de leguminosas em Paragominas, Para." Belem: EMBRAPA-CPATU. p. 2. (EMBRAPA-CPATU. Pesquisa em Andamento, 34). Dias Filho, M. B., and E. A. S. Serrao. 1981. "Introducao e avaliacao de gramineas forrageiras na regiao de Paragominas, Estado do Para." Belem, EMBRAPA CPATU. p. 14. (EMBRAPA-CPATU. Circular Tecnica, 17). Dias Filho, M. B., and E. A. S. Serrao. 1982. "Introducao e avaliacao de leguminosas forrageiras na regiao de Paragominas, Pa." Belem, EMBRAPA-CPATU. p. 18. (EMBRAPA-CPATU. Circular Tecnica, 29). Dias Filho, M. B., and E. A. S. Serrao. 1987. "Limitacoes de fertilidade do solo na recuperacao de pastagem degradada de capim coloniao (.Panicum maximum Jacq.) em Paragominas, na Amazonia oriental." Belem, EMBRAPA/CPATU, p. 19. (EMBRAPA-CPATU. Boletim de Pesquisa, 87). Dias Filho, M. B., M. Simao Neto, and E. A .S. Serrao. 1989- "Utilizacion de roca fosforica parcialmente acidulada y superfosfato simple en el estabelecimiento de Brachiaria brizantha cv. Marandu." Pasturas Tropicales 11: 25-8. Dias Filho, M. B., and M. Simao Neto. 1992. "Eficiencias agronomica e economica de um fosfato parcialmente acidulado em Brachiaria brizantha cv. Marandu em solo de floresta na Amazonia." Pesquisa Agropecuaria Brasileira 27: 395-401. Diez, J. A., A. Polo, C. C. Cerri, and F. Andreux. 1991. "Influencia do pousio e da pastagem sobre a dinamica de nutrientes em oxissolos recentemente desflorestados na Amazonia oriental." Pesquisa Agropecuaria Brasileira 26: 77-83. Dobereiner, J., J. M. Day, and P. J. Dart. 1972. "Nitrogen activity in the rhizosphere of sugarcane and some other tropical grasses." Plant Soil 37: 191-196.
Linking Biogeochemical Cycles to Cattle Pasture Management and Sustainability in the Amazon Basin Dunnin, F. X. 1987. "Run-off and drainage from grassland catchments." In: Managed grasslands. Analytical studies. Ecosystems of the world series 17B, ed. R. W. Snaydon (Elsevier, Amsterdam), pp. 205—213. Eden, M. J., P. A. Furley, D. F. McGregor, W. Milliken, and J. A. Ratter. 1991. "Effect of forest clearance and burning on soil properties in northern Roraima, Brazil." Forest Ecology and Management 38: 283-290. Embrapa. 1980. "Centro de Pesquisa Agropecuaria do Tropico Umido, Belem. Projeto de Melhoramento de Pastagens da Amazonia Legal—PROPASTO;" Relatorio Tecnico 1976/1979. Belem, Brazil. Ewel, J., C. Berish, B. Brown, and N. Price. 1981. "Slash and burn impacts on a Costa Rica wet forest site." Ecology 62: 816-829. Falesi, I. C. 1976. "Ecossistema de pastagem cultivada na Amazonia brasileira." Belem, EMBRAPA-CPATU. p. 193. (EMBRAPA-CPATU. Boletim Tecnico, 1) Feigl, B. J., J. Melillo, and C. C. Cerri. 1995. "Changes in the origin and quality of soil organic matter after pasture introduction in Rondonia (Brazil)." Plant and Soil 175: 21-29. Feigl, B. J., P. A. Steudler, and C. .C. Cerri. 1995. "Effects of pasture introduction on soil CO2 emissions during the dry season in the state of Rondonia, Brazil." Biogeochemistry 31: 1-14. Firestone, M. K., and E. A. Davidson. 1989- "Microbial basis of NO and N2O production and consumption in soil." In: Exchange of Trace Gases Between Terrestrial Ecosystems and the Atmosphere, eds. M. O. Andreae, and D. S. Schimel (John Wiley and Sons, New York), pp. 7-21. Fisher, M. J., I. M. Rao, M. A. Ayarza, C. E. Lascano, J. I. Sanz, R. J. Thomas, and R. R. Vera. 1994. "Deep-rooted grasses store carbon in South American soils." Nature 371: 236-238. Frost, P. G. H., and F. Robertson. 1985. "The ecological effects of fire in savannas." In: Determinants of Tropical Savannas, ed. B. H. Walker (CSIRO, Zimbabwe), pp. 93-140. Grace, J. 1994. "Responses of trees to stress." Scottish Forestry 48 (2): 90-95. Garcia-Oliva, F., I. Casar, P. Morales, andj. M. Maass. 1994. "Forest-to-pasture conversion influences on soil organic carbon dynamics in a tropical deciduous forest." Oecologia 99: 392-396. Giovannini, G., S. Lucchesi, and M. Giachetti. 1990. "Effects of heating on some chemical parameters related to soil fertility and plant growth." Soil Science 149: 344-50. Hecht, S. B. 1982. "Deforestation in the Amazon Basin: Magnitude, dynamics and soil resource effects." Studies in Third World Societies 13: 61-101. Italiano, E. C., E. de Moraes, and A. do C. Canto. 1982. "Fertilizacao de pastagens de capim coloniao em degradacao." Manaus, EMBRAPA-UEPAE de Manaus. p. 3. (EMBRAPA-UEPAE de Manaus. Comunicado Tecnico, 31) Jipp, P. H., D. C. Nepstad, D. K. Cassel, and C. R. de Carvalho. 1998. "Deep soil moisture storage and transpiration in forests and pastures of seasonally-dry Amazonia." Climatic Change 39: 395^412. Kauffman, J. B., D. L. Cummings, D. E. Ward, and R.
103
Babbitt. 1995. "Fire in the Brazilian Amazon: 1. Biomass, nutrient pools, and losses in slashed primary forests." Oecologia 104: 397-408. Keller, M., and W. A. Reiners. 1994. "Soil-atmosphere exchange of nitrous oxide, nitric oxide, and methane under secondary succession of pasture to forest in the Atlantic lowlands of Costa Rica." Global Biogeochemical Cycles 8: 399-409. Keller, M., E. Veldkamp, A. M. Weitz, and W. A. Reiners. 1993. "Effect of pasture age on soil trace-gas emissions from a deforested area of Costa Rica." Nature 365: 244-246. Ledgard, S. F, and W. M. H. Sanders. 1982. "Effects of nitrogen fertiliser and urine on pasture performance and the influence of soil phosphorus and potassium status." New Zealand Journal of Agricultural Research Agric. Res. 25: 541-547. Ledgard, S. F., K. W. Steele, and W. M. H. Sanders. 1982. "Effect of cow urine and its major constituents on pasture properties." New Zealand Journal of Agricultural Research 25: 61-68. Leon, L. A., and L. L. Hammond. 1985. "Phosphorus limitations and management considerations." In: Land in Tropical America, eds. T. T. Cochrane, L. G. Sanchez, L. G. de Azevedo, J. A. Porras, and C. L. Garver (Centro Internacional de Agricultura Tropical (CIAT)-Centro de Pesquisa Agropecuaria dos Cerrados (CPAC), Planaltina, D. F., Brazil), pp. 105-110. Ludlow, M. M., M. J. Fisher and J. R Wilson. 1985. "Stomatal adjustment to water deficits in three tropical grasses and a tropical legume grown in controlled conditions and in the field." Australian Jounal of Plant Physiology 12: 131-149. Luizao, R. C. C., T. A. Bonde, and T. Rosswall. 1992. "Seasonal variation of soil microbial biomass-the effects of clearfelling a tropical rainforest and establishment of pasture in the Centaral Amazon." Soil Biology and Biochemistry. 24: 805-813. Luizao, F., P. Matson, R. Livingston, and P. M. Vitousek, 1989. "Nitrous oxide flux following tropical land clearing." Global Biogeochemical Cycles 3: 281-285. Matson, P. A., P. M. Vitousek, J. J. Ewel, M. J. Mazzarino, and G. P. Robertson. 1987. "Nitrogen transformations following tropical forest feeling and burning on a volcanic soil." Ecology 68: 491-502. McWilliam, A.-L. C., O. M. R. Cabral, B. M. Gomes, J. L Esteves, and J. M. Roberts. 1996. "Forest and pasture leaf-gas exchange in south-west Amazonia." In: Amazonian deforestation and climate, eds. J. H. C. Gash, C. A. Nobre, J. M. Roberts, and R. L. Victoria (John Wiley & Sons, London), pp. 266-285. Mengel, K., and E. A. Kirkby. 1987. Principles of plant nutrition. International Potash Institute, Bern, Switzerland. Montagnini, F., and, R. Buschbacher. 1989. "Nitrification rates in two undisturbed tropical rain forests and three slash-and-burn sites of the Venezuelan Amazon." Biotropica 21: 9-14. Moraes, J. E, B. Volkoff, C. C. Cerri, and M. Bernoux. 1996. "Soil properties under Amazon forest and changes due to pasture installation in Rondonia, Brazil." Geoderma 70: 63-81. Neill C, M. C. Piccolo, P. A. Steudler, J. M. Melillo,
104
B. J. Feigl, and C. C. Cerri. 1995. "Nitrogen dynamics in soils of forests and active pastures in the western Brazilian Amazon basin." Soil Biology and Biochemistry. 27: 1167-1175. Neill, C., B. Fry, J. M. Melillo, P. A. Steadier, J. F. Moraes, and C. C. Cerri. 1996. "Forest and pasture-derived carbon contributions to carbon stocks and microbial respiration of tropical pasture soils." Oecologia 107: 113-119Neill, C., J. M. Melillo, P. A. Steudler, C. C. Cerri, J. F. L. de Moreaes, M. C. Piccolo, and M. Brito. 1997. "Soil carbon and nitrogen stocks following forest clearing for pasture in the southwestern Brazilian Amazon. Ecological Applications 7: 1216-1225. Nepstad, D. C. 1989. Forest regrowth in abandoned pastures of eastern Amazonia: limitations to tree seedlings survival and growth. Ph.D. Thesis, Yale University, New Haven. Nepstad, D. C., C. Uhl, and E. A. S. Serrao. 1991. "Recuperation of a degraded Amazonian landscape: forest recovery and agricultural restoration." Ambio 20: 248-255. Nepstad, D. C., C. J. R. de Carvalho, E. A. Davidson, P. Jipp, P. Lefebvre, G. H. Negreiros, E. D. da Silva, T. A. Stone, S. Trumbore, and S. Vieira. 1994. "The role of deep roots in the hydrological and carbon cycles of Amazonia forests and pastures." Nature 372: 666-669. Nguyen, M. L., and K. M. Goh. 1992. "Nutrient cycling and losses based on a mass-balance model in grazed pastures receiving long-term superphosphate applications in New Zealand. 1. Phosphorus." Journal of Agricultural Science, Cambridge 119: 89-106. Piccolo, M. C., C. Neill, and C. C. Cerri. 1994. "Natural abundance of 15N in soils along forest-to-pasture chronosequences in the water Brazilian Amazon basin." Oecologia 99: 112-117. Rosswall, T. 1982. "Microbiological regulation of the biogeochemical nitrogen cycle." Plant and Soil 67: 15-34. Roberts, J. M., O. M. R Cabral, J. P. Costa, A.-L. C McWilliam, and T. D. de A. Sa. 1996. "An overview of the leaf area index and physiological measurements during ABRACOS." In: Amazonian deforestation and climate, eds. J. H. C. Gash, C. A. Nobre, J. M. Roberts, and R. L. Victoria (John Wiley and Sons, London), pp. 287-306. Robbins, G. B., J. J. Bushell, and G. M. McKeon. 1989. "Nitrogen immobilization in decomposing litter contributes to productivity decline in ageing pastures of green panic (Panicum maximum var. trichoglume)." Journal of Agricultural Science, Cambridge 113: 401-406. Salati, E., R. Sylvester-Bradley, and L. Victoria. 1982. "Regional gains and losses of nitrogen in the Amazon basin." Plant and Soil 67: 367-376. Sanchez, P. A. 1976. "Soil management in shifting cultivation areas." In: Properties and Management of soils in the tropic, ed. P. A. Sanchez (John Wiley & Sons, New York), pp. 347-412. Sanchez, P. A., 1987. "Management of acid soils in the humid tropics of Latin America." In: Management of acid
Moacyr B. Dias-Filho et al. tropical soils for sustainable agriculture: proceedings of an IBSRAM inaugural workshop, eds. P. A. Sanchez, E. R. Stoner, and Pushparajah (Bankok, Thailand), pp. 63-107. Sanchez, P. A., J. H. Villachica, and D. E. Bandy. 1983. "Soil fertility dynamics after clearing a tropical rainforest in Peru." Soil Science and Social American Journal 47: 1171-1178. Serrao, E. A. S., I. C., Falesi, J. B. da Veiga, andj. F. Teixeira Neto. 1979- "Productivity of cultivated pastures in low fertility soils of the Amazon of Brazil." In: Pasture production in acid soils of the tropics, eds. P. A. Sanchez, and L. E. Tergas (Centro Internacional de Agricultura Tropical, Cali, Colombia), pp. 195-225. Serrao, E. A. S., and J. M. Toledo. 1990. "The search for sustainability in Amazonian pastures." In: Alternatives to Deforestation: Steps Toward Sustainable Utilization of Amazon Forests, ed. A. B. Anderson (Columbia University Press, New York), pp. 195-214. Seubert, C. E., P. A. Sanchez, and C. Valverde. 1977. "Effects of land clearing methods on soil properties of an ultilsol and crop performance in the Amazon jungle of Peru." Tropical Agriculture. 54: 307-321. Skole, D. S., W. Chomentowski, W. A. Salas, and A. D. Nobre. 1994. "Physical and human dimensions of deforestation in Amazonia." Bioscience 44: 314-322. Smyth, J., andj. B. Bastos. 1984. "Alteracoes na fertilidade de um Latossolo Amarelo Alico pela queima da vegetacao." Revista Brasileira de Ciencia do Solo. 8: 127-132. Spain, J. M., and J. G. Salinas. 1985. "A reciclagem de nutrientes nas pastagens tropicais." In: Reciclagem de nutrientes em agricultura de baixos insumos, ed. P. Cabala-Rosand (CEPLAC/SBCS, Ilheus, Brazil), pp. 259-299. Street, J. 1982. "Changes of carbon inventories in live biomass and detritus as a result of the practice of shifting agriculture and the conversion of forest to pasture: case studies in Peru, New Guinea and Hawaii." International seminar on geography and the third world. Department of Geography, University Kebangsaan Malaysia Bangi, Selangor, Malaysia. Teixeira, L. B., 1987. "Dinamica do ecossistema de pastagem cultivada em area de floresta na Amazonia Central." Tese de doutorado. INPA/FUA, Manaus, Brazil. Tiessen, H., E. Cuevas, and P. Chacon. 1994. "The role of soil organic matter in sustaining soil fertility." Nature 371: 783-785. Toledo, J. M., and V. A. Morales. 1979. "Establishment and management of improved pastures in the Peruvian Amazon." In: Pasture production in acid soils of the tropics, eds. P. A. Sanchez, and L. E. Tergas (Centro Internacional de Agricultura Tropical, Cali, Colombia), pp. 177-194. Trumbore, S. E., E. A. Davidson, P. B. de Camargo, D. C. Nepstad, and L. A. Martinelli. 1995. "Below-ground cycling of carbon in forests and pastures of eastern Amazonia." Global Biogeochemical Cycles 9: 515-528. Uhl, C., R. Buschbacher, and E. A. S. Serrao. 1988. "Abandoned pastures in Eastern Amazonia, I: Patterns of plant succession." Journal of Ecology 76: 663-681. Veldkamp, E. 1994. "Organic carbon turnover in three trop-
Linking Biogeochemical Cycles to Cattle Pasture Management and Sustainability in the Amazon Basin ical soils under pasture after deforestation." Soil Science and Social American Journal 58: 175—180. Vieira, I. C. G, C. Uhl, and D. C. Nepstad. 1994. "The role of the shrub Cordia multispicata Cham, as a 'succession facilitator' in an abandoned pasture, Paragominas, Amazonia." Vegetatio 115: 91—94. Vitousek, P. M., and R. L. Sanford. 1986. "Nutrient cycling in moist tropical forests." Annual Review of Ecology and Systematics 17: 137-167.
105
Wright, H. A., and A. W. Bailey. 1982. "Soil and water properties." In: Fire ecology. United States and Southern Canada, eds. H. A. Wright, and A. W. Bailey Qohn Wiley and Sons, New York), pp. 24-48. Wright, I. R., J. H. C. Gash, H. R. da Rocha, W. J. Shuttleworth, C. A. Nobre, G. T. Maitelli, C. A. G. P. Zamparoni, and P. R. A. Carvalho. 1992. "Dry season micrometeorology of central Amazonian ranchland." Q. J. Roy. Met. Soc. 118: 1083-1099.
7 Nutrient Considerations in the Use of Silviculture for Land Development and Rehabilitation in the Amazon Horencia Montagnini
Tropical plantations serve diverse economic, social, political, and ecological functions. With considerably higher yields than managed native forests, tropical and subtropical plantations make substantial contributions to world timber and pulp production (Wadsworth 1983, Evans 1992). Tree plantations can also be a source of cash, savings, and insurance for individual farmers. Plantations may help stabilize rural populations in regions where shifting agriculture is the predominant land use. In combination with subsistence and commercial crops (agroforestry) or cattle (agrosilvopastoral systems), plantations have been used as tools in rural development projects worldwide. Plantations are often seen as alternatives to deforestation as they can provide products that otherwise would be taken from natural forests (Fearnside 1990, McNabb et al. 1994). Nutrient cycling characteristics of tropical plantations differ from those of natural forests in a number of ways. Natural forests are adapted to ecological niches by intricate and effective physiological adaptations of growth in the environment (see Cuevas, this volume). Instead, tropical plantations are simplified, generally monospecific ecosystems that occupy the site for a limited period of time that can range from 4-12 years (for biomass, pulpwood, or fuelwood) up to 2040 years (timber). In many instances plantations are composed of species that are exot-
ic to the region, or even when indigenous, are new to the particular plantation site. Since plantation tree species have been generally selected for production of timber or other aboveground tree parts, they tend to maintain a smaller fraction of total tree biomass nutrients in roots than natural forests (Vogt et al. 1997). In rain forests growing on poor soils, high tree productivity is in part due to the existence of important nutrient conserving mechanisms mediated by the root system (Cuevas, this volume). The smaller biomass of plantation root systems may thus make them more susceptible to nutrient and water stress. Smaller root systems may also make plantation forests more susceptible to disturbances from strong winds and pathogens that attack aerial parts (Vogt et al. 1997). Nutrient demands by plantation trees vary from season to season and with the developmental age of the stand (Drechsel and Zech 1993). During the life of the plantation, large quantities of nutrients are returned to the soil by above- and belowground litter, harvest residues, stem flow, and throughfall. These elements are mineralized in the soil and may be used by the trees or by associated crops. Deposition of certain mineral elements from the atmosphere can also be important, sometimes even compensating from losses in stemwood removal at harvest (Bruijnzeel 1989)- Retranslocation of absorbed nutrients
Nutrient Considerations in the Use of Silviculture for Land Development and Rehabilitation in the Amazon
can be also of considerable magnitude, often supporting sustained growth in the short and long term (Nambiar 1984). Plantations are frequently managed intensively, with silvicultural inputs, for maximum wood or biomass production in the shortest time. Plantation management (e.g., thinning, pruning, coppicing, fertilizing) profoundly affects stand structure, crown characteristics, and plantation micro environment, all of which in turn can affect nutrient cycling processes. Other forestry operations, especially during harvest and those associated with the establishment of the following crop (biomass removal, soil disturbance, residue disposal by burning, intense site preparation such as ploughing, discing or blading, weed control, and fertilization) have profound influences on the productivity of a site (Lundgren 1980, Sanchez et al. 1985). The effects of perturbation on nutrient losses are greatly accelerated in short rotations of intensively managed plantations, and these considerations have often led to the concept of economic v. ecological rotations (Nambiar 1984, Drechsel and Zech 1993). Fast growing tropical tree plantations can incorporate considerable amounts of nutrients in their biomass over a relatively short period of time. Site fertility declines can limit sustained plantation forestry in tropical regions: soil fertility can be decreased through excessive removal of nutrients in living biomass, particularly if nutrients in tree crowns are lost through harvest or site preparation (Jorgensen and Wells 1986, Perry and Maghembe 1989)- Site fertility declines can be more serious if the trees are harvested in short rotation schemes, and if whole-tree harvesting is applied. This can be particularly serious when plantations are established on soils that are inherently poor, as occurs over vast areas of the Amazon basin. If plantation species are chosen with knowledge of their nutrient-use efficiencies
107
and recycling capacities, they can be highly productive and even serve a function in ecosystem restoration projects. In particular, mixed-species designs can offer product diversification and can complement nutrient resource use with positive impacts on soil nutrients (Alvim 1989, Silva and Uhl 1992, Montagnini et al. 1995). In addition, rotation length and harvest techniques can be adjusted to minimize nutrient losses and maintain site productivity in the long term. This chapter considers large-scale industrial tree plantations, as well as small- and medium-scale plantations for timber and land rehabilitation. Aspects that influence plantation sustainability are emphasized, and suggestions are given regarding plantation design and management.
Plantations in the Amazon Tree plantations in the Amazon are used at small, medium, and large scales, each of which serves different functions, such as rural development, land rehabilitation, or industrial purposes. The first major entrepreneurial effort in Amazonian tree plantations was the establishment of native rubber (Hevea brasiliensis) in 1928 at the Fordlandia Estate owned by Henry Ford on the lower Tapajos river in the state of Para, Brazil. These plantations were abandoned in the late 1940s due to low latex productivity that resulted from a high incidence of the South American leaf blight (Rankin 1985). After this failure, large-scale tree plantations were not undertaken again in the region until 1968 when Jari Forestal e Agropecuaria was established, also in Para. Other early attempts at plantation establishment in the Brazilian Amazon include small plantations of Brazil nut (Bertholletia excelsa) near Manaus in 1931, which had generally low nut yield but good tree growth, and scattered efforts to cultivate the valuable Meliaceae. Of the Meliaceae, Swietenia macrophylla (mahogany)
108
and Cedrela odoratd), were planted most often, but neither these nor other species were successful, as all were stricken by the shoot borer Hypsipyla grandella (Palmer 1977). Plantations have been sometimes recommended as a potentially viable land use for the Amazon region (Alvim 1989, Hoyos et al. 1992, McNabb et al. 1994). Silvicultural plantations have also been promoted as an alternative means of supplying wood and paper needs while also reducing pressure for additional rain forest clearing (Fearnside 1990). However, as in other tropical humid regions of the world, many authors have raised concerns about the ecological and economic sustainability of large-scale plantations in the Amazon basin: fast-growing exotics planted in monospecific stands deplete soil nutrients in a few rotations and often suffer from pest and disease problems (Johnson 1976, Fearnside and Rankin 1980, Rankin 1985, Palmer 1977, Russell 1987). In the Brazilian Amazon, hardwood species have been restricted to experimental settings, but fast-growing species for pulp, plywood, and sawlogs have been planted in an increasing number of commercial plantations that are partially or fully owned by foreign investors. Examples include 20,000 ha of Caribbean pine (Pinus caribaea) in the Amapa territory, 500,000 ha of Caribbean pine near Portel in the state of Para, and the highly publicized Jari project, also in Para.
Nutrient Cycling in Industrial Plantations of Fast- Growing Exotics: the Jari Project Located along the Jari River (a tributary 300 km from the mouth of the Amazon), the 1.6 million ha Jari estate was originally purchased by the American multimillionaire Daniel Ludwig in 1967. In the first fifteen years of the project, more than 100,000 ha of forest were cleared and planted with exotic
Florencia Montagnini
species, mostly Gmelina arborea, Pinus caribaea var. hondurensis, and Eucalyptus spp. (McNabb et al. 1994). In 1982 Ludwig sold the forestry operation to a consortium of Brazilian businesses, taking a substantial loss on the original investment (Jordan, this volume). Now called the Monte Dourado Forestry Company, this estate has become a modern forestry enterprise similar to those established in southeastern Brazil. Current forestry, agricultural, and mining uses encompass about 28% of the property, while the remaining 72% is still relatively undisturbed rain forest (McNabb et al. 1994). Controversy has surrounded the capacity of the soils to support long-term plantation productivity at Jari. For example, Irion (1981) suggested that the soil would be completely exhausted after the second generation (14-20 years) even though the Gmelina arborea plantations had reached a height of 20 m in 7 years in the first generation, and the other species (Pinus caribaea and Eucalyptus deglupta) had shown few signs of nutrient deficiency at the time. Other authors have suggested that soil fertility could be improving under the forest plantations at Jari. For example, company researchers showed that in the plantations there was no economic response to the application of mineral fertilizers in the early years (Palmer 1977). Likewise, Greaves (1979) stated that the fertility and moisture retention of the soils was apparently improving under the forest plantations, making reference to the thick organic layer developed under both pine and Gmelina species. Russell (1987) examined whole ecosystem nutrients within a native forest, a 6-month pine plantation, a 9-5-year pine plantation, an 8.5-year Gmelina plantation, and a 1.5-year second generation pine plantation that was established on a previous 8.5-year Gmelina plantation. A synthesis of the changes in nutrient stocks during the course of plantation establishment and growth is presented
Nutrient Considerations in the Use of Silviculture for Land Development and Rehabilitation in the Amazon
in Figure 7.1. At the end of the first rotation of Gmelina (8.5 years) or Pinus caribaea (9.5 years), total plant biomass was about 40-60% of that in the virgin forest. Most of the losses
Fig. 7.1 Total nutrient stocks and in plant biomass plus soil in rain forest, newly planted Pinus caribaea (6 months old), P. caribaea, and Gmelina arborea plantations at the end of the first rotation (9.5 and 8.5 years old, respectively), and second rotation P. caribaea (1.5 years old) at Jari (Sanchez et al. 1985). H = harvest loss from trees taken when clearing the rain forest for the plantations; L = leaching. Total nutrient stock is defined as the sum of all the nutrients in plant biomass (aboveground, litter, detritus, roots) plus total N, available P (extracted by the Mehlich method), and exchangeable K, Ca, and Mg in the top meter of the soil.
109
were accounted for by the dry matter extracted by harvest, as well as by the disturbance caused by site preparation. The plantations of all ages contained about 60% of the total N stock of the rain forest. Most losses occurred shortly after rain forest clearing, but thereafter the plantations maintained a relatively constant level of N. The ecosystem, therefore, lost 40% of its N, and then reached a new equilibrium level. Stocks of P ranged from 76% to 116% of the rain forest values, with the decrease in P at the start of the rotation largely accounted for by the P removed in the first rotation harvest. Significant losses of K occurred after rain forest clearing, with a decrease to about 32% of that in the forest before clearing. Most of the losses were accounted for by the removal at harvest, as well as by rapid leaching losses recorded during this period, an expected result since K is a highly mobile ion and it is readily lost through leaching. Afterwards, there were slight increases, to about 40% of the rain forest values. Ca stocks decreased to about 56% of the rain forest values upon planting of the first pine rotation. Losses were again accounted for by Ca removal at forest harvest and small amounts that were leached. This reduced level of Ca remained relatively stable with Pinus, but increased to above preclearing levels with Gmelina. Magnesium stocks decreased with age in Pinus, but mature Gmelina plantations maintained a steady level of about 75% of that in the rain forest. Most of the ecosystem nutrient losses occurred during the plantation establishment phase because of the removal of forest debris at clearing and soil disturbance during mechanized operations. Despite the potential improvement in Ca levels and maintenance of Mg levels as mentioned for Gmelina, both the extraction of nutrients during harvesting, and leaching losses prior to canopy closure, lead to a depletion of key nutrients, particularly potassium, that must be replaced by fertilization if yields are to be maintained (Sanchez et al. 1985, Russell 1987).
no Likewise, Spangerberg (1994) reported on nutrient removal by eucalyptus at Jari and stressed the need for fertilization after the first rotation. Studies were made on first rotation, 4.5-year-old plantations of Eucalyptus urograndis (a hybrid of E. urophylla and E. grandis) in order to calculate nutrient losses due to the removal of wood and bark during harvesting for pulpwood. Plantation trees exhibited no nutrient deficiencies, although foliar concentrations of Ca seemed to be rather low. Average nutrient losses due to removal of wood and bark amounted to 64.6% for N, 54.2% for P, 76.3% for Ca, 57.2% for K, and 60.7% for Mg, of the whole tree biomass. If the same nutrient removal was assumed in the next rotation, there would be a deficiency of Ca, and compensations of up to 250 kg Ca/ha would be necessary. Results of experimental research on nutrient dynamics in industrial plantations elsewhere in the Amazon also point to the need for supplemental fertilization to sustain yields in the medium to long term. In experimental plantations at the National Institute for Amazonian Research (INPA) in Manaus, results of studies by Magalhaes et al. (1986) showed how height of several tree species growing in 3-year-old plantations correlated strongly with soil fertility parameters (exchangeable bases, organic matter, exchangeable Al, total Zn and Mn). However, nutrient requirements vary among plantation species and with plantation age. For example, in experiments using 4 doses of lime (calcium carbonate) during establishment of plantations on acid soils at La Selva, Costa Rica, only 1/3 of the species exhibited a significant response in growth (Soto et al. 1996). The species tested were all native to the region, and they were probably adapted to growing on the acid soil conditions of the experimental area. In addition, responses to fertilizer also vary with plantation age. In the early stages of stand development prior to canopy closure, the annual rate of nutrient accumu-
Florencia Montagnini
lation increases rapidly and tree growth is very dependent on current nutrient uptake. Mineral deficiencies are frequently observed and responses to fertilizer application are common during this stage. Once the canopy has closed, the reduction in rate of nutrient uptake is associated with attaining maximum foliage biomass, high internal retranslocation of mobile nutrients, as well as increasing nutrient recycling via litterfall. This decreases the nutrient contribution by the soil, thus fertilizer responses are unlikely during this stage (Drechsel and Zech 1993). The type, amounts, and timing of fertilization used in plantations need to be adjusted according to species requirements, ideally using results of field fertilization experiments. In summary, in high-yield, short-rotation plantation forestry in rain forest sites such as Jari, and other locations with similar ecological conditions, nutrients are likely to become limiting after the second or third rotation; therefore, yields will not be maintained without extra fertilizer inputs. Repeated fertilization can upset soil nutrient balances and alter microbial populations, and it would also pose problems for the economics of producing low-value commodity timbers (see also Jordan, this volume). In spite of the lessons from Jari, other large scale plantation development plans continue in the Brazilian Amazon. If developed, the Grande Carajas Program would consume large areas of tropical forest in the eastern Amazon in order to obtain charcoal for the smelting of pig-iron from the Carajas' mines. The area of Eucalyptus plantation to produce the required amount of charcoal would be over 700,000 ha-more than 10 times the area of E. deglupta already cultivated in the Jari Project (Fearnside 1989).
Small-Scale Plantations for Timber Nutrient cycling in timber plantations in the Amazon has not been studied as much as
Nutrient Considerations in the Use of Silviculture for Land Development and Rehabilitation in the Amazon
industrial plantations, yet these systems have great potential to provide a more sound land use option for the region: they can produce high-quality timber, help to restore degraded lands, and contribute to rural development. Because there are still large areas of native forest available for management and extraction of valuable timber species, plantations of exotic and native hardwoods remain small and mostly experimental. A lack of information on the silvicultural management for some of the most preferred timber species, which are often difficult to grow in open plantations, also limits the current expansion of timber plantations in the Amazon. For example, in recent experiences from the Peruvian Amazon, populations of Hypsipyla grandella (the mahogany shoot borer) grew rapidly in the rainy season as food availability increased from new growth sprouted on the host trees (Cedrela odorata, C. fissilis, and Swietenia macrophylla). C. odorata was more susceptible to the pest than S. macrophylla (Yamazaki et al. 1990). It was suggested that planting trees in small cleared areas and avoiding clean weeding may reduce the ease with which the pest finds its host species. Several other native species have been tried with success in plantations in the Peruvian Amazon. Results of 20 years' experience with plantations of 9 exotic and 104 native species at Jenaro Herrera, near the Ucayali river 200 km upstream from Iquitos, were summarized by Claussi et al. (1992). As a result of the failure of some exotics and attacks by Hypsipyla on the Meliaceae, the plantation "work concentrated on non-Meliaceae native species. Cedrelinga catenaeformis and Simarouba amara [Quassia simarouba] were the two best timber species, with mean annual increments of over 1.5 cm in diameter and 1.5 m in height at 10 years of age. Two species suitable for rural construction work, Guatteria elata and G. hyposericea, had mean annual increments of over 1.5 cm
111
in diameter and 1.5 m in height. Other suitable timber species recommended by these authors include Parkia multijuga, Ormosia spp., and Sclerolobium off. tinctorium. Recommendations on plantation establishment included species mixtures, such as C. catenaeformis with Q. simarouba and Guatteria elata, and agroforestry combinations with agricultural crops. In the Brazilian Amazon, the earliest experiences reported are those from timber plantations at EMBRAPA's Curua-Una station in Santarem, Para (Pedroso 1973a, 1973b; SUDAM 1979). These experiences provided data on species trials established since 1959, as well as information on silvicultural methods and regeneration techniques developed for establishing plantations of indigenous and exotic species in the Amazon region. Fourteen to eighteen years after planting, tree growth of 47 native and 18 exotic species showed that promising native species included Anacardium giganteum, Bagassa guianensis, Bertholletia excelsa, Buchenavia huberi, Didymopanax morototoni, Goupia glabra, Jacaranda copaia, Parkia multijuga, Simarouba amara, Vatairea guianensis, Virola cuspidata, and Vochysia maxima (SUDAM 1979). Performance of Pinus caribaea var. hondurensis and Eucalyptus spp. were also good. Other species recommended for their inclusion in tree improvement programs were the natives Carapa guianensis, Platonia insignis, Schizolobium amazonicum, and Virola surinamensis and exotics such as Gmelina arborea and Terminalia ivorensis (Pitcher 1976). Cordia alliodora has also been recommended as a good prospect for commercial, large-scale plantations (Carpanezzi et al. 1982). Important lessons regarding plantation management strategies have been learned from the results of pioneering studies in the eastern Amazon. From their experiences in Paragominas (in Para), Nepstad et at. (1991), and Uhl et al. (199D suggested that burning
112
Florencia Montagnini
The experiences in Paragominas, Para, Brazil mentioned in the previous section and experiments at EMBRAPA stations constitute isolated, pioneering examples of this type of silvicultural plantation in the Amazon. Keys to the success of these systems are correct species choice and design. Choice of species should be based on the nutrient use efficiency and growth rate of each species. Recent studies on the natural impact of native trees on soil fertility and nutrient cycling in the Atlantic forest region of Bahia, and in the Atlantic humid lowlands of Costa Rica, help to illustrate these points. During trials in Bahia, Brazil, 15 out of 20 tree species planted in monospecific stands had positive effects on at least 5 soil fertility parameters 13-15 years after planting. Several species contributed to increased soil carbon and nitrogen levels in the topsoil: Inga affinis, Parapiptadenia pterosperma, Plathymenia foliolosa (leguminous, N fixing species), Caesalpinia echinata, Copaifera luscens (leguminous, non N fixing), Eschweilera ovata, and Pradosia lactescens (Montagnini et al. 1994). In experiments at La Selva Biological Station in Costa Rica, improved soil conditions were found just after tree canopy closure in pure plantations of timber trees compared to pasture lands. The highest values for soil carbon, total nitrogen, and phosphorus were found under Vochysia ferruginea, a valuable timber species native to Central America. After three years the soils under pure stands of native species had soil fertility values Plantations as a Tool for Land similar to those found in 20-year-old natural Rehabilitation and Rural Development secondary forests originated after abandonment of cattle pastures (Montagnini and Plantations for land rehabilitation are Sancho 1994). generally grown at a small to medium scale In the same region of Costa Rica, the use because they usually require intensive initial of improved fallow systems yielded land valcare, often including external inputs (mycor- ues of $5-12 thousand per hectare at a 5% rhiza inoculation, use of fertilizers or herbi- real interest rate, after inflation (Montagnini cides). In some cases, initial intercropping and Mendelsohn 1996). This system involved with annual plants can help to offset costs. planting native tree species to replenish soil
before planting would decrease competition by grass. In trials with 27 native species, they concluded that the addition of fertilizer or manure was not essential for establishment. They also suggested planting trees in patches of already established shrubs, such as Cordia multispicata, which presumably will offer a more favorable microclimate for tree establishment. Among the best performers in these trials were fruit trees such as Anacardium occidentale (caju), and Bixa orellana, as well as timber trees such as Swietenia macrophylla. As seen, a considerable number of valuable timber species can be grown in plantations with relative success in the Amazon region, and a few silvicultural guidelines are available to aid in their establishment and management. Tree plantations, especially with indigenous species, can contribute to soil restoration and facilitate natural forest regeneration (Guariguata et al. 1995). Plantations of exotics can also be beneficial on many degraded sites and can facilitate natural regeneration in some circumstances, as explained in the next section. Over large areas of the tropics, with pressing human needs for land, the restorative effects of tree plantations can be realized if local people are willing to plant trees for their products such as timber. Overall, in the Amazon basin this is probably the most important role of silvicultural plantations, the most unrealized one, and the one in need of greater research and promotion.
Nutrient Considerations in the Use of Silviculture for Land Development and Rehabilitation in the Amazon
and provide economically valuable timber. Although the experiences are site specific, the species used have broad distribution in Latin America. Therefore, the systems would be transferable in other areas with similar ecological and economic conditions.
Reforestation of mine spoils Under the highly degraded conditions present in mine spoils, serious nutrient deficiencies limit tree growth and require heavy use of external inputs to ensure success of restoration projects. In experiences on reforestation of bauxite mining sites at Porto Trombetas (western Para), low levels of soil organic matter, N, P, K, Ca, and Mg were found to be the principal factors for reduced growth of planted trees (Ferraz 1993). The plantations present great variability in growth and degree of vegetation cover, even in areas of apparently homogeneous soils. The restored areas receive surface soil as an amendment before reforestation. This topsoil
113
is stored after removing it to allow for bauxite mining, thus the length and conditions of soil storage may affect its suitability to serve as bed for the newly planted seedlings. Additions of mycorrhizae and supplemental nutrients are often necessary. This increases the per hectare cost of reforestation although levels remain reasonable for the mining company (Knowles, personal communication). In the past, reforestation relied on the establishment of monospecific plantations, often with exotics (e.g., Eucalyptus, Pinus, Acacia spp.) or a limited number of native tree taxa for which seeds are readily available and silvicultural practices have been developed (Knowles and Parrotta 1995). Also at Porto Trombetas, about 160 native forest species in mixed plantings were evaluated for their suitability for forest restoration on bauxite mine land over a 14-year period. Observations over 600 ha of plantings have yielded information on ecological characteristics of the species and cost-effective propagation methods. The plantations of exotics and natives are expected to catalyze natural forest succession in the understory and thus accelerate the rate at which species-rich native forest stands develop on severely degraded lands.
Nutrient Cycling and Plantation Design and Management The impacts of trees on soil fertility depend on nutrient recycling characteristics such as litter chemistry and decomposition rates. Tree litter can be used as mulch with different outcomes: a fast mulch decomposition rate may accelerate the growth of associated crops on poor soils, while in other cases a more persistent litter may provide a steady source of nutrients and a better soil cover Fig. 7.2 Nitrogen and phosphorus use efficien- year round. In the experiments in Costa cy of 4-year-old stands of 4 indigenous timber Rica described in the previous section, high species grown in pure plantation at La Selva rates of litterfall and slower decomposition Biological Station, Costa Rica. resulted in high litter accumulation and high
114
soil organic matter under V. ferruginea, making this species well suited for protecting soils against erosion. In contrast, litter from another species of Vochysia, V. guatemalensis, may be especially important for Ca and Mg recycling (Montagnini et al. 1993). Although in this experiment the litter of Hieronyma alchorneoides was less abundant than the other 3 species, it had a relatively faster decomposition rate and higher nutrient content. These characteristics promoted fast nutrient recycling, especially of N, Ca, Mg, K, and P. The ability of a species to produce large amounts of biomass with less nutrients may be an important consideration in choosing species for degraded, nutrient-poor sites. When put in context with nutrient recycling characteristics of a species, Nutrient Use Efficiency (NUE) values, calculated as the annual biomass increments per nutrients in annual leaf litterfall, can indicate appropriate system design and management to maintain productivity and recover or conserve nutrients over the long term. Results from experiments in Bahia, Brazil, suggest that Bombax macrophyllum and Plathymenia foliolosa, with overall high NUE values, would grow well on relatively nutrient-poor soils, and thus could be good alternatives for reforestation of degraded sites following the abandonment from agriculture and pasture that is frequent in the region (Montagnini 1994). B. macrophyllum stands tended to accumulate high amounts of litter under its canopy while P. foliolosa stands had relatively high amounts of organic matter and total N in the topsoil in comparison with adjacent areas of secondary forest (Montagnini et al. 1994). These features indicate that these species may be well suited for soil rehabilitation, including increasing soil organic matter content and protecting against soil erosion. Also at Bahia, species such as Buchenavia grandis and Hymenaea aurea, with overall lower NUE values, would be most appropri-
Florencia Montagnini
ate for agroforestry combinations where crops could benefit from nutrients recycled from litter. Species choices and designs should be suited to the kind of soil rehabilitation desired, a goal that should be tightly connected with the future use expected from the rehabilitated land. At La Selva, Vochysia ferruginea showed comparatively low efficiency values for N and P (Fig. 7.2), confirming the beneficial role of this species in recycling organic matter and positively impacting soil fertility as shown above, while the higher NUE values shown for V. guatemalensis illustrate the relatively large allocation of nutrients to stem biomass and comparatively low recycling in leaf litter. The relatively low efficiency (high recycling) of N and P found for Stryph-nodendron microstachyum and Hyeronima alchorneoides (Fig. 7.2) was also shown in experiments where maize grown with mulch of these species grew better and absorbed more N and P than with mulch of other species (Montagnini et al. 1993). Nutrient allocation in trees can impact soil nutrients in plantations when different parts of the trees are harvested and removed. When biomass nutrient allocation was compared among species at La Selva, the forest floor appeared to be an important compartment for long-term recycling of N, with marked differences among species (Fig. 7.3a). Similar results were obtained with respect to Ca and Mg, which also had relatively low use efficiency values (Montagnini 1994). If the forest floor is burned or collected for fuelwood, a substantial loss of organic matter and nutrients may occur, while if the litter is left after harvest, it represents a significant reservoir for the next tree rotation. In contrast, most species had higher use efficiency of P, and this nutrient was found in higher proportion in live tree biomass (Fig. 7.3b); a similar pattern was found for K (Montagnini 1994). Roots were also important for P, in comparison with
Nutrient Considerations in the Use of Silviculture for Land Development and Rehabilitation in the Amazon
115
Fig. 7.3 Nitrogen and phosphorus stocks in above ground biomass, forest-floor litter and roots of 4-year-old stands of 4 indigenous timber species grown in pure plantation at La Selva Biological Station, Costa Rica.
116
forest-floor P. Roots of tropical trees appear to decay at slower rates than leaf tissue, which means that they may function as a longerterm storage mechanism for nutrients (Bloomfield et al. 1993). The biomass of fine roots in tropical plantations is generally 2 to 4 times lower than in adjacent secondary forest of similar age (Lugo 1992, Vogt et al. 1997). Certain indigenous species that tend to maintain a relatively large proportion of living tree biomass nutrients in roots, while still producing high stem biomass increments (such as V. ferruginea and H. alchorneoides), would be preferred in sites where nutrient losses are a major concern. This can be especially critical for P and K, often mentioned as the nutrients which are most likely to be depleted from soils with subsequent rotations (Wadsworth 1983, Bowen and Nambiar 1984, Bruijnzeel 1984). Altering the rate of nutrient removal in products is probably one of the most important design considerations in planning sustainable plantations (Wang et al. 1991). Variation between species in nutrient allocation and the parts of the tree removed from the site will determine the extent to which the nutrients are removed during harvesting. For example, whole tree harvests of V. guatemalensis -would result in substantial nutrient removals, especially of Ca and Mg, while the harvest of H. alchorneoides would result in large removals of P and K (Montagnini and Sancho 1994). This can be assessed through nutrient and biomass sampling and estimation, and harvesting guidelines can be developed that reduce the extent of nutrient losses. The degree to which harvest losses will impact on forest production will depend on the amount of nutrients stored in remaining pools and the rate at which they are mineralized from these pools. Finally, a key issue in nutrient management of tree plantations is to find the best compromise between ecological requirements
Florencia Montagnini
and rotation length so as to ensure adequate nutrient supply in the long term. Due to increasing nutrient use efficiency in the stands with time, lengthening the rotation age of fast growing species will reduce net nutrient uptake from soil reserves (Drechsel and Zech 1993).
Other considerations for management and design The recovery of the productive capacity of soils is frequently expensive, thus the techniques involved must produce financial returns to ensure they are adopted by local farmers. In soil rehabilitation projects, species such as B. macrophyllum and P. foliolosa (Bahia) or V. guatemalensis (Costa Rica) should be grown first as they can generally tolerate more impoverished conditions than other species while producing valuable timber and protecting soils against erosion. Other species, such as B. grandis and H. aurea (Bahia), or H. alchorneoides (Costa Rica), with lower nutrient use efficiency and potentially producing higher nutrient cycling, might be planted at the same time, or underplanted later. The calculation of NUE and its interpretation in context with the recycling capacities of the species could contribute to species selection and system design focusing on the potential role of each species on productivity and soil rehabilitation. Mixed-species designs can be more advantageous than tree monocultures for site nutrient rehabilitation if systems are planned to complement each species' nutrient demands and effects. For the fastest-growing, light-demanding species included in a mixture, tree productivity can be higher than in monospecific stands, while for the slowergrowing, shade-tolerant species of generally more valuable timber, the mixture offers a more adequate growth environment than a single-stratum, open plantation (Montagnini et al. 1995). In experiments with mixed and
Nutrient Considerations in the Use of Silviculture for Land Development and Rehabilitation in the Amazon
117
Fig. 7.4 Aboveground biomass and roots (0-15 cm soil depth) of 3-year-old stands of 4 indigenous tree species: Terminalia amazonia, Hieronyma alchorneoides, Albizia guachapele, and Virola koschnyi, grown in pure plots, and a mixture of the 4 species at La Selva Biological Station, Costa Rica. pure species plantations at La Selva, Costa Rica, mixed stands of four indigenous species ranked first or second in terms of above ground biomass, in comparison with pure stands of each species (Fig. 7.4). The data for roots in Fig. 7.4 were only for the top 15 cm layer of mineral soil; it would be interesting to see if roots explore the soils in a more efficient way in mixed than in pure stands. The advantages to mixed forest designs in terms of longterm effects on ecosystem nutrients have yet to be demonstrated, but mixed systems show promise as they may be preferred to diversify product outputs, decrease risks of failure, and contribute to species and landscape diversity.
Combinations of Plantation Trees with Agricultural Crops or Cattle (Agroforestry) In plantations for timber, with relatively slower growth and longer rotations than pulp or energy species, the first years of plantation establishment can be costly, and combinations with agricultural crops or cattle in agroforestry or agrosilvopastoral systems often can help to defray these costs. Based on available information on the agroecological characteristics of humid tropical regions, Alvim (1989) recommended agroforestry systems that include black
118
pepper, cacao, rubber, and some successful silvopastoral systems similar to those in the cacao region of Bahia and in the Amazon region of Brazil. Likewise, Sombroek (1992) pointed to timber plantations and agroforestry systems as approaches for the reclamation of degraded and abandoned areas. The potential of agroforestry as an ecologically and economically sound land use alternative in the Amazon is also discussed by Jordan (this volume). A variety of experimental settings involving forestry and agroforestry practices have been established by EMBRAPA over the last 10-15 years in uplands (terra firme) near Santarem, on the Tapajos river, a tributary of the Amazon. For example, along the Santarem/Cuiaba road, research has been initiated to help small land holders to adopt more environmentally sustainable farming practices and to obtain additional sources of capital. Fast growing species with high economic value have been introduced into farming areas. Systems include combinations of corn, banana, freijo (Cordia goeldiana), and mahogany (Swietenia macrophylla) (Brienza-Junior and Yared 1991). Other agroforestry models involve combinations of crops, fruit trees, and timber tree species designed to provide immediate returns and longer-term investment for the small holders. Food crops associated with tree establishment reduced tree plantation costs and decreased the frequency of crop weeding. For example, the trees tatajuba (Bagassa guianensis), parapara (Jacaranda copaia), and freijo were planted in a mixture with cowpea (Vigna unguiculatd) with good results in terms of crop yields and tree growth. The use of agrosilvopastoral systems is an alternative to the ranching system operated over large areas of Amazonia, which results in progressive loss of soil fertility. Such agrosilvopastoral systems have been introduced to the Paragominas region of eastern Para using the forestry species
Florencia Montagnini
Eucalyptus tereticornis, B. guianensis and Schizolobium amazonicum, the crop Zea mays, and the forage species Brachiaria brizantha, B. humidicola and Panicum maximum. Performance of the forestry and forage species has been satisfactory and the maize yield was greater than the regional average. The cattle, at a low stocking rate, grazed normally on the site and caused no harm to the planted trees (Brienza-Junior and Jared 1991). The sustained management of tree plant-ations and tree-crop combinations (agroforestry) are potentially biologically and socially feasible alternatives for soils unsuitable for local agriculture because of severe degradation. Tree plantations or agroforestry also represent productive uses of lands that have poor regeneration due to a lack of nearby sources of forest propagules. As the area of degraded land expands, there is an increasing emphasis on the planting of tree species which can grow in such conditions and still yield potentially profitable products (such as timber and fuelwood) as well as environmental benefits (such as soil conservation and watershed protection). The choice of appropriate species is the key to success of such systems. Ideally, species should be indigenous to the region, have fast growth, good economic value, and positive impacts on soil fertility. Additionally, management systems must be compatible with local farming and cultural practices and avoid increasing labor requirements beyond acceptable levels. These conditions should aid the adoption of these systems by local farmers. Because of the importance of nutrients in soils of humid tropical regions, nutrient cycling studies in timber plantations and agroforestry systems, including those using valuable timber species, are needed to determine appropriate designs and management so that systems and practices can be promoted throughout the Amazon basin.
Nutrient Considerations in the Use of Silviculture for Land Development and Rehabilitation in the Amazon
Conclusions Though the total area in the Amazon basin dedicated to plantation silviculture is relatively small compared to other managed systems, plantations have been encouraged as an alternative land use for the region, and their area has been increasing steadily over the last decade. Large scale plantations with fast-growing exotics planted in monospecific stands can deplete soil nutrients in a few rotations and often suffer from pest and disease problems. Most of the ecosystem nutrient losses occur during plantation establishment because of the removal of forest debris at clearing and soil disturbance during mechanized operation. Long-term productivity of short rotation plantations can generally be sustained with the application of fertilizers, in heavily subsidized systems much like those in southern Brazil and in several industrialized countries of the temperate region. In comparison with natural forests, plantations tend to have a smaller standing biomass, and a smaller proportion of their biomass in the root system. Therefore, belowground nutrient cycling in plantations is potentially diminished compared with natural forests. Management guidelines generally focus on the aboveground component of plantations since the ultimate goal is the production of wood and other aboveground tree products. There is an important need to understand the role of roots and root symbionts in acquiring nutrients and water in plantations since they control carbon allocation patterns within plants.
119
Nutrient cycling in small-scale timber plantations in the Amazon has not been the subject of extensive study. However, using longer rotations of native species, or trees in combination with agriculture, these systems are a potentially more sound land use option for the region, as they can produce high quality timber, contribute to the restoration of degraded lands, and serve a rural development purpose. A considerable number of valuable timber species can be grown in plantations in the Amazon region, and a few silvicultural guidelines are available to aid in their establishment and management. Further information on the silvicultural management of preferred timber species would enhance the successful establishment of these plantations. Choice of species for land rehabilitation should be based on their nutrient use efficiencies (NUE) and growth rates. Recent studies show the recycling abilities and ameliorating effects of indigenous species on soils and suggest guidelines for their use in economically and ecologically sound land use systems. Species with high NUE values can produce high biomass on poor sites, while species with low efficiencies would be preferred for nutrient recycling. Mixed-species designs can be more advantageous than tree monocultures for site nutrient rehabilitation if systems are planned to complement each species' nutrient demands and effects. Acknowledgments: Victoria Derr provided useful comments and editorial remarks.
Literature Cited Alvim, P. 1989."Tecnologias appropriadas para a agricultura nos tropicos umidos." Agrotropica 1: 5-26. Bloomfield, J., K. A. Vogt, and D. J.Vogt. 1993. "Decay rate and substrate quality of fine roots and foliage of two
tropical tree species in the Luquillo Experimental Forest, Puerto Rico." Plant and Soil 150: 233-245. Brienza-Junior, S., and J. A. G. Yared. 1991. "Agroforestry systems as an ecological approach in the Brazilian
120
Amazon development." Forest Ecology and Management 45: 319-323. Bowen, G. D., and E. K. S. Nambiar. 1984. Nutrition of plantation forests. Academic Press, New York. pp. 516. Bruijnzeel, L. A. 1984. "Immobilization of nutrients in plantation forests of Pinus merkusii and Agathis dammara growing in volcanic soils in central Java, Indonesia." In: Proceedings of International Conference on Soils and Nutrition of Perennial Crops, eds. A. Tajib, and E. Pushparajah (Malaysian Soil Science Society, Kuala Lumpur, Malaysia), pp. 19-29. Bruijnzeel, L. A. 1989. "Nutrient content of bulk precipitation in south-central Java, Indonesia." Journal of Tropical Ecology 5: 187-202. Carpanezzi, A. A., M. Kanashiro, I. A. Rodrigues, S. BrienzaJunior, and L. C. T. Marques. 1982. "Informacoes sobre Cordia alliodora (R. & P.) Oken na Amazonia Brasileira." Unidade Regional de Pesquisa Florestal Centro Sul, EMBRAPA, Brazil. No. 10: 105-115. Claussi, A., D. Marmillod, and J. Blaser. 1992. Descripcion silvicultural de las plantaciones forestales de Jenaro Herrera. Instituto de Investigaciones de la Amazonia Peruana, Iquitos, Peru. pp. 334. Drechsel, P., and W. Zech. 1993. "Mineral nutrition of tropical trees." In: Tropical Forestry Handbook, Volume 1, ed. L. Pancel (Springer-Verlag, New York), pp. 516-567. Evans, J. 1992. "Plantation forestry in the tropics." In: Tree planting for industrial, social, environmental, and agroforestry purposes. 2nd. ed. Clarendon Press. Oxford. England, pp. 403. Fearnside, P. M., and J. M. Rankin. 1980. "Jari and development in the Brazilian Amazon." Interciencia 5: 146-156. Fearnside, P. M. 1989. "The charcoal of Carajas: a threat to the forests of Brazil's eastern Amazon region." Ambio 18: 141-143. Fearnside, P. M. 1990. "Predominant land uses in Brazilian Amazonia." In: Alternatives to Deforestation: Steps Towards Sustainable Use of the Amazon Rain Forest, ed. A. B. Anderson (Columbia University Press, New York), pp. 233-251. Ferraz, J. B. S. 1993- "Soil factors influencing the reforestation on mining sites in Amazonia." In: Restoration of tropical Forest Ecosystems, eds. H. Lieth, and M. Lohman (Kluwer Academic Publishers, The Netherlands), pp. 47-52. Greaves, A. 1979. "Gmelina-. large scale planting, Jarilandia, Amazon basin." Commonwealth Forestry Review. 58: 267-269. Guariguata, M. R., R.. Rheingans, and F. Montagnini, 1995. "Early woody invasion under tree plantations in Costa Rica: implications for forest restoration." Restoration Ecology 3: 252-260. Hoyos, J. B, F. J. Bruseke, A. K. O. Homma, J. B. Veiga, N. M. Silva, and W. L. Overal. 1992. Desenvolvimento sustentavel: um novo caminho? J. B. Hoyos, editor. Serie Universidade e Meio Ambiente. Universidade Federal do Para, Nucleo de Meio Ambiente, Belem, Brazil, pp. 119. Irion, G. 1981. Holzplantage im Urwald? aturwissnschaften 68: 133-138. Johnson, N. E. 1976. "Biological opportunities and risks associated with fast-growing plantations in the tropics." Journal of Forestry 74: 206-211.
Florencia Montagnini Jorgensen, J. R., and C. G. Wells. 1986. "Tree nutrition and fast-growing plantations in developing countries." International Tree Crops Journal 3: 225-244. Knowles, O. H., and J. A. Parrotta. 1995. "Amazonian forest restoration: an innovative system for native species selection based on phenological data and field performance indices." Commonwealth Forestry Review 74: 230-243. Lugo, A. E. 1992. "Comparison of tropical tree plantations with secondary forests of similiar age." Ecological Monographs 62: 1-41. Lundgren, B. 1980. Plantation forestry in tropical countries-physical and biological potentials and risks. Swedish University of Agricultural Sciences. International Rural Development Centre." Rural Development Studies 8. Uppsala, pp. 134. Magalhaes, L. M. S., W. E. H. Blum, and N. P. Fernandes. 1986. "Crescimento de Eucalyptus deglupta Blume em solos de diferentes texturas. Caracteristicas edaficonutricionais de plantios florestais na regiao de Manaus." 1. Acta Amazonica. 16-17: 509-521. McNabb, K., J. Borges, and J. Welker. 1994. Jari at 25. "An investment in the Amazon." Journal of Forestry 92: 21-26. Montagnini, F., K. Ramstad, and F. Sancho. 1993. "Litterfall, litter decomposition and the use of mulch of four indigenous tree species in the Atlantic lowlands of Costa Rica." Agroforestry Systems 23: 39-61. Montagnini, F. 1994. "Rehabilitating forest ecosystems in the humid tropics: Recent experiences from Latin America." In: Proceedings from IUFRO International Symposium on Growth and Yield of Tropical Forests. Sept. 26^Oct. 1, 1994 (Tokyo University of Agriculture and Technology. Fuchu, Tokyo, Japan), pp. 224-234. Montagnini, F., and F. Sancho. 1994. "Above-ground biomass and nutrients in young plantations of four indigenous tree species: implications for site nutrient conservation." Journal of Sustainable Forestry 1: 115-139. Montagnini, F., A. Fanzeres, and S. G. da Vinha. 1994. "Studies on restoration ecology in the Atlantic Forest region of Bahia, Brazil." Interciencia 19: 323-330. Montagnini, F., and R. Mendelsohn. 1996. "Managing forest fallows: improving the economics of swidden agriculture." Ambio 26(2): 118-123. Montagnini, F., E. Gonzalez, R. Rheingans, and C. Porras. 1995. "Mixed and pure forest plantations in the humid neotropics: a comparison of early growth, pest damage and establishment costs." Commonwealth Forestry Review 74: 306-314. Nambiar, E. K. S. 1984. "Plantation forests: their scope and perspective on plantation nutrition." In: Nutrition of Plantation Forests, eds. G. D. Bowen and E. K. S. Nambiar (Academic Press, New York), pp. 1-15. Nepstad, D., C. Uhl, and E. A. S. Serrao. 1991. "Recuperation of a degraded Amazonian landscape: forest recovery and agricultural restoration." Ambio 20: 248-255. Palmer, J. R. 1977. "Forestry in Brazil-Amazonia." Commonwealth Forestry Review 56: 115-130. Pedroso, L. M. 1973a. Informacoes sobre o atual comportamento de especies exoticas na regiao do Medio Amazonas. SUDAM-Document 5: 21-31.Superintendencia do Desenvolvimento da Amazonia (SUDAM), Belem, Para, Brazil.
Nutrient Considerations in the Use of Silviculture for Land Development and Rehabilitation in the Amazon
Pedroso, L. M. 1973b. Alguns aspectos sobre o florestamento e reflorestamento na Amazonia. SUDAM Document 5:35-49. Superintendent do Desenvolvimento da Amazonia (SUDAM), Belem, Para, Brazil. Perry, D. A., and J. Maghembe. 1989. "Ecosystem concepts and current trends in forest management: time for reappraisal." Forest Ecology and Management 26: 123-140. Pitcher, J. A. 1976. Forestry development and research. Brazil. A tree improvement programme for Amazonia. FAO-Report. 1976, No. FO: DP/BRA/71/545 Technical Report 3. pp. 42. Rankin, J. M. 1985. "Plantations." In: Key Environments, eds. G. T. Prance and T. E. Lovejoy (Amazonia. Pergamon Press, New York), pp. 379-387. Russell, C. E. 1987. "Plantation forestry. Case study No. 9: The Jari project, Para, Brazil." In: Amazonian Rain Forests. Ecosystem Disturbance and Recovery. Ecological Studies 60, ed. C. F. Jordan (Springer-Verlag, New York), pp. 76-89. Sanchez, P. A., C. A. Palm, C. B. Davey, L. T. Szott, and E. C. Russell. 1985- "Tree crops as soil improvers in the humid tropics?" In: Attributes of trees as crop plants, eds. M.G.R. Cannell, and J.E. Jackson (Institute of Terrestrial Ecology, Natural Environmental Research Council, Abbots Ripton, Huntingdon, England), pp. 327-350. Silva, N. M, and Uhl, C. 1992. "Forest management for timber production: A sustainable use for the Brazilian Amazon." Anais da Academia Brasileira de Ciencias 64 (suppl. 1): 89-95. Sombroek, W. G. 1992. "Strategies for protection of soil fertility in the Amazon Region." In: Sustainable land use systems and human living conditions in the Amazon region: proceedings of a meeting held in Bonn, Germany, 1-2 November 1991, eds. H. Jaenicke, and P. Flynn (Commission of the European Communities, Luxembourg), pp. 43-46. Soto, G., Alvarado, A. y Montagnini, F. 1996. "Encalado de especies forestales nativas." In: Suelos: «;Puede la
121
Agricultura Sostenible ser Competitiva? Memorias, X Congreso Nacional Agronomico y de Recursos Naturales/II Congreso de Suelos. Volumen III. Ed., ed. F. Bertsch, W. Badilla, and y E. Bornemisza (Universidad Estatal a Distancia/Ed. de la Universidad Nacional, San Jose, Costa Rica), pp. 144. Spangenberg, A. 1994. Nahrstoffvorrate und exporte von Eucalyptus urograndis, Plantagen in Ostamazonien (Jari), Brasilien. Gottingen Beitrage zur Land und Forstwirtschaft in den Tropen und Subtropen No.93. Gottingen, Germany; Institut fur Pflanzenbau und Tierhygiene in den Tropen und Subtropen. SUDAM. 1979. Caracteristicas silviculturais de especies nativas e exoticas dos plantios do Centro de Technologia Madeireira-Estacao Experimental de Curua-Una. Superintendencia do Desenvolvimento da Amazonia, Belem, Brazil, pp. 351. Uhl, C., D. Nepstad, J. M. Cardoso da Silva, and I. Vieira. 1991. "Restauracao da floresta em pastagems degradadas." Ciencia Hoje 13: 22-31. Vogt, K., H. Asbjornsen, A. Ercelawn, F. Montagnini, and M. Valdes. 1997. "Roots and mycorrhizas in plantation ecosystems." In: Management of Soil, Nutrients and Water in Tropical Plantation Forests, eds. E. K. S. Nambiar, and A. G. Brown.(ACIAR/CSIRO/CIFOR. ACIAR, Canberra, Australia), pp. 247-296. Wadsworth, F. H., 1983. "Production of usable wood from tropical forests." In: Tropical Rain Forest Ecosystems. Structure and Function. Ecosystems of the World 14A, ed. F. B. Golley (Elsevier, New York), pp. 279-288. Wang, D., F. H. Bormann, A. E. Lugo, and R. D. Bowden,. 1991. "Comparison of nutrient-use efficiency and biomass production in five tropical tree taxa." Forest Ecology and Management 46: 1-21. Yamazaki, S., A. Taketani, K. Fujita, C. P. Vasques, and T. Ikeda, 1990. "Ecology of Hypsipyla grandella and its seasonal changes in population density in Peruvian Amazon forest." Japan Agricultural Research Quarterly 24: 2, 149-155.
8 Extractive Reserves and Participatory Research as Factors in the Biogeochemistry of the Amazon Basin I. Foster Brown, Karen A. Kainer, Andrea S. Alechandre, Eufran do Amaral
The word Amazonia conjures up diverse images, ranging from an exotic jungle to resources for development to a vast web of ecosystems that interact with global element cycles-the focus of this book. This chapter examines the biogeochemical role of extractive reserves, a relatively new land use type within Amazonia in which nontimber forest extraction is the defining human activity. The chapter also provides examples of how participatory research with local communities can enhance the quality of the results and improve their transmission to society. Humans have been a part of the Amazon for the past several thousand years. Amerindian activities have affected forest structure in significant manners by selective planting and clearing (Balee 1989) and by increasing fire frequency, particularly during mega-El Nino events (Meggers 1994). During the last few centuries, neo-Europeans have tragically reduced native indigenous populations by several million and made wide-scale transformations in the tropics of the Americas (Crosby 1993, Ribeiro 1996). The booms in rubber extraction in the late 1800s and during World War II brought waves of nonindigenous migrants to Brazilian Amazonia (Dean 1989). More recently, largescale implantation of cattle ranching and colonization projects, and to a lesser degree, mining activity, have accelerated change in Amazonian landscapes (Schmink and
Wood 1992). In addition, the ensuing road network and infrastructure left in the wake of these recent activities increased access to primary forest, precipitating further deforestation. By 1996, about 52 million hectares, nearly the size of France, had been deforested in Brazilian Amazonia (INPE 1998). At the average rate of deforestation from 1992 to 1996 (1.9 million hectares per year), another area equivalent to this figure will be added by the year 2025, a time frame within the career of many reading this book. Continuation of the present trends will result in an increasing savannization of the Amazonian region, with pastures, secondary forests, and crop lands expanding into areas once occupied by closed-canopy forests. This phenomenon may also be called the "Africanization" of Amazonia because most of the pastures are planted with grasses imported from Africa, such as Bracharia brisanthum, which are notably different in their response to rainfall patterns and to fire than the forests that they replace. This human-induced transformation leads to major impacts on biogeochemical cycles, many of which can be considered deleterious and are detailed in the other chapters. The tragedy of this type of forest conversion in Amazonia is that in spite of its costs in environmental terms, little economic, social, or cultural benefit accrues to local populations. The resultant landscape is one
Extractive Reserves and Participatory Research as Factors in the Biogeochemistry of the Amazon Basin
123
that is economically as well as ecologically research that involves local researchers and impoverished (Anderson 1990). This crisis communities in the research process in an has forced critical examination of alterna- attempt to enhance local capabilities for tive land use strategies that encourage improved resource management. simultaneous economic use and biological Extractive Reserves and conservation of existing forests. Extractive reserves are one of the few Biogeochemical Cycling land use alternatives in Brazilian Amazonia that reconciles traditional resource use patFrom a biogeochemical perspective, land terns that maintain forest cover with subsis- uses such as extractive reserves that maintence and cash needs of local populations. tain most forest cover and ecosystem funcFirst established in 1990 in western tion intact are least likely to modify element Brazilian Amazonia, extractive reserves are cycling. Extractive reserves are actually government-owned, protected lands desig- comprised of a mosaic of land uses, includnated for sustainable extraction and conser- ing pasture areas, shifting cultivation plots vation of renewable natural resources by (including various successional stages), and resident populations with a tradition of mature forest. The impact on biogeocheminon-timber extraction (Allegretti 1989, 1990, cal cycles of each of these land uses can be 1994). This land use strategy is unique in evaluated independently as has been done that it institutionalizes collective land use in previous chapters. However, in the case areas for nonindigenous populations, and of extractive reserves, the relative percentassigns management responsibilities to age of land devoted to pasture and shifting local people who have a long-term stake in cultivation plots is small in comparison to maintenance of the resource base. the area covered by mature forests. Extractive reserves have been praised as an Nontimber forest product extraction forms important land use alternative that combines the basis for extractivist livelihoods; hence, sustainable development and conservation of for economic reasons, most forest cover forest cover (Fearnside 1989, Alegretti 1994, is maintained. In addition, shifting cultivaAnderson 1992). They have also been tion and pasture areas within reserves are criticized as having limited chances for also relatively small and within close proxeconomic success (Browder 1992a, 1992b, imity to mature forest. Therefore, once Homma 1989) and for their limitations as abandoned, these agricultural plots convert conservation units (Redford and Stearman rapidly to secondary forest. 1993), particularly for megafaunal diversity Extractive reserves provide a new concept (Almeida 1992). The authors view them as for Amazonia. Most implemented models for evolving social experiments that will take Amazonian development and conservation years, if not decades, to fully evaluate their are based on a preservationist/developmenenvironmental, political, and socioeconomic talist dichotomy (Foresta 1991). In this strengths and weaknesses. dichotomy, forests and their ecosystem The purpose of this chapter is twofold: components are to be preserved in parks first, to analyze the role of extractive and biological reserves. Areas designated reserves as a potential factor in Amazonian for development are exploited at varying biogeochemistry, using changes in forest degrees of intensity that usually involve cover and land use as drivers, and the car- the complete elimination of forest cover bon cycle as the example; and second, to (pasture and annual crops) or partial extracdiscuss a participatory model for scientific tion and degradation (selective logging). This
I. Foster Brown et al.
124
Table 8.1 Distribution of extractive reserves and settlements in Brazilian Amazonia.
Extractive Settlements (INCRA)* Project Name
Decree and Date of Creation
Municipality/State
Area (ha)
Population Capacity
Antimary Canary
Port. 1055/98 de 28.07.88
Boca do Acre/AM Bujari/AC Epitaciolandia/AC
260,227 8054
4340
Port. 011/97 of 10/03/97 Port. 158/88 of 08/03/88
Chico Mendes -(Cachoeira) Limoeiro Porto Dias Porto Rico Remanso Riozinho Santa Quiteria Santo Antonio Mourao Terrua Maraca Praialta/Piranheira
24898
135 340
Port. 11/98 of 19/03/98
Bujari/AC
11,150
Resol. 40/89 of 20/10/89 Resol. 43/91 of 11/07/91 Port. 472/87 of 04/06/87 Resol. 39/89 of 20/10/89 Port. 886/88 of 24/06/88 Resol. 1855/92 of 20/08/92
Acrelandia/AC Epitaciolandia/AC Capixaba/AC Sena Madureira/AC Brasileia/AC Eirunepe/AM
22,145 7530 39,570 35,896
Port. 01/89 of 04/01/89 Port. 017/97 of 28/04/97
Pauini/AM Mazagao/AP
2320 5340
Port. 42/97 of 22/08/97
Nova Efigenia/PA Subtotal for extractive settlements
139,236 363,500 22,000 999,936
17,695
44,205 21,525
185 415 230 790 600 750 1000
1250
Extractive Reserves (IBAM A)** Name
Decree and Date of Creation
Principle Municipalities
Area (ha)
Estimated Population
Alto Jurua
98.862/90- 23/01/90
506,186
6000
Chico Mendes
99.144/90-12/03/90
970,570
7500
Rio Cajari Ciriaco Quilombo do Frexal Mata Grande Rio Ouro Preto Extreme Norte do -Tocantins
99.145/90-12/03/90 534/92-20/05/92 536/92-20/05/92 532/92-20/05/92 99.166/90-12/03/90 535/92-20/05/92
Cruzerio do Sul.Taumaturgo de Azevedo/AC Rio Branco, Xapuri, Brasileia, Assis Brasil/AC Laranjal doJarimasaga/AP Imperatriz/MA Imperatriz/MA Imperatriz-Joao Lisboa/MA Guajara-Mirim/RO Augustinopolis Sampaio/TO8280
481,650 7050 9542 10,450
5000 1150 900 1500 3410
Subtotal for extractive reserves Total of reserves and settlements
2,198,311 3,199,691
204,583 2000
27,460 46,155
Source: Data from INCRA * The National Institute for Agrarian Reform (INCRA) is responsible for establishing extractive settlements. ** The Brazilian Institute for Environment and Natural Renewable Resources (IBAMA) has governmental responsibility for management of extractive reserves. Population estimates for extractive reserves based on an average of 5 persons/ family. Population estimates are of variable accuracy and differences between areas reported in decrees and digitized from maps can vary by several percent. State extractive reserves from Rondonia are not included.
Extractive Reserves and Participatory Research as Factors in the Biogeochemistry of the Amazon Basin
preserved/exploited dichotomy is useful for maintenance of biodiversity of rare species, particularly megafauna such as jaguars and harpy eagles that have large natural ranges and require minimal forest disturbance. But between the end points of the landuse spectrum (national parks and extensive pastures/ intensive agriculture), lies a sizable portion of Brazilian Amazonia where human activity exists, but at relatively low intensities. It is for this region where lessons learned from extractive reserves have implications for how human populations and conservation of ecosystem function can coexist. Extractive reserves are not the panacea for forest conservation of all of Amazonia; the appropriate conditions for such reserves exist at maximum for 25 percent of Brazilian Amazonia (Meneses 1994). Nevertheless, such an area would exceed that of France. At present, extractive reserves administered by IBAMA, the Brazilian Institute for Environment and Natural Renewable Resources, and extractive settlements administered by INCRA, the National Institute for Agrarian Reform, cover about three million hectares, representing about one percent of the closed canopy forest in Brazilian Amazonia (see Table 8.1). From a carbon perspective, these reserves and settlements store about 6 x 108 Mg (tons) C (400 Mg/ha x 0.5 Mg C/Mg biomass x 3 x 106 ha). This is equivalent to about 10 percent of the global annual flux of carbon to the atmosphere from fossil fuel sources. Currently, the state of Acre in western Brazilian Amazonia has nearly half the area of extractive reserves and settlements in Amazonia (Table 1). Of the 1.7 million ha in reserves and settlements within the state, Chico Mendes Extractive Reserve is the largest, covering 970,000 ha. It is located in the Brazil nut-rich southeastern portion of Acre and it is this extractive region that is the focus of our subsequent discussions.
125
Extractivist Maintenance of Forest Cover Approximately 94% of the total area within Chico Mendes Extractive Reserve is retained in mature forest designated for nontimber forest extraction (Brown et al. 1992). This defining activity of extractive reserves provides the principle cash-generating products of the reserves, primarily rubber and Brazil nuts, and a wide variety of raw materials for home consumption such as fuelwood, wild fruits, medicines, game, and construction and artisan materials (Kainer and Duryea 1992). The provision of both cash and subsistence products makes the standing forest valuable to extractivist families, motivating them to maintain a high percentage of their land in mature forest rather than convert it to other forms of land use. The forest is also a key component in shifting cultivation, the mode of crop production in Acre's extractive reserves. Each family clears and burns approximately 0.25 to 2 ha of secondary or primary forest each year in preparation for a 3- to 6-year cropping cycle involving rice, corn, beans and finally manioc. When the forest is burned, some of the Ca, K, and Mg in the biomass is converted to ash and deposited on the soil surface, and perhaps more importantly, burning dramatically raises soil pH, increasing phosphorous availability (Nye and Greenland 1964, Ewel et al. 1981). After the cropping cycle, some of the land may be converted to pasture, or more typically, successional vegetation invades the site, thus beginning a 10- to 15-year fallow period. The fallow helps sustain agriculture by positively affecting two key elements: (1) restoration of soil fertility, and (2) decline in agricultural pest populations (Ewel 1986). Therefore, the mature forest and subsequently secondary forest, or fallow, provides critical biological services to the extractivists. Without these services, sustained crop production would become prohibitively costly in terms of labor
126
I. Foster Brown et al.
The basic equation is: as well as other energy inputs such as fertilizers and pesticides. Carbon content in _ carbon in biomass + The enormous value Acre's extractivists a landscape unit carbon in soil (1) place on the forest is exemplified in the well-documented and highly publicized Most research focuses on the time rate of rubber tapper resistance movement. During change of these stocks, stratified by vegetathe wholesale conversion of forest to tion type. Using the chain rule (see appenpasture during the 1970s and 1980s, many dix), we can derive the following relationship extractivist families in Acre risked their lives where the time rate of change of carbon in when threatened with expulsion from their the landscape unit is a function of four forested lands (Allegretti 1990, Hecht and groups of fluxes. These fluxes are listed as aCockburn 1989). Indeed, it was this d for each of the vegetation types movement that eventually led to the in the landscape unit: creation of the first extractive reserves in Change in carbon content in a landscape Acre. Simply stated, deforestation eliminates unit = the resource base from which the extractivist (a) Carbon in vegetation x migration of livelihood depends. vegetation type +
Carbon cycling and extractive reserves
(b) Carbon in secondary regrowth x area of vegetation type +
To help understand the effect of extrac(c) Erosion x carbon content of soil tive reserves on biogeochemical cycles, we in vegetation type + have selected carbon as the keystone reference element. The cycling of carbon drives (d) In situ changes of soil carbon x area of vegetation type, (2) the cycling of many other elements such as nitrogen and potassium. Carbon composes about half the dry biomass of tropical summed over all the vegetation types in the forests and consequently reflects ecosystem landscape unit. These fluxes are a product of natural and structure. Carbon stocks in biomass can vary more than ten-fold between primary anthropogenic factors, as discussed in the forest ecosystems and the agroecosystems other chapters. Flux (a) includes deforestathat replace them. Soil organic carbon also tion, natural migration of vegetation types, strongly affects the physical and chemical and pasture expansion. In the case of deforproperties of most tropical soils; it modifies estation where pasture substitutes forest, the the ion exchange capacity of soils, nitrogen rates of migration of forest and pasture are stocks, soil structure, and soil water reten- equal in magnitude but opposite in sign. The tion. Because much of the soil carbon is flux from this process equals the difference located close to the soil surface, its stocks in carbon content between forest and can be affected by changing land use. pasture. As the carbon content of tropical We can apply differential calculus and pastures is typically an order of magnitude the law of mass conservation to derive a less than that of closed canopy forests, this conceptual model for carbon dynamics in difference is roughly equal the original forest a forested landscape unit where carbon carbon content, when averaged over several is sequestered in either biomass (alive or burning seasons. dead but identifiable) and soil carbon (see The term "secondary regrowth" in flux (b) appendix for derivation). for forests also includes natural daily,
Extractive Reserves and Participatory Research as Factors in the Biogeochemistry of the Amazon Basin
seasonal, decadal, or secular changes in forest biomass that can affect the carbon content of the forest. For the purpose of this chapter, we will first examine how human activities affect these fluxes, estimating the relative importance of the various anthropogenic fluxes of equation (2) for Brazilian Amazonia; the landscape unit for our estimates is therefore the entire original forested region (Table 8.2). Then, we will demonstrate how extractive reserves and extractivist activities are related to these broader anthropogenically induced fluxes. The relative magnitudes of the fluxes described in equation (2) can be compared graphically in Fig. 8.1. Although the erosion flux may be important on a local basis, it probably represents more a redistribution of carbon between the terrestrial and fluvial systems than a flux to the atmosphere. The dominant gas flux is from deforestation. Consequently, land uses that maintain forest cover and/or reduce the deforestation rate will have a major impact on the anthropogenic fluxes, thereby affecting the total
127
atmospheric carbon flux in the Amazon basin. It is in this context that extractive reserves, and other land uses that maintain forest cover, have their importance for biogeochemical cycling. Extractive reserves can reduce deforestation by two mechanisms. First, local residents limit deforestation in their specific colocacoes or landholdings within the reserve in which they reside. Currently, federal regulations governing extractive reserve activities restrict deforestation to 10 % of the extractivist's landholding (Brasil 1995). As a result, a typical extractivist household that manages a landholding of 350 ha, acts as a guardian for more than 50,000 Mg of carbon of their managed forest (Brown et al. 1992). The fate of that carbon is largely determined at the household level, where decisions are made as to whether the high-biomass closed-canopy forest or a converted low-biomass agroecosystem provides a higher quality of life for that household. A second mechanism for reducing deforestation results from the buffering effect of reserves, limiting the extension of cattle
Fig. 8.1 A graphical representation of carbon fluxes directly associated with anthropogenic activities in Brazilian Amazonia, based on Table 8.2. Note the importance of the deforestation flux, calculated as committed emissions.
I. Foster Brown et al.
128
Table 8.2 Estimates of relative importance of the various anthropogenic fluxes of equation (2) for Brazilian Amazonia. Component
Estimate*
Source
Other estimates*
Flux (a): Deforestation rate
1.9 x loV"1
Original forest mass
400 Mg ha'1
INPE (1998), average for 1992-1996 One significant figure estimate, based on other estimates
464 Mg ha'1 Fearnside (1997), 424 Mg ha"1 Brown et al. (1992)
Carbon content of biomass (primary and secondary forests)
0.5 Mg C (Mg biomass)'1
Idem
0.48 Mg C (Mg biomass)'1 Carvalho et al. (1995)
Committed carbon emissions from biomass burning/decomposition
+400 x 106 Mg C yr1
Result of calculation, to one significant figure
+261 x 106 Mg C yr1 Fearnside (1997)
Regrowth rate of secondary forests
4 Mg ha'1 yr1
Salomao et al. (1996)
4-8 Mg ha'1 yr1 Saldarriaga et al. (1988) 6 Mg ha"1 yr"1 Schroeder and Winjum (1995)
Area of secondary forests
26 x 106 ha
one-half of deforested area in 1996 (INPE 1998)
17.5x 106 ha Schroeder and Winjum (1995)
Sink of carbon by secondary forests
- 50 x 106 Mg C yr1
Result of calculation, to one significant figure
-29 xlO6 Mg C yr1 Fearnside and Guimaraes (1996)
Flux (b):
Flux (c): Erosion loss
???
Flux (d): In situ change in soil carbon per unit area
-1 to 2 Mg C^ha-V calculated from Trumbore over a decade et al. (1996)
Area
52 x 106 ha
deforested area in 1996 (INPE 1998)
Flux
-50 to 100 x 106 Mg C yr1
Result of calculation, to one significant figure
*+ signifies source to atmosphere, - signifies sink from atmosphere.
Extractive Reserves and Participatory Research as Factors in the Biogeochemistry of the Amazon Basin
ranching and colonization projects. Differences between forest cover in Chico Mendes Extractive Reserve and the surrounding countryside are marked (Fig. 8.2), and the eastern boundary of the reserve was drawn precisely to halt the expansion of the ranches. At the scale of this image, extractivist clearings for shifting cultivation plots and pastures do not appear. The buffering role of these reserves only increases in importance as frontier expansion continues. The Chico Mendes Extractive Reserve, for example, borders the Brazilian Highway BR-317 that someday will connect western Brazilian Amazonia with a Peruvian seaport.
The Role of Research in Extractive Reserve Viability Although Acrean extractivists have been utilizing and managing their forests for decades, they are constantly challenged to successfully address ever-changing constraints and opportunities in the local, regional, national, and global socioeconomic and environmental arena. These spatial and temporal factors are multiple (Table 8.3), and their complexity is aggravated by their interactions as can be seen by citing a few examples. From a regional perspective, some parts of the Chico Mendes Extractive Reserve are dominated by extensive (101 to 104 hectares) patches of bamboo (Guadua sp). In these patches, transportation is difficult and Brazil nut trees are nearly absent; not surprisingly, these areas have the lowest population densities in the Reserve. In this case, regional environmental factors affect not only abundance of economic species, but also inhibit transportation of goods and general communication. On a global scale, success of rubber plantations in diverse areas such as Malaysia and southern Brazil have drastically altered the price of natural rubber, making extractive rubber production less lucrative
129
(Homma 1992). At these same scales, Acrean rubber tappers, lead by Chico Mendes in the mid 1980s, forged alliances with the global environmental movement that brought them international recognition and eventually led to the creation of extractive reserves (Hecht and Cockburn 1989, Keck 1995). Technological innovations, road networks, and changing national aspirations have all affected extractivist activities. In addition to environmental and socioeconomic drivers operating at various spatial and time scales, extractivists are faced with changing political and conceptual challenges. For example, the Chico Mendes Extractive Reserve joined over forty seringals into one
Fig. 8.2. Landsat TM image of the central portion of the Chico Mendes Extractive Reserve in eastern Acre, Brazil. The white line outlines the boundary of the reserve. North is to the top of the image, which covers about 90 km x 90 km. Xapuri is the local urban center. The light areas are deforested lands, principally large cattle ranches. The national highway BR-317 threads through the center of the deforested lands along the eastern side of the image. Data is from August 1989, TM bands 3, 4, and 5.
I. Foster Brown et al.
130
Table 8.3 Some factors affecting the viability of extractive reserves as functions of temporal and spatial scales. Socioeconomic Factors*
Environmental Factors*
Local
Household structure, local community organization, access to information
Distribution and abundance of economic species, topography, soil type
Regional
Transport network, education and health services, regional organizations and markets
Large-scale dieoffs of dominant species (e.g. bamboo), faunal distributions, alteration by smoke of Photosynthetically Active Radiation (PAR)
National
Policy decisions (e.g. rubber subsidies, creation of reserves), inflation control, recession
Global
Commodity prices, political pressure from other countries
Climate changes, particularly of rainfall patterns (e.g. El Nino)
Seasonal
Cash inputs and outputs
Excessive rainfall, dry season, burning activity
Annual
Market fluctuations, health of individual household members, fluctuating strength of representative political and economic organizations
Fluxes in productivity of economic species (e.g. Brazil nut), pest and predator populations
Decadal
Changes in political climate, household labor availability by life cycle
Recruitment of economic species, El Nino events
Generational
Education and cultural transmission
Scale Spatial
Temporal
* Note that factors also interact across space and time as well as across environmental and socioeconomic domains.
political entity. Extractivists in this reserve are faced with forging a new identity in order to make such an entity work. The rising importance given to biodiversity conservation has provoked the debate as to whether the conservation role of these extractive reserves is to be a byproduct of their economic activities or should be subsidized
by national and global beneficiaries (Almeida 1994, Carvalho and Brown 1996). These issues are complex and research can help understand that complexity and tease out the more critical points. While extractivists have made heady gains in their struggle for land rights, access to social services, and cultural dignity (Keck 1995),
Extractive Reserves and Participatory Research as Factors in the Biogeochemistry of the Amazon Basin
there still exists a pressing need for sustainable economic options and technical assistance in extractive reserves. The scientific community can play an important role in facilitating the identification, testing, and evaluation of the myriad production and forest conservation strategies open to these rain forest peoples. Moreover, scientists can further enhance research effectiveness by embracing a participatory approach in their research efforts, collaborating with both the extractive community and local professionals who contribute to reserve viability.
Participatory research as one form of scientific inquiry The decision to adopt a participatory approach in our research reflected the premise that the responsibility of environmental scientists is to provide reliable information that will help society resolve its environmental problems. While the premise may be laudatory, it is slippery to apply. For example, the typical means of supplying new information to society takes the form of scientific papers published in English in international journals (e.g. Brown et al. 1992, Kainer et al. 1998). In Brazilian Amazonia, we have found, however, that information in this format does not easily reach the local and regional decisionmakers, few of whom read English or have effective access to scientific literature. Also, those who have the most at stake, the local communities, require a different mode of information transfer. We also considered that involvement of extractivists in our research was essential. Ultimately, the extractivists are the ones who will make the decision to retain the majority of their land under forest cover or convert it to other types of managed land use less favorable for maintenance of natural biogeochemical cycles. In addition, by proactively seeking collaboration with local professionals, we felt that we could use our knowledge,
131
contacts, and financial resources to learn from and train a cadre of local individuals and organizations interested in the fate of extractive reserves. In adopting a participatory approach, we expected that the research process, not simply the findings, would provide a rich, immediate learning experience for all who make forest management decisions affecting extractivist living conditions and conservation of their environment. These considerations led us to adopt a more participatory approach, the principles of which are listed in Table 8.4. The following two case histories illustrate aspects of a more participatory approach to research that we have used in an attempt to enhance economic viability in extractive reserves while minimizing disruption of forest cover. The objectives of our research were both result and process oriented. First, our goal was to provide information to those who make decisions concerning reserve management. Second, we wanted to execute our research in such a way as to build the capacity of those who are most intimate with the day-to-day management of the reserves so that they may use the information. The first case concerns research directed to enriching the forest with Brazil nut trees, and the second addresses mapping as a key activity for local land use control.
Brazil Nut Research Brazil nut (Bertholletia excelsa Humbolt & Bonpland) is arguably the most important species to many extractivists in eastern Acre due to its multiple indigenous uses and the cash income generated from nut sales. Ecologically, the importance of Brazil nut within the local ecosystem is equally impressive (Ortiz 1995), and further, this species exemplifies the functional role of large trees on the biogeochemical cycle in extractive reserves. As the epithet excelsa implies, Brazil nut is a giant in the Amazon forest. It is a
132
I. Foster Brown et al.
Due to the fundamental role of Brazil nut dominant, upper-canopy tree often reaching 50 m in height, found on nonflooded lands in extractivist livelihoods, and ultimately in (terra firme) in the Amazon basin and the the success of extractive reserves, the Guianas (Prance 1990). In eastern Acre, Brazil National Council of Rubber Tappers called nut trees comprised 20% of the total biomass for research on Brazil nut ecology, processof a 5 ha survey of the primary forest ing, and political economy (ECOTEC and (F. Brown, personal observation). Clearly, on CNS 1990). One specific recommendation was mass alone, Brazil nut trees figure prominent- to increase Brazil nut densities through enrichly in the broader biogeochemical cycles of ment plantings (CNS et al. 1991)- In response, participatory research was carried out in 1993 the region. Human interaction with Brazil nut is long- to 1994 in three extractive communities near standing. Throughout its range, Brazil nut Xapuri, Acre, to assist extractivists in deteroften has an erratic distribution, occurring in mining which sites within the reserves are the stands of 50 to 100 individuals with each most appropriate for establishing Brazil nut stand separated from another by distances of seedlings. The specific research objectives of up to 1 km (Prance and Mori 1979). These the doctoral study were to examine and comdistribution patterns have been attributed to pare environmental and socioeconomic conAmerindian interventions (Miiller et al. 1980, straints to seedling establishment in forest Balee 1989). Even the type specimen of the gaps, shifting cultivation plots, and pastures first recorded description of the species made (Kainer et al.1998). Aside from these specific by Humboldt and Bonpland in 1807 came research objectives, the scientists had a from a tree in a Venezuelan churchyard that second participatory agenda that included the may have been cultivated from a Brazilian following modest development and empowseed source (Prance 1990). Brazil nut erment objectives. Carry out the research "on-site" in order to currently provides a host of raw materials used for household consumption in extrac- insure that research treatments, results, and tive reserves from food and medicinal prod- recommendations better respond to the socioecoucts to children's games. Most critically, nut nomic realities of the extractive reserves. sales are currently the greatest source of cash Selection of the specific landholdings for income for the many extractivist households, seedling establishment took place in a generparticularly those living in southeast Acre al community meeting wherein the research proposal was introduced and feedback solicited. (Campbell 1996). Table 8.4 Working guidelines for participatory research in extractive reserves. 1. Seek situations where local organizations are effective, available, and interested enough to incorporate the results of the research (multiplier effect) 2. Carry out research "on-site" if experimental results are to be applied locally 3. Clarify both researcher and participant expectations 4. Develop a base of trust by becoming involved in community functions, even if unrelated to the research 5. Become as fluent as possible in local language and culture 6. Include as many local institutions as possible in the research process 7. Include a training component for building capacity & interest of local professionals 8. Return research results in a timely manner and in forms appropriate for local consumption * Note that not all items are necessary nor desirable in all research situations.
Extractive Reserves and Participatory Research as Factors in the Biogeochemistry of the Amazon Basin
133
Objectives, treatments, and procedures were EMBRAPA (Agronomic Research Institute of discussed in detail to insure that all volunteer Brazil) participated in various phases families clearly understood responsibilities if of the research. In addition, a local they chose to participate. nongovernmental organization, PESACRE The on-site nature of the research and (Agroforestry Systems Research and inclusion of socioeconomic variables great- Extension Group of Acre), was a key ly affected final results and recommenda- collaborator during the entire study. tions. Based on environmental variables Intensively train local professionals in the alone (light, water, and nutrients), shifting research process. Trained researchers in the cultivation plots and pastures appear to be Amazon are few, and a young, local agronoequally suitable for enrichment plantings. my student participated fully in the entire Employing solely socioeconomic criteria field research process. She subsequently (labor considerations and overall compati- received a grant to accompany the laboratobility with the extractivist system), forest ry and analysis phases of the research at the gaps are the best enrichment sites. University of Florida. During this period, an However, upon consideration of both envi- extension booklet of the Brazil nut research ronmental and socioeconomic variables, was also designed. shifting cultivation plots are clearly the best Expose extractivists to the research experience. sites for enrichment plantings, as confirmed Day to day exposure to field studies takes by growth results. the mystery out of research. This gives extracCarrying out research on site also assisted tivists a fuller understanding of how research in our understanding of the practical results and conclusions are reached, and it constraints of extractivist living as well as validates their own local experimentation. For important cultural and language elements. To example, one extractivist who was not even enhance our interactions with extractivists, directly involved in the research has we lived with extractivist families throughout constructed a tree nursery and is outplanting the study period, participating in as many his seedlings into his shifting cultivation plots. community functions as possible. In addition, as extractivists are exposed to Include as many local institutions as possible research, foreigners, and urban Brazilians, in the research process. Despite the fact that they are broadening their understanding of the rural poor are a major target group for where they fit within the global scheme. This many governmental and nongovernmental empowerment experience gives them greater organizations, there is limited exposure and social and technical skills for managing the understanding of rural realities on the part extractive reserves, and helps them refine of most of these local, urban-based their ability to negotiate with the outside research, development, and conservation world, thus increasing control of their own organizations. In addition to facilitating development. linkages between extractivists and the Return results in an appropriate and timely institutions interested in and/or responsible manner. Meetings about research progress for supporting extractive reserve success, were held periodically, and preliminary the personal exposure and involvement by results were returned one year into the study individuals of these organizations creates a at a final community meeting that included a powerful human connection that often band, dance, and banquet with traditional leads to further personal commitment. foods made of Brazil nuts. One and a half Professionals from CAEX (Agroextractive years later, final results and recommendaCooperative of Xapuri), IBAMA, and tions were returned in oral form to the
134
professional research and extension community in Acre, the local nongovernmental organization most responsible for extractivist needs (CAEX), and the extractivists in the reserves. An extension booklet outlining the study in a simple, didactic manner was also distributed at this time to the extractivist community. Clearly, these guidelines and the manner in which this one research project was carried out does not ensure improved forest management in the reserves, but we believe that the participatory approach we adopted contributed greatly to the observed positive outcomes. Of equal importance, we feel that the science itself was fortified by using a participatory approach. We were easily able to tap into local knowledge and experience that improved treatments and provided ecological insights into the biology and cultivation of Brazil nut trees.
Land use mapping using satellite imagery Spatial location forms part of the human experience. Extractivists use mental maps to navigate within their own landholdings, colocacoes, typically covering several hundred hectares. Contiguous landholdings compose a seringal or former rubber estate. These former estates range from a few thousands to tens of thousands of hectares in size. Extractivists normally know their own rubber estate extremely well and possibly neighboring ones; however, the larger extractive reserves cover several hundred thousand hectares, well beyond the mental maps of individual extractivists. The National Council of Rubber Tappers realized that management of these reserves would require geographic information of human populations, resources, and land use, in short, maps. They requested help in mapping the Chico Mendes Extractive Reserve (Brown et al. 1995).
I. Foster Brown et al.
Much of the information available from residents in the reserve is referenced to their landholdings. In order to create detailed thematic maps, our first step has been to map the location of about 700 landholdings living within the reserve. Naming the landholdings, however, can only be done with the active participation of community residents. We have found this participation to be most effective when residents gather for meetings with union representatives or during other similar activities. Once extractivists understand that these maps will enable them to communicate with other extractivists, union representatives, and government officials, we began the mapping exercise using Landsat TM images (Fig. 8.3). Extractivists superimpose their "mental maps" with the image and typically produce a map of clearings over thousands of hectares in a few hours. There have been several positive spinoffs from incorporating extractivists into the research process. These mapping visits have spawned demand in local communities for more training in mathematics in order to calculate areas for agroforestry systems. The use of Global Positioning System (GPS) units for locating landholdings stimulated extractivists to learn about coordinate systems. Frequently, we used inflatable globes that illustrated not only the logic behind latitude and longitude, but also helped extractivists understand the geographical context of their reserve. Two local professionals who engaged in the mapping activities have continued on for their Masters degrees, using their experiences as a basis for their dissertations. The mapping of the Chico Mendes. Extractive Reserve has served two purposes. It has helped generate information necessary for the management of the reserve, the initial goal of the research. More importantly, however, it has stimulated dozens of extractivists to learn more about their reserve and
Extractive Reserves and Participatory Research as Factors in the Biogeochemistry of the Amazon Basin
135
Fig. 83 Example of local professionals building the capacity of local extractivists in the use of satellite images for mapping. Black and white images and tracing paper are used for the exercise. On wall at back is a preliminary map of landholdings of the Chico Mendes Extractive Reserve made in collaboration with The Technological Foundation of the State of Acre (FUNTAC) and the Remote Sensing Center of IBAMA.
encouraged them to develop skills that are not only useful for mapping, but also in deciding their own fate. During the continued mapping of the Chico Mendes Extractive Reserve, we anticipate that extractivists will continue to take more responsibility for the mapping and posterior quality control, as well as to expand this mapping to other areas.
Ecological research with local participation The creation of extractive reserves alone will not resolve the issues of ecosystem impoverishment and alteration of biogeochemical cycles in Amazonia (Nepstad et al. 1992). These complex issues need to be addressed at several scales simultaneously (Carvalho and Brown 1996). Nevertheless, the explicit managerial position of extractivists as stewards of the reserves, and therefore the forests they inhabit, suggests that a more collegiate relationship between researcher and local resident is warranted. Extractive reserves therefore provide rich opportunities for participatory research in which scientists and local communities may act together for improved conservation and development.
There is no fixed manner in which participatory research should be conducted. The examples above are simply two ways in which ecological research "was carried out with a certain degree of participation by the extractivist community and local professionals. To our knowledge, little ecological or biogeochemical research has been carried out in a participatory manner, but there exists a great body of experiences with local participation in agronomic research (Hildebrand 1986, Chambers et al. 1989, Quiros et al. 1991). Documentation of participation in development work (Cohen and Uphoff 1977) and discussion on how to carry it out in the field (Bunch 1982) has an even longer history. Participatory conservation is a relatively new offshoot of these efforts, and Little (1994) provides a review of some issues and experiences relevant to the incorporation of local participation in conservation work. Specific methods and tools, as well as thoughtful discussion on local participation and empowerment, are becoming increasingly available for those working with rural peoples (FAO 1990, Slocum et al. 1995, Pretty et al. 1995), and can be adapted for use in scientific research. There is an opportunity to incorporate such participatory research in the program for
136
economic and ecological zoning of Brazilian Amazonia, under the ageis of federal and state agencies. As the scale of this zoning becomes more detailed, local communities can make significant contributions to thematic mapping and, through such activity, learn to participate more effectively in zoning decisions that will affect their lives.
Appendix. Derivation of the Model for carbon fluxes We can apply the rules of differentiation to derive a model of the change in carbon content of a landscape unit covered by vegetation types of differing carbon contents. Equation (A-l) expresses the carbon content of the unit as the sum of carbon in the biomass and in the soil of the vegetation types:
I. Foster Brown et al.
(A-l) with respect to time results in the following first-order differential equation that can be simplified by integrating the biomass component over the soil depth z, so that Bv is in more conventional values of mass/area, and assuming that fv is not a function of time:
Equation (A-2) summarizes several processes. The first term includes
, the rate of
migration of vegetation types for which the carbon flux is proportional to the biomass Bv. The second term,
, is often
considered secondary regrowth, but also can include daily, seasonal, decadal, or secular changes in forest biomass (Keller et al. 1996). The third term, -
where C is the mass of carbon in a threedimensional landscape unit, Av is the area covered by vegetation type v, Zv is the depth of soil to which the vegetation type affects carbon storage, Bv is the biomass of the vegetation type (above-and belowground) in mass per soil volume, fv is the fraction of the biomass that is carbon in vegetation type v, Pv is the soil density, and Sv is the fraction of the soil that is carbon (mass/mass). All the values are averaged for the vegetation type. The total carbon is then the sum (D of the carbon in vegetation and soil of the various vegetation types that compose the landscape unit. Gaseous and dissolved carbon are considered to be of negligible mass. What is of interest are the time rate of change of carbon in the landscape and its forcing functions. Differentiating equation
5,,
, represents
primarily mass transport of soil and associated carbon via erosion or deposition associated with the vegetation type v, while the last term, , is the carbon flux induced by increased decomposition or in situ accumulation of soil carbon in vegetation type v. While equation (A-2) describes an instantaneous situation, the data available are typically over discrete time intervals. The time interval chosen will influence the conclusions. For example, averaged over one year, the fraction of biomass not burned in the initial fire has a great influence on flux estimates (Kauffman et al. 1995). However, when averaged over several years and repeated burns, the fraction of biomass not burned initially has no effect on the flux estimate
Extractive Reserves and Participatory Research as Factors in the Biogeochemistry of the Amazon Basin
137
Literature Cited Allegretti, M. H. 1989. "Reservas extrativistas: uma proposta de desenvolvimento da floresta Amazonica." Para Desenvolvimento 25: 3-29. Allegretti, M. H. 1990. "Extractive reserves: an alternative for reconciling development and environmental conservation in Amazonia." pp. 252-264. In: Alternatives to Deforestation: Steps toward Sustainable Use of the Amazon Rain Forest, ed. A. B. Anderson (Columbia University Press, New York), p. 281. Allegretti, M. H. 1994. "Reservas extractivistas: parametros para uma politica de desenvolvimento sustentavel na Amazonia." In: O destina da floresta: reservas extrativistas e desenvolvimento sustentavel na Amazonia, ed. R. Arnt (Relume-Dumara, Rio de Janeiro, Brazil), pp. 17-47. Almeida, M. W. 1992. "Reservas extractivistas como estrategia de conservacao de fauna." In: Manejo da vida silvestre para a conservacao na America Latina (Relatorio Tecnico, Belem, Para, Brazil), pp. 7-14. Almeida, M. W. 1994. "As reservas extractivistas e o valor da biodiversidade." In: O destina da floresta: reservas extrativistas e desenvolvimento sustentavel na Amazonia, ed. R. Arnt (Relume-Dumara, Rio de Janeiro, Brazil), pp. 259-276. Anderson, A. B., editor. 1990. Alternatives to deforestation: steps toward sustainable use of the Amazon rain forest. Columbia University Press, New York. Anderson, A. B. 1992. "Land-use strategies for successful extractive economies in Amazonia." Advances in Economic Botany 9: 67-77. Brazil, Ministerio do Meio Ambiente, dos Recursos Hidricos e da Amazonia Legal. 1995. Piano de utilizacao da Reserva Extrativista Chico Mendes. Institute Brasileiro do Meio Ambiente e dos Recursos Renovaveis. Brasilia, p. 32. Browder, J. O. 1992a. "Social and economic constraints on the development of market-oriented extractive reserves in Amazonian rain forests." Advances in Economic Botany 9: 33-41. Browder, J. O. 1992b. "The limits of extractivism." BioScience 42: 174-182. Balee, W. 1989. "The culture of Amazonian forests." Advances in Economic Botany 7: 1-21. Brown, I. F., D. C. Nepstad, I. Pires, and L. Luz e A. Alechandre. 1992. "Carbon storage and land-use in extractive reserves, Acre, Brazil." Environmental Conservation. 19(4): 307-315. Brown, I. F., A. Alechandre, H. S. Y. Sassagawa, and M. A. de Aquino. 1995. "Empowering local communities in land-use management: The Chico Mendes Extractive Reserve, Acre, Brazil." Cultural Survival Quarterly. Winter: 54-57. Bunch, R. 1982. Two Ears of Corn. World Neighbors, Oklahoma City, Oklahoma. Campbell, C. 1996. Forest, field and factory: changing livelihood strategies in two extractive reserves in the Brazilian Amazon. Ph.D. Dissertation, University of Florida, Gainesville. Carvalho, F. M. M., and I. F. Brown. 1996. "Polarization of biotic and economic wealth: the world, the tropics, and
Brazil." International Journal of Environment and Pollution 6(2): 160-171. Carvalho, Jr., J.A., J. M. Santos, J. C. Santos, M. M. Leitao, and N. Higuchi. 1995. "A tropical rain forest clearing experiment by biomass burning in the Manaus region." Atmospheric Environment 29(17): 2301-2309. Chambers, R. A. Pacey, and L. A. Thrupp, editors. 1989. Farmer first: farmer innovation and agricultural research. Intermediate Technology Publications, London. CNS (Conselho Nacional dos Seringueiros), Institute de Estudo Amazonicos, and The Ford Foundation. 1991. Seminario: alternativas economicas para as reservas extrativistas. Universidade Federal do Acre, Rio Branco, Acre, Brazil. Cohen, J., and N. Uphoff. 1977. Rural development participation: concepts and measures for project design, implementation and evaluation. Rural Development Committee, Cornell University, Ithaca, New York. Crosby, A. W. 1993- Ecological imperialism: The biological expansion of Europe, 900-1900. Cambridge University Press, New York. Dean, W. 1989. Brazil and the struggle for rubber. Cambridge University Press, Cambridge, England. ECOTEC (Tecnologias ecologicas) and CNS (Conselho Nacional de Seringueiros). 1990. Projeto Castanha: estrategias para aumentar os retornos economicos de populacoes extrativistas atraves do beneficiamento descentralizado da castanha. Unpub. Report. Ewel, J. 1986. "Designing agricultural ecosystems for the humid tropics." Annual Review of Ecology and Systematics. 17: 245-271. Ewel, J., C. Berish, B. Brown, N. Price, and J. Raich. 1981. "Slash and burn impacts on a Costa Rican wet forest site." Ecology 62: 816-829. FAO. 1990. The community's toolbox: the idea, methods and tools for participatory assessment, monitoring and evaluation in community forestry. Community forestry field manual 2. FAO (Food and Agriculture Organization of the United Nations, Rome. Fearnside, P. M. 1989- "Extractive reserves in Brazilian Amazonia." BioScience 39: 387-393. Fearnside, P. M. 1997. "Greenhouse gases from deforestation in Brazilian Amazonia: net committed emissions." Journal of Climate Change 35: 321-360. Fearnside, P. M., and W. M. Guimaraes. 1996. "Carbon uptake by secondary forests in Brazilian Amazonia." Journal of Forest Ecology & Management 80: 35-46. Foresta, R. A. 1991. Amazon conservation in the age of development. University of Florida Press, Gainesville, p. 366. Hecht, S., and A. Cockburn. 1989. The fate of the forest: developers, destroyers and defenders of the Amazon. Verso, London, Great Britain. Hildebrand, P. E., editor. 1986. Perspectives on farming systems research and extension. Lynne Rienner Publishers, Boulder, Colorado. Homma, A. K. O. 1989. "Reservas extrativistas: uma opcao de desenvolvimento viavel para a Amazonia?" Para Desenvolvimento 25: 38-48. Homma, A. K.O. 1992. "The dynamics of extraction in
138
Amazonia: a historical perspective." Advances in Economic Botany 9: 23-31. INPE. 1998. Divulgacao das estimativas oficiais do desflorestamento bruto na Amazonia Brasileira-1995, 1996 e 1997. Instituto Nacional de Pesquisas Espaciais. Sao Jose dos Campos, SP, Brazil. Kauffman, J. B., D. L. Cummings, D. W. Ward, and R. Babbitt. 1995. "Fire in the Brazilian Amazon. 1. Biomass, nutrient pools, and losses in slashed primary forests." Oecologia 104: 397-408. Kainer, K. A., and M. L. Duryea. 1992. "Tapping women's knowledge: plant resource use in extractive reserves in Acre, Brazil." Economic Botany 46: 408-425. Kainer, K. A. M. L. Duryea, N. C. de Macedo, and K. Williams. 1998. "Brazil nut seedling establishment and autecology in extractive reserves of Acre, Brazil." Ecological Applications 8: 397-410. Keck, M. E. 1995. "Social equity and environmental politics in Brazil: lessons from the rubber tappers of Acre." Comparative politics 27: 409-424. Keller, M., D. A. Clark, D. B. Clark, A. M. Weitz, and E. Veldkamp. 1996. "If a tree falls in the forest ...." Science 273: 201. Little, P. D. 1994. "The link between local participation and improved conservation: a review of issues and experiences." In: Natural connections: perspectives in community-based conservation, eds. D. Western, and R. M. Wright (Island Press, Washington, D.C., USA), pp. 347-372. Meggers, B. J. 1994. "Archeological evidence for the impact of mega-Nino events on Amazonia during the past two millennia." Climatic Change 28: 321-338. Menezes, M. A. 1994. "As reservas extrativistas como alternativa ao desmatamento na Amazonia." In: O destine da floresta: reservas extrativistas e desenvolvimento sustentavel na Amazonia, ed. R. Arnt (Dumara Dist. Rio de Janeiro), pp. 49-72. Miiller, C. H., I. A. Rodrigues, A. A. Muller, and N. R. M. Muller. 1980. "Castanha-do-Brasil: resultados de pesquisa." EMBRAPA, Centre de Pesquisa Agropecuario do Tropico Umido. Miscelanea 2. Belem, Para, Brazil. Nepstad, D. C., I. F. Brown, L. Luz, A. Alechandre, and V. Viana. 1992. "Biotic impoverishment of Amazonian forests by rubber tappers, loggers, and cattle ranchers." Advances in Economic Botany 9: 1-14. Nye, P. H. and D. J. Greenland. 1964. "Changes in the soil after clearing tropical forest." Plant and Soil 21: 101-112. Ortiz, E. G. 1995. Es o no es nuez. Americas. Organization of American States, Washington D.C. Prance, G. T. 1990. Bertholletia. In: Lecythidaceae-Part II: the zygomorphic-flowered New World genera
I. Foster Brown et al. (Couroupita, corythophora, Bertholletia, Couratari, Eschweilera, & Lecythis). Flora Neotropica Monograph 21, eds. S. A. Mori, and G.T. Prance (New York Botanical Garden, Bronx, New York.), pp. 114-118. Prance, G. T. and S. A. Mori. 1979- Lecythidaceae-Part I: the actinomorphic-flowered New World Lecythidaceae (Asteranthos, Gustavia, Grias, Allantoma, & Cariniana). Flora Neotropica Monograph 21. New York Botanical Garden, Bronx, New York. Pretty, J. N., I. Guijt, J. Thompson, and I Scoones. 1995. A trainer's guide for participatory learning and action. International Institute for Environment and Development, London. Quiros, C. A., T. Garcia, and J. A. Ashby. 1991. Farmer evaluations of technology: methodology for open-ended evaluation. Instructional unit No. 1. Centre Internacional de Agricultura Tropical, Cali, Colombia. Redford, K. H., and A.M. Stearman. 1993- "Forest-dwelling native Amazonians and the conservation of biodiversity: interests in common or in collision?" Conservation Biology 7: 248-255. Ribeiro, D. 1996. O Povo Brasileiro. A formacao e o sentido do Brasil. Segunda Edicao. Companhia das Letras. Sao Paulo, Brazil, p. 475. Salomao, R. de P., D. C. Nepstad, and I. C. G. Vieira. 1996. "Como a biomassa de florestas tropicais influi no efeito estufa?" Ciencia Hoje 21(123): 39-47. Saldarriaga, J. G., D. C. West, M. L. Tharp, and C. Uhl. 1988. "Long-term chronosequence of forest succession in the upper Rio Negro of Colombia and Venezuela." Journal of Ecology 76: 938-958. Schmink, M. 1992. "Building institutions for sustainable development in Acre, Brazil." In: Conservation of neotropical forests: working from traditional resource use, eds. K. H. Redford, and C. Padoch (Columbia University Press, New York, New York, USA) pp. 276-297. Schmink, M., and C. Wood. 1992. Contested Frontiers in Amazonia. Columbia University Press, New York. Schroeder, P. E., and J. K. Winjum. 1995. "Assessing Brazil's carbon budget: II. Biotic fluxes and net carbon balance." Forest Ecology and Management 75: 87-99. Slocum, R. L. Wichhart, D. Rocheleau, and B. ThomasSlayter, editors. 1995. Power, process, and participation: tools for change. Intermediate technology publications, London. Trumbore, S. E., E. A. Davidson, P. B. de Camargo, D. C. Nepstad, and L. A. Martinelli. 1995. "Belowground cycling of carbon in forests and pastures of Eastern Amazonia." Global Biogeochemical Cycles 9(4): 515-528.
9 The Recovery of Biomass, Nutrient Stocks, and Deep Soil Functions in Secondary Forests Daniel Nepstad, Paulo R. S. Moutinho, and Daniel Markewitz
Secondary forests cover approximately one third of the 0.5 million km2 of the Brazilian Amazon that have been cleared for agriculture (Houghton et al. 2000, Fearnside and Guimaraes 1996). These forests counteract many of the deleterious impacts of forest conversion to agriculture and cattle pasture. They absorb carbon from the atmosphere, they reestablish hydrological functions performed by mature forests, and they reduce the flammability of agricultural landscapes. Secondary forests transfer nutrients from the soil to living biomass, thereby reducing the potential losses of nutrients from the land through leaching and erosion. They also allow the expansion of native plant and animal populations from mature forest remnants back into agricultural landscapes. The study of forest recovery has focused on aboveground processes, primarily biomass accumulation. The few studies that have examined the recovery of belowground functions in Amazon secondary forests have been restricted to the upper meter or less of soil (e.g. Buschbacher et al. 1988). A review of our knowledge of secondary forest recovery is needed that incorporates accumulating evidence that approximately half of the region's forests rely upon root systems extending to depths of several meters to maintain evapotranspiration during prolonged seasonal drought
(Nepstad et al. 1994, Jipp et al. 1998, Nepstad et al. 1999a, Hodnett et al. 1997; see also Richter and Markewitz 1995). This discovery demands a conceptual shift in our approach to forest recovery on abandoned land. Are secondary forests capable of regrowing deep root systems, thereby recovering hydrologic functions and fire resistance of the mature forest? At what rate does this recovery take place? How does this ability to tap a large soil volume change our thinking about the role that nutrient shortages play in restricting secondary forest recovery? In this chapter, we begin to address these questions with the goal of furthering a mechanistic understanding of forest recovery on abandoned Amazonian lands. Our analysis focuses on three measures of secondary forest development: biomass accumulation, nutrient accumulation, and hydrological recovery. We choose biomass accumulation, because it is the best integrative measure of secondary forest development, it is the basis for estimates of carbon sequestration by secondary forests, and it is the most frequently measured secondary forest parameter. An analysis of nutrient accumulation allows us to examine the commonly held assumption that nutrient shortages limit rates of secondary forest recovery (e.g. Gehring et al. 1999, Tucker et al. 1998). Although hydrological recovery in secondary forests is poorly documented in
140
the literature, we discuss it here because of its considerable importance in reducing the deleterious effects of deforestation, .Osuch as stream flooding and deforestationprovoked changes in Amazonian rainfall (Nobre et al. 1991, Shukla et al. 1990, Victoria et al. 1991).
The Two Phases of Secondary Forest Development: Establishment and Growth There are two very distinctive phases in the development of secondary forests on abandoned land. In the first stage, which we call "secondary forest establishment," trees and vines colonize, and begin to overtop, the matrix of grasses, forbs and shrubs that typically occupy abandoned agricultural land. Trees and vines can establish in this grass/forb/shrub vegetation through the germination of seeds and the sprouting of roots that persist in the soil from prior to deforestation, and through the germination of seeds that are dispersed into the grass/forb/shrub vegetation from nearby forests (Nepstad et al. 1991, 1996a, Silva et al. 1996). Land use intensity determines the trajectory of secondary forest recovery largely through its influence on this establishment phase. Once trees and vines establish and overtop the grass/forb/shrub vegetation, the secondary forest enters the "growth" phase, in which the factors that control the rate of forest development are poorly understood.
Secondary forest growth: patterns of biomass accumulation Secondary forests accumulate biomass over time through the growth of woody plants. Biomass accumulation can be measured by monitoring permanent plots in secondary forests over a period of several years, or by using the chronosequence
Daniel Nepstad et al.
technique, in which the biomass of several secondary forests of different ages are compared at a single point in time. With rare exception, our knowledge of secondary forest biomass accumulation is based on the much faster chrono-sequence approach. We have assembled published estimates of secondary Amazon forest biomass as the basis for describing the general patterns of biomass accumulation (Fig. 9-1)- This review is restricted to estimates of aboveground biomass, since few studies have documented belowground biomass. Brown and Lugo (1990) have reviewed biomass accumulation in tropical secondary forests for all of the tropics. Three general observations emerge from this literature review. First, the rate of biomass accumulation is highly variable. After eight to ten years of development, secondary forests range from 10 to 90 metric tons of aboveground biomass (dry weight) per hectare, for an average annual accumulation rate of ~0 to 10 tons per hectare (Fig. 9.la). Second, the rate of biomass accumulation varies in part as a function of the history of the land on which it grows. At one extreme, the highest rates of accumulation were measured in secondary forests growing on land that was abandoned after a single slash-and-burn agricultural cycle, involving forest felling, burning, two years of crop production, and abandonment (Fig. 9-la, Saldarriaga et al. 1988). At the other extreme, the lowest rates of accumulation were measured in forests on land that had been converted to cattle pasture, grazed heavily for several years, burned for weed control several times, and herbicided and/or scraped with a bulldozer (for example, the sites with the lowest biomass accumulation after 3, 4 and 8 years of abandonment, Fig. 9.1a, Uhl et al. 1988). The intensity of land-use history explained much of the variation in rates of biomass accumulation in secondary forests on
The Recovery of Biomass, Nutrient Stocks, and Deep Soil Functions in Secondary Forests
abandoned cattle pastures, on similar soils, near Paragominas, Para (Uhl et al. 1988), and in rates of forest recovery on abandoned agricultural sites near San Carlos, Venezuela (Uhl et al. 1982, Uhl et al. 1990). The third pattern that is discernible from this review is the decline of accumulation rate with time (Fig. 9-lb). For example, 80-year-old forests on abandoned crop fields in the Venezuelan Amazon had accumulated only 150 Mg ha'1 of aboveground
141
biomass (~2 Mg ha'1 yr1, on average), and had the same amount of biomass as 60-yearold forests in the same region (Fig. 9-lb, Saldarriaga et al. 1988). After nearly a century of development, these forests had accumulated only 60% of the biomass of mature forests. In the Bragantine Zone east of Belem, where slash-and-burn agriculture has been practiced for a century, a 40-yearold forest had accumulated only 90 tons of aboveground biomass, which is one fourth
Fig. 9.1 Aboveground biomass estimates for Amazonian secondary forests, (a). Data for 0 to 20year-old forests, (b). Data for 0 to 80-year-old forests. The data points for forests that are 60 years and older all come from a single study of secondary forests on abandoned agricultural clearings in the Upper Rio Negro region of Venezuela (Saldarriaga et al. 1988). Other sources of data are: Aide 1993; Alves et al. 1997; Brown et al. 1992; Fearnside and Guimaraes 1996; Guimaraes and Fearnside unpublished manuscript, Moutinho 1998b; Nepstad 1989; Salomao et al. 1998; Uhl et al. 1988.
142
Daniel Nepstad et al.
that of a neighboring mature forest of land use effects on mechanisms of sec(Salomao et al. 1998). The long-term trajec- ondary forest establishment. With each additory of aboveground biomass accumulation tional land management practice-burning, in secondary forests is unknown, but the planting of crops and forage grass, appears to be much lower than it is during weeding, grazing, herbiciding, reburning, the first few years of recovery. bulldozing-the availability of tree propagSecondary forests withdraw carbon from ules is diminished, and the probability is the atmosphere through the accumulation reduced that trees will survive and grow if of biomass, counteracting the liberation of they are able to establish (Uhl et al. 1982, carbon to the atmosphere through the 1990; Nepstad et al. 1991, 1996b). Let us cutting and burning of mature forest. The consider the extreme cases to illustrate this magnitude of this carbon sequestration can point. If a forest is cut down and burned but be estimated by multiplying the area of never planted or grazed (e.g. Uhl and secondary forest by the average rate at Jordan 1984, Saldarriaga et al. 1988), then which secondary forest accumulates carbon. the establishment of secondary forest is The area of secondary forest is approxi- rapid. Tree and liana propagules are mately 170,000 km2 if we assume that one abundant in this situation, both buried in third of the area deforested by late 1998 the soil as seeds and sprouting stumps and (INPE 1998) now supports secondary forest roots, and carried into the cut-over site by (Fearnside & Guimaraes 1996). The average birds, bats, ground mammals, and wind. rate of biomass accumulation in above- There are few obstacles to the subsequent ground biomass is approximately 4 Mg yr1, survival and growth of these tree and liana which is 2 Mg C yr1 (Fig. 9.la). Approx- seedlings and shoots. imately 0.03 Pg of carbon are, therefore, At the other extreme, consider our study incorporated into aboveground biomass site, the Fazenda Vitoria, in eastern each year by these forests, which is approx- Amazonia, where the forest was cut down imately one tenth of the carbon released and burned, followed by planting with from the region through deforestation (0.27 pasture forage grass, grazing by cattle for Pg, Fearnside 1997), and a smaller portion eight years, and repeated burning and of the total carbon released through defor- weeding. Upon abandonment, this heavily estation, logging, and fire (Nepstad et al. used site was virtually devoid of tree and 1999a). This carbon sequestration is short liana propagules (Nepstad et al. 1996a), and lived, however, in that most secondary it supported dense vegetation that was a forests are eventually cut and burned, mixture of grasses, shrubs, and forbs. Root releasing carbon back to the atmosphere length density of this abandoned pasture (Fearnside and Guimaraes 1996, Moran et vegetation was four times higher than that al. 1994). The net effect of secondary forests of primary forest, and presented a high level on the carbon balance of the Brazilian of competition to young tree and liana Amazon appears to be small (Fearnside and shoots (Nepstad et al. 1996a). The Fazenda Guimaraes 1996). Vitoria site also supported a community of seed- and seedling-eating ants and rodents, Mechanisms of biomass accumulation which greatly reduced the survivorship of tree and liana seeds and seedlings. Grazing The considerable variation in the rate of had compacted the upper few centimeters biomass accumulation in secondary forests of soil, further reducing the probability of can be explained, in part, as a consequence seedling success on this abandoned site
The Recovery of Biomass, Nutrient Stocks, and Deep Soil Functions in Secondary Forests
143
(Moutinho 1998a, b; Nepstad et al. 1991, found no correlation between rates of 1996). Secondary forest establishment on aboveground biomass accumulation in this site depended upon the facilitation of secondary forests on abandoned pastures tree and liana establishment by small islands and soil fertility in the Paragominas region. of shrubs, such as Solanum crinitum and Similarly, the rate of biomass accumulation Cordia multispicata, which attracted seed- in secondary forests in the Zona Bragantina, carrying birds and bats onto the abandoned east of Belem, showed little response to site, ameliorated the harsh physical condi- several forms of fertilization, including tions of the abandoned pasture, and shaded complete nutrient solution (Gehring et al. out the weedy grass/shrub/forb vegetation 1999). We believe that an understanding of (Vieira et al. 1994, Nepstad et al. 1996a, the role of soil fertility as a determinant of Silva et al. 1996). Patches of secondary aboveground biomass accumulation in forest develop around these shrub islands, secondary forests requires information on and eventually coalesce into larger expans- the ability of secondary forests to absorb es of secondary forest. This process of facil- nutrients below the frequently used soil itation of tree and liana establishment, and sampling depth of 0.5 to 1.0 m, as we prescoalescence of secondary forest islands, can ent here for the secondary forest of Fazenda take several years on sites with histories of Vitoria, near Paragominas, eastern Para. heavy use, delaying the rate at which Nutrient Uptake on Deep Soils: biomass is accumulated. As the islands of trees and lianas begin The Case of Fazenda Vitoria to coalesce on the abandoned site, other Much of what we know about the nutrient factors may determine the trajectory of biomass accumulation in secondary forests stocks of secondary forests is based on soil with histories of heavy use. The most impor- measurements made to depths of only 0.1 to tant factor determining this trajectory is 1.0 m (Uhl and Jordan, 1984, Buschbacher certainly fire. Large areas of secondary et al., 1988, Koutika et al. 1997, Neill et al. forest burn each year in Amazonia (Nepstad 1997). However, water balance studies and et al. 1999b), setting back forest recovery direct measurements of deep soil moisture processes by killing aboveground tissues and roots provide evidence that forests and releasing to the atmosphere some of the across much of seasonally dry Amazonia depend upon root systems that extend well nutrients contained in biomass. A second factor that may determine the beyond this conventional sampling depth to rate of biomass accumulation is species absorb water during the dry season composition. Secondary forests on sites with (Nepstad et al. 1994, Jipp et al. 1998, histories of intensive land use, and that are Hodnett et al. 1997, Holscher et al. 1997). far from sources of tree seeds, may be The occurrence of root systems extending to depauperate in some of the fast growing 18 m depth in Amazonian forests (Nepstad pioneer species that are responsible for et al. 1994) demands a re-examination of rapid biomass accumulation during initial our thinking about the nutrient stocks of these ecosystems, and the recovery of these secondary forest growth (Uhl et al. 1988). One of the most commonly invoked nutrient stocks in secondary forests. If the impediments to secondary forest growth is rooting zone of Amazonian forests extends low nutrient availability (e.g. Tucker et al. to several meters depth, instead of several 1998), although hard evidence is lacking. centimeters depth, are these forests less Buschbacher et al. (1988), for example, vulnerable than previously believed to nutri-
Daniel Nepstad et al.
144
ent losses through leaching? Are the soil nutrient stocks that are available to these forests much greater than previous studies have estimated? Is the "direct nutrient cycling" described in aseasonal forests of the Venezuelan Amazon (Herrera et al. 1978) applicable primarily to regions of extreme soil acidity, and low fertility? Although a full response to these questions is beyond the scope of this chapter, we discuss here the potential of deep soil nutrient supplies to contribute to the nutrient stocks of secondary forests near Paragominas. Little direct evidence of deep root uptake of nutrients is available in the Amazon or elsewhere. Within our eastern Amazonian research site at the Fazenda Vitoria in Paragominas, Para, however, we have made measurements of soil and vegetation that help us bound estimates of the potential for deep root uptake of nutrients by secondary forests. Specifically, we have (1) estimated
the losses of the aboveground nutrient capital in the secondary forests and inferred a future need for nutrient supply, (2) evaluated the stocks of nutrients in the soil profile to 8 m depth and inferred a need for nutrient uptake from deep soil, (3) measured the density of fine roots and associated arbuscular mycorrhizae to estimate the forest's capacity to tap these nutrients, and (4) measured nutrient concentrations in deep soil solution to quantify the potential nutrient uptake through mass flow of water. To assess the potential need for deep soil nutrients we first evaluate the loss of aboveground nutrients during forest clearing. If through various mechanisms (i.e. soil surface charge) the nutrients from the primary forest were retained on site, no new nutrients •would be required to regenerate a secondary forest of equal stature. Or similarly, if losses were small, surface soil might contribute a sufficient quantity of nutrient such that deep
Table 9.1 Total nutrient stocks in mature and secondary forest aboveground biomass and in the upper 20 cm (meanil SD)of Oxisol (Haplustox) soils within the Fazenda Vitoria, Paragominas, Brazil.
c* (kgha-i)
N* (kg ha-i)
p* (kg ha-1)
K* Ca* (kg ha-i) (kgha-i) 343
487
131 75 ±14 45±21
Mg* (kg ha-i)
Mature forest
Biomass**
130,000
1368
41
soil (cm)
0-10
24,553 ±888
2238 + 132
194+8
34±9
257 + 100
10-20
16,895 ±606
1574 + 10
213 + 2
20±5
94+61
Total
171,448 + 1494
5180±142
448 ±10
397 ±14 838 + 161
251±35
Secondary forest
Biomass***
33,487
187
5
104
172
28
soil (cm)
0-10
29,102±4,360
2,582±334
210±27
40+8
691 ±88
144 ±29
10-20
16,792 + 1,492
1,705+113
205 + 13
24±6
327 ±85
88±11
Total
79,381 ±5852
4,474+447
420 ±40
168 + 14 1,190±173 260 ±40
* Soil contents are estimated from composite samples along three transects in each land use, C, N, and P are totals, K, Ca, and Mg are Mehlich III exchangeable (Markewitz and Davidson, unpublished data, 1997). ** Mature forest biomass was estimated by Nepstad (1989). Nutrient contents are estimated using average concentrations from other mature Amazonian forests. *** Secondary forest biomass was estimated by allometry using DBHs from ten permanent plots. Equations used are from Uhl et al., 1988 and Markewitz and Davidson unpublished data, 1996. Nutrient concentrations were estimated on site (Markewitz and Davidson, unpublished data, 1997).
The Recovery of Biomass, Nutrient Stocks, and Deep Soil Functions in Secondary Forests
root uptake would not be required. At the Fazenda Vitoria, we estimated nutrient stocks in both aboveground biomass of mature and secondary forest, and in the upper 0 to 20 cm of soil (only in this upper portion did we find significant changes from land use conversion). We hypothesized that nutrients not reaccumulated in the secondary forest or retained in the upper surface soils were lost from the forest ecosystem and that these losses will have to be contributed from new sources (i.e. atmosphere or soils) during forest regeneration (Table 9-1). Within this ecosystem comparison fairly large amounts of C and N have been lost during conversion, as have lesser amounts of P and K. Both Ca and Mg appear to have been fully retained on the soil exchange sites of the upper surface soil. In a future forest, C and N will most likely be supplied from atmospheric sources while Ca and Mg should require no additional inputs (this assumes that there is not a continued leaching loss of these elements from the site). For P and K, however, there is a clear need to identify potential soil sources for nutrient resupply.
145
Soil nutrients easily extractable with salt or dilute acid solutions are used to infer the quantities of plant nutrients available to regenerating forests (Jackson 1958, Mehlich 1978). In this highly weathered, low-charge Oxisol, concentrations of Mehlich III (0.2 M CH3COOH, 0.25 M NH4NO3, 0.015 M NH4F, 0.013 M HNO3, and ImM EDTA) extractable P are extremely low throughout the upper 8 m of soil (Table 9.2). In fact, the total stock of Mehlich III available P throughout the 8 m (13.4 kg ha"1) would not be sufficient to replace the apparent loss of 28 kg ha"1 during conversion of primary to secondary forest. These extremely low soil concentrations of extractable P would tend to confer a strong advantage to those trees able to explore a vast soil volume in hopes of beating their competitors to the available soil P pool. In the case of Mehlich III exchangeable K the full 8 m of the soil is not necessary to resupply the apparent loss of 229 kg ha"1 but a complete depletion of the upper two meters would be necessary. Naturally the other option is a smaller depletion of K throughout the 8 m profile.
Table 9.2 Soil chemical and physical parameters for an Oxisol (Haplustox) from Paragominas, Brazil, 1996 (meanilSD). Depth*
0-10 10-20 20-50 50-100 100-200 200-300 300-400 400-500 500-600 725-750 825-850
BD
3
(g cm" )
Mehlich III P (pg g'1)
1.02 ±0.03 0.83 + 0.07 1.18 ± 0.02 0.43 ± 0.03 1.26 ± 0.02 0.21 ± 0.07 1.28 ±0.02 0.17 + 0.12 1.31 ± 0.02 0.09 ± 0.05 1.32 ± 0.02 0.13 ± 0.08 1.31 ± 0.01 0.08 ± 0.04 1.31 ± 0.01 0.09 ± 0.10 1.26 ± 0.07 0.11 ± 0.10 1.30 ± 0.04 0.17 ± 0.13 1.31 ± 0.02 0.12 ± 0.04
NaOH P (pg g'1)
Total P (Pg g"1)
Exch K (cmolc kg'1)
Nonex Mehlich K III P 1 (cmolc kg" ) (kg ha"1)
54.1 ±3.3 27.0 + 2.7 16.9 + 0.7 14.6 ±0.6 9.3 ± 0.3 6.3 ± 0.1 3.9 ± 0.4 3.2 ± 0.6 3.0 ± 0.3 2.3 ± 0.2 2.2 ± NA*
209+13 181 ± 6 162 ± 6 146+10 142 ± 11 140 ± 20 143 ±5 156 ± 11 160 ± 19 181 + 23 180 ± 10
0.117 ±0.016 0.055 ± 0.013 0.038 ± 0.035 0.020 ± 0.020 0.012 ± 0.013 0.006 ± 0.004 0.004 ± 0.002 0.004 ± 0.002 0.003 ± 0.001 0.003 ± 0.003 0.004 ± 0.003
0.397 ± 0.028 0.349 ± 0.079 0.337 ± 0.037 0.342 ± 0.053 0.282 ± 0.036 0.236 ± 0.019 0.214 ± 0.035 0.184 ± 0.035 0.167 + 0.050 0.138 ± 0.069 0.172 ± 0.008
0.8 0.5 0.8 1.1 1.2 1.7 1.0 1.2 1.4 2.2
1.6
Exch K (kg ha"1)
46.7 25.4 56.2 50.1 61.5 31.0 20.5 20.5 14.8 15.3 20.5
Sources: Bulk Density (BD) data are from Nepstad (1989). Soil chemical data are from Markewitz and Davidson (unpublished, 1997), Exch. K is also by Mehlich III, Total P and nonexchangeable K are estimated from H2SO4/H2O2 digests. * Upper 0-20 cm samples are composite samples from secondary forest (n = 3). Samples below 20 cm are averages of two auger holes from four land uses (n = 8). ** NA-data is not available.
146
For both P and K there is some further question as to the availability of other soil extractable fractions. In the case of P, for example, fractionation schemes with various alkaline extracts have been used to quantify slowly available pools of inorganic or organic P. The P extractable with 1M NaOH (Tiessen and Moir, 1993), for example, is believed to be associated with the large pool of Fe and Al oxides present in these Amazonian Haplustox. This extractable P fraction currently only identifies a potential additional source of soil P, however, since
Daniel Nepstad et al.
its rate of availability is poorly quantified. In the case of K, non-exchangeable pools of K that are recoverable with strong extractants (Sparks 1987) and are often associated with micaceous interlayers may also provide a slowly available plant supply. The micaceous component constitutes only a small fraction of these highly weathered soils (O. Chadwick, unpublished data, 1995) but could provide an additional source of longterm K supply. If these larger pools of P and K were available to plants in the upper soils this would limit the need for deep root
Fig. 9.2 Fine-root biomass (0-1 mm diameter) to 6 m depth in adjacent mature forest, secondary forest on abandoned cattle pasture (16 years old), and active cattle pasture planted with Brachiaria brizantha forage grass. Each data point is the average value for at least 24 1.5-kg soil samples taken using long-handled auger borings. Roots were separated from the soil following Nepstad et al. (1994). Contamination of deep soil samples was avoided by lining the boring hole with PVC tubing prior to sampling.
The Recovery of Biomass, Nutrient Stocks, and Deep Soil Functions in Secondary Forests
uptake. On the other hand, the limited availability of these pools might again confer a strategy of exploring vast soil volumes to sequester P and K throughout the 8 m profile. The availability of scarce plant nutrients in deep soil is only relevant to our understanding of secondary forest nutrient acquisition if plants are able to absorb these nutrients with deeply penetrating root systems. Enticing evidence for such an extensive approach to nutrient acquisition is supported by the distribution of fine-root biomass (diameter = 0-1 mm) in the secondary forest, as compared to that in the neighboring mature forest and active cattle pasture (Fig. 9.2). Fine root biomass to 6 m depth is virtually identical in mature and secondary forests, and is > 10 times greater at depth than in the active cattle pasture. Hence, after 16 years of recovery, some of the trees, lianas and palms of the secondary forest had re-established root systems to at least 6 m depth. But are there sufficient fine roots at depth to absorb significant amounts of soil nutrients? In both mature and secondary forests, there is a hundred-fold decline in fine-root biomass from the soil surface to 6 m depth, which means that fine root length density at 6 m is less than 1 cm of root per 100 cm3 of soil. Moreover, the few roots that occur at depth are concentrated in patches of soft soil that comprise approximately 1% of the soil volume, and that show no nutrient enrichment (Carvalheiro and Nepstad 1996). Given the low mobility of phosphorus in the soil, it is unlikely that such a sparse, patchy root system could absorb substantial amounts of this scarce nutrient. Mycorrhizal associations can increase the absorptive capacity of sparse root systems. Fungal hyphae can penetrate the soil surrounding the infected root, absorbing P that is otherwise too far from the root surface for absorption to take place. Our
147
measurements of arbuscular mycorrhizal infection of live fine roots excavated at the Fazenda Vitoria study site show that fine roots at all depths are infected, and that the rate of infection is higher in secondary forests than in mature forests (Fig. 9-3). Hence, secondary forest plant species may have the potential to absorb deep soil nutrients of low soil mobility, such as P, despite the low length densities of their root systems. In fact, they may have greater capacity to absorb deep soil nutrients than mature forests because of higher mycorrhizal infection rates. Finally, nutrients in soil solutions within the lower portion of the soil profile can be incorporated during water uptake, a hydrologic process that has been well established within the lower soils of our eastern Amazonian forest (Nepstad et al, 1994; Jipp et al., 1998). The nutrients in solution deep in the soil profile may be derived from various sources. These infiltrating nutrients might be derived directly from atmospheric inputs of wet or dry deposition, they may be derived from exchange or decomposition processes in upper soil horizons, or they might be derived at depth from solidsolution interactions with the mineral elements present or from root exudates. The ability to identify the source of deep solution nutrients is limited. For some elements such as Ca, however, investigators have tried to use natural isotopic tracers, i.e. Sr 86/87, with some success (Miller et al., 1993; Bailey et al., 1996). Regardless of the source of deep soil nutrients, deep root uptake will increase the total nutrient pool available for growth. In the case of atmospheric inputs, deep roots will improve the retention efficiency of secondary forests, increasing the "tightness" of the nutrient cycle or in the second case, incorporating inputs derived from subsoils, deep roots are providing an additional nutrient pool for regeneration.
148
In the secondary lands of Paragominas, we have measured the soil solution concentrations from tension lysimeters at 0.25, 3.0, and 7.0 m. Tension lysimeters tend to collect soil waters that have infiltrated slowly through the soil profile and are in more intimate contact with soil minerals. For our elements of interest, P and K, there is a general decrease in the concentration of these elements with depth (Table 9.3). At 7.0 m depth, soil solution concentrations are less than or equivalent to those input in bulk
Daniel Nepstad et al.
precipitation (bulk precipitation indicates funnels for collection that are always open to the atmosphere). These direct comparisons of solution concentrations, however, do not account for the evapoconcentration effect of plant transpiration. As plants absorb and transpire water infiltrating through the soil, we expect nutrient concentrations to increase unless the plant root also takes up the nutrients. Thus, if we could compare the content (as opposed to the concentration) of K in kg ha'1 passing
Fig. 9.3 Arbuscular mycorrhizal associations in fine roots of adjacent mature forest, secondary forest on abandoned cattle pasture (16 years old), and active cattle pasture planted with Brachiaria brizantha forage grass. Each data point is the average value for six root samples, with 75 1-cm root segments stained and analyzed per sample following Phillips and Hayman (1970). Percent infection refers to the percent of each root segment's length in which fungal hyphae or arbuscles were visible.
The Recovery of Biomass, Nutrient Stocks, and Deep Soil Functions in Secondary Forests
by a given soil depth, these data would indicate an actual depletion of nutrients on a content basis. An estimate of the uptake of P and K available in the mass-flow of soil water driven by plant transpiration in this lower portion of the soil profile (4 to 8 m) can be derived using the hydrologic estimates from Jipp et al. (1998). Over a 4-year period (1991 to 1994) the depletion of volumetric water content averaged approximately 47 cm of water in this 4 m soil layer. This water uptake times the nutrient concentration at that depth indicates a small uptake of P but a potentially important uptake of K (Table 9.3). In both cases this nutrient uptake through mass-flow improves the "tightness" of the nutrient cycle but does not indicate if this is a retention of atmospheric deposition or an input of deep soil nutrients. The biogeochemical and biological evidence for deep root activity presented here for subsoil horizons of these highly weathered Oxisols provides a strong impetus for improving our quantitative understanding of deep soil nutrient supply. Our data reveal the potential importance of deep soil nutrients in the biogeochemistry of secondary forest ecosystems, and calls attention to the need for additional studies that incorporate a more complete concept of the soil rooting volume.
149
Hydrological Recovery In addition to providing access to the nutrients of deep soil layers, deeply penetrating root systems also enable Amazonian forests to absorb large amounts of deep soil water. This deep soil water uptake function allows mature forests in seasonally-dry Amazonia (where most secondary forests are found) to maintain their leaf canopies during dry periods of 3 to 6 months (Nepstad et al. 1995, Jipp et al. 1998, Hodnett et al. 1997). The remarkable capacity of mature forests to tap water stored more than 8 m below the soil surface allows these forests to pump water vapor into the atmosphere throughout the dry season, converting radiant energy into latent heat, and providing moisture for cloud formation and rain events downwind, to the west (Nepstad et al. 1996b, Nobre et al. 1991). Are secondary forests able to recover root systems that are more than ten meters deep? Water balance measurements from two studies conducted in Amazonia suggest that the answer to this question is "yes." After 15-18 years of development, the secondary forest at the Fazenda Vitoria evapotranspired at a rate virtually identical to that of a neighboring mature forest during a 4-year study period (Jipp et al. 1998). This study,
Table 9.3 Volume-weighted mean solution concentrations and fluxes for secondary forest in Paragominas, Para, Brazil. Solution
H2O (cm)
Bulk Precipitation 155 T 25 cm**** T 300 cm 47 T 750 cm
Conductivity
pH
(US cm'1)
5.9
5.10
25.3 14.6 14.9
5.30 5.26 5.16
Alkalinity** Oteq L'1)
Total P*** (Mg L'1)
K
P
0*g L'1)
(kg ha-1) (kg ha'1)
23.8 82.6 44.8 37.0
2
0.32
0.03
4.96
3 3 1
2.33 0.48 0.35
0.01
2.26
K
Samples were collected biweekly from May 1996 to May 1997. Hydrologic water volumes are averages from 1991 to 1994 and indicate precipitation inputs and water evapotranspired between 300 and 750 cm. Alkalinity is estimated by titration to a pH endpoint of 4.5. Total P = total dissolved P determined as molybdate reactive after filtration and persulfate digestion. T = tension lysimeter, the numbers indicate the soil depth for each collection in cm.
150
Daniel Nepstad et al.
which calculated evapotranspiration and Because of greater leaf shedding during deep soil drainage based on periodic meas- the dry season, secondary forests are more urements of soil water content to eight flammable than mature forests (Uhl and meters depth using Time Domain Reflect- Kauffman 1990, Nepstad et al. 1995, ometry, found that the secondary forest Fig. 9.4). A key element of forest flammabilwithdrew soil moisture from the deepest ity in Amazonia is the rate at which the (6-8 m) soil interval at a similar rate to that organic debris on the forest floor dry during of the mature forest. By recovering deep soil rainless periods, which is largely a function moisture uptake, the secondary forest also of the amount of sunlight that penetrates the had lower rates of deep soil water drainage leaf canopy. Hence, the very important role than a neighboring cattle pasture. We esti- played by Amazonia's mature forests in mated that no soil water drained into the blocking the spread of escaped agricultural water table from Paragominas mature and fires appears to take more than two decades secondary forests in 1992 and 1993, during to fully recover because of the lower resistandfollowing a severe El Nino drought, ance of secondary forests to droughtwhile 1 mm day1 of soil water drained from induced leaf-shedding (Fig. 9.4). a neighboring cattle pasture (Jipp et al. Further support of rapid recovery of 1998). The dry season desiccation of deep hydrological functions in secondary forest clay soils by forests reduces the amount of comes from the Zona Bragantina, east of soil water draining down to the water table Belem. Here, Holscher et al. (1997) used a and into streams, and therefore serves as an Bowen ratio approach to measure evapoimportant buffer against flooding. transpiration in a 2.5-3.5-year-old secondary The Paragominas secondary forest recov- forest on abandoned crop land, and found ered the hydrological functions of the an average daily rate (3-9 mm d"1) very mature forest even though it had not fully close to that of mature forests receiving recovered the mature forest's resistance to similar amounts of radiation. During the dry dry season leaf shedding (Fig. 9-4). During season, this secondary forest was absorbing the severe drought of the 1992 El Nino water from below 3 m depth, and therefore event, -when only 90 mm of rain fell on had recovered a portion of its deep soil Paragominas during a 5-5-month period, the water uptake capacity. leaf area of the mature forest declined by The recovery of deep soil water uptake in 15%, of the secondary forest by 25%, and of the Paragominas secondary forest was the pasture by 55%. Leaf shedding was possible because of the re-establishment of greater in the secondary forest than in the deep root systems following pasture abanmature forest even though the tensions that donment (Fig. 9-2). The root systems of developed in the xylem sap were greater in secondary forest trees, vines, and palms the mature forest for understory trees (min- rapidly penetrate to at least 8 m depth imum of -3-0 MPa) than in the secondary during the first 15 years of regrowth. We forest (-1.4 MPa, Fig. 9.4). Extremely identified one third as many "morphosdrought-tolerant trees of the mature forest, pecies" of roots to 8 m depth in the such as the Lecythis idatimon, Tachigalia secondary forest as in the neighboring myrmecophilia and Sterculia pruriens of mature forest, with a prevalence of vine and the understory (Fig. 9.4), may have to estab- palm roots in the secondary forest. The vine lish in the secondary forest before it can Davilla kunthii, for example, penetrates to at fully recovery the canopy resistance to leaf least 8 m depth by the time its stem has shedding of the mature forest. attained 1 m height (Restom 1998). Vines in
The Recovery of Biomass, Nutrient Stocks, and Deep Soil Functions in Secondary Forests
Fig. 9.4 Soil water content, leaf area, plant water stress, flammability and daily rainfall of mature forest, secondary forest and cattle pasture during the severe 1992 dry season. As deep soil water was depleted during this measurement period (a, b), severe drought stress developed in some trees of the mature forest (d), but the loss of green leaf area was lower in the mature forest than in the secondary forest and cattle pasture (c). Because of this capacity to retain leaves despite severe water stress, the mature forest is rarely susceptible to fire even during a severe dry season such as this (e). Plant-available soil water was measured from 0 to 2 m depth (a) and from 2 to 8 m depth (b) using Time Domain Reflectometry sensors imbedded in the walls of deep soil shafts (Nepstad et al. 1994, Jipp et al. 1998). The percent of maximum leaf area was measured based on monthly observations of 20 tagged branches in each of 10 to 13 species of trees and lianas (n = 3 individuals per species). Predawn leaf water potential (d), which is a direct measure of the amount of tension in the xylem sap (more negative numbers signify greater drought), was measured for the same trees and lianas as the leaf area measurements, using a pressure chamber. Flammability was estimated using rainfall data and the fine fuel drying rates reported by Uhl and Kauffman (1990) (e). Daily rainfall (f).
151
152
this secondary forest accounted for 4% of total basal area, and 8% of annual evapotranspiration (Restom 1998; Restom and Nepstad, in press).
Mechanisms of Rapid Root Penetration into Deep Soil: The Cutter Ant Hypothesis How are root systems able to penetrate more than 8 m into the soil during the first 10 to 15 years of secondary forest recovery? This question is particularly challenging considering the high bulk densities of these soils (ca. 1.3 g cm"3, Table 9.2), and their high resistance to penetration (4 to 6 MPa cnr2). Moutinho (1998b) hypothesizes that leaf cutter ants play an important role in facilitating the penetration of woody root systems following land abandonment in Amazonia. Cutter ant populations (e.g. Atta sexdens) often explode following deforestation, increasing their nest density ~30-fold above that of the mature forest. Secondary
Daniel Nepstad et al.
forests can harbor 10 enormous (50-m2) nests per hectare (Vasconcelos and Cherrett 1994, Moutinho 1998b). Although they are known primarily for the presumably deleterious effects on plants associated with their practice of clipping leaves and twigs to supply their subterranean fungus gardens, cutter ants may have a beneficial effect on plants because of their deep soil "tillage" activities. These ants excavate channels and chambers to soil depths of > 6 m, depositing organic matter and loose soil in many of these chambers, facilitating both the penetration of root systems to deep soil and the proliferation of root systems once they have penetrated (Fig. 9-5). Moreover, cutter ants reduce the resistance to penetration of deep, intact soil that has not been excavated, further facilitating the penetration and proliferation of deep root systems (Fig. 9.5). The influence of cutter ants on the belowground recovery of secondary forest may go beyond their role as tillers of deep soil. The soil beneath ant nests is also enriched in
Fig. 9.5 Soil resistance to penetration and fine-root length density in the soil beneath cutter ant nests (Atta sexdens) and beneath neighboring forest without nest influence.
The Recovery of Biomass, Nutrient Stocks, and Deep Soil Functions in Secondary Forests
nutrients. Soil from 1-3 m depth beneath cutter ant nests is enriched in Ca (3-4-fold), K (7-35-fold) and Mg (2-3-fold) when compared to soil of the same depths that lies beyond the influence of the nest mound (Moutinho 1998b). The competing cutter ant effects of defoliation v. deep soil tillage and nutrient enrichment upon secondary forest growth are difficult to measure, but point to the potential role of a soil-excavating insect in explaining the rapid rates of deep root system recovery in secondary forests of Amazonia.
Conclusions Despite their considerable importance in the ecological recovery of impoverished agricultural landscapes, our understanding of secondary forests in Amazonia remains superficial. We know that they generally accumulate biomass aboveground at the average rate of approximately 4 tons per hectare per year, and that the establishment stage of this regrowth can be limited by barriers to tree establishment and survival. We know that these forests have the potential to
153
re-establish deep root systems within a decade or two, thereby recovering important hydrological and, perhaps, nutrient uptake functions. We continue to lack, however, a mechanistic understanding of the factors that control the rates of these various aspects of forest recovery. The recent discovery that cutter ants may facilitate secondary forest recovery reveals the primordial state of our understanding of these under studied ecosystems. Acknowledgments: This chapter was supported by grants from the U.S. Agency for International Develop-ment, the National Science Foundation, and the A.W. Mellon Foundation to the Woods Hole Research Center. Further support was provided by the Conselho Nacional de Pesquisa (CNPq), to P. Moutinho. E. Cheng and V. Mendes Jr. assisted with data collection. K. Schwalbe and G. Carvalho assisted with data analysis and graphics. Dan Richter provided laboratory facilities. The Woods Hole Research Center and the Amazon Institute of Environmental Research (IPAM) provided institutional and logistical support.
Literature Cited Aide, T. M. 1993- "Patterns of leaf development and herbivory in a tropical understory community." Ecology 74(2): 455-466. Alves, D. S., J. V. Soares, S. Amaral, E. M. Mello, S. A. Almeida, O. F. Da Silva, and A. M. Silveira. 1997. "Biomass of primary and secondary vegetation in Rondonia, Western Brazilian Amazon." Global Change Biology 3: 451-461. Bailey, S. W., J. W. Hornbeck, C. T. Driscoll, and H. E. Gaudette. 1996. "Calcium inputs and transport in a basepoor forest ecosystem as interpreted by Sr isotopes." Water Resources Research 32: 707-719. Brown, I. F., D. C. Nepstad, I. d. O. Pires, L. Luz, and A. S. Alechandre. 1992. "Carbon storage and land-use in extractive reserves, Acre, Brasil." Environmental Conservation 19: 307-315. Brown, S. and A. E. Lugo. 1990. "Tropical secondary forests." Journal of Tropical Ecology 6:1-32. Buschbacher, R., C. Uhl, and E. A. S. Serrao. 1988. "Abandoned pastures in eastern Amazonia II. nutrient stocks in the soil and vegetation." Journal of Ecology 76: 682-699Carvalheiro, K., and D. C. Nepstad. 1996. "Deep soil hetero-
geneity and fine root distribution in forests and pastures of eastern Amazonia." Plant and Soil 182: 279-285. Fearnside, P. M. 1997. "Greenhouse gases from deforestation in Brazilian Amazonia: net committed emissions." Climatic Change 35: 321-360. Fearnside, P. M., and W. M. Guimaraes. 1996. "Carbon uptake by secondary forests in Brazilian Amazonia." Forest Ecology & Management 80: 35-46. Gehring, C. M. Denich, M. Kanashiro, and P. L. G. Vlek. 1999. "Response of secondary vegetation in Eastern Amazonia to relaxed nutrient availability constraints." Biogeochemistry 45(3): 223-241. Herrera, R., T. Merida, N. Stark, and C. Jordan. 1978. "Direct phosphorus transfer from leaf litter to roots." Naturwissenschaften 65: 208-209. Hodnett, M. G., J. Tomasella, A. d. O. Marques Filho, and M. D. Oyama. 1997. "Deep soil water uptake by forest and pasture in central Amazonia: predictions from long-term daily rainfall data using a simple water balance model." In: Amazonian Deforestation and Climate, eds. J. H. C. Gash, C. A. Nobre, J. M. Roberts, and R. L. Victoria (John Wiley & Sons, New York), pp. 79-100, 611 pages.
154
Holscher, D., T. D. de A. Sa, T. X. Bastos, M. Denich, and H. Folster. 1997. "Evaporation from young secondary vegetation in eastern Amazonia." Journal of Hydrology 193: 293-305. Houghton, R. A., D. L. Skole, C. A. Nobre, J. L Hackler, K. T. Lawrence, and W. H. Chomentowski. 2000. "Annual fluxes of carbon from deforestation and regrowth in the Brazilian Amazon." Nature 403: 301-304. INPE (Institute Nacional de Pesqui&s Espaciais). 1998. Desflorestamento. Sao Paulo, Brazil: Sao Jose dos Campos. Jackson, M. L. 1958. Soil chemical analysis, Prentice-Hall, Englewood Cliffs, NJ. pp. 82-109. Jipp, P., D. Nepstad, K. Cassle, and C. Reis de Carvalho. 1998. "Deep soil moisture storage and transpiration in forests and pastures of seasonally-dry Amazonia." Climatic Change 39 (2-3): 395-412. Koutika, L., F. Bartoli, F. Andreux, C. C. Cerri, G. Burtin, T. Chone, and R. Philippy. 1997. "Organic matter dynamics and aggregation in soils under rain forest and pastures of increasing age in the eastern Amazon Basin." Geoderma 76: 87-112. Mehlich, A.. 1978. "New extractant for soil test evaluation of phosphorus, potassium, magnesium, calcium, sodium, manganese and zinc." Communications in Soil and Plant Analysis 9: 455-476. Miller, E. K., J. D. Blum, and A. J. Friedland. 1993. "Determination of soil exchangeable-cation loss and weathering rates using Sr isotopes." Nature 362: 438-441. Moran, E., E. Brondizio, P. Mausel, and Y. Wu. 1994. "Deforestation in Amazonia: land use change from ground and satellite level perspectives." BioScience 44: 329-338 . Moutinho, P. R. S. 1998a. "Impactos da formacao de pastagens sobre a fauna de formigas: conseqtiencias para a recuperacao florestal na Amazonia oriental." In: Floresta Amazonica: Dinamica, regeneracao e manejo, eds. C. Gascon, and P. R. Moutinho (Manaus-Amazonas: INPA), pp. 155-170. Moutinho, P. R. S. 1998b. O papel das sauvas (Atta sexdens) na sucessao florestal em pastagens abandonadas na Amazonia. Ph.D. Thesis, Universidade Estadual de, Campinas, Sao Paulo. Nepstad, D. C., A. Verissimo, A. Alencar, C. Nobre, P. Lefebvre, P. Schlesinger, C. Potter, P. Moutinho, E. Lima, M. Cochrane, and V. Brooks. 1999a. "Large-scale impoverishment of Amazonian forests by logging and fire." Nature 398: 505-508. Nepstad, D. C., A. Moreira, and A. Alencar. 1999b. Flames in the Rainforest: Origins, Impacts and Alternatives to Amazonian Fire. Brasilia, Brazil: World Bank, Pilot Program for the Conservation of the Brazilian Rainforest. Nepstad, D. C., C. R. de Carvalho, E. A. Davidson, P. Jipp, P. Lefebvre, G. H. Negreiros, E. D. da Silva, T. Stone, S. Trumbore, and S. Vieira. 1994. "The role of deep roots in the hydrological and carbon cycles of Amazonian forests and pastures." Nature 372: 666-669. Nepstad, D. C., P. Jipp, P. Moutinho, G. Negreiros, and S. Vieira. 1995- "Forest recovery following pasture abandonment in Amazonia: Canopy seasonality, fire resistance and ants." In: Evaluating and Monitoring the Health of LargeScale Ecosystems, ed. D. Rapport (NATO ASI Series, New York: Springer-Verlag), pp. 333-349.
Daniel Nepstad et al. Nepstad, D. C., C. Uhl, C. A. Pereira, and J. M. C. da Silva. 1996a. "A comparative study of tree establishment in abandoned pasture and mature forest of eastern Amazonia." Oikos 76: 25-39. Nepstad, D. C., P. R. Moutinho, C. Uhl, I. C. Vieira, andj. M. C. da Silva. 1996b. The ecological importance of forest remnants in an eastern Amazonian frontier landscape. In: Forest Patches in Tropical Forest Landscapes, eds. J. Schelhas and R. Greenberg (Washington D.C.: Island Press), pp. 133-150. Nepstad, D. C., C. Uhl, and E. A. S. Serrao. 1991. "Recuperation of a degraded Amazonian landscape: forest recovery and agricultural restoration." Ambio 20 (6): 248-255. Nepstad, D. C. 1989. Forest recovery following pasture abandonment in eastern Amazonia. Ph.D. Dissertation, Yale University. Neill, C., J. M. Melillo, P. A. Steudler, C. C. Cerri, J. F. L. de Moraes, M. C. Piccolo, and M. Brito. 1997. "Soil carbon and nitrogen stocks following forest clearing for pasture in the southwestern Brazilian Amazon." Ecological Applications 7: 1216-1225. Nobre, C. A., P. J. Sellers, and J. Shukla. 1991. "Amazonian deforestation and regional climate change." Journal of Climate 4: 957-988. Phillips, J. M., and D. S. Hayman. 1970. "Improved procedures for cleaning roots and staining parasitic and vesicular-arbuscular mycorrhizal fungi for rapid assessment of infection." Transactions of the Mycological Society 55: 158-160. Restom, T. G., and D. C. Nepstad. In press. "Contribution of vines to the evapotranspiration of a secondary forest in eastern Amazonia." Plant and Soil. Restom, T. G. 1998. "Recuperacao do sistema radicular profundo em uma floresta secundaria na Amazonia oriental." Floresta Amazonica: Dinamica, Regeneracao e Manejo. C. Gascon, and P. Moutinho, 145-153. Manaus, Amazonas: Ministerio da Ciencia e Tecnologia Institute Nacional de Pesquisa da Amazonia. Richter, D. D, and D. Markewitz. 1995 "How deep is soil?" BioScience 45: 600-609. Saldarriaga, J. G., D. C. West, M. L. Tharp, and C. Uhl. 1988. "Long-term chronosequence of forest succession in the upper Rio Negro of Columbia and Venezuela." Journal of Ecology 76: 938-958. Salomao, R., D. C. Nepstad, and I. Vieira. 1998. "Biomassa e estoque de carbono de florestas tropicais primarias e secundarias." In: Floresta Amazonica: Dinamica, Regeneracao e Manejo, eds. C. Gascon, and P. Moutinho (Manaus, Amazonas: Ministerio da Ciencia e Tecnologia Institute Nacional de Pesquisa da Amazonia), pp. 99-119. Shukla, J., C. A. Nobre, and P. Sellers. 1990. "Amazon deforestation and climate change." Science 247: 1322-1325. Silva, J. M. C. da; C. Uhl, and G. Murray. 1996. "Plant succession, landscape management, and the ecology of frugivorous birds in abandoned Amazonian pastures." Conservation Biology 10(2): 491-503. Sparks, D. L. 1987. "Potassium dynamics in soils." Advances in Soil Science 6: 1-63. Tiessen, H., and J. O. Moir. 1993. "Characterization of available P by sequential extraction." In: Soil sampling and methods of analysis, eds. M. R. Carter (Lewis Pub., Boca Raton, FL), pp. 75-86.
The Recovery of Biomass, Nutrient Stocks, and Deep Soil Functions in Secondary Forests Tucker, J. M., E. S. Brondizio, and E. F. Moran. 1998. "Rates of forest regrowth in Eastern Amazonia: a comparison of Altamira and Bragantina regions, Para State, Brazil." Interciencia 23(2): 64-73. Uhl, C., and J. B. Kauffman. 1990. "Deforestation, fire susceptibility and potential tree responses to fire in the eastern Amazon." Ecology 71(2): 437-449. Uhl, C., R. Buschbacher, and E. A. S. Serrao. 1988. "Abandoned pastures in Eastern Amazonia, I: Patterns of plant succession." Journal of Ecology 76: 663-681. Uhl, C., C. F. Jordan, K. Clark, H. Clark, and R. Herrera. 1982. "Ecosystem recovery in Amazon caatinga forest after cutting, cutting and burning and bulldozer clearing treatments." Oikos 38: 313-320. Uhl, C., and C. F. Jordan. 1984. "Succession and nutrient dynamics following forest cutting and burning in Amazonia." Ecology 65(5): 1476-1490. Uhl, C., D. Nepstad, R. Buschbacher, K. Clark, B. Kauffman, and S. Subler. 1990. "Studies of ecosystem response to
155
natural and anthropogenic disturbances provide guidelines for designing sustainable land-use systems in Amazonia." In: Alternatives to deforestation: steps toward sustainable use of the Amazon rain forest, ed. A. B. Anderson (Columbia University Press, New York), pp. 24-42. Vasconcelos, H. L., and J. M. Cherrett. 1994. "Changes in leaf-cutting ant populations (Formicidae: Attini) after the clearing of mature forest in Brazilian Amazonia." Studies on Neotropical Fauna and Environment 30 (3): 107-113. Victoria, R. L., L. A. Martinelli, J. Mortatti, and J. Richey. 1991. "Mechanisms of water recycling in the Amazon Basin: isotopic insights." Ambio. 20(8): 384-387. Vieira, I. C. G., C. Uhl, and D. C. Nepstad. 1994. The role of the shrub Cordia multispicata Cham, as a "succession facilitator" in an abandoned pasture, Paragominas, Amazonia. Vegetatio 115: 91-99.
10 The Interface Between Economics and Nutrient Cycling in Amazon Land Development Carl F. Jordan The Nutrient Problem Most of the terra firme soils in the Amazon are highly weathered, highly leached, have low capacity for retaining nutrients against the continual leaching and weathering of the tropical climate, and are classified as Oxisols and Ultisols, soil types with extremely low fertility (see Cuevas, this volume). The naturally occurring forests of the region maintain a high production of wood and leaves through very efficient recycling of nutrients from decomposing litter to roots in a roothumus layer on top of the mineral soil or near its surface. The decomposing litter is important not only as a source of nutrients, but as a source of organic acids which prevent phosphorus fixation in the iron- and aluminumrich soils of the Amazon. When forests on Amazonian terra firme soils are cut and burned, and the soils used for agriculture, litter, and humus are rapidly oxidized and destroyed. As a result, the potassium remaining from the original forest is quickly leached, the nitrogen is volatilized, and the phosphorus is immobilized in the mineral soil. This is one of the most important reasons that crop production can be carried out for only a few years under shifting cultivation. It is not just small scale agriculture that is limited by the low fertility of Amazonian soils. In the past, almost all types of development that destroy the nutrient conserving mechanisms of the forest have suffered financially. Two examples are given here to illustrate.
Economics and Nutrient-Poor Soils: Case Studies Pulp plantations In 1967, one of the largest conversions of tropical forest to pulp plantation began near the junction of the Jari and Amazon rivers, in the state of Para, Brazil (Time, 1976). The "Jari" project was initiated and financed by Daniel K. Ludwig, one of the world's richest men, and owner of numerous international corporations. Ludwig had anticipated a global shortage of wood fiber for pulp, and to meet this shortage, he and his advisors selected a site that they believed had high potential for pulp production (Time 1979, Kinkead 1981). By 1981, the total investment in the 12,000 km2 tract of land was approximately $1 billion (Kinkead 1981). Ludwig's advisors recommended melina (.Gmelina arborea) as the best species to plant. It is a species that often grows well in plantations, and can produce high-quality pulp (Greaves 1979). Ludwig ordered the entire area planted in melina, without ever having his advisors examine the soils at Jari for their suitability for this species (Palmer 1986). Only after growth of melina was 40% below target were soils examined. Although company officials blamed poor management as the cause (Hartshorn 1981), the most important factor was the low fertility of the soils
The Interface Between Economics and Nutrient Cycling in Amazon Land Development
Oordan and Russell 1989). Sioli (1973) believed that the reluctance of administrators to mention poor soils as the reason for failure of Amazonian projects. Jari was their faith in the myth of the Amazon as a source of unbridled productivity. In 1982, a majority interest in Jari was sold to a consortium of 27 Brazilian companies for $280 million (Fearnside and Rankin 1982, Time 1982). A question remains as to who finally paid for the $720 million loss sustained by Ludwig when he sold Jari to the Brazilian investors. Ludwig's parent corporation, International Bulk Carriers of New York may have been able to take the loss as a tax deduction against U.S. income tax, or, Ludwig may have defaulted on part of his debt to Brazilian banks Qordan and Russell 1989). In 1983, Jari reported a $2.1 million operating profit (International Society for Tropical Forests 1985). This represents a rate of return of 0.2% on the original investment of 1 billion. Even for the Brazilian consortium who paid $280 million for Jari, the rate of return is only about 0.8% on their investment. Because annual sales at Jari are greater than annual expenses such as salaries and maintenance, the project should continue to operate in the near future. However, eventually the major capital item, the pulp mill, will become obsolete. Willingness of investors to recapitalize will depend on Jari's history of investment return.
Ranching Economic reform was a major factor behind the 1964 military coup in Brazil. One of the steps taken by the new government was the establishment of a Super-intendency for the Development of the Amazon (SUDAM). Its mission was to stimulate agricultural development in the Amazon, a region perceived to be rich in developmental potential (Buschbacher et al. 1987). One of
157
the steps taken was to initiate a tax incentives program to promote more corporate investment in the Amazon basin. The purpose of the SUDAM tax incentives program was to mobilize companies in Sao Paulo and other parts of Brazil to reinvest their taxable incomes in projects in the Amazon (Foweraker 1981). The plans for development in the Amazon had various motivations, such as opportunities for landless peasants and national security. However, the most important concern seemed to be integration of the Amazon region into the national economy of Brazil. Through this "manifest destiny," the new agricultural frontier in the Amazon was to provide a solution to vital economic questions and thus to help legitimize the governmental regime (Hecht 1984). Cattle ranching seemed to be a direction for development that promised relatively easy profits. Compared with the other agricultural options in the region such as pepper, cacao, and rubber plantations, ranching seemed relatively easy to initiate and maintain (Hecht 1984). SUDAM hoped that over 500 large cattle ranches would be established under this program in the Amazon regions of Brazil. One of the first companies to take advantage of the new program was the King Ranch of Texas. In 1968, King Ranch, in collaboration with the SwiftArmour Company of Brazil, was granted authorization to establish a 180,000 acre cattle ranch near the town of Paragominas, in the state of Para (Davis 1977). Soon many other ranches also became established in the area (Foweraker 1981). The increasing conversion of rain forest to pasture was sharply criticized by environmental scientists. Because of the very low fertility of the soils, maximum carrying capacity the first year following conversion of forest to pasture was predicted to be only about 0.4 head of cattle per hectare. Rapid loss of nutrients and fixation of phosphorus caused the maximum possible stocking rate
158
to decline to half that by the third year (Fearnside 1979). Most of the pastures were overgrazed, resulting in trampling and compaction of soils, and loss of topsoil and nutrients through erosion. Weeds, many unpalatable, poisonous, and much better adapted to degrades soils became established and out-competed the pasture grasses. Conversion of forest to pasture was ranked least desirable of all types of development activities in Amazon rain forest areas (Goodland 1980). By the middle to late 1970s, pasture degradation by grazing was widespread. As a result of one to two decades of various combinations of clearing, burning, grazing, weedings, and mechanical scraping, many pastures in the Paragominas region of the eastern Amazon appeared to be in very poor condition (Uhl and Buschbacher 1985). By 1978, about 85% of the ranches in Paragominas had failed, according to the Director of the Para State Cattlemen's Cooperative (Hecht 1984). By 1980, many of the pastures had been abandoned for several years, and some of them were degraded to the point that questions arose about whether a forest could ever be re-established on the sites, or whether there was a permanent conversion to heath-like vegetation (Buschbacher et al. 1987).
"Sustainable Agriculture" In 1982, Science reported that technology had been developed which permitted continuous production of annual crops in acid, infertile soils of the Amazon basin (Sanchez et al. 1982). The studies in Yurimaguas Peru showed that three grain crops can be produced annually with appropriate fertilizer inputs. Twenty-one crops were harvested from one field in 8 1/2 years, with an average yield of 7.8 tons of grain per hectare. Soil properties were said to improve with continuous cultivation.
Carl F. Jordan
The "technology" used to attain this continuous yield was primarily the addition of inorganic fertilizers. The economic feasibility of agricultural development which relies on hauling fertilizer to remote jungle areas has been questioned (Jordan 1987). If the purpose of the agriculture is to produce crops for the marketplace, how would it be possible for peasant farmers in remote jungles to compete against others close to centers of population? The costs of hauling fertilizers in, and crops out, would be prohibitive. If the purpose of the agriculture were merely for subsistence, then it would seem that paying for importation of fertilizers into the jungle would not be rational. Where costs of transportation are high relative to the value of the crop, it would make more sense for the farmer to import a kilogram of manioc than it would to import a kilogram of fertilizer so he could grow manioc.
Recent Trends Beginning in the early 1990s, a series of papers emerged which suggested that despite the problems due to low fertility of the soils, agriculture and ranching in the Amazon could be profitable. Mattos and Uhl (1994) reported that many ranchers in the eastern Amazon region near Paragominas, Para, have succeeded in making a profit through specialization. Many small operators are specializing in calf/milk production, while larger operators often increase profits by specializing in range fattening of calves. Another important trend they reported is that of intensification of land use on ranches. As pastures age, planted grasses usually loose their vigor and are gradually replaced by weedy herbs, shrubs, and trees. Some ranchers are bringing these old pastures back into production by tilling, reseeding with better adapted forages, and fertilizing. The investment to restore these pastures is considerable, but returns are generally 3-10 times greater than those for less intensive approaches.
The Interface Between Economics and Nutrient Cycling in Amazon Land Development
159
They conclude that "for the first time, deficits in tion. The tax that a property owner should pay each year is a certain percentage of the soil nutrients are being corrected." value of the land. The land value is deterIntensification has yielded financial gains for farmers as well as for ranchers. Intensive mined after subtracting the value of any approaches to agriculture can result in investments made on the property and the increased yields, higher net returns, and greater value of standing timber. An important job and tax generation (Toniolo and Uhl 1995). feature of the tax is the discounts designed to They found that intensive approaches to agri- stimulate production. Two discounts of up to cultural production, such as the cultivation of 45% each, are available. The first is for black pepper, oranges, or vegetables are more "degree of land utilization" and the second is lucrative than traditional extensive approaches, for the "intensity of land utilization." The such as shifting cultivation, but the capital need- discount for the degree of land utilization is ed to initiate these more intensive activities is determined by the percentage of the land considerable and frequently not available to the eligible for production on a property that is actually in production. The second discount small producers. Almeida and Uhl (1995) suggest that based on the intensity of land utilization is Brazil's Rural Land Tax (Imposto Territorial calculated by dividing the observed producRural) has an important role in the increase tivity of an area by a "minimum acceptable in specialization and intensification of ranch- productivity value" and multiplying this ing and agriculture. The tax was instituted in quotient by the discount for degree of land 1964 to encourage economic growth in the utilization. Alameida and Uhl criticize present rural sector by keying tax breaks to produc- enforcement as not emphasizing enough the
Fig. 10.1 Economic value of a parcel of land in the Amazon as a function of distance from markets where products from that parcel could be sold. Reprinted from Fig. 2.4 in Schneider, 1995, with permission of The World Bank.
160
intensification aspect. With greater rewards for intensification, less forest would be cut down for conversion to extensively graze pasture.
Profitability of Amazonian agriculture Are tax incentives sufficient to overcome the natural disadvantages due to soils in the Amazon region, and to render agriculture and ranching profitable? A recent report from the World Bank (Schneider 1995) presents a theory to explain why agriculture and ranching in some areas of the Amazon is becoming profitable, despite the need for intense fertilization and weeding to overcome natural soil deficiencies. The theory is summarized in Fig. 10.1. In this figure, the horizontal axis is the distance from a particular piece of land to markets for agricultural products or beef. In the eastern Amazon region, Belem is a traditional market, and Paragominas is becoming one. "Distance from Markets" is not necessarily linear distance "as the crow flies," but rather the logistical difficulty of getting supplies from the market and produce to the market. Logistical difficulty often is proportional to linear distance, but not always. A farm 100 km from the market on a paved road is logistically closer to the market than a farm 40 km away on a dirt road that becomes impassible during rainstorms. "Net Present Value" of the land is simply its going market price. Net present value is determined in part by the opportunity cost of labor, that is, the value of the labor necessary for economic income; the cost of defending property rights to that land; the opportunity cost of capital, that is, the value of the capital necessary to produce an economic income; and the equilibrium price of the land, that is, what bidders in the market place will pay after consideration of other factors. The economic frontier is those areas most remote from markets. In the Amazon, these
Carl F. Jordan
are settled by landless peasants who practice shifting cultivation. Such areas are not only distant from markets, they are also distant from any type of infrastructure such as roads, railroads, or even small cities. Even if the land is legally owned by a corporation, an individual investor, or by the government, the land is essentially "free" to the peasant. The cost of keeping out the squatters is more than the land is worth to the official owners. The only cost to the peasant in such areas is his and his family's labor. As infrastructure begins to develop in such regions, the logistical "distance" to markets begins to decrease, and we enter the "Zone of Conflict and the Emergence of Government." As a road is built across a region, property rights begin to be an issue. If the land officially belongs to someone or some institution, that person may try to evict the squatter, either forcibly or through negotiations. Obtaining a legal title to a piece of land through "squatter's rights" requires an investment of time and money on the part of the peasant, and many prefer to sell their "rights," or to just give them up and move on to the next frontier. Land is given up or sold not for ecological reasons of decreasing productivity (although that may be a factor), but mainly for economic considerations. Often, land is bought up by speculators, or by companies engaged in speculation hoping to profit from the increase in value brought about by increasing governmental infrastructure in the area. New capitalist owners of the land often do not begin managing their land right away. Costs of importing fertilizers, herbicides, etc., and of exporting crops are still too high for the land owner to use the land profitably. The land owner assumes however, that the government eventually will build and improve roads into the region, and establish governmental services such as health care, schools, and market support in local villages. Meanwhile, the land lies abandoned.
The Interface Between Economics and Nutrient Cycling in Amazon Land Development
Usually substantial governmental services do eventually emerge in a frontier region. As a result, the logistical distance to an existing market decreases, or a new market may emerge. At this point, it becomes worthwhile for the owner to begin investing in infrastructure and supplies that will cause the land to yield a profit. The investor may build a ranching complex or a farm, buy trucks and tractors, hire labor, and import agricultural chemicals. At first, profits are small, but as transportation logistics improve, and the market becomes larger and more economically diversified, profits increase. Because the operation is logistically close to the market, transportation costs fall, and the entrepreneur can successfully compete in the emerging market against producers in other regions. Profitability on lands close to the markets is increased by intensification and by specialization. This is not a phenomena that is exclusive to the Amazon. Rather, intensification and specialization is at the heart of the capitalist system and has been the essence of "progress" in all developing regions of the world. When agriculture or ranching is begun on a piece of land, intensification of factors of production will always result in an increase in profit. Eventually, intensification per unit area of land may reach a limit. Increases in fertilization beyond a certain point may not result in corresponding increases in corn production. Many areas in Midwestern United States are at that point, at least with regard to fertilization. However, other types of intensification such as more powerful insect control may increase yields even further. With intensification comes specialization. When there are many factors of production involved in the eventual production of a crop, it becomes impossible for a single farmer or rancher to master the details of efficiently using all aspects of production, and investing in the resources to carry them out. Specialization begins, because it is more feasible for
161
the individual farmer or rancher to invest in, and to master, just a single component of the entire process of cattle raising or production of high-yield rice. Figure 10.1 could be as applicable to the United States in the nineteenth century as it is to Brazil in the twentieth century. There would be only one important difference. Because native soil fertility in Midwestern United States is much higher than in the rain forest area of Brazil, the logistical "distance to markets" could be further, for each respective phase of development. The production costs in the United States would be lower, because less fertilizer would be needed. With lower production costs, capital investment becomes profitable at relatively longer distances from the market. There is, of course, a negative aspect to intensification. Pesticides can poison native wildlife, agricultural workers, and consumers of food. Fertilizer runoff can cause eutrophication. Herbicide resistance can emerge in weeds, and make weed control extremely difficult. We can scarcely speculate upon the downside of genetic engineering. Intensification of agriculture in the United States began in earnest only after World War II, and it took a generation before the environmental side-effects of intensification were realized. One of the most important things that we have learned about intensification is that it is dependent upon energy subsidies, chiefly petroleum subsidies. Cheap fertilizers, pesticides, herbicides, and other agricultural chemicals are based upon technology that depends upon petroleum. Equally important are petroleum based tractors, planes, and trucks to deliver these inputs into the field. And now that the environmental costs of these subsidies are being uncovered, the energetic cost of protection or of cleaning up after accidents makes reliance upon petroleum even higher. Modern agriculture in the United States is sustainable only as long as cheap petroleum
162
is sustainable. And while the supplies of petroleum decrease, the costs associated with environmental protection against the effects of intensification increase. Farmers in the Amazon are headed down the same path of unsustainability as are farmers in the United States.
Alternatives An important challenge for ecologists working in the Amazon is to devise more sustainable systems of farming and forestry that at the same time yield some economic profit, although perhaps not as great a profit as "mining" the system as is currently done. Rain forests produce wood and leaves at a phenomenal level. However, most of the production is consumed by bacteria, fungi, and other decomposers. The trick to taking economic advantage of rain forest areas is to devise a system whereby consumption of organic products by bacteria and fungi is substituted for by extraction by humans. The best way to do that is to devise a system that has a structure and function similar to the naturally occurring rain forest. That way, the services of nature can be conserved, especially the ones involved with recycling of nutrients mentioned above. When rain forests are cut down, for example, and converted to fields of rice or corn, the nutrient recycling mechanisms are destroyed, and as a result, the productivity of the field quickly declines. A much better approach would be to devise a production system whereby the nutrient recycling mechanisms are left intact. Extractive reserves for Brazil nuts and rubber are examples of management approaches that are compatible with sustainability of harvest (Brown et al. this volume). The trees are not destroyed when the nuts and the rubber are harvested, and consequently, the nutrient recycling mechanisms of the forest remain intact. A problem with extractive reserves however, is
Carl F. Jordan
that the trees are so dispersed that the task of collecting the nuts and the latex makes the operation uncompetitive on the world market. Ways must be found to produce agricultural products more intensively, while still having a production unit that maintains the nutrient conserving mechanisms of an intact forest. There are many potentially useful combinations of species that could take advantage of the services of nature, but to successfully design the combinations, it is important to know the potential role of each species in the system. For example, in plantations of leguminous trees over an understory of coffee, the legumes fix nitrogen, which is made available to the coffee trees as the leaves fall. In addition, the legumes may have a deeper root system that can extract cations leached down into the lower soil horizons. Another potentially useful combination would include pigeon pea (Cajanus cajan) whose roots exude an organic acid that can solubilize phosphorus bound by the iron and aluminum in tropical soils (Ae et al. 1990). Yet another challenge is to determine the role of silica, which is present in high concentrations in many native tropical trees. It may be an adaptation to the low phosphorus levels in many tropical soils. Silicates can replace phosphates bound with iron and aluminum in the soil, thereby releasing the phosphates in soluble form (D'Hoore and Coulter 1972). The benefits of silicate fertilizers in promoting growth of tropical grasses such as sugar cane are well known (Fox et al. 1967). Selective harvesting of trees can be a sustainable use of the forest, if a relatively few trees are taken, and care is taken in harvesting and skidding so that the remaining trees, especially seedlings and saplings of valuable species, are not damaged. A major problem with selective harvest of valuable trees in the Amazon is their depletion. In some regions, rosewood no longer is to be
The Interface Between Economics and Nutrient Cycling in Amazon Land Development
found. Plantation forestry is practiced in some regions, such as the Jari plantation in Para. However, plantation forestry is more like agriculture than it is like natural forest management, and the growth of monocultures in plantations suffers from the same problems of soil nutrient deficiencies as does agriculture. The greatest challenge to tropical ecologists and foresters is to devise ways to reforest the degraded, abandoned pastures of the Amazon with mixed species plantations of native hardwoods. One major obstacle to this is the lack of availability of many seeds, and the inability of many species to keep dormancy (Janzen and Vasquez-Yanes 1991). Another problem is the low rate of growth of the seedlings of many tropical hardwoods, leading to the impression that productivity will be low. A challenge to scientists working in tropical ecosystems is to devise agriculturally sustainable systems that yield an economic profit. An agriculturally sustainable system in the sense used here is one in which there occur nutrient cycling mechanisms as part of the crop structure. This, almost by definition, means a
163
production system which resembles the structure of a native forest. There are a number of such systems already in existence. One combination that is not infrequent is an overstory of rubber trees with an understory of coffee or cocoa. In some cases, vanilla vines are grown with the cocoa trees as support. In some cases, nitrogen fixing trees are used as an overstory species. These systems however, are almost always less economically profitable in the short run than are exploitative approaches that "mine" the system. In order to encourage a more sustainable approach to resource management, it will be necessary to subsidize the approach, at least until transition costs can be overcome. Further, to make sustainable systems more competitive with exploitative approaches, subsidies for ranching and other unsustainable land use must be eliminated. Designing sustainable production systems based upon an understanding of ecosystem function will be the easy part. Weaning the economy from its focus on short-term profits and from its dependence upon an unsustainable source of petroleum will be the hard part.
Literature Cited Ae, N., J. Arihara, K. Okada, T. Yoshihara, and C. Johansen. 1990. "Phosphorus uptake by pigeon pea and its role in cropping systems in the Indian subcontinent." Science 248: 477-480. Almeida, O. T. de, and C. Uhl. 1995. Brazil's rural land tax. Land Use Policy 12: 105-114. Buschbacher, R. J., C. Uhl, and E. A. S. Serrao. 1987. "Largescale development in eastern Amazonia." In: Amazonian Rain Forests, ed. C. F. Jordan (Springer- Verlag, N. Y.), pp. 90-99. Davis, S. H. 1977. Victims of the Miracle. Cambridge University Press, Cambridge. D'Hoore, J., and J. K. Coulter. 1972. "Soil silicon and plant nutrition." In: Soils of the Humid Tropics (National Academy of Sciences, Washington, D.C.), pp. 163-173. Fearnside, P. M. 1979. "Cattle yield prediction for the Transamazon highway of Brazil." Interciencia 4: 220-226. Fearnside, P. M., and J. M. Rankin. 1982. "The new Jari: Risks and prospects of a major Amazon development." Interciencia 7: 329-339.
Foweraker, J. 1981. The Struggle for Land: A Political Economy of the Pioneer Frontier in Brazil from 1930 to the Present Day. Cambridge University Press. Cambridge. Fox, R. L, J. A. Silva, O. R. Younge, D. L. Plucknett, and G. D. Sherman. 1967. "Soil and plant silicon and silicate response by sugar cane." Soil Science Society of America Proc. 31: 775-779. Goodland, R. J. A. 1980. "Environmental ranking of Amazonian development projects in Brazil." Environmental Conservation 7: 9-26. Greaves, A. 1979. "Gmelina large scale planting, Jarilandia, Amazon basin." Commonwealth Forestry Review 58: 267-269. Hartshorn, G. 1981. Report to Institute of Current World Affairs, December 1981, on Activities as a Forest and Man Fellow, Sponsored by That Institute. Institute of Current World Affairs, Wheelock House, Hanover, New Hampshire. Hecht, S. B. 1984. "Cattle ranching in Amazonia: Political and ecological considerations." In: Frontier Expansion in
164 Amazonia, eds. M. Schmink, and C. H. Wood (Univ. of Florida Press, Gainesville, Florida), pp. 366-398. Inter. Soc. Trop. Forestry. 1985. Jari among top 100 agribusiness firms. ISTF (Bethesda, Maryland) News 6: 7. Janzen, D. H., and C. Vasquez-Yanes. 1991. "Aspects of tropical seed ecology of relevance to management of tropical forested wildlands." In: Rain Forest Regeneration and Management, eds. A. Gomez-Pompa, T. C. Whitmore, and M. Hadley (UNESCO and The Parthenon Publishing Co, Carnforth, U.K.), pp. 137-157. Jordan, C. F. 1987. "Conclusion." In: Amazonian Rain Forests, ed. C. F. Jordan (Springer Verlag, N. Y.), pp. 100-121 Jordan, C. F., and C. E. Russell. 1989. "Jari: A pulp plantation in the Brazilian Amazon." Geojournal 19.4: 429-435. Kinkead, G. 1981. "Trouble in D. K. Ludwig's jungle." Fortune magazine, April 20, 1981: 102-117. Mattos, M. M., and C. Uhl. 1994. "Economic and ecological perspectives on ranching in the Eastern Amazon." World Development 22: 145-158. Palmer, J. R. 1986. Jari: lessons for land managers in the tropics. Ms presented at the International Workshop on Rainforest Regeneration and Management, Guri Venezuela, 24-28 November, 1986, under auspices of the UNESCO Man and the Biosphere Program. Proceedings 1986.
Carl F. Jordan Sanchez, P. A., D. E. Bandy, J. H. Villachica, and J. J. Nicholaides. 1982. "Amazon basin soils: management for continuous crop production." Science 216: 821-827. Schneider, R. R. 1995. Government and the Economy on the Amazon Frontier. World Bank Environment Paper No. 11. Washington, D.C. Sioli, H. 1973. "Recent human activities in the Brazilian Amazon region and their ecological effects." In: Tropical Forest Ecosystems in Africa and South America: A Comparative Review, eds. B. J. Meggars, E. S. Ayensu, and W. D. Duckworth (Smithsonian Institution Press, Washington, D.C.), pp. 321-24. Time. 1976. "Ludwig's wild Amazon kingdom." Time magazine, Nov. 15, 1976: 59-59A. Time. 1979. "Billionaire Ludwig's Brazilian gamble." Time magazine, Sept. 10, 1979: 76-78. Time. 1982. "End of a billion-dollar dream." Time magazine, Jan. 25, 1982: 59. Toniolo, A., and C. Uhl. 1995. "Economic and ecological perspectives on agriculture in the Eastern Amazon." World Development 23: 959-973. Uhl, C., and R. Buschbacher. 1985. "A disturbing synergism between cattle ranch burning and selective tree harvesting in the Eastern Amazon." Biotropica 17: 265-268.
11 Carbon Storage in Biomass and Soils Martial Bernoux, Paulo M. A. Graga, Carlos C. Cerri, Philip M. Fearnside, Brigitte J. Feigl, Marisa C. Piccolo
Carbon dioxide and methane integrate biogeochemical cycles of C and constitute, together with nitrous oxide, the main trace gases responsible for the greenhouse effect. Increasing interest in the global consequences of climate change has prompted the global scientific community to deepen their studies about the global C stocks and the interrelations among its different compartments. As main compartments, soils and phytomass (living and nonliving) have received special attention. Many authors proposed a quantification of C stored in soils and proposed to study their role as both a source and sink of carbon (Post et al. 1982, Eswaran et al. 1993, Sombroek et al. 1993, Batjes 1996). The world's mineral soils are estimated to contain about 1500 Pg C (Post et al. 1982, Eswaran et al. 1993, Batjes 1996), while the biomass of plants is estimated to be comprised between 560 and 835 Pg C (Whittaker and Likens 1975, Bouwman 1990). Tropical forests account for between 20 and 25% of the world terrestrial C (Brown and Lugo 1982, Dixon et al. 1994). The Amazon contains the largest expanse of native tropical ecosystems and has a direct influence on global biogeochemical cycles, especially the C cycle. The C stored in phytomass is of importance because of its quantity and its potential to be released easily. Carbon in soil is proved to be important because soil organic carbon (SOC) is
intimately involved in virtually all biological processes, and organic matter (OM), even when present in small amounts, is an extremely important soil constituent.
Carbon in Soil Soil types Two Brazilian soil classes, Latossolos and Podzolicos, make up 73% of the total area of the Legal Amazon Basin of Brazil (Prado 1996, Jacomine and Camargo 1996) (Fig. 11.1). More precisely, only three dystrofic soil types, Podzolico Vermelho Amarelo (Acrisol), Latossolo Amarelo (xanthic Ferralsol), and Latossolo Vermelho Amarelo (orthic Ferralsol) cover approximately 60% of the total, and are therefore of prime interest. The remainder is distributed between 13 additional classes. Only 6, however, represent more than one percent, and only 2 of which are more than 5%: Plintossolos (Inceptisols, Oxisols, and Alfisols) and Gleissolos (Entisols and Inceptisols). The Brazilian Latossolos correspond to well drained Oxisols in the U.S. Soil Taxonomy, and the FAO-UNESCO soil map legend identifies them as Ferralsols. In the French classification they are defined as "Sols Ferrallitiques," commonly "fortement desatures, typiques." The Podzolicos belong to the Alfisols (when eutrofic) and to the Kandisols
166
(when dystrofic) orders of the Soil Taxonomy, and most of them fall into the Acrisols, Nitosols, and Lixisols (Luvisols) of the FAO-UNESCO map legends (Van Wanbeke 1992). The French classification identifies most of them as "Sols ferrallitiques" and the rest as "Sols ferrugineux." Most Alfisols, Kandisols and Oxisols belong to low-activity clay soils (i.e., cation exchange capacity < 24 cmolc kg"1 clay). Kandisols usually occupy younger geomorphic surfaces than the Oxisols with which they are often associated in landscapes (Moraes et al. 1996). These soils are mineral soils with thickness often > 2 m. Nepstad et al. (1994) estimated that 36% of the Brazilian Legal Amazon remains evergreen even under severe moisture regime, which provides further evidence for the existence of very deep soils.
Martial Bernoux et al.
Carbon stocks and Amazonian estimates Moraes et al. (1995) gave the first reliable estimates of C in soil for the Legal Amazon Basin. Their work was based on 1162 soil profile descriptions (28 soil types) from the RADAMBRASIL soil survey, which was carried out during the 1970s and early 1980s. Briefly, the carbon concentration reported was determined using methods derived from Walkley and Black (1934). Carbon stocks by horizon, expressed in kilograms per square meter, were calculated by multiplying bulk density, C concentration and thickness of the horizon. Moraes et al. (1995) estimated mean soil bulk densities for each soil type using data reported for 474 soil horizons. They gave estimates of C stocks in the 0-20 cm and the 0-100 cm layers. The carbon stock for the
Fig. 11.1 Soils map of the Brazilian Legal Amazon.
Carbon Storage in Biomass and Soils
167
Table 11.1 Variability in organic carbon content for the soil units (stone free soil) of the Brazilian classification*. Brazilian Classification U.S. Soil Taxonomy Classification **
Symbol n
Mean
Median
SD
SE
(kg C . m-2)
(kg C . m-z)
(kg C . m-2)
(kg C . m-2)
Areias Quartzosas
Psamments
9.43
8.38
4.23
0.62
Psamments
AQ AQH
43
Areias Quartzosas Hidromorficas
11
9.39
8.27
5.44
1.41
Brunizens Avermelhados
Chernozems
BA
8
15.57
13.22
Cambissolos Tropicais distroficos
Inceptisols
CTd
34
7.75
7.06
8.05 3.12
2.55 0.44
Cambissolos Tropicais eutroficos
Inceptisols
CTe
14
6.84
5.74
4.06
0.96
Gleissolos distroficos
Aquic suborders
Gd
46
12.24
1.27
Aquic suborders
Ge
11
7.2
10.33 6.81
9.38
Gleissolos eutroficos
2.67
0.74
Plintossolos Hidromorficos
Oxisols, Ultisols
LH
43
7.77
6.29
4.58
0.65
Latossolos Amarelos
Oxisols
LA
87
8.49
7.84
3.55
0.37
Latossolos Roxos
Oxisols
LR
3
21.65
10.4
19.68
11.36
Latossolos Vermelho Amarelos
Oxisols
LVA
155 10.51
9.42
4.88
0.39
Latossolos Vermelho Escuros
Oxisols
LVE
35
9.3
6.96
4.98
0.83
Planossolos
Alfisols, Ultisols, Mollisols
P
8
9.47
7.34
7.41
2.34
Podzois
Spodosols
Po
4
11.42
10.02
4.7
2.35
Podzois Hidromorficos
Spodosols
PH
13
18.53
8.84
25.45
6.17
Podzolicos Vermelho Amarelo
Ultisols
PVA
315 9.5
8.35
4.92
0.26
Podzolicos Vermelho Amarelos eutroficos Ta***
Alfisols
PVAa
29
7.58
7.68
2.27
0.38
Podzolicos Vermelho Amarelos eutroficos Tb***
Alfisols
PVAb
49
8.73
7.8
3.95
0.49
Podzolicos Vermelho Amarelos plinticos
Ultisols
PVAp
25
9.41
8.5
4.54
0.91
Solonetz-Solodizados/ Solonchak
Aridisols
S
1
2.32
Solos Aluviais eutroficos
Entisols (fluvents)
SAe
0.7
SAd
2.69
0.7
Solos Concrecionarios
not classified
SC
10
5.25
1.36
Solos Litolicos distroficos
Lithic subgroups
SLd
31
13.69 8.62
6.19 7.03 11.92
2.41
Entisols (fluvents)
11 11
7.24
Solos Aluvias distroficos
6.72
5.89
1.06
Solos Litolicos eutroficos
Lithic subgroups
SLe
5
11.6
5.04
9.37
Terras Roxas distroficas
Ultisols
TRd
6
15.04
15.76
4.7
4.19 1.66
Terras Roxas eutroficas
Alfisols
TRe
11.97
9.74
6.62
1.77
Vertissolos Eutroico
Vertisols
V
9 1
6.77
11.81
n is the number of samples per soil unit, SD the standard deviation, and SE the standard error. Soils were mapped in the Brazilian classification and only the higher taxa can be correlated accurately with the U.S. Soil Taxonomy (Beinroth 1975). Ta = cation exchange capacity >24 cmolc kg'1 clay; Tb = cation exchange capacity <24 cmolc kg'1 clay.
168
Martial Bernoux et al.
Brazilian Amazon basin was calculated for each soil type by multiplying the mean C stocks by the area of each soil type from a digitized soil map at scale 1:5,000,000 (Empresa Brasileira de pesquisa Agropecuaria, 1981). The authors determined that 47 Pg of carbon is stocked in the first soil meter, of which 44% is in the upper 20 cm. Cerri et al. (2000) also showed that if the standard deviation is used for evaluating the accuracy of estimates, the associated error would be 11.6 Pg of carbon, i.e. 24.5% of the mean value. The use of means is subject to caution, even more when the number of samples is reduced. As the sample size for most of the soil types is less than 30, outlying values may have a marked influence on the mean. The calculation with the median values, because it is less sensitive to extreme values,
leads to a lower global estimate of 41 Pg for the 0-100 cm layer and 23.4 Pg for the 0-20 cm layer. Carbon stocks did not exceed 22 kg Cnr2 for the 0-100 cm layer, and for the three major soil types the variation was less, ranging from 8.5 to 10.5 kg Cm'2. These estimates were similar to others for tropical soils based on less extensive regional data (Brown and Lugo 1982, Post et al. 1982), and to global averages for tropical Oxisols, Ultisols, and Alfisols (Kimble et al. 1990, Batjes 1996). Moraes et al. (1995) neglected the volume of the soil fraction superior to 2 mm. Using a soil dataset made up with soil profile descriptions in Rondonia (Occidental Brazilian Amazon) we calculated that considering the > 2 mm fraction would decrease carbon stocks (at 1m
Table 11.2 Soil carbon content for the 0-100 cm layer of the main soil type of the western Amazon, calculated using an exponential modelisation of the vertical carbon profile. Brazilian Classification
Symbol
Areias Quartzosas AQ Areias Quartzosas Hidromorficas AQH BA Brunizem Avermelhado Ca Cambissolo Tropical alico Cambissolo Tropical distrofico Cd Cambissolo Tropical eutrofico Ce GH Gley Humico GPH Gley Pouco Humico PI Lateritas Latossolo Amarelo LA LR Latossolo Roxo Latossolo Vermelho Escuro LVE Latossolos Vermelho Amarelos LVA Podzol Hidromorfico PH Podzolico Vermelho PVAd Amarelo distrofico Podzolico Vermelho PVE Amarelo eutrofico SA Solos Aluvias SL Solos Litolicos Terra Roxa Estruturada TRE
n
Mean
Min
Max
SD
Median
(kg C.m-2)
(kg C.m-2)
(kg C.m-2)
(kg C.m-2)
(kg C.m-2)
17
8.48
3
10.29 12.15 6.90 8.55 9.48 17.15 7.35 6.75 7.76 8.34 7.16
19.66 14.88 20.15 10.97 13.42 23-54
4.02 6.21 4.26 2.60 4.40 7.87 5.87 2.18
2 188
7.95 8.31 6.51
4.64 3.22 7.00 2.48 4.02 5.28 8.76 4.06 2.12 5.07 7.45 3.85 4.10 1.50 1.60
8
5.96
3.91
10.13
2.14
5.06
10 2
5.98 7.41 8.58
4.10 7.16 4.54
11.69 7.66
2.12 0.36
16.13
3.45
5.24 7.41 7.86
8 27 4 5 5
16 18 30 3 12 51
Source: Adapted from Bernoux et al. (1998b).
15
25-19 11.64 24.74 11.62 9-83 12.16 19-60 15.12 23.44
4.89 1.61 1.30
6.71 12.76 12.69 6.74 8.37 6.04 16.79 7.06 5.78 7.80
2.29 3.13 9.63 3.05
7.73 6.59 7.23 8.31 5.85
Carbon Storage in Biomass and Soils
depth) from 3 to 6 % for all soil types, except for the Podzolico Vermelho Amarelo Distrofico (PVAd) type where the decrease reached values between 15 and 20% The PVAd type alone represents near 21.5 % of the Legal Amazon Basin. Table 11.1 reports the mean C densities by main soil type (considered stone free) used by Moraes et al. (1995) and completed by Cerri et al. (2000). For all Amazonian countries, Sombroek (1992) using profiles from Brazil (Rondonia State) and also from Peru (Yurimaguas) and Colombia (Araraquara) determined that "maybe 95% of the upland terra firme soil has total C contents varying from 8 to 13 kg Cm-2." This is consistent with the global estimates of 14.5 kg Crrr2 stocked in Ferralsols and 7.1 kg Cm~2 in nonhumic Acrisols determined by Sombroek et al. (1993). Batjes (1996) recently gave global estimates for xanthic Ferralsols (equivalent to the Brazilian Latossolo Amarelo) and orthic Ferralsols (Brazilian Latossolo Vermelho Amarelo) of 8.2 ± 3.2 kg Crrr2 and 9-6 ± 5.1 kg Cm~2, respectively (variations are based on CV). All of theses estimates support the hypothesis that estimates for the Brazilian Amazon can be extrapolated to the entire Amazon to furnish an indicative estimate of C Amazonian pools. More recently Bernoux (1998), Bernoux et al. (1998b) and Cerri et al. (2000) studied an area of 334,000 km2 of the western Brazilian Amazon basin. These authors applied a first correction assuming that the soil fraction > 2 mm is carbon free. Soil bulk densities were often lacking in previous studies, and soil carbon content not always determined for several horizons. Bernoux (1998) proposed several methods to estimate the missing information. A stepwise multiple regression procedure was developed to predict soil bulk density from other properties using the data of 323 soil horizons spread over the whole Amazon basin (Bernoux et al., 1998c). Clay and carbon contents were the best predictors to estimate soil bulk density. For instance, in the case of
169
Latossolos (Oxisols) the use of carbon and clay contents as predictors resulted in a percentage of explained variation near 70%. The following equations were used to estimate soil bulk densities for Latossolos and Podzolicos: 1. Latossolos: BD = 1,419 - 0,0037 Clay% 0,061 OC% for A horizons BD = 1,392 - 0,0044 Clay% for B horizons 2. Podzolicos: BD = 1,133 - 0,041 OC% + 0,0026 Sand% for A horizons BD = 1,718 - 0,0056 Clay% 0,068 pHwater for B horizons The problem of missing soil carbon data was resolved using "power" (Bennema 1974) and "exponential" (Arrouays and Pelissier 1994) type equations to model soil carbon distribution with depth. Bernoux et al. (1998a) determined that both models were well adapted for tropical forested Oxisols, and that the "exponential" model is more precise and produces better data. Bernoux et al. (1998b) extend the use of the exponential model to the other Amazonian soil type, mainly the Podzolicos (Table 11.2). Cerri et al. (2000) evaluated the regional carbon stocks for the 0-30 and 0-100 cm layers using two different approaches. The first approach is based on carbon stocks averaged by soil type multiplied by the extent of these soil types. It corresponds to the methodology used by Moraes et al. (1995). The second approach is based on interpolation of carbon stocks using geostatistics. Geostatistical treatments were carried out using the theory first presented by Matheron (1965) and applied to pedometrics by Burgess and Webster (1980a, b). Resulting estimates are very similar, ranging from 2100 to 2400 Tg C in the top meter of soil with the first approach. The geostatistics furnish an estimate of 2220 Tg C, with an associated
Martial Bernoux et al.
170
error of only 13%, compared to a 40% error in the case of soil map-based methods. Bernoux et al. (1998b) proposed a third methodology running geostatistical interpolation separately on the parameters of the exponential model fitted for each carbon profile. This method appeared to be precise, giving an estimate of 2400 Tg C (0-100 cm layer), but with the advantage of estimating soil C stored at any given depth. These results suggest a mean soil C content down to 1 m between 6.2 and 7.2 kg Cm'2.
labile pool. But the kinetics after 10 years slow down and (13C values increase slowly because of a very resistant forest-derived C pool, which represents up to 50% of the original C under forest. Different authors (Moraes et al. 1996, Moraes et al. 1995, Bernoux et al. 1998d) showed that the initial decrease after burning is followed by a regular increase leading to a recovery of the initial level under forest and even to a 20% increase after 20 years of very well managed pasture.
Carbon in Phytomass
Dynamics of soil carbon An important question is, how does the soil C pool react when native ecosystems are modified? Land use practices affect soil C stocks by modifying inputs to soil, as well as the decomposition rate of soil organic matter (SOM). One of the most important changes in a forest Oxisol if used for agriculture, or converted to pasture for ranching purpose, is the change in the C content (Moraes et al. 1996). Recently, Shang and Tiessen (1997) pointed out that organic matter in tropical Oxisols is quite labile. They showed that a rapid degradation of this pool is possible under cultivation and, therefore, organic matter monitoring and management should be a priority. Fearnside and Barbosa (1998) reviewed the soil C changes from conversion of the Brazilian Amazon forest into pasture. These authors pointed out that pasture soils are a net sink or a net source of C depending on their management, and that most pastures are under management practices resulting in a net C source. When forest is cut and converted to pasture, 13C isotope techniques can be used to estimate soil organic matter turnover and different types of C pools, from very labile/active to very resistant/passive (see Bernoux et al., 1998d for the theory). Commonly, the forest-derived C declines sharply in the first years of pasture installation, due to rapid mineralization of a very
Prior to determining the carbon in phytomass it is necessary to determine the total biomass. Forest biomass is a basic factor in studies related to nutrient cycling and carbon release from terrestrial systems following deforestation. This is now one of the more important global research themes due to concerns about possible humandriven climatic changes. Biomass estimates have an enormous influence in modeling based on emissions of carbon from deforestation. The use of inaccurate values for this parameter will drastically affect the reliability of the results of the models. The Brazilian Legal Amazon is about 5 x 106 km2 and represents 60% of Brazil's territory, with an area originally occupied by forests of approximately 4 x 106 km2 (Fearnside 1997a). This immense forest area possesses high potential to affect the carbon flows among the terrestrial and atmospheric systems. For forest areas of large extension, as is the Amazon case, estimating the forest biomass should be a priority.
Estimation methods for Amazonian forest biomass: direct versus indirect methods One of the largest problems in quantifying the carbon stock of Amazon forest is the difficulty of obtaining a reliable estimate of
Carbon Storage in Biomass and Soils
the representative mean biomass for the region. Amazonian biomass estimates have generated extensive academic debates (Brown and Lugo 1984, 1992a,b; Fearnside 1985, 1986, 1992a, 1993; Lugo and Brown 1986; Brown et al. 1989). In spite of the substantial effort to improve estimates, some uncertainties remain in relation to their reliability (Brown et al. 1995). Forest biomass in the Amazon has been estimated through direct and indirect methods. Independent of the method used, one should take into consideration that all components of forest biomass must be quantified. Besides the trees that are the main component of the forest, other components to be included are vines (lianas), understory plants, litter, roots, palms, etc.
171
Direct methods or destructive harvest Direct methods consist of the complete weighing of biomass contained in the plots and subsequent extrapolation to hectares (ha). Reducing the forest to small loads capable of being weighed manually (Fig. 11.2) is necessary (Klinge et al. 1975, Klinge and Herrera 1983, Fearnside et al. 1993, Carvalho Jr. et al. 1995, Graca 1997). Although it is generally more reliable, this process is quite difficult and requires a great expense of time in relation to area to be sampled when compared with indirect methods. Direct measurements for weighing biomass are important in estimating other forest components not available in indirect inventories of commercial wood volume. Direct measurements are also capable of generating useful information in the quantification of carbon transfers by burning, such as biomass burning efficiency, charcoal formation, and unburned wood residues subject to decay.
Indirect methods
Fig. 11.2 Weighing of biorrass by direct method in Ariquemes, Rondonia.
Indirect methods, based on allometric inference, from measurements of the diameter at breast height (DBH) and the height of the trees to obtain wood volumes, is the main method adopted to estimate biomass in the Amazon. Forest biomass estimates are made through regression analysis, where several fitting curves are tested to obtain an ideal model that can be applied to the trees. These models are calibrated by direct weighing of the biomass from a subsample of trees (see for example, Jordan and Uhl 1978, Higuchi et al. 1994, Brown et al. 1995), and could also include other compartments besides trees, such as vines or understory. Indirect methods are broadly adopted in the forestry industry to evaluate the volume of commercial wood. The employment of technically more sophisticated techniques, such as remote
Martial Bernoux et al.
172
sensing using NDVI (normalized difference vegetation index) from LANDSAT-TM, aerial measuring using airborne lasers, and radar data using images of SAR (synthetic aperture radar), have all been applied in the quantification of forest aboveground biomass (Sader 1987, Nelson et al. 1988, Sader et al. 1989). These methods are considered indirect and use regressions to fit an equation that can better predict the biomass from the signals received by the sensor. Technical problems in relation to employment of these methods in tropical forests have sometimes been disabling and there is a great deal of uncertainty in the numbers obtained. Honzak et al. (1996) mentioned that one limitation of remote sensing is that only the spectral reflectance of the upper layers of the canopy are detected by the sensor. Another difficulty
pointed out by Brown et al. (1989) field access to check the measurements. Line-intersect sampling (LIS) is another indirect method that is recognized for its ease of use and precision in quickly evaluating log volume in slashed forest areas. This method was first developed for conifer forests in temperate regions (see for example, Warren and Olsen 1964, de Vries 1974, Van Wagner 1968, Kaiser 1983). Now, this method is being used in the tropics to estimate the biomass volume within slash and burn areas of the Amazon (Kauffman et al. 1995, Fearnside et al. 1993, Graca 1997). This method reduces spatial variability by allowing biomass to be evaluated in the same plot after burning, which is not possible using the destructive method (Fig. 11.3).
Table 11.3 Estimates of total aboveground biomass in Amazonia, sampling areas up to 1 ha Forest Type
Local
Method
Biomass Author (t/ha)
Dense Dense Dense Open Dense Open Dense Dense Dense Open Dense Dense Dense Open Open
Tucurui, Para Manaus, Amazonas Manaus, Amazonas Belo Monte, Para Tome-Acu, Para Samuel, Rondonia Sul do Para Sul de Roraima Paragominas, Para Ariquemes, Rondonia Altamira, Para Maraba, Para Jacunda, Para Sta. Barbara, Rondonia Jamari, Rondonia
Direct* Direct Direct Direct Indirect** Indirect Indirect Indirect Indirect Direct/indirect *** Direct/indirect Indirect/direct**** Indirect/direct Indirect/direct Indirect/direct
340 425 265 143 173 285 185 228 300 313 263
435 292 290 362
Revilla Cardenas et al., 1982 Carvalho et al., 1995 Fearnside et al., 1993 Revilla Cardenas, 1987 Araujo et al., 1997 I. F. Brown et al., 1995 Higuchi et al., 1994 Higuchi et al., 1994 Uhl et al., 1988 Graca, 1997 Fearnside et al., nd-a Kauffman et al., 1995 Kauffman et al., 1995 Kauffman et al., 1995 Kauffman et al., 1995
* Direct method refers to the destructive harvesting of biomass in sampling and weighing. ** Indirect method consisted of the estimate of live aboveground biomass through relationships, allometrics, using models of regression equations. Other components of the forest that are not standing trees were not included. This is generally estimated directly (understory, litter, vines, fallen logs, etc.). *** Direct/indirect methods refers to the estimate of biomass by destructive harvest, supplemented for branches and trunks < 10 cm in diameter by indirect methods of line-intersect sampling. Used to estimate the biomass in areas of felled forest. *** Indirect/direct methods used the nondestructive method of sampling by planar intersection (planar intersect models) for branches and trunks of trees, and the destructive method to determine litter biomass, root layer, seedlings, and sprouts.
Carbon Storage in Biomass and Soils
Estimates at regional versus local scales Regardless of which method is used, high variability of biomass exists among the different types of forests in the Amazon and also within the same forest type (Table 11.3). This variability is the main impediment to estimating the average biomass of the Amazon region using surveys of forest biomass on a local scale. For example, biomass stocks using the direct method vary from 143 to 340 t/ha, with a difference of 197 t/ha among them. For the indirect method the range is from 173 to 300 t/ha. Another factor that can affect the reliability of biomass estimates, mainly in the direct method, is the size of the sampling area. Brown et al. (1995) observed that the high variability in small plots is explained by the lack of large trees which will produce a value
Fig. 11.3 Evaluating the biomass using the line-intersect sampling.
173
below the global average, and the few plots with big trees will tend to augment their biomass disproportionately. Derived estimates of large-scale surveys based on forest volume are perhaps the most useful technique to estimate the average forest biomass for the Amazon region. However, these estimates tend to treat homogeneously the different forest types when wood volume data are transformed into forest biomass. The high heterogeneity of the forest, a consequence of the great diversity of tree species, is one of main difficulties in obtaining a representative average biomass for Amazonia. Forest inventories, such as those done by RADAMBRASIL in the 1970s and by FAO during the 1950s and 1960s, have been used by some researchers to obtain more precise estimates for forest biomass in Amazonia (Brown et al. 1989, Brown and Lugo 1992a; Bohrer and Campos 1993; Fearnside 1992b, 1997a). These inventories were done to evaluate commercial wood volume in the Amazon, including trees with DBH > 31-8 cm in the case of RADAMBRASIL and > 25 cm of DBH in the FAO surveys. Estimates using data from these inventories need several adjustment factors so that the commercial wood biomass can be expanded to terms of total biomass. A part of the uncertainty related to the different biomass estimates in Amazonia is the exclusion of some biomass components. Thus, Fearnside (1992a) mentioned several adjustment factors to be adopted for the estimates of Brown and Lugo (1992a) based on the forest surveys by RADAMBRASIL and FAO. Logging is also a factor that should be considered in estimates of forest biomass in terms of its contribution to global warming. Biomass estimates using data at the time of the forest inventories do not reflect the current tendencies in a great part of the region. Deforestation in the 1970s generally referred to forest burning without wood
174
Martial Bernoux et al.
removal before the forest was felled. From the 1990s, in consequence of improvement of road access and of increases in timber prices, the removal of any commercial wood before slashing and burning became common practice. However, biomass postlogging should only be used if logging is explicitly included in the calculations. Currently, explicit treatment of logging is rare in this type of calculation, and there is a strong tendency to error if a reduced biomass post-logging is used in calculations that omit explicit estimates for logging (Fearnside 1994). The most recent estimates for total biomass in Amazonia from inventories on a regional scale vary from 227 t/ha to 464 t/ha (Table 11.4).
Phytomass and carbon stocks Forests The vegetation classification of the Brazilian Institute of Environment and Renewable Natural Resources (IBAMA) (Table 11.5; Brazil, IBGE and IBDF 1988) is
more general than that used by the RADAMBRASIL Project (Brazil, Projeto RADAMBRASIL 1973-1983). The 1:5,000,000scale IBAMA map indicates 48 vegetation types occurring in the Legal Amazon (including nonforest vegetation types), whereas 145 types appear on the RADAMBRASIL 1:1,000,000 maps. The RADAMBRASIL map code is not consistent among the volumes into which the survey is divided, the same map code sometimes referring to somewhat different vegetation types in different volumes. Equivalents between the IBAMA and RADAMBRASIL vegetation classifications are given elsewhere for each RADAMBRASIL volume (Fearnside, nd-a). Brazil's Legal Amazon region contains 19 forest types under the IBAMA classification. Seventy-nine "ecoregions" (an IBAMA forest type within one of the region's nine states) are indicated on the IBAMA vegetation map as present in the region. Biomass estimates in each ecoregion (Table 11.6) are updated from an earlier set of estimates (Fearnside 1994) based on additional data, including weighted mean values for the basic density
Table 11.4 Current estimates for forest biomass in Amazonia from surveys on a regional scale. Total Biomass (t/ha)
Database Source
Forest Type
227*
RADAMBRASIL
Dense
Brown and Lugo, 1992
298*
FAO
Dense
Brown and Lugo, 1992
272**
RADAMBRASIL
All forest
Fearnside, 1992a
320**
RADAMBRASIL
Dense
Fearnside, 1992a
372***
RADAMBRASIL/FAO
All forest
Fearnside, 1992b
394**
RADAMBRASIL/FAO
All forest
Fearnside, 1992b
232*
RADAMBRASIL
Dense and open
Bohrer and Campos, 1993
464**
RADAMBRASIL/FAO
All forest
Fearnside, 1997a
434***
RADAMBRASIL/FAO
All forest
Fearnside, 1997a
Only aboveground biomass Including above- and belowground biomass. Estimate for forest biomass deforested in 1990 in the Brazilian Legal Amazon
Reference
Carbon Storage in Biomass and Soils
175
Table 11.5 Vegetation Types in the Brazilian Legal Amazon. Category
Code
group
Subgroup
Class
Dense Forest
Da-0 Db-0
Ombrophyllous forest Ombrophyllous forest
Dense forest Dense forest
Aluvial Amazonian Lowland Amazonian
Dm-0 Ds-0
Ombrophyllous forest
Dense forest
Montane Amazonian
Ombrophyllous forest
Dense forest
Submontane Amazonian
Nondense
Aa-0
Ombrophyllous forest
Open
Alluvial
Forest
Ab-0
Ombrophyllous forest
Open
Lowland
As-0 Cs-0 Fa-0
Ombrophyllous forest Open Seasonal forest Deciduous Seasonal forest Semideciduous Seasonal forest Semideciduous Woody oligotrophic vegetation of swampy and sandy areas Woody oligotrophic vegetation of swampy and sandy areas Woody oligotrophic vegetation of swampy and sandy areas
Fs-0 La-0 Ld-0 Lg-0
Submontane Submontane Alluvial Submontane Open arboreal Dense arboreal Grassy-woody
LO-0
Areas of ecological tension and contact
Woody oligotrophic vegetation of swampy and sandy areas Ombrophyllous forest
ON-0
Areas of ecological tension and contact
Pf-0 SM-0
Areas of pioneer formations Areas of ecological tension and contact
SN-0 SO-0
Areas of ecological tension and contact Areas of ecological tension and contact
Ombrophyllous forest — seasonal forest Fluvio-marine influence Savanna—dense Ombrophyllous forest Savanna—seasonal forest Savanna—Ombrophyllous forest
of wood in each forest type in accordance with the volume of each species present (Fearnside 1997a). The wood volume data for density means are based on forest inventories done for the RADAMBRASIL Project. Approximately 90% of the almost 3000 ha of wood volume data used for the biomass estimates comes from the RADAMBRASIL surveys, while the remainder comes from surveys conducted in the 1950s by FAO (see Fearnside 1994, nd-a). The biomass estimates in Table 11.6 represent total biomass (above- and belowground, living and nonliving, including palms and nontree components) for unlogged forests
originally present (i.e., prior to European contact) in the Legal Amazon. All values are expressed as oven-dry biomass. The regional average is 463 t/ha, of which 354 t/ha is aboveground; nonliving biomass represents 28 t/ha for above-ground biomass and 4 t/ha for belowground biomass (Fearnside, nd-a). Adjustments for biomass removals through logging and for the location of deforestation through 1990 lower the average total biomass of forests still standing in 1990 to 462 t/ha, while the intense logging prior to felling and adjustments for the location of clearing activity in 1990 lower the average total biomass for forests
176
Martial Bernoux et al.
Table 11.6 Forest biomass per hectare: means by ecorerions, vegetation type, and state (t/ha)*. Category Category
Code
Dense Forest
Non-dense Forest
Brazilian State ** AC AP AM Da-0 Db-0 Dm-0 Ds-0 Dense forests Aa-0 Ab-0 As-0 Cs-0 Fa-0 Fs-0 La-0 Ld-0 Lg-0 LOO ON-0 Pf-0 SM-0 SN-0 SO-0
Nondense Forests All Forests
451
438 391 436
471 583 434 613 601
MA
513 495 344 478 493
MT
RO
RR
TO/GO
Area (Weighted Mean)
303
412 585 434 518 522
310
419 417 448 429 431
492
458 456 458
434 466 507
424 455
PA
402 402
536
379
536 457 445 424 403 373
370
370 379 418
546
418 457 431 431 437 409
479 410
379 398
335 398
379 398
434 456 387 379 370 420 435 435 435 490 388 435 435 396 431 422
419 384 370 403
361 379
383
395 431
383 360
472
457 431 431 500 431
456 431 379
111 286
478 510 416 506 502
398
408 582
453
408
478
393
392
379
395
422
386
591
487
436
393
469
400
428
367
463
Values in italics are for ecoregions where no sample exists: values are based on the mean in sampled plots for the same vegetation type in other states. Source: Fearnside, nd-a.; Brazilian state codes are: AC = Acre; AP = Amapa; AM = Amazonas; MA = Maranhao; MT = Mato Grosso; PA = Para; RO = Rondonia; RR = Roraima; TO = Tocantins, and GO =Goias.
cleared in that year to 406 t/ha (Fearnside nd-b, Fearnside 1997b). The area of each ecoregion is given in Table 11.7, based on measurement of the IBAMA map. Carbon stocks can be calculated by multiplying the area of each vegetation type (in hectares) by its respective biomass per hectare, and by the carbon content of the vegetation. Carbon content of primary forest vegetation is approximately 50% of dry weight (Fearnside et al. 1993, Higuchi and Carvalho 1994). The foregoing discussion refers only to "primary" (also known as mature or virgin
forests). The replacement landscape in deforested areas, including secondary forest vegetation, had an average total biomass (dry matter, including belowground and nonliving components) of 43.5 t/ha in 1990 in the 410,000 km2 that had been deforested by that year for uses other than hydroelectric dams (Fearnside 1996). Secondary forests, pasture, and agricultural biomass have carbon contents of approximately 45%. The C stock originally contained in biomass for all types of forests present in Legal Amazonia is 87.8 Pg. Dense forests contributed 55.6% (48.8 Pg) of this total, and
Carbon Storage in Biomass and Soils
177
Table 11.7 Area originally present of each forest ecoregion in the Brazilian Legal Amazon *. Brazilian State" Category Code Dense
Da-0
Forest
Db-0
AC
AM
9011
164,876
16,408 2184
615,203
Dm-0
Ds-0
AP
518
113
10,181
99,220
178,103
Subtotal 16,926 110,528 968,363 Nondense Aa-0
10,591
Forest
114,380
Ab-0
MA
MT
22,586
PA
RO
RR
TO/GO
76,570
2704
3326
2610
164,091
2066
10,248
3,418 413,345
14,607
83,692
24,574
23,154
657,424
19,377
117,927 5665
805
2273
3,055
3666
736
Fa-0
3554
Fs-0
24,317
34,373 817,682 1,943,938
79,417 366,496
41,064 124,620 286,271
37,555
Cs-0
20,661
23,154
211,052
As-0
259,097 832,786
1,988
65,748
TOTAL
77,794
8430
5386
7718
1041
1216
535,886
115
9903 3554
1328
34,404
La-0
14,979
970
Ld-0
37,405
10,967
15,949 48,372
Lg-0
9663
9767
19,430
LO-0
172,607
30,184
202,791
3045
178,906
ON-0 Pf-0
1,823
2089 384
SM-0 SN-0
SO-0
4226
Subtotal 124,971 6,049
1082 27,350
6570
577,441
12,709
168,069 2991 3894
4801
142,778 27,812
4781 21,932
384
22,124
Total All 141,897 116,577 1,545,804 37,283 Forests
7806
59,734
904
14,465
198,392
4286
6551
146,203
486,198 386,893 160,363 69,594 23,675 509,352 1,044,317 179,740 187,521 29,340
1,847,893 3,791,831
Source: Fearnside and Ferraz (1995). * Areas in km2 measured from 1:5,000,000 vegetation map (Brazil, IBGE and IBDF, 1988). These areas do not reflect losses due to recent deforestation. ** Brazilian state code are: AC = Acre; AP = Amapa; AM = Amazonas; MA = Maranhao; MT = Mato Grosso; PA = Para; RO = Rondonia; RR = Roraima; TO = Tocantins and GO = Goias.
nondense forests contributed 44.4% (39.0 Pg). If the forest area as of 1990 is considered (3.38 x 106 km2), the carbon stock contained in the biomass of all forest types is 78.0 Pg. Areas deforested as of 1990 are covered by secondary forests (including old secondary forests approximately pre-1960), pasture and agriculture having a carbon stock of 0.8 Pg.
Nonforest vegetation Areas covered by each vegetation type in Amazonia were estimated by Graca (1997) using a digitized image generated by scanning a phytoecological Legal Amazon map from SUDAM (Superintendency of Development in the Amazon) together with IBGE (Brazilian Institute of Geography and
178
Statistics) at a scale of 1:2,500,000 (Brazil, SUDAM and IBGE 1989). This map was used to calculate the total average biomass of nonforest vegetation types, weighted by area of each vegetation type. Figure 11.4 shows the main Legal Amazon vegetation types in the digitized map. The nonforest types in Legal Amazonia (not including the vegetation of deforested areas), in agreement with the classification done by SUDAM and IBGE, are characterized by savannas (closed), grasslands, campinas and pioneer formations divided into a total of 11 categories. A detailed presentation of this classification is shown in Table 11.8. Estimates of average total biomass weighted by the area covered by each vegetation type in Legal Amazonia were made using available data, based on the map described above, including alterations in forest cover to 1988.
Martial Bernoux et al.
The database for calculations to estimate the biomass was generated from surveys of volume of firewood by the RADAMBRASIL project. One hundred and sixty one datasets of firewood volume were used for nonforest types (dense and open savannas; 0.5 ha per sample). The biomass of the vegetation types not included by the surveys of RADAMBRASIL was estimated from small isolated surveys of ecological interest done in Amazonia or other tropical areas. Parameters to derive biomass estimates from data from firewood volume by RADAMBRASIL and also other categories not included are shown in Table 11.9. Results of biomass averages weighted by area occupied by each type of nonforest vegetation are shown in Table 11.10, along with an estimate of potential carbon contained in each biomass type. Mean total biomass for nonforest vegetation types is 33-1 t/ha and carbon stocks equal 1.2
Fig. 11.4 Main vegetation types in Brazilian Legal Amazon.
Carbon Storage in Biomass and Soils
179
Table 11.8 Nonforest vegetation types in the Brazilian Legal Amazon.
Category
Code Group
Savanna
Sa Sd
Other Nonforest
Subgroup
Sg Sp ST
Savanna Cerrado Savanna Cerrado Savanna Cerrado Savanna Cerrado Areas of ecological tension and contact
Ta Td Tp
Steppe-like savanna Steppe-like savanna Steppe-like savanna
Pa
Areas of pioneer formations Woody oligotrophic vegetation of swampy and sandy areas Ecological refugium High altitude
Lg Rm
Pg C. Savanna has a total average biomass of 28.4 t/ha and represents 86.3% of the area occupied by nonforest vegetation, with a carbon stock of 0.9 Pg.
Conclusions: The Future-How to Improve the Estimates Most of the available estimates are based on a simple approach: characterization and
Roraima grassland Roraima grassland Roraima grassland
Class Open arboreal Dense arboreal Grassy-woody Parkland Savanna—steppe-like savanna Open arboreal Dense arboreal Parkland Fluvial influence Grassy-woody Montane grassy-woody
delineation of a map unit, individual or multiple local sampling of each map unit, and finally integration over the map unit of the representative value. Moreover the error associated with the estimate is no more than a standard error calculated from the sample. This approach implies that each map unit is independent of its neighbors; it also ignores internal spatial variability. Therefore, the notion of a continuum in soil and vegetation
Table 11.9 Parameters to derive the estimate of total biomass of nonforest vegetation types frorr RADAMBRASIL data. Factor Multiplier
Source
Conversion of firewood volume for 0.39* biomass
Fearnside, 1992b
Adjust for aboveground biomass**
1.12***
Fearnside, 1992b; Kauffman et al., 1994
Adjust for underground biomass
1.56****
Seiler and Crutzen, 1980
390 kg dry weight/stere (m3 stacks firewood, including air spaces among pieces) for savanna (cerrado) in Carajas (value used by Fearnside 1992b). An increment is assumed in aboveground biomass of 3 t/ha of grass for open savanna (Kauffman et al. 1994). A multiplier factor of 1.12 is assumed for the fraction of 0-10 cm used for forest and that firewood is larger than 10 cm diameter (adopted by Fearnside 1992b). Assumed that underground biomass is the same as 36% of total biomass (used by Seiler and Crutzen 1980).
180
Martial Bernoux et al.
Table 11.10 Biomass and carbon stocks of nonforest vegetation in Amazonia. Code
Area (km^)
Total biomass (t/ha)
Savannas Sa* Sd*
364,505.5
33.6 60.2 11.4
Sg** Sp** St* Ta*
47,268.9 34,532.3 171,756.6 8,040.0 167.6
Td* Tp** Subtotal
2,323.5 9,612.5 638,206.9
Other nonforests 81,592.2 Pa***
Subtotal
19,025.1 966.7 101,584.0
All categories
739,790.9
Lg**** Rm*****
Mean biomass weighted by area (t/ha)
Carbon stock (106 1) 611.6 142.2 19.7 100.6 18.8
11.7 46.9 33.6 60.2
28.4
0.3 7.0 5.6 905.8
62.4
244.9 43.8 1.0 289.6
33.1
1195.5
11.7
66.7 46.0 22.2
Vegetation types estimated from RADAMBRASIL. Values for Td and Ta were extrapolated from estimate values, respectively, of Sd and Sa. ST was estimated from the average between Sd and Sa. Sp and Sg were estimated, respectively, from values of aboveground biomass of "campo sujo" and "campo limpo" found by Kauffman et al. (1994), for a region of closed savanna near Brasilia. Value for Sp was extrapolated for Tp type. Biomass of roots was estimated assuming the value of 1.6 for the root/stem ratio adopted by Schroeder and Winjum (1995) for vegetation in the savanna/grassland class. Value of total biomass for Pa was from Olson et al. (1983) for wetlands in the tropics, assuming mean carbon content of 0.45. Lg was estimate from aboveground biomass for "campina" (bana), for San Carlos do Rio Negro in Venezuela (Kauffman et al., 1988) and belowground biomass of "campina" near Manaus (Klinge, 1976). Value of total biomass for Rm from Olson et al. (1983) for areas of grass-woody vegetation of tropical mountain complexes, assuming mean carbon content of 0.45.
cover is not taken into account. Recent its components. For example, biomass of approaches such as geostatistic-based palms and vines contributed 10.8% to total studies (Bernoux et al. 1998b, Cerri et al. aboveground biomass in an open forest in 2000) suggest that these techniques, if Ariquemes, Rondonia (Graca, 1997) and applied at continental scales, would give 3.7% for dense forest in Manaus (Fearnside a more precise estimate of C stored in soil et al., nd-b). Also necessary are better estiand vegetation. mates of forest root biomass, mainly of Concerning more precisely the forest thick roots. Root biomass data are quite rare biomass, better estimates can be made by for the Amazon, currently the only data obtaining better data for each forest type, existing are for Manaus (Amazonas), Jari A large variation exists among forests in (Para), Paragominas and Trombetas (Para) relation to percent of contribution of each of (Klinge et al. 1975; Klinge and Rodrigues
Carbon Storage in Biomass and Soils
1973; Russell 1983, D. Nepstad; personal communication, cited by Fearnside 1994). Intrinsic forest factors such as the percent of bark and hollow trees should be studied better, as well as the estimates for carbon content of each forest type. The nonforest biomass also needs better refinement, as it contributes on the order of 2.6% of the total stock of carbon contained in vegetation biomass of the Legal Amazon (Graca, 1997). Brown and Lugo (1992), Fearnside (1992b), and Graca (1997) used the average apparent general density for wood trees in the Amazon (or other tropical forests) to estimate biomass of different forest types. It is known that the density of trees can vary between species and within species. Recently, Fearnside (1997a) recalculated the biomass for Amazonia using an average of apparent wood density weighted by area occupied by each forest type. However, factors of biomass expansion and biomass adjustments are still treated in great part in a homogeneous manner in those calculations. However, biomass estimates for the Amazon region using surveys on a regional scale are currently the most reliable. To improve these data, a treatment more refined for different forest formations that comprise the forests would be needed. Estimates through indirect methods should improve the regression equations to estimate the biomass of larger trees, which represent
181
most of the biomass. The distribution of the biomass within the diameter classes of forest trees is quite distorted. Martinelli et al. (1994) found that only 15 trees with DBH greater than 55 cm represented about 50% of aboveground biomass. This same tendency was found by Graca et al. (1999) for an open forest in Ariquemes, Rondonia, where 7 trees in 1 ha represented 30% of aboveground biomass. Thus, new biomass estimates should prioritize the measure of big trees, improving the reliability of regression equations with wood density and form factor for those trees. Brown et al. (1995) related that measurements and biomass component considerations limit the accuracy of current biomass estimates by at least 10%. Acknowledgments: Research support at CENA was provided by the Fundacao de Amparo a Pesquisa do Estado de Sao Paulo (FAPESP) with contracts 94/6046-0 and 95/1451-6, and by the Fundacao Coordenacao de Apercoamento de Pessoal de Nivel Superior (CAPES-MEG) 'with grant number 2129/95. P. M. Fearnside and P. M. L. A. Graca thanks the National Council of Scientific and Technological Development (CNPq Al 350230/97-98) and the National Institute for Research in the Amazon (INPA PPI 5-3150) for financial support. Summer Wilson and Christian Feller provided valuable comments on the manuscript.
Literature Cited Araujo, T. M., J. A. Carvalho Jr., N. Higuchi, A. C. P. Brasil Jr., and A. L. A. Mesquita. 1997. 'Estimativa de taxas de liberacao de carbono em experimento de queimadas no Estado do Para." Anais da Academia Brasileira de Ciencia, 69(4): 575-584. Arrouays, D., and P. Pelissier. 1994. "Modeling carbon storage profiles in temperate forest humic loamy soils of France." Soil Science. 157: 185-192. Batjes, N. H. 1996. "Total carbon and nitrogen in the soils of the world." European Journal of Soil Science. 47: 151-163.
Beinroth, F. H. 1975. "Relationships between U.S. soil taxonomy, the Brazilian soil classification system, and FAO/UNESCO soil units." In: Soil Management in Tropical America, eds. E. Bornemisza and A. Alvarado (North Carolina State Univ., Raleigh), pp. 92-108. Bennema, J. 1974. "Organic carbon profiles in Oxisols." Pedologie XXIV, 2: 119-146. Bernoux, M. 1998. Conteudo de carbono dos solos da Amazonia ocidental e mudancas decorrentes da conversao da floresta em pastagens-Stocks de carbone des sols de 1'Amazonie occidentale et leur dynamique lors de la
182
conversion de la foret en paturage . Doctorate Thesis of University of Sao Paulo-Brazil and University of OrleansFrance (Co-tutelle), Piracicaba, SP, p. 121. Bernoux, M., D. Arrouays, C. Cerri, and H. Bourennane. 1998a. "Modeling vertical distribution of carbon in Oxisols of the Western Brazilian Amazon (Rondonia)." Soil Science. 163(12): 941-951. Bernoux, M., D. Arrouays, C. Cerri, P. M. de A Graga, B. Volkoff. and J. Trichet. 1998b. "Estimation des stocks de carbone des sols du Rondonia (Amazonie bresilienne)." Etudes et Gestion des Sols. 5(1): 31-42. Bernoux, M., D. Arrouays, M. Volkoff, C.C. Cerri, and C. Jolivet. 1998c. "Bulk densities of Brazilian Amazon soils related to other soil properties." Soil Science and Society of America Journal 62(3): 743-749. Bernoux, M., C.C. Cerri, C. Neill, and J. L Moraes. 1998d. "The use of stable carbon isotopes for estimating soil organic matter turnover rates." Geoderma. 82(1-3): 43-58. Bohrer, C. B. de, and A. N. S. Campos. 1993. "Estimativa da biomassa florestal na Amazonia Legal a partir de dados proveniences de inventario florestal." In: CONGRESSO FLORESTAL BRASILEIRO, 7., Curitiba, 1993. Anais. Curitiba: SBS, SBEF. pp. 511-512. Bouwman, A. F. 1990. Soils and the Greenhouse effect. Proceedings of the International Conference on Soils and the Greenhouse Effect. John Wiley & Sons, New York, USA. Brazil, Institute Brasileiro de Geografia e Estatistica (IBGE) and Institute Brasileiro do Desenvolvimento Florestal (IBDF). 1988. Mapa de Vegetacao do Brasil. Scale 1:5,000,000. Institute Brasileiro do Meio Ambiente e dos Recursos Naturais Renovaveis (IBAMA). Brasilia, Brazil. Brazil, Superintendencia do Desenvolvimento da Amazonia (SUDAM) and Fundacao Instituto Brasileiro de Geografia e Estatistica (IBGE). 1989. Amazonia Legal: Mapa de Vegetacao. Scale 1:2.500.000. Ministerio do Interior (MI) and Secretaria do Planejamento e Coordenacao da Presidencia da Republica (SEPLAN). Brasilia, Brazil. Brazil, Projeto RADAMBRASIL. 1973-1983. Levantamento de Recursos Naturais, Vols. 1-27. Ministerio das Minas e Energia, Departamento Nacional de Producao Mineral (DNPM), Rio de Janeiro, Brazil. Brown, S., and A. E. Lugo. 1982. "The storage and production of organic matter in tropical forests and their role in the global carbon cycle." Biotropica 4(3): 161-187. Brown, S. and A. E. Lugo. 1984. "Biomass of tropical forests: A new estimate based on forest volumes." Science 223: 1290-1293. Brown, S. and A. E. Lugo. 1992a. "Aboveground biomass estimates for tropical moist forests of the Brazilian Amazon." Interciencia 17(1): 8-18. Brown, S. and A. E. Lugo. 1992b. "Biomass of Brazilian Amazonian forests: The need for good science." Interciencia 17(4): 201-203. Brown, S., A. J. R. Gillespie, and A. E. Lugo. 1989. "Biomass estimation methods for tropical forests with applications to forest inventory data." Forest Science 35(4): 881-902. Brown, I. F., L. A. Martinelli, W. W. Thomas, M. Z. Moreira, C. A. Cid Ferreira, and R. A. Victoria. 1995. "Uncertainty in the biomass of Amazonian forests: An example from Rondonia, Brazil." Forest Ecology and Management 75: 175-189.
Martial Bernoux et al. Burgess, T. M., and R. Webster. 1980a."Optimal interpolation and isarithmic mapping of soil properties, I, The semi-variogram and punctual kriging," Journal of Soil Science. 31: 315-331. Burgess, T. M., and R. Webster. 1980b."Optimal interpolation and isarithmic mapping of soil properties, II, Block kriging," Jounal of Soil Science. 31: 333-341. Carvalho Jr, J. A., J. M. Santos, M. M. Leitao, and N. Higuchi. 1995. "A tropical rainforest clearing experiment by biomass burning in the Manaus region." Atmospheric Environment 29(17): 2301-2309Cerri C. C., M. Bernoux, D. Arrouays, B. Feigl, and M. C. Piccolo. 2000. "Carbon pools in soils of the Brazilian Amazon." In: Global Climate Change and Tropical Ecosystems, eds. R. Lai, J. M Kimble, and B. A. Stewart (CRC press, Boca Raton, Advances in Soil Science) pp. 33-50. De Vries, P. G. 1974. "Multi-stage line intersect sampling." Forest Science 20: 129-133Dixon, R. K., S. Brown, R. A. Houghton, A. M. Solomon, M. C. Trexler, and J. Wisniewski. 1994. "Carbon pools and flux of global forest ecosystems." Science 263: 1985-190. Empresa Brasileira de pesquisa Agropecuaria. 1981. Mapa de Solos do Brasil, escala 1:5,000,000. Eswaran, H., E. van Den Berg, and P. Reich. 1993. "Organic carbon in soils of the world." Soil Science and Social American Journal 57: 192-194. Fearnside, P. M. 1985- "Brazil's Amazon forest and the global carbon problem." Interciencia 10(4): 179-186. Fearnside, P. M. 1986. "Brazil's Amazon forest and the global carbon problem: reply to Lugo and Brown." Interciencia 11(2): 58-64. Fearnside, P. M. 1992a. Greenhouse gas emissions from deforestation in the Brazilian Amazon. Carbon Emissions and Sequestration in Forests: Case Studies from Developing Countries. Vol. 2, LBL-32758, UC-402. Climatic Change Division, Environmental Protection Agency, Washington, D.C. and Energy and Environment Division, Lawrence Berkeley Laboratory (LBL), University of California (UC), Berkeley, U.S.A. p. 73Fearnside, P. M. 1992b. "Forest biomass in Brazilian Amazonia: comments on the estimate by Brown and Lugo." Interciencia 17(1): 19-27. Fearnside, P. M. 1993. "Biomass of Brazil's Amazonian forests: Replay to Brown and Lugo revisited." Interciencia 18(1): 5-7. Fearnside, P. M. 1994. "Biomassa das florestas Amazonicas brasileiras." pp. 95-124 In: Anais do Seminario Emissao X Sequestro de CO2: Uma Nova Oportunidade de Negocios para o Brasil. Companhia Vale do Rio Doce (CVRD), Rio de Janeiro, p. 221. Fearnside, P. M. 1996. "Amazonian deforestation and global warming: Carbon stocks in vegetation replacing Brazil's Amazon forest." Forest Ecology and Management 80: 21-34. Fearnside, P. M. 1997a. "Wood density for estimating forest biomass in Brazilian Amazonia." Forest Ecology and Management 90(1): 59-89. Fearnside, P. M. 1997b. "Greenhouse gases from deforestation in Brazilian Amazonia: Net committed emissions." Climatic Change 35: 321-360. Fearnside, P. M. nd-a. Biomass of Brazil's Amazonian forests (in preparation).
Carbon Storage in Biomass and Soils Fearnside, P. M. nd-b. Tropical forest logging and management: Implications for global warming, (in preparation). Fearnside, P. M., and R. I. Barbosa. 1998. "Soil carbon changes from conversion of forest to pasture in Brazilian Amazonia." Forest Ecology and Management 108 (1-2): 147-166 Fearnside, P. M., and W. M. Guimaraes. 1996. "Carbon uptake by secondary forests in Brazilian Amazonian." Forest Ecology and Management 80: 35-46. Fearnside, P. M., P. M. L. A. Graca, N. Leal Filho, F. J. A. Rodrigues, and J. M. Robinson. 1999. "Tropical forest burning in Brazilian Amazonia: measurement of biomass loading, burning efficiency and charcoal formation at Altamira, Para," Forest Ecology and Management 123(1): 65-79Fearnside, P. M., P. M. L. A. Graca, and F. J. A. Rodrigues. nd-b. Burning of Amazonian rainforests: Burning efficiency and charcoal formation in forest cleared for cattle pasture near Manaus, Brazil (in preparation). Fearnside, P. M., N. Leal Filho, and F. M. Fernandes. 1993. "Rainforest burning and global carbon budget: biomass, combustion efficiency, and charcoal formation in the Brazilian Amazon." Journal of Geophysical Research (Atmospheres) 98(D9): 16733-16743. Graca, P. M. L. A.. 1997. Conteudo de carbono da biomassa florestal na Amazonia e alteracoes apos a queima. Masters thesis in forest sciences, Escola Superior de Agricultura "Luiz de Queiroz," Universidade de Sao Paulo, Piracicaba, Sao Paulo, p. 105. Graca, P. M. L. A., P. M. Fearnside, C. C. Cerri. 1999. "Burning of Amazonian forest in Ariquemes, Rondonia, Brazil: Biomass, charcoal formation and burning efficiency." Forest Ecology and Management 120 (1-3):179-191. Higuchi, N., and J. A. de Carvalho, Jr. 1994. "Fitomassa e conteudo de carbono de especies arboreas da Amazonia." pp. 125-153 In: Anais do Seminario Emissao X Sequestro de CO2, Companhia Vale do Rio Doce (CVRD), Rio de Janeiro, Brazil, p. 221. Higuchi, N., J. M. Santos, M. Imanga, and S. Yoshida. 1994. Aboveground biomass estimate for Amazonian dense tropical moist forests. Memorial Faculty of Agriculture of Kagoshima University 30:43-54. Honzak, M., R. M. Lucas, Amaral do I., P. J. Curran, G. M. Foody, and S. Amaral. 1996. "Estimation of the leaf area index and total biomass of tropical regenerating forests: comparison of methodologies." In: Amazonian deforestation and climate, eds. J. H. C. Gash, C. A. Nobre, J. M. Roberts and R. L. Victoria (John Wiley, Chichester), pp. 365-381. Jacomine, P. K. T., and M. N. Camargo. 1996. "Classificacao pedologica nacional em vigor." In: O solo nos grandes dominios morfoclimaticos do Brasil e o desenvolvimento sustentado, eds. V. H. Alvarez, L. E. F. Fontes, and M. P. F. Fontes (Sociedade Brasileira de Ciencia do Solo, Univ. Federal of Vicosa, Vicosa, Mato Groso, Brazil), pp. 675-693. Jordan, C. F., and C. Uhl. 1978. "Biomass of a "tierra firme" forest of the Amazon basin." Oecologia Plantarum 13(4): 387-400. Kaiser, L. 1983. "Unbiased estimation in line-intercept sampling." Biometrics 39: 965-976. Kauffman, J. B., D. L. Cummings, D. E. Ward, and R. Babbit.
183
1995. "Fire in the Brazilian Amazon: 1. Biomass, nutrient pools, and losses in slashed primary forests." Oecologia 104: 397- 408. Kauffman, J. B., D. L. Cummings, and D. E. Ward 1994. "Relationships of fire, biomass and nutrient dynamics along a vegetation gradient in the Brazilian cerrado." Journal of Ecology 82: 519-531. Kauffman, J. B., C. Uhl, and D. L. Cummings. 1988. "Fire in the Venezuelan Amazon 1: fuel biomass and fire chemistry in the evergreen rainforest of Venezuela." Oikos 53: 167-175. Kimble, J. M., H. Eswaran, and T. Cook. 1990. "Organic carbon on a volume basis in tropical and temperate soils." In: International Congress of Soil Science, Kyoto, 1990, Transactions. Kyoto: International Society of Soil Science, v. 5, pp. 248-253Klinge, H. 1975. "Root mass estimation in lowland tropical rain forests of Central Amazonia, Brazil: III. Nutrients in fine roots from giant humus podsols." Tropical Ecology 16(1): 28-38. Klinge, H., and R. Herrera. 1983. "Phytomass structure of natural plant communities on spodosols in southern Venezuela: the tall Amazon caatinga forest." Vegetatio 53: 65-84. Klinge, H., W. A. Rodrigues, E. Bruning, and E. J. Fittkau. 1975. "Biomass and structure in a Central Amazonian rain forest." In: Tropical ecological Systems: trends in terrestrial and aquatic research, eds. F. B. Golley, and E. Medina (Springer-Verlag, New York), pp. 115-122. Klinge, H., W. A. Rodrigues, E. Bruning, and E. J. Fittkau. 1975. "Biomass and structure in a Central Amazonian rain forest." In: Tropical ecological Systems: trends in terrestrial and aquatic research, eds. F. B. Golley, and E. Medina, (Springer-Verlag, New York), pp. 115-122. Klinge, H., and W. A. Rodrigues. 1973. "Biomass estimation in a central Amazonian rain forest." Acta Cientifica Venezoelana 24: 225-237. Klinge, H., W. A. Rodrigues, E. Brunig, and E. J. Fittkau. 1975. "Biomass and structure in a Central Amazonian rain forest." In: Tropical ecological Systems: trends in terrestrial and aquatic research, eds. F. B. Golley, and E. Medina, (Springer-Verlag, New York), pp. 115-122. Lugo, A. E. and S. Brown. 1986. "Brazil's Amazon forest and the global carbon problem." Interciencia 11: 57-58. Martinelli, L. A., M. Z Moreira, I. F. Brown, R. L. Victoria. 1994. Incertezas associadas as estimativas de biomassa em florestas tropicais: O exemplo de uma floresta situada no Estado de Rondonia." pp. 197-221 In: Anais do Seminario Emissao X Sequestro de CO2: Uma Nova Oportunidade de Negocios para o Brasil. Companhia Vale do Rio Doce (CVRD), Rio de Janeiro, p. 221. Matheron, G. 1965. Les Variables Regionalisees et Leur Estimation, p. 306, Masson, Paris. Moraes, J. F. L. 1991. Conteudos de carbono e nitrogenio e tipologia de horiontes nos solos da bacia amazonica. Mestrado Dissertation, CENA-USP, Piracicaba, Brazil. Moraes, J. L., C.C. Cerri, J. M. Melillo, D. Kicklighter, C. Neill, D.L. Skole, and P. A. Steudler. 1995. "Soil carbon stocks of the Brazilian Amazon basin." Soil Science America Journal 59: 244-247. Moraes, J. F. L., C. C. Cerri, B. Volkoff, and M. Bernoux. 1996. "Soil properties under Amazon forest and changes
184
due to pasture installation in Rondonia, Brazil." Geoderma 70: 63-81. Nelson, R., R. Swift, and W. Krabill. 1988. "Using airborne lasers to estimate forest canopy and stand characteristics." Journal of Forestry 86(10): 31-38. Nepstad D. C, C. R. de Carvalho, E. A. Davidson, P. H. Jipp, P. A. Lefebvre, G. H. Negreiros, E. D. da Silva, T. A. Stone, S. E. Trumbore, and S. Vieira. 1994. "The deep-soil link between water and carbon cycles of Amazonian forests and pastures." Nature v. 372, p. 666-669. Olson, J. S., J. A. Watts, and L. J. Allison. 1983. Carbon in live vegetation of major world ecosystems. Rep. DOE/NBB-0037, U.S. Department of Energy, Office of Energy Research, Carbon Dioxide Research Division, Washington, D.C., p. 164. Post, W. M., W. R. Emmanuel, P. J. Zinke, and A. G. Stangenberger. 1982. "Soil carbon pools and world life zones." Nature 298: 156-159. Prado, H. do. 1996. Manual de classificacao de solos do Brasil. 3rd ed. Jaboticabal, FUNEP. Revilla Cardenas, J. D. 1987. Relatorio: Levantamento e analise da fitomassa da UHE de Kararao, Rio Xingu, p. 181. Institute Nacional de Pesquisas da Amazonia, Manaus, Brazil. Revilla Cardenas, J. D., F. L. Kahn, and J. L. Guillaumet. 1982. Estimativa da fitomassa do reservatorio da UHE de Tucurui. In: Projeto Tucurui, Relatorio Semestral JanuaryJune 1982, vol 2, Limnologia, macrofitas, fitomassa, degradacao de fitomassa, doencas endemicas, solos, pp. 1-11, Institute Nacional de Pesquisas da Amazonia, Manaus, Brazil. Russell, C. E. 1983. Nutrient cycling and productivity of native and plantation forests at Jari Florestal, Para, Brazil. Dissertacao de Ph.D. em Ecologia, University of Georgia, Athens. Seiler, W., and P. J. Crutzen. 1980. "Estimates of gross and net fluxes of carbon between the biophere and the atmosphere from biomass burning." Climatic Change 2: 207-247. Sader, S. A. 1987. "Forest biomass, canopy struture and species composition relationships with multipolarization L-band syntetic aperture radar data," Photogrammetric Engineering and Remote Sensing. 53(2): 193-202.
Martial Bernoux et al. Sader, S. A., R. B. Waide, W. T. Lawrence, and A. T. Joyce. 1989. "Tropical forest biomass and successional age class relationships to a vegetation index derived from Landsat TM data." Remote Sensing Environmental 28: 143156. Schroeder, P. E., and J. K. Winjum. 1995. "Assessing Brazil's carbon budget: I. Biotic carbon pools." Forest Ecology and Management 75: 77-86. Shang, C., and H. Tiessen. 1997. "Organic matter lability in a tropical Oxisol: evidence from shifting cultivation, chemical oxidation, particle size, density, and magnetic fractionations." Soil Science. 162: 795-807. Sombroek, W. G. 1992. Biomass and carbon storage in the Amazon ecosystems. Interciencia 17: 269-272. Sombroek, W. G., F. O. Nachtergaele, and A. Hebel. 1993. "Amounts, dynamics and sequestration of carbon in tropical and subtropical soils." Ambio 22: 417-426. Uhl, C., R. Buschbacher, and E. A. S. Serrao. 1988. "Abandoned pastures in Eastern Amazonia. I. Patterns of plant succession." Journal of Ecology 76: 663-681. USDA. 1996. Chemical analyses, organic carbon (6A) Walkley-Black modified acid-dichromate organic carbon (6A1) Pages 219-222. Soil survey laboratory methods manual. Soil Survey Investigations Report na 42, Version 3.0, January 1996. U.S. Department of Agriculture, Washington, D.C., USA. Van Wanbeke, A. 1992. Soils of the tropics: properties and appraisal. McGraw-Hill, New York, p. 343. Van Wagner, C. E. 1968. "The line intersect method for forest fuel sampling." Forest Science, 14(1): 20-26. Walkley, A., and I. A. Black. 1934. "An examination of the Degtjareff method for determining soil organic matter and a proposed modification of the chromic titration method." Soil Science. 37:29-38. Warren, W. G. and P. F. Olsen. 1964. "A line intersect technique for assessing logging waste." Forest Science 10(3): 267-276. Whittaker, R. H., and G. E. Likens. 1975. "The biosphere and man." In: Primary production of the biosphere. Ecological Studies 14, eds. H. Lieth, and R. H. Whittaker (SpringerVerlag, Heidelberg, Berlin, New York).
12 Terrestrial Inputs to Amazon Streams and Internal Biogeochemical Processing Michael E. McClain and Helmut Elsenbeer
Enormous meandering rivers are the most remarkable fluvial feature of the Amazon landscape, but these rivers are only the largest component of a much denser network of streams which finely dissects and drains the basin. In terms of combined length and total amount of lotic habitat, streams dominate over their more visible downstream counterparts; this dominance is especially dramatic for first- and secondorder streams which alone may account for greater than 80% of total channel length in meso-scale Amazon drainag basins (Table 12.1). The flow of Amazon streams emerges directly from the extensive forests and savannas that compose the basin. Biogeochemical cycles in streams are thus intricately associated with processes operat-
ing in adjoining riparian and upland ecosystems. Terrestrial processes regulate the input of organic and inorganic species to stream systems, and the chemistry of inflowing waters determines, to some extent, the nature of subsequent reactions and even the composition of the stream's biological community (Fittkau 1971). Undisturbed Amazon streams are thought to experience virtually no primary production (Walker 1995), thus most inputs of energy, as well as nutrients, must ultimately derive from terrestrial sources. This connection is particularly acute in first-order streams where there is no upstream input and all water, particulates, and solutes derive from immediately adjacent to the stream. Pathways linking the two systems include groundwater runoff, surface and subsurface storm
Table 12.1. Distribution of streams in the 3300 km2 Cueiras basin of the central Amazon. Data hand digitized from 1:100,000 scale map provided by the Brazilian Ministerio do Exercito. Stream Order
Number of Streams
Total Length (km)
% of Total Length
1
584 121 29 10 2 1
1699
65 19 8 4 2 2
2
3 4 5 6
501 207
103 46 61
186
runoff, wetland seepage, direct litterfall, and litter blow-in (Fig. 12.1). These pathways are active across the entire Amazon basin, but their relative importance may vary regionally (McClain and Richey 1996, Elsenbeer and Lack 1996). Riparian ecosystems continue to influence the biogeochemistry of downstream reaches, but as streams become rivers upstream and in-channel influences become increasingly dominant. Streams and the corridors through which they flow also play a crucial role in regional-scale biogeochemical cycles. Greater than 90% of all terrestrial to lotic transfers in the Amazon basin occur in streams of order 6 and less. Thus, organic and inorganic species moving from terrestrial systems to large rivers and ultimately to the ocean must first pass through streams, where rates of material cycling and processing are rapid.
Michael E. McClain and Helmut Elsenbeer
Stream corridors also contain abundant fringing wetlands, which are sources of methane (CH4) and perhaps other greenhouse gases. Junk and Furch (1993) estimated that wetlands along the margins of small Amazonian streams may cover a combined area of one million km2, or nearly three times the area of massive floodplains bordering the region's major rivers. There has been very little consideration of the role of small streams in basin-wide land-atmosphere exchanges, but it is clear that their contributions must be addressed. From prehistoric times, the human occupation of Amazonia has been concentrated along major transport avenues. For the greatest part of the basin's occupation history, these avenues were large rivers, which traverse and connect all parts of the basin along natural and free-flowing
Figure 12.1. Illustration of the major pathways linking terrestrial and stream ecosystems in the Amazon basin and the main physicochemical processes operating in each ecosystem. Processes identified by an asterix have been examined in Amazon stream systems, although the number of these studies and diversity of terrain types examined is generally very limited.
Terrestrial Inputs to Amazon Streams and Internal Biogeochemical Processing
courses. Humans settled along large rivers and relied upon them as sources of food and water for drinking, cooking, and bathing and as repositories for wastes. During the second half of the twentieth century, however, roads have been cut across the region in straight lines, connecting region to region and supplanting rivers as the major avenue of immigration and occupation. Roads brought immigrants into the region in unprecedented numbers, this time settling along roads from which they cut homesteads back into the forest. As ever, water is crucial to their everyday needs, but most often the closest water source is a small stream rather than a large river. Because of their smaller size and drainage area, streams are acutely sensitive to anthropogenic disturbance. Thus, as colonization continues, the most extreme impacts to water quality are likely to occur in countless small streams throughout the basin. Furthermore, it is the cumulative impact of these alterations that is likely to produce the first detectable change in the water quality of large rivers of the basin. In view of these varied roles of small streams in the larger Amazonian ecosystem, current and recent research has primarily focused on four over-arching questions. 1. What are the principal flowpaths linking terrestrial and lotic systems, and what processes control the composition and flux of particulates and solutes moving to streams? 2. How are terrestrial inputs processed within the lotic environment, and how does this processing support the trophic structure of lotic systems? 3. What is the role of streams in regional to continental elemental cycles? 4. What are the biogeochemical consequences of anthropogenic activities in the Amazon (that is, how do these activities alter the natural functioning of stream
187
systems), and what management practices best minimize adverse effects? In this chapter we will present the progress made toward answering these questions. We will also suggest where research is necessary to fill remaining gaps in our understanding.
The Amazon Stream The label "Amazon" stream is deceiving in its suggestion of uniformity. In the vast region of Amazonia one will encounter a heterogeneity of stream types which is probably representative of a great part of the heterogeneity that occurs across the entire globe. Beginning at Andean glaciers, 5000+ m.a.s.l., Amazon streams flow in tangled braids down broad U-shaped
Figure 12. 2. Photo of the Achumani river at 4000 m elevation. The Achumani (l6'30"S; 68'02"W) passes near La Paz, Bolivia, and feeds into the Beni River system.
188
valleys (Fig. 12.2), there joining with other Amazon streams that emerge from the tundra bogs of slightly lower elevations. Together these streams cut deep V-shaped valleys as they cascade down the precipitous slopes of the Andean Cordillera, combining to form rivers loaded with nutrientrich sediments (Fig. 12.3). At the base of the mountains these rivers deposit much of their suspended loads across an expansive, forest-covered, depositional plain, from which emerge Amazon streams with lower gradients and smaller, but still nutrient rich, loads. The suspended mineral contributions of the Andes continue to move downstream, but beyond the piedmont region they are confined to the mainstem Amazon River and a few of its major tributaries, the Madeira, Purus, lea, and Japura. Outside of these river corridors, across the more than 4 million km2 of lowland Amazonia, occur still more variable Amazon streams, which
Michael E. McClain and Helmut Elsenbeer
drain ancient and mineral depleted geologic formations. To the north and south lie Precambrian shields covered by forests, grasslands, and all forms of intermediate vegetation. In the central basin, a nearly complete cover of forest masks the heterogeneity of soils beneath. From all of these terrains emerge streams carrying unique imprints of the landscape, variable in their loads of nutrients and organic matter and supporting diverse biological communities. Eventually, this bewildering complexity of streams is homogenized into the mainstem Amazon River and its unified contribution is delivered to the Atlantic Ocean. Although the characteristics of Amazon streams are strongly heterogeneous, their biogeochemistry is governed by common sets of processes, and a general framework for evaluating biogeochemical dynamics may be applied. In this chapter we draw upon existing theories such as the River Continuum Concept (Vannote et al. 1980, Cummins et al. 1995) and the Patch Dynamics Concept (Pringle et al. 1988) to frame our understanding.
The Historical Roots of Biogeochemical Stream Research in Amazonia
Figure 12.3. Photo of the Coroico River near Caranavi, Bolivia (15'50"S; 67'33"W).
Biogeochemical investigations of Amazonian stream waters began during the second half of the nineteenth century with an investigation of the nature of blackwater rivers (Reindl 1903). This investigation included the chemical composition of samples from the mainstem Amazon and the Negro River, an interpretation of the difference between black- and Whitewater rivers in terms of calcium and organic matter contents, and a comparison of blackwater rivers worldwide. In the second half of the twentieth century, a German-Brazilian collaborative effort resulted in the establishment of chemical limnological laboratories at the newly established Institute Nacional de Pesquisas da
Terrestrial Inputs to Amazon Streams and Internal Biogeochemical Processing
Amazonia (INPA), in Manaus. Through this collaboration, it was Sioli (1954, 1955) who first documented the generally ion-poor condition of stream waters in the central Amazon region and linked this condition to a scarcity of nutrient reserves in central Amazonian forests. He also correctly attributed the coloring of blackwater streams to processes operating in Podzolic soils of the basin, thereby revising the erroneous notion that clear stream waters turned black upon flowing over the red-brown roots of certain plants or upon flowing into flooded (Igapo) forests (Sioli 1964). This early phase of German-Brazilian biogeochemical research was motivated primarily by the goal of explaining the ecological structure of stream and lake biological communities. Work focused on delineating spatial patterns in the abundance of nutrients in relation to local edaphic and geologic characteristics. Fittkau (1967, 1971) summarized the early accomplishments of this effort in his ecological classification of the Amazon basin into three distinct provinces: (1) the central Amazon province where soils and stream waters are extremely nutrient poor, (2) the northern and southern shield province where soils and stream waters are moderately nutrient poor, and (3) the Andean province where soils and waters are nutrient rich. Increasing nutrient abundance in waters of these provinces were theorized to promote greater densities of more complex biological communities. In the 1970s, research into controls on stream biogeochemistry took a major step forward with the initiation of systematic studies into the sequential processes affecting water as it moves through the forest ecosystem and into the stream. Brinkmann and Santos (1971, 1973) led the way through studies of major cations in rainwater, stemflow, throughfall, groundwater, and stream water at the Ducke Reserve, north of the city of Manaus. They concluded that
189
more magnesium and calcium were moving in the hydrological cycle of central Amazonian forests than had previously been thought (mainly in stemflow and throughfall), but that these nutrients were effectively retained in forests rather than being transported on to groundwater and streams. These findings seemed to support the hypothesis of direct nutrient recycling in Amazonian forests, which had been suggested by Sioli (1964) and later elaborated by Went and Stark (1968) and Stark (1971a, 1971b) [see Cuevas, this volume]. Nortcliff and Thornes (1978) and Nortcliff et al. (1979) explained the low ion/nutrient content of stream waters by suggesting that stream flow originates largely from flow within soil macropores. They reasoned that, as the residence time of draining waters in these macropores is relatively brief, only small quantities of mineral-derived nutrients accumulate. In support of this hypothesis, they later showed that cation concentrations (Mg, Ca, K, Na) in soil water generally increased in smaller soil pores (Nortcliff and Thornes 1989). This mechanistic explanation of low stream nutrient concentrations added further strength to the already popular biological explanations for closed nutrient cycling in Amazon forests. With its intuitive appeal and apparent experimental backing, the notion of closed nutrient cycling based on low stream nutrient concentrations persisted into the 1990s, when Forti and Neal (1992) arrived at the same conclusion using a field sampling approach similar to that of Nortcliff and Thornes. Elsenbeer et al. (1995), however, questioned the sampling strategies of previous studies and therefore the observational basis for the link between low stream nutrient concentrations and direct nutrient recycling in adjacent forests. They demonstrated that a nutrient "leakage" occurs during stormflow conditions. Therefore, any sampling design that ignores rainfall event-
190
activated hydrologic flowpaths and stormflow itself is prone to underestimate nutrient losses, and hence to inadvertently infer tight nutrient recycling. Much of the understanding gained during the first 30 years of systematic biogeochemical research in Amazonia was summarized in various chapters of the now classic book, The Amazon: Limnology and Landscape Ecology of a Mighty River and its Basin (Sioli 1984). Building upon the understanding gained during these early years of stream research in Amazonia, researchers in the past decade have probed more deeply into the controls on stream biogeochemistry and the consequences of human activities. The remainder of this chapter will focus on these more recent advances.
Michael E. McClain and Helmut Elsenbeer
has not yet been quantified, but it is reasonable to assume that its rate is similar to that in riparian forests. In a survey of the available literature, McClain and Richey (1996) found a consistent rate of 0.7 kg rrr2 yr1 across forests ranging from montane (Veneklaas 1991) to terra firme riparian (Franken et al. 1979, Luizao 1989) to flooded forests (Adis et al. 1979, Irmler 1982). Direct litterfall inputs are continuous throughout the year but do vary seasonally. In contrast, mass wasting represents truly episodic inputs of paniculate matter to streams, such as in the form of landslides in mountainous regions. Despite their geographic concentration in the Andean portion of the Amazon basin, they codetermine the chemical and mineralogical
Terrestrial/Lotic Linkages Question: What are the principal flowpaths linking terrestrial and lotic systems, and what factors control the composition and flux of particulates and solutes moving to streams? The terrestrial signals that streams receive depend both on the pathways along which incoming water and materials travel and on the chemical composition and reactivity of materials encountered along the way. Specific pathways include (refer again to Fig. 12.1.): 1. direct channel input (litterfall, mass wasting, precipitation, throughfall), 2. overland flow, 3. lateral subsurface flow (throughflow, subsurface stormflow), 4. groundwater flow (including seeps and springs on hillslopes and wetlands). Litterfall and mass wasting represent material transport pathways, influenced by hydrologic conditions but predominantly driven by structural variables in trees and hillslopes. Direct litterfall to stream channels
Figure 12.4. Mass wasting in the Amazonian lowland of Peru (75°5'W, 10°13'S).
Terrestrial Inputs to Amazon Streams and Internal Biogeochemical Processing
properties of Whitewater rivers and contribute to the "Andean" characteristics of alluvial soils as far downstream as Manaus. Contrary to popular belief, mass wasting is not restricted to the high-relief Andean and subandean Cordilleras. Bank failure and collapsing gully headwalls provide episodic sediment inputs to streams even in lowland rain forests (Fig. 12.4). The remaining pathways linking terrestrial and lotic ecosystems correspond to hydrologic flowpaths. Groundwater flow, and its interaction with lithologic units, controls the flux and composition of baseflow in streams, but baseflow is also influenced by the relative proportion of wetlands in a landscape. It is conceivable, too, that direct litterfall alters baseflow solute and paniculate flux and chemistry. Rainfall activates direct precipitation, throughfall, overland, and lateral subsurface flowpaths. These, in turn, impart still different terrestrial signals on streamflow, thus transforming the baseflow signal into a stormflow signal. The stormflow signal is somewhat influenced by direct channel inputs, but
191
mainly controlled by overland flow and lateral subsurface flow. These two flowpaths convey terrestrial signals that reflect both forest canopy and floor, and different soil layers. However, they are activated neither in all ecosystems nor by all rainfall events. The independent variables controlling hydrologic inputs of terrestrial signals to streams are (1) soilscape (soil structure/ composition plus topography, Buol et al. 1989), (2) lithology, and (3) rainfall characteristics (amount, frequency, intensity). Soil structure and topography exert primary control over the hydrologic framework of an ecosystem and interact with rainfall characteristics to define the actual pathways of an ecosystem. In parallel, soil composition and lithology control the chemical composition of water moving along these pathways. Given the diversity of soils, landforms and precipitation patterns in the Amazon basin, we may expect equally diverse, and possibly quite contrasting, patterns of flowpaths and terrestrial signals (Fig. 12.5). In the following paragraphs, we examine these linkages by applying the above
Figure 12.5. Expected hillslope runoff pathways at La Cuenca, western Amazonia, and Reserva Ducke, central Amazonia. Arrow widths are proportional to flow along each pathway. Greater proportional amounts of lateral flow are expected in the Ultisols of La Cuenca in comparison with the Oxisols of Reserva Ducke. From Elsenbeer and Lack, 1996.
Michael E. McClain and Helmut Elsenbeer
192
framework to three catchments that represent two rather contrasting Amazonian soilscape systems. Two of them, Lago Calado and Reserva Ducke in central Amazonia, are in an Oxisol-dominated soilscape with concave slopes and riparian wetlands, whereas the third, La Cuenca in western Amazonia, is in an Ultisol-dominated soilscape with rectilinear slopes and a negligible wetland portion. We begin by examining the terrestrial/lotic linkage provided by groundwater. As explained above, the terrestrial signal in the flowpath groundwater reflects predominantly lithology, but also the topographic origin, that is whether it is hillslope or riparian wetland groundwater. The lithologic diversity of the Amazon basin is reflected in the contrasting groundwater chemistry of Lago Calado and La Cuenca, as shown in the columns "seep" of Table 12.2. The silica and cation content of groundwater from the Tertiary Barreiras formation (Lago Calado, Reserva Ducke) is substantially lower than that from the sandstones and siltstones of the western Amazonian Tertiary "red beds" (La Cuenca). This difference also
shows in the pH difference of the two sites, and it is even more striking when looking at groundwater from deep wells at Lago Calado. The dominant influence of lithology on groundwater is still obvious in the riparian wetlands, with the exception of potassium whose concentration ig practically identical at Reserva Ducke and La Cuenca, that is, independent of lithology. Regarding the control of streamwater chemical composition by groundwater flow during baseflow conditions, the respective comparisons are consistent for the western Amazonian site, but inconclusive for the central Amazonian sites (Table 12.3). Both at Lago Calado and Reserva Ducke, the silica content in baseflow is higher than that in groundwater. To account for this discrepancy, a source high in silica must be postulated. For Lago Calado, Williams et al. (1997) refer to such a source as "bank seepage," with a silica concentration of 82.6 pM. "Bank seepage" might be groundwater outflow in which case a rather variable groundwater silica concentration must be assumed. This additional source, however, does not explain the lower baseflow con-
Table 12.2. The chemical composition of groundwater in three Amazonian catchments. OiM) Solute
Seepage Waters* Lago** Calado
Silica Calcium Magnesium Potassium PH
62.0 12.1
6.5 3.4 4.9
La Cuenca*** 112 (106-115) 50 (47-55) 9.1 (8.2-9.9) 12.0 (11.5-12.8) 5.9 (5.8-6.0)
Wetland Groundwater* Reserva Ducke**** La Cuenca*** nd
13.7 3.7 30 nd
15 0.5 4.5 nd nd
90 (87-94) 30 (29-32) 11.5 (11.1-11.9) 29.4 (26.9-30.9) 4.5 (4.4-4.6)
* All units in uM, except pH. ** Lago Calado: from Williams et al., 1997, Table 12.1. We assume that their "spring seepage" reflects the same groundwater flowpath and a similar sampling situation as "seep" in La Cuenca (Elsenbeer et al., 1995). *** La Cuenca: from Elsenbeer et al., 1995. The median value is shown along with upper and lower 95% confidence limits in parentheses. **** Reserva Ducke: data from Nortcliff and Thornes (left column), 1978, and Brinkmann and Dos Santos (right column), 1970, 1971, 1973 (well IV1 only, a sampling situation comparable to "wetland" at La Cuenca).
Terrestrial Inputs to Amazon Streams and Internal Biogeochemical Processing
193
Table 12.3. The chemical composition of streamwater (baseflow) in three Amazonian catchments.*
Solute (/iM)
Lago Calado**
Reserva Ducke***
La Cuenca****
Silica
90.2
nd
2.7 1.2 1.5 4.7
0.75 0.7 0.6 4.6
nd 0.2 1.2
104.0 (96.5-109.8)
Calcium
Magnesium Potassium pH
1.5 nd
34.5 <0.5 <0.8 nd nd
39.5 (37.9-41.9) 10.7 (9.9-11.5) 20.2 (18.7-22.0) 5.8 (5.7-5.9)
All units in uM, except pH. Lago Calado: from Williams et al., 1997 (left column), Table 12.1, and Lesack, 1993a (right column), Table 12.1. Reserva Ducke: from Nortcliff and Thornes, 1978 (left column), and Brinkmann and Dos Santos (right column) 1970, 1971, 1973. La Cuenca: from Elsenbeer et al., 1995. The median value is shown along with upper and lower 95% confidence limits in parentheses.
centrations, because its values are similar to the ones given in Table 12.2. A similar complication arises for Reserva Ducke, whereas results from La Cuenca point to baseflow as a mixture of hillslope and wetland groundwater. The latter contribution is particularly striking in the case of potassium. Regardless of these inconsistencies and explanatory gaps, the terrestrial signal in baseflow reflects the contrasting lithology of the two ecosystems, as well as the lithologyindependent influence of wetlands. The lithologic control of baseflow chemistry extends also to dissolved organic carbon (DOC). In a one-year comparison of clearwater (Oxisol on Barreiras formation) and blackwater stream systems (Spodosol on sand deposits), McClain et al. (1997) reported consistent groundwater DOC contents of near 3000 pM in the blackwater system and 100 to 300 pM in the clearwater system (Fig 12.6). Baseflow in the two systems reflected this large difference in DOC content, which McClain et al. (1997) attributed to the different DOC sorption characteristics of Spodosols and Oxisols. Wetlands also are key factors in the DOC concentration and composition of streams
draining Oxisols of the Barreiras formation. In analogy to La Cuenca where potassium contents were higher in baseflow than in groundwater (Tables 12.2 and 12.3), McClain et al. (1997) reported higher baseflow DOC concentrations compared to groundwater and attributed this difference to seepage from riparian wetlands. In fact, baseflow was shown to be compositionally more similar to wetlands than to riparian groundwater. These findings are consistent with results from Guyot and Wasson (1994) who showed an increase in the DOC concentration of Andean rivers upon entering the lowland Llanos region with its abundant wetlands (Fig 12.7). They also agree trend-wise with the results from Lago Calado (Williams et al. 1997). The results from western (Elsenbeer et al. 1995) and central (McClain et al. 1997, Williams et al. 1997) Amazonia suggest that an index combining a clear "wetland signal," for example DOC and potassium, and a distinct hillslope groundwater signal, for example silica, might be a useful indicator of the relative proportions of wetlands and uplands in an Amazon catchment or drainage basin.
194
The relatively straightforward terrestrial/ lotic linkages characterizing baseflow conditions become more complicated during rainfall events, which activate additional, short-lived flowpaths imparting distinct terrestrial signals. This is reflected in stormflow solute dynamics manifest in concentrationdischarge relationships. In many, if not most, case studies, these relationships form the basis for inferring flowpaths. Table 12.4 compares the stormflow solute dynamics of two central (Lago Calado, Bacia Modelo) and one western (La Cuenca) Amazonian sites. The difference in stormflow chemistry between the two soilscape systems (Oxisol v. Ultisol) is even more pronounced than that of baseflow chemistry. Increases in calcium, magnesium, and sodium concentrations with increasing discharge at Lago Calado are in direct conflict with corresponding concentration-discharge relationships
Michael E. McClain and Helmut Elsenbeer
at La Cuenca. These unexpected differences raise the question, does rainfall activate different flowpaths at the two sites, or do similar flowpaths expose water to different materials? Overland flow and lateral subsurface flow (throughflow) are the flowpaths most influential in controlling stormflow chemical signals. At La Cuenca these pathways have been studied in detail. However, similar systematic studies are lacking for the central Amazonian sites. A certain amount of speculation will therefore be required. The concentration-discharge patterns presented in Table 12.4 require a Ca- and Mg-rich stormflow source at Lago Calado, but a Ca- and Mg-poor source at La Cuenca. Concurrently, these same sources must both generate positive concentration-discharge patterns for K. For the western Amazonian site, Elsenbeer et al. (1995) showed that
Figure 12.6. Time-series plot of DOC concentrations in different hydrologic compartments of the two sites. Results are plotted within the same 12-month period only for illustrative purposes. Campina water was actually monitored during the year following monitoring at Barro Branco. The annual occurrence of rainy and dry seasons is shown. Crosses mark sample dates. From McClain et al.(1997).
Terrestrial Inputs to Amazon Streams and Internal Biogeochemical Processing
195
Table 12.4. Concentration responses to increased discharge during storms at three Amazonian sites. Solute
Lago Calado
La Cuenca
Bacia Modelo
Silica Calcium Magnesium Sodium Potassium pH
No information Increase Increase Increase Increase No trend
Decrease Decrease Decrease Decrease Increase/flat Decrease
Decrease Increase Increase Decrease Increase Decrease
Sources: Lago Calado: from Lesack (1993a); La Cuenca: from Elsenbeer et al. (1996); Bacia Modelo: from VegasVillarubia et al. (1994); one storm only.
overland flow originating on hillslopes is primarily responsible for the high K and low Ca, Mg, and Si concentrations in stormflow. Indeed, stormflow chemical composition approached that of overland flow (Elsenbeer and Lack 1996). Overland flow "amplifies" the strong K signal in throughfall by its contact with the forest floor (Fig. 12.8). Its Ca, Mg, and Si signals, in contrast, reflect the mixing of throughfall and soil water, both sources with lower concentrations of these solutes than groundwater or baseflow.
At Lago Calado (Lesack 1993b) and Reserva Ducke (Franken 1979), both soil physical data and anecdotal evidence suggest that overland flow does not occur on hillslopes surrounding the stream. Instead, overland flow appears limited to near-stream wetlands (Nortcliff and Thornes 1981). Under this scenario, overland flow either simply transmits an already strong throughfall signal (Forti and MoreiraNordemann 1991, Lesack 1993a) or, in analogy to La Cuenca, amplifies it also
Figure 12.7. Plot of DOC concentration versus elevation for rivers of the Bolivian Amazon basin. From Guyot and Wasson, 1994.
196
Michael E. McClain and Helmut Elsenbeer
Table 12.5. The chemical composition of soil water in three Amazonian catchments.* Solute
Depth
Lago Calado
Silica
0.3 m 0.6 m 0.3m 0.6m 0.3m 0.6 m 0.3m 0.6 m 0.3m 0.6 m
nd
Calcium Magnesium Potassium pH
41.1 nd 1.4 nd 0.8 nd 1.1 nd 3.9
Reserva Ducke nd nd
11.6 20.4 15.2 17.9 25.7 41.4 nd nd
La Cuenca 52.8 (47.4-54.9) 46.2 (42.9-49-1) 11.0 (9.7-12.8) 11.0 (9.5-12.2) 3.7 (3.3-4.1) 3.1 (2.5^3.3) 7.4 (6.6-7.9) 6.1 (5.6-6.9) 4.3 (4.2-4.5) 4.3 (4.2-4.5)
Sources: Lago Calado: from Williams et al., 1997, Table 12.1. The sampling depth was at 0.5m, the sampling device was a zero-tension lysimeter. Reserva Ducke: from Nortcliff and Thornes, 1978, Table IV, average of upper site and lower site. Soil water was sampled at a tension of 600mb. La Cuenca: from Elsenbeer et al., 1995. Soil water was sampled at a tension of 700 mb. The median value is shown along with upper and lower 95% confidence limits in parentheses. * All units in uM, except pH.
with respect to magnesium and calcium. Amplification of the throughfall signal is supported by Williams et al. (1997), who reported Ca, Mg, and K concentrations in wetland seepage at Lago Calado which are almost identical to the "seep" values compiled in Table 12.3. According to Table 12.5, the observed concentration-discharge relationships for Lago Calado are not compatible with a discernible soil water signal. The results from Reserva Ducke, included for completeness sake only, are entirely at odds, not only with the lithologically and pedologically similar site at Lago Calado, but also with the western Amazonian site. As the preceding discussion demonstrates, our opening question about terrestrial/lotic linkages and factors controlling them can be adequately answered only in small catchments from Western and Central Amazonia. The spatial representativeness of these catchments is unknown, as is their position in the spectrum of potential terrestrial/lotic linkages in Amazonian soilscapes. In view
of the lithological, geomorphological, and pedological diversity of the Amazon basin, this spectrum could be quite wide. However, until a spatially more refined picture emerges the western and central Amazonian soilscape types and their terrestrial/lotic linkages discussed could be viewed as prototypes. At present, the following features stand out. 1. The relative proportions of wetlands and uplands is crucial in determining the groundwater-mediated terrestrial signal in streams, not only during baseflow conditions but also in response to precipitation events; the wetland signal appears to be independent of the regional geology. 2. Terrestrial/lotic linkages in the form of fast near-surface flowpaths, such as overland flow, require certain soil conditions together with rainfall patterns. Where these are met, a "hillslope" terrestrial signal, with some resemblance to a wetland signal, is transmitted to the stream channel.
Terrestrial Inputs to Amazon Streams and Internal Biogeochemical Processing
These features are neither Amazon-specific nor a peculiarity of the humid tropics. Future attempts to study hydrologic processes and solute dynamics and fluxes in the Amazon basin require a careful site selection to adequately reflect its geo-ecological diversity.
In-Channel Processing Question: How are terrestrial inputs processed within the lotic environment, and how does this processing support the trophic structure of lotic systems? Upon entering the stream environment, and during downstream transport, terrigenous solutes and particulates participate in a wide array of physicochemical and biological reactions. Over the length and time scales of flow within streams, these reactions may be irreversible and result in progressive modifications to stream biogeochemistry, such as the continual decomposition of incoming labile organic compounds and the corresponding accumulation of degraded and refractory organic compounds. Alternatively, reactions may be cyclic in nature, such that individual nutrient species repeatedly cycle through organic and inorganic phases in a spiraling effect (Newbold et al. 1995). We can be confident that the turnover time of the most refractory forms of organic matter in Amazonian streams are greatly in excess of their transport time out of the system (Hedges et al. 1986), but we do not yet know what proportion of organic matter input to streams is remineralized in the system. In terms of element spiraling lengths, no published studies have yet quantified this parameter in Amazonian streams. There has been some progress, however, in examining the structure and dynamics of detrital food chains, as well as certain other biotic and abiotic processes governing the biogeochemistry of Amazon streams.
197
Litter input to central Amazon streams accumulates in banks behind debris dams and at the bottom of pools of slow-moving water. These litter banks, in turn, support dynamic microbial and invertebrate communities, which process the accumulated litter through a well-evolved detrital food chain (Walker 1995). At the base of this chain are imperfecti, hyphmycete fungi, and bacteria, which colonize the new litter and excrete hydrolyzing enzymes to begin the breakdown sequence. Simultaneously, fungi take up nutrients from the stream water, thereby increasing the nutrient content of the colonized litter mass. Invertebrate scrapers
Figure 12.8. Plots of discharge and a) K+ and b) SiO2 concentrations over the course of a dry season storm event at La Cuenca, Peru,Types of water sampled include throughfall (T), overland flow (OF2), soil water (SW), fioodplain roundwater (FGW), seeps (seep), and stream water (stream). From Elsenbeer and Lack (1996).
198
and shredders feed on the nutrient-enriched litter mass, and as they are consumed by larger predators, the nutritional capital transferred from the forest is assimilated throughout the trophic assemblage of the stream. At each link in the aquatic food chain there are small losses of organic and inorganic compounds to the water column, but their efficient recycling precludes any significant downstream buildup of nutrient concentrations (Walker 1995). The physical breakdown of leaves and other organic debris does, however, produce finer and more transportable particulates which may be carried to downstream communities. Organic matter input to streams in dissolved form also contributes to the energy and nutritional needs of stream organisms, but these contributions have not yet been investigated. Upon noting that the detrital communities of blackwater streams contained higher densities of organisms than those of clearwater streams, Walker (1990) suggested that DOM may act as an additional carbon and nutrient source for fungi. However, DOM mineralization in streams is more often attributed to the action of benthic bacteria contained in biofilms coating streambed sediments and submerged woody debris (Meyer 1990). In lowland Amazon streams where waters are acidic and depleted in base cations, silica has an expanded biological component to its chemistry. In the lower sections of the Taruma-Mirim stream of the central Amazon, spiculae of the sponge, Drulia braunii, are the dominant form of particulate silica (Chauvel et al. 1996), and all along central Amazonian streams certain riparian plants incorporate elevated concentrations of silica into their leaves and wood (Vasconcelos et al. 1993). In shallow or clear waters beneath openings in the overhanging canopy, gelatinous tufts of the alga, Batrachospermum (Rhodophyceae), appear on submerged wood (Walker 1995).
Michael E. McClain and Helmut Elsenbeer
Konhauser et al. (1992) observed abundant benthic diatoms on the surface (Naviculd) and shallow interior (Synedra and Eunotia) of a wood sample from the margin of the Negro River. Navicula was also observed on submerged leaves, but in lesser abundance. The algae were thought to be extracting nutrients from the substrates as well as using them for support. Konhauser et al. (1992) suggested that diatoms were sufficiently abundant to influence bulk silicon concentrations in the Negro River. Furthermore, Konhauser and Fyfe (1993) reported on the bioconcentration of 17 trace metals in the tissues of benthic diatoms from the Negro River, Solimoes River, and the Carajas area. Concentration factors of greater than 106 were reported for Ti and Zr, while Hg was concentrated by a factor of less than 1000. The algae strip metals from solution in proportion to their dissolved concentrations. Thus, algae exhibited the highest trace metal concentrations in the Solimoes due to its relatively high solute load. Similar patterns are to be expected in algae occurring across the range of streams composing the basin. As another example of microbial controls on biogeochemistry in Amazonian riverine systems, Konhauser et al. (1993) reported on the formation of authigenic Fe/Al silicates in association with epiphytic and apisammic bacteria in the Solimoes River. All bacterial populations examined, regardless of their physiology, formed identical mineral phases, suggesting to the authors that the biomineralization was a surface process associated with the anionic nature of the cell walls. No mineral formation was observed in the solute-poor waters of the Negro River, possibly due to metal complexation with the relatively high concentrations of DOM. In the heterogeneous landscape of Amazonia, mixing of waters from disparate stream types may exert a strong control on
Terrestrial Inputs to Amazon Streams and Internal Biogeochemical Processing
downstream changes in stream biogeochemistry. A dramatic illustration of this control may be seen where blackwater streams mix with their white- or clearwater counterparts. Once again, investigations of these processes have not been centered in small streams, but rather in rivers of the region. However, similar processes may be expected to operate where streams of varying character mix. Leenheer and Santos (1980) presented evidence suggesting that downstream of the confluence of the Negro and Branco rivers in central Amazonia, suspended grains of kaolinite derived from the Branco complexed with dissolved humic substances derived from the Negro. The complexes then flocculated and were deposited. An essential feature of this process is a strong shift in pH of the river carrying the kaolinite. At the Negro-Branco confluence, kaolinite from the Branco experiences a sudden shift in pH from 6.6 to 4.9. The greater H+ concentration in the river water drives the surface charge of the kaolinite nearer to zero, thereby promoting complexation with hydrophobic dissolved humic compounds, which are likewise protonated and therefore depolarized. In a more recent study, however, Filoso (1996) found no evidence of flocculation in the Negro River downstream of its confluence with the Branco. Mixing processes are not confined to the confluence of different stream types but may also occur along a single stream reach. Chauvel et al. (1996) noted increased deposition of kaolinite-organic matter complexes with increasing distance downstream in the Taruma-Mirim stream. They attributed the deposition to a progressive increase in the mass of floes and a corresponding tendency for more permanent deposition. In addition to increased deposition downstream, Chauvel et al. (1996) documented a downstream increase in the organic content of sediment and a shift in the composition of
199
sediment organic matter. Both the ratio of dissolved fulvic acid to humic acid and the C/N ratio of sedimentary organic matter decreased downstream. Compositional shifts were attributed to the progressive partitioning of N-rich fulvic acids onto clay surfaces via a mechanism similar to that proposed by Hedges et al. (1986). Our examination of biological and physicochemical processing of terrigenous materials in Amazon streams thus reveals a sequence of events which profoundly alters certain components of the original biogeochemical signal input from the forest. Litter falling into streams appears to be the fundamental energy source for the majority of macroscopic organisms, which form diverse communities in banks of accumulated litter. Microscopic organisms such as algae and bacteria strongly influence, and possibly determine, levels of metals and Si in the stream water via uptake and surface catalyzed precipitation processes. Benthic bacteria certainly influence the speciation of redox sensitive elements like C, N, S, Fe, and Mn, but specific experiments are lacking. In summary, we have made only modest progress toward completely answering the research question posed in this section, and much work remains before the extent of terrigenous material processing in streams is known.
Role of Streams in Regional Cycles Question: What is the role of streams in regional to continental elemental cycles? Streams form a crucial link in larger scale biogeochemical cycles, both through down-stream transport of terrigenous materials and gas exchange between saturated soils of stream corridors and the atmosphere. Surface runoff represents the primary pathway by which organic and inorganic materials are transferred from continents to
200
the ocean (Martin et al. 1979), and streams are the first interval in this pathway. Previous sections have presented our current understanding of how terrigenous materials are input to and processed within Amazon stream corridors. Key to our regional perspective is the realization that streams do not function simply as conduits for downstream transfers. They are highly reactive environments, where the dissolved and paniculate load discharged downstream may be strikingly different from the original signal input at the terrestrial/lotic interface. Especially within the ion-depleted waters of the lowland Amazon basin, it is probably incorrect to assume that any bioactive element behaves conservatively in its biogeochemistry. Stream corridors may play a significant role in regional trace gas emissions to the atmosphere due to their large areal extent and relative abundance of anaerobic environments. Junk and Furch (1993) estimated that saturated or near-saturated soils of Amazon stream corridors cover a combined area of one million km2, or roughly three times the area covered by floodplains of the basin's major rivers. Where topographic depressions occur in the corridors, the water table may intersect the ground surface and produce shallow wetlands. Gas fluxes from these wetlands have not been examined in detail; however, a preliminary measurement of CH4 flux from a fringing wetland of the Barro Branco catchment yielded a value of 70 mg CH4 nr2 d"1 (McClain, unpublished data), which is similar to values reported by Devol et al. (1990) for the mainstem floodplain. Interestingly, measurements of CH4 fluxes from the surface of near-saturated streamside soils yielded values near zero, suggesting that complete saturation is necessary before fluxes to the atmosphere become significant. The abundance of small fringing wetlands has not been quantified across the Amazon, but if they occupy even a few
Michael E. McClain and Helmut Elsenbeer
percent of the estimated one million km2 of stream corridors, their emissions of CH4 could prove quite significant in basinwide calculations. Livingston et al. (1988) measured N2O fluxes from soils in the stream corridor of the Barro Branco catchment, and although their results indicated that corridor fluxes were equal to those of the adjoining hillslope and ridge, the flux of N2O per unit of N mineralization in corridor soils was twice that of adjoining upland soils. Conversely, a preliminary measure of the flux of N2O from the fringing wetland at Barro Branco yielded a value only half that measured from soils, which is to be expected given the solubility of N2O in water. Keller and Reiners (1994) reported that soil N2O fluxes in Costa Rica increased significantly with increasing soil moisture content. One would therefore expect greater fluxes from stream corridor soils which have shallow water tables and are frequently flooded during the rainy season. Elevated fluxes were not found by Livingston et al. (1988), but given that stream corridor soils may cover as much as 20% of the basin's area, it would be advisable to investigate their possible contributions in more detail. Streams, as a result of their position as the first interval in horizontal regional transfers and the large combined area of their floodplains, are certain to play a significant role in regional biogeochemical cycles. This role, however, is still entirely uninvestigated across the region.
The Effects of Anthropogenic Activities Question: What are the biogeochemical consequences of anthropogenic activities to streams in the Amazon (that is, how do these activities alter the natural functioning of stream systems), and what management practices best minimize adverse effects?
Terrestrial Inputs to Amazon Streams and Internal Biogeochemical Processing
201
We have until now focused our discussion fluxes of NO3, NH4, Na, K, Ca, Mg, Cl, SO4, on the functioning of undisturbed stream Al, Fe, Mn, and DOC in the throughflow of systems, but the most pressing research a cut-and-burn experimental plot relative to questions in the Amazon today relate to a forested control plot in terra firme rain widespread modification of natural systems forest 80 km SW of Manaus, Brazil. by human activities. The course of forest Overland flow, which occurred only in the conversion in Amazonia is highly complex deforested plot, carried elevated concentraand driven by land uses ranging from tions of the these same ions as well as PO4, extractivist activities like selective logging total dissolved P, total dissolved N, dissolved and harvesting of forest fruits, nuts, and inorganic carbon, and Si. Higher ion game to complete forest conversion to concentrations in runoff were relative to pasture, agricultural fields, and industrial those in rainfall (Williams et al. 1997). mining sites (Nepstad et al. 1997). Rivers Concentrations of the above list of ions and streams feel the effects of these changes (excluding Cl, SO4, Fe, Al, and Mn) on land, but water courses are also directly remained elevated in groundwater and impacted via human activities such as stream water relative to the forested placer mining, industrial fishing, reservoir catchment. In another study, Uhl and Jordan creation, and waste discharge. (1984) documented increases in of NO3, K, Between 1978 and 1996, it is estimated Ca, and Mg in shallow throughflow of a cut that 345,000 km2 of forest in the Brazilian and burned plot of terra firme rain forest Amazon basin was converted to pastures of near San Carlos, Venezuela. At the San African forage grasses, "degraded" pastures Carlos site, nutrient concentrations in of shrubs and nonforage herbs, and agricul- throughflow returned to preburn concentratural fields (INPE 1997). Each of these land tions after 2 years, while at the site near uses is preceded by complete deforestation Manaus, concentrations were approaching and burning of slash, which produce pre-burn values after only 1 year. profound changes in the biogeochemistry of An important finding of the study by the sites by altering the stocks and flows of Williams et al. (1997) is that excess nutrients energy, water, and elements (Kauffman et mobilized at the soil surface following burning do not move conservatively al. 1995, Fernandez et al. 1997). The net effects of land conversion gener- through the subsurface hydrologic system ally reduce the nutrient retention capacity and to the adjoining stream. The behavior of of altered systems and therefore increase inorganic N species illustrates well the array the availability of elements to be input to of processes impacting nutrient concentrastreams. Destruction of the vegetation also tions along flowpaths. Nitrate concentrareduces evapotranspiration and leads to tions in throughflow following burning greater fluxes of water over and through the were greatly in excess of concentrations in soil system. Leaching is therefore enhanced the adjoining stream. Increased N mineralat the very time •when retention mechanisms ization rates in decomposing litter and soil are impaired. The net effect is increased humus is the suggested source of high NO3 nutrient fluxes along surface and subsurface concentrations in throughflow (Williams et flowpaths. In the one study to have al. 1997). The authors suggest that increased investigated this phenomenon in various nitrification rates may be linked to system hydrological compartments of an disturbance, which results in an increase in Amazonian catchment (Braco do Mota), pH and accumulation of NH4 via continued Williams et al. (1997) documented increased mineralization in the absence of NH4 uptake
202
Michael E. McClain and Helmut Elsenbeer
by plant roots. Excess NO3 concentrations although the intact riparian buffer strip at are reduced, however, in underlying the site shielded the stream from direct groundwater. This reduction may result erosive inputs, the authors observed that at from dilution by low-NO3 waters from other other sites without buffer strips, overland sources, but more likely concentrations flow, and its accompanying sediment load, decrease due to NO3 - removal processes in ran directly into streams. This resulted in deeper soil horizons (Williams et al. 1997). increased stormflow sediment loads in the From groundwater to stream water, NO3 deforested locations (Williams and Melack concentrations decreased further, with the 1997). Increased stream sediment loads in most dramatic decrease occurring as response to Amazon deforestation and land groundwater moved from the deforested development have also been documented in hillslope into the still-forested riparian zone. other tropical systems (Wiersum 1984, Williams et al. (1997) attribute this decrease Rijsdijk and Bruijnzeel 1990). to a combination of denitrification in The coverage of urban area is a relatively anaerobic riparian groundwaters, NO3 limited but growing threat to streams in uptake by aggrading riparian vegetation, Amazonia. Streams flowing through the and dilution by low-NO3 groundwater orig- urban areas of Manaus, Brazil, have been inating from deeper flowpaths. Interestingly, severely impacted by destruction of streamthe same pattern of dramatic decreases in side habitats and uncontrolled discharge of NO3 concentration at the hillslope-riparian industrial and human waste into stream interface was observed by McClain et al. channels (Silva and Silva 1993). Solid waste (1994) in an undisturbed catchment north of has become so thick as to clog culverts and Manaus. Together these studies illustrate the block passage of water through the system. pronounced role that intact riparian zones Human wastes contribute harmful amounts play in buffering stream systems against of reactive organic matter, as evidenced upland NO3 influxes under natural and by the extreme biochemical oxygen demand disturbed conditions. measured in Igarape Quarenta. Upon incuDeforestation may also lead to increased bation, the already low levels erosion from catchments, depending upon of oxygen in the stream (2.1 mg/1) were the mode of forest cutting and subsequent reduced to zero within 5 hours (Silva use of the land (Bruijnzeel 1990). Removal and Silva 1993). Uncontrolled industrial of the forest and litter layer and the discharges into streams are manifest in high disintegration of soil root structures initially levels of heavy metals in the flesh of fish. destabilizes hillslope soils. With time, Again in Igarape Quarenta, Silva (1992) vacated root casts collapse, humus decom- reported levels of Zn, Cr, Cd, and Cu in the poses, and the soil becomes more compact- muscle and liver of Tamoata (Hoplostemum ed and less permeable. Overland flow littorale) which were in excess of regulatory events may then become more frequent, limits. In response the extensive pollution of especially given the greater intensity of the system, the species diversity of Igarape rainfall at the soil surface once the forest Quarenta is only 27% (12 v. 44) that of canopy is removed. This process of Igarape Candiru, which lies just outside the compaction is greatly exacerbated by the urban area of Manaus. In fact, the two introduction of heavy equipment or cattle to systems share only four fish species in the site (Bruijnzeel 1990). At Braco do Mota, common, suggesting a shift in species Williams et al. (1997) documented overland composition as well as a decrease in species flow only in the deforested plot, and numbers (Silva 1992).
Terrestrial Inputs to Amazon Streams and Internal Biogeochemical Processing
Anthropogenic activities thus impact Amazon streams to varying degrees depending on the specific type of activity. Responses often include increased surface and subsurface fluxes of solutes and sediments from uplands following slash and burn activities. But these increased fluxes were shown to only slightly impact streams when a riparian buffer strip was left along the stream margin. Placer mining results in greater suspended sediment loads in stream and rivers, and perhaps more importantly contamination of the system by mercury. And finally, unchecked urban expansion was shown to have a devastating effect on stream biogeochemical and ecological conditions. Anthropogenic threats to the health and integrity of Amazon streams are the most immediate issue confronting stream researchers today. Thus, a call for additional research forms the basis of the final section of this chapter.
Research Needs Across the Amazon region, many fascinating discoveries remain to be made 'concerning the ecology of tropical stream systems, and considerable research will be necessary before we gain a truly comprehensive understanding of natural systems. However, given the momentum of current anthropogenic activities sweeping the region, we suggest that the scientific community focus a substantial proportion of its limited time and resources on pressing problems related to development. Large-scale colonial, resource, and industrial development characterize current trends in Amazonia. With regards to stream ecosystems, we have barely begun to investigate the impacts of these activities on biological communities and flows of nutrients through the system. Useful information may be gained from northern temperate studies, but the extreme degree of alteration in these regions have led scientists there to focus
203
more on restoration of degraded systems. In Amazonia, the main issue is protection of as yet undisturbed systems. Stream biogeochemical research in Amazonia should address questions that better enable humans to minimize adverse impacts to streams along which they live. We know that certain basic biogeochemical conditions must be maintained for stream communities to thrive. With this knowledge in mind, we recommend that applied stream research be concentrated in the following two priority areas. 1. Determine, in collaboration with the biological and hydrological scientific communities, tolerances of stream ecosystems to perturbation. This includes tolerance to physical perturbations, such as increased storm flow volumes, which may destroy habitat architecture, and tolerances, or thresholds, of stream biota to higher amplitude fluctuations in temperature, oxygen, and nutrient levels. 2. Investigate measures to maintain stream conditions within the tolerance ranges of native stream biota as development of the surrounding region progresses. An example of such a measure is the maintenance or establishment of forested buffer zones along stream margins. Quantitative needs with regard to buffer zones include correlations between their physical and biological characteristics (dimensions, geomorphology, soils, hydrogeology, vegetation type, etc.) and their capacity to mitigate adverse impacts (e.g. trap sediment eroding from uplands and/or reduce high nutrient concentrations in inflowing groundwater). These priorities represent a call for immediate action in response to a management crisis, but more basic stream research in Amazonia also suffers from a lack of common direction or any governing principle. We offer the following stepwise plan as one example of a research protocol
204
that could be implemented across the basin. Field research within this plan is to be carried out at the scale of small catchments, which is the appropriate scale at which to generate information useful in the management of streams. But the development of a more regional understanding of stream processes must rely on quantitative models. 1. Identify and delineate unique soilvegetation, or hydro-biogeochemical, response units across the basin. 2. Investigate, through monitoring and experimental manipulations, key biogeochemical processes and their signals in streams draining these response units. 3. Develop quantitative and predictive models of these biogeochemical processes and of related hydrochemical signals in streams. Step 1 of this plan recognizes that, given the geoecological diversity of the Amazon basin, the existing knowledge of biogeochemical processes in small catchments is far from adequate. We suggest the identification of key environments based on soil and vegetation types, because existing research has indicated that identifiable soil-vegetation units have unique biogeochemical processes and hydrological pathways, and hence unique hydrochemical fingerprints in the streams that drain them. We expect that the same type of disturbance will affect different soil-vegetation units in different ways. Hence, any serious attempt at formulating a basin-wide understanding of stream biogeochemistry must be preceded by small-catchment studies in key environments yet to be identified. Due to the scarcity of soil information, lithological and geomorphologic information is expected to serve as a substitute. In the absence of potentially appropriate radar systems (Israelsson et al. 1994), the best possible use
Michael E. McClain and Helmut Elsenbeer
must be made of optical systems to classify vegetation, as demonstrated by Foody and Hill (1996). With respect to biogeochemical process studies (see below), information about the canopy chemistry of thus classified forests is desirable, with an emphasis on, but not limited to, nitrogen and lignin. Promising approaches developed elsewhere (for example Zagolski et al. 1996) should be given high priority in Amazonia. The extent to which hydrobiogeochemical response units identified in such ways are also distinct with respect to hydrochemical signals must be verified by an appropriate sampling approach. Having identified and delineated key environments, Step 2 should begin by addressing the issue of which biogeochemical processes should be investigated, and which hydrochemical signals, including biomarkers, should be employed to link with these processes. In the widest sense, this issue encompasses ecosystem element fluxes, with specific signals reflecting "losses" of particular elements. For a basic characterization of selected undisturbed key environments in terms of enrichment or depletion of elements, an approach similar to Konhauser et al. (1993) is desirable. This involves a simple comparison of atmospheric inputs (wet and dry) and streamwater outputs. From such an approach, one may construct enrichment/depletion histograms for elements lacking a significant gaseous phase. These "black-box" representations of biogeochemical processes in a given key environment permit (a) an assessment of disturbances, and (b) an assessment of the influence of scale, and of in-channel processes. We expect both disturbances and catchment size to modify the shape of enrichment/depletion histograms. The appropriate experimental approaches to be implemented in Step 2 are paired and nested catchment studies. Process studies are not only a complement to input-output
Terrestrial Inputs to Amazon Streams and Internal Biogeochemical Processing
studies, but a prerequisite for the assessment of the effect of disturbances. Two specific research issues should be considered. First is the elucidation of the "geo-bio" connection through an interdisciplinary effort (Moldan and Cerny 1994), with a new emphasis on the link between soil biodiversity and biogeochemical cycling (Beare et al. 1995). To provide an element-specific example, nitrogen-related research should consider the storage, availability, and mobilization of nitrate in subsoils with appreciable anion exchange capacity (potentially in Oxisol soilscapes), and the role of dissolved organic nitrogen, in relation to leaf litter chemistry, in ecosystem N fluxes (Northup et al. 1995). Second is the propagation and modification of hillslope signals of biogeochemical processes towards streams: the terrestrial/lotic linkage. Based on the experience of Lago Calado and La Cuenca, a coordinated, well-defined sampling and monitoring approach of flowpaths should be implemented in the research catchments of the selected response units. The emphasis is on coordination to avoid the pitfalls of the past which have entailed a virtual incomparability of Amazonian hydrochemistry results (Elsenbeer and Lack 1996). To ensure the correct identification of biogeochemical process signals, however, the traditional suite of tracers must be expanded considerably, to include, at the least, low molecular weight organics. Alternately, sample volumes have to be sufficiently large to permit concentration of DOC and its subsequent analysis with regards to molecular mass and functional groups. The potential of particulate organic and inorganic matter as signals to pinpoint processes must be further explored. For example, sterol biomarkers in suspended POC might be useful to document the extent of in-channel processing of hillslope signals (Li et al. 1995), whereas the strontium isotopic composition as a function of parti-
205
cle size of suspended matter might be a potential signal for land use changes involving erosion (Douglas et al. 1995). Development of a thorough and useful understanding of Amazon stream biogeochemistry, of course, cannot rely exclusively on monitoring or experimental smallcatchment studies. Hence, Step 3 of our general research plan emphasized the need to develop quantitative models of biogeochemical in streams. Clearly, Steps 2 and 3 of this plan should be implemented simultaneously, as each of these research activities complements the other. In contrast to the plethora of model applications in highlatitude regions in the wake of surface water acidification, similar efforts related to land use change in Amazonia have not been undertaken, with one exception (Forti et al. 1995). A wide choice of models is available, for a variety of time steps and scales (see Ball and Trudgill 1995, for an overview), and a coordinated effort at establishing research catchments in key environments must include a consensus as to which models to employ to ensure a coordinated data collection. A comparatively simple starting point, to link stream hydrochemical signals to flowpaths and/or source areas, would be linking a mixing model to a hydrologic framework, (e.g. TOPOG or TOPMODEL). The evolution of the signal itself, i.e. its link to biogeochemical processes, is amenable to modeling with SAFE or PROFILE (Sverdrup et al. 1995). Nearly 50 years of active research into the biogeochemistry of Amazon streams have yielded considerable insight into the functioning of these diverse and biologically rich ecosystems. We have come to appreciate the crucial roles of streams as habitat, bioreactors, and conduits for large-scale material transfers across the basin and to the ocean. However, our process-based understanding of Amazon streams remains painfully limited, and as a
206
community we remain relatively unprepared to face the management challenges that will dominate the initial decades of the twenty-first
Michael E. McClain and Helmut Elsenbeer
century. Now is the time for concerted international research efforts to fill the most compromising of these gaps in our understanding.
Literature Cited Adis, J., K. Furch, and U. Irmler. 1979. "Litter production of a Central-Amazonian blackwater inundation forest." Tropical Ecology 20: 236-245. Ball, J., and S. Trudgill. 1995. "Overview of solute modelling." In: Solute Modelling in Catchment Systems, ed. S. T. Trudgill 0ohn Wiley & Sons, Chichester), pp.356. Beare, M. H., D. C. Coleman, D. A. Crossley Jr., P. F. Hendrix, and E. P. Odum. 1995. "A hierarchical approach to evaluating the significance of soil biodiversity to biogeochemical cycling." In: The Significance and Regulation of Soil Biodiversity, eds. H. P.Collins, G. P. Robertson, and M. J. Klug (Kluwer, Netherlands), pp. 5-22. Brinkmann, W. L. F., and A. dos Santos. 1970. "Natural waters in Amazonia: III. Ammonium molybdate-reactive silica." Amazoniana II: 443-448. Brinkmann, W. L. F., and A. dos Santos. 1973. "Natural waters in Amazonia: VI soluble calcium properties." Acta Amazonica 3: 33-40. Brinkmann, W. L. F., and A. dos Santos. 1971. "Natural waters of Amazonia: V. Soluble magnesium properties." Turrialba 21: 459-465. Bruijnzeel, L. A. 1990. Hydrology of Moist Tropical Forests and Effects of Conversion: A State of Knowledge Review. UNESCO, Paris, p. 224. Buol, S. W., F. D. Hole, and R. J. McCracken. 1989. Soil Genesis and Classification. Iowa State University Press, Ames, USA. Chauvel, A., I. Walker, and Y. Lucas. 1996. "Sedimentation and pedogenesis in a central Amazonian blackwater basin." Biogeochemistry 33: 77—95. Cummins, K. W., C. E. Gushing, and G. W. Minshall. 1995. "Introduction: An overview of stream ecosystems." In: River and Stream Ecosystems, eds. C. E. Gushing, K. W. Cummins, and G. W. Minshall (Elsevier, New York), pp. 1-8. Devol, A. H., J. E. Richey, B. R. Forsberg, and L. A. Martinelli. 1990. "Seasonal dynamics of methane emissions from the Amazon River floodplain to the troposphere." Journal of Geophysical Research 95: 16417-16426. Douglas, G. B., C. M. Gray, B. T. Hart, and R. Beckett. 1995. "A strontium isotopic investigation of the origin of suspended paniculate matter (SPM) in the MurrayDarling River system, Australia." Geochim. Cosmochim. Acta 59: 3799-3815. Elsenbeer, H., and A. Lack. 1996. "Hydrological pathways and water chemistry in Amazonian rain forests." In: Advances in Hillslope Processes, eds. M. G. Anderson, and S. M. Brooks (John Wiley & Sons, New York), pp. 939-959. Elsenbeer, H., A. Lack, and K. Cassel. 1995. "Chemical fingerprints of hydrological compartments and flow
paths at La Cuenca, western Amazonia." Water Resources Research 31: 3051-3058. Elsenbeer, H., A. Lack, and K. Cassel. 1996. "The stormflow chemistry at La Cuenca, western Amazonia." Interciencia 21: 133-139. Fernandes, E. C. M., Y. Biot, C. Castilla, A. C. Canto, J. C. Matos, S. Garcia, R. Perin, and E. Wanderli. 1997. "The impact of selective logging and forest conversion for subsistence agriculture and pastures on terrestrial nutrient dynamics in the Amazon." Ciencia e Cultura 49: 34-47. Fittkau, E. J. 1967. "On the ecology of Amazonian rainforest streams." Atas do Simposio sobre a Biota Amazonica 3: 97-108. Fittkau, E. J. 1971. "Okologische Gliederung des AmazonasGebietes auf geochemischer Grundlage." Miinster. Forsch. Geol. Palaorit. 20/21: 35-50. Foody, G. M., and R. A. Hill. 1996. "Classification of tropical forest classes from Landsat TM data." International Journal of Remote Sensing, 17: 2353-2367. Forti, M. C., and L. M. Moreira-Nordemann. 1991. "Rainwater and throughfall chemistry in a "terra firme" rain forest: central Amazonia." J. Geophys. Res. 96: 7415-7421. Forti, M. C., and C. Neal, 1992. "Hydrochemical cycles in tropical rainforests: An overview with emphasis on central Amazonia." Journal of Hydrology. 134: 103-115. Forti, M. C., C. Neal, and A. Jenkins. 1995. "Modeling perspective of the deforestation impact in stream water quality of small preserved forested areas in the Amazonian rainforest." Water, Air, and Soil Pollution 79: 325-337. Franken, M., U. Irmler, and H. Klinge. 1979. "Litterfall in inundation, riverine, and terra firme forests of Central Amazonia." Tropical Ecology 20: 225-235. Franken, W. 1979. "Untersuchungen im Einzugsgebiet des zentralamazonischen Urwaldbaches "Barro Branco" auf der "terra firme," 1. Abflussverhalten des Baches." Amazoniana IV: 459^66. Guyot, J. L., and J. G. Wasson. 1994. "Regional pattern of riverine dissolved organic carbon in the Bolivian Amazonian drainage basin." Limnology and Oceanography 39: 452^58. Hedges, J. I., J. R. Ertel, P. D. Quay, P. M. Grootes, J. E. Richey, A. H. Devol, G. W. Farwell, F. W. Schmidt, and E. Salati. 1986. "Organic carbon-14 in the Amazon River system." Science 231: 1129-1131. INPE (Instituto Nacional de Pesquisas Espaciais). 1997. Deforestation in Brazilian Amazonia, Sao Jose dos Campos. Irmler, U. 1982. "Litterfall and nitrogen turnover in an Amazonian blackwater inundation forest." Plant and Soil 67: 355-358. Israelsson, H., J. Askne, and R. Sylvander. 1994. "Potential
Terrestrial Inputs to Amazon Streams and Internal Biogeochemical Processing of SAR for forest bole volume estimation." International Journal of Remote Sensing 15: 2809-2826. Junk, W. J., and K. Furch. 1993. "A general review of South American floodplains." Wetlands Ecology and Management 2: 231-238. Kauffman, J. B., D. L. Cummings, D. E. Ward, and R. Babbitt. 1995. "Fire in the Brazilian Amazon: 1. Biomass, nutrient pools and losses in slashed primary forests." Oecologia 104: 397-408. Keller, M., and W. A. Reiners. 1994. "Soil-atmosphere exchange of nitrous oxide, nitric oxide, and methane under secondary succession of pasture to forest in the Atlantic lowlands of Costa Rica." Global Biogeochemical Cycles 8: 399-409. Konhauser, K. O., and W. S. Fyfe. 1993- "Biogeochemical cycling of metals in freshwater algae from Manaus and Carajas, Brazil." Energy Sources 15: 595-608. Konhauser, K. O., W. S. Fyfe, F. G. Ferris, and T. J. Beveridge. 1993. "Metal sorption and mineral precipitation by bacteria in two Amazonian river systems: Rio Solimoes and Rio Negro, Brazil." Geology 21: 1103-1106. Konhauser, K. O., H. Mann, and W. S. Fyfe. 1992. "Prolific organic SiO2 precipitation in a solute-deficient river: Rio Negro, Brazil." Geology 20: 227-230. Leenheer, J. A., and U. M. Santos. 1980. "Consideracoes sobres os processes de sedimentacao na agua preta acida do rio Negro (Amazonia Central)." Acta Amazonica 10: 343-355. Lesack, L. F. W. 1993a. "Export of nutrients and major ionic solutes from a rain forest catchment in the Central Amazon basin." Water Resources Research 29: 743-758. Lesack, L. F. W. 1993b. "Water balance and hydrologic characteristics of a rain forest catchment in the Central Amazon Basin." Water Resources Research 29: 759-773. Li, W., J. Dagaut, and A. Saliot. 1995. "The application of sterol biomarkers to the study of sources of paniculate organic matter in the Solo River system and the Seraya River, Java, Indonesia." Biogeochemistry 31: 139—154. Livingston, G. P., P. M. Vitousek, and P. A. Matson. 1988. "Nitrous oxide flux and nitrogen transformations across a landscape gradient in Amazonia." Journal of Geophysical Research 93: 1593-1599. Luizao, F. J. 1989. "Litter production and mineral element input to the forest floor in a central amazonian forest." Geojournal 19.4: 407-417. Martin, J. M., J. D. Burton, and D. Eisma (eds.). 1979. River Inputs to Ocean Systems. UNEP & UNESCO. McClain, M. E., and J. E. Richey. 1996. "Regional-scale linkages of terrestrial and lotic ecosystems in the Amazon basin: a conceptual model for organic matter." Archiv fur Hydrobiologie Suppl. 113: 111-125. McClain, M. E., J. E. Richey, J. A. Brandes, and T. P. Pimentel. 1997. "Dissolved organic matter and terrestrial-lotic linkages in the central Amazon basin of Brazil." Global Biogeochemical Cycles 11: 295-311. McClain, M. E., J. E. Richey, and T. P. Pimentel. 1994. "Groundwater nitrogen dynamics at the terrestriallotic interface of a small catchment in the Central Amazon basin." Biogeochemistry 27: 113-127. McClain, M. E., J. E. Richey, and R. L. Victoria. 1996.
207
"Andean contributions to the biogeochemistry of the Amazon River system." Bulletin de 1'Institut Francais des Etudes Andines 24: 425-437. Meyer, J. L. 1990. "Production and utilization of dissolved organic carbon in riverine ecosystems." In: Organic Acids in Aquatic Ecosystems, eds. E. M. Perdue, and E. T. Gjessing (John Wiley & Sons, Ltd., Dahlem Konferenzen), pp. 281-299. Moldan, B, and J. Cerny (ed.). 1994. Biogeochemistry of Small Catchments. A Tool for Environmental Research. SCOPE 51, John Wiley & Sons, Chichester. Nepstad, D. C, C. A. Klink, C. Uhl, I. C. Vieira, P. Lefebvre, M. Pedlowski, E. Matricardi, G. Negreiros, I. F. Brown, E. Amaral, A. Homma, and R. Walker. 1997. "Land-use in Amazonia and the Cerrado." Ciencia e Cultura 49: 73-86. Newbold, J. D., B. W. Sweeny, J. K. Jackson, and L. A. Kaplan. 1995. "Concentrations and export of solutes from six mountain streams in northwestern Costa Rica." Journal of the North American Benthological Society 14: 21-37. Nortcliff, S., and J. B. Thornes. 1978. "Water and cation movement in a tropical rainforest environment." Acta Amazonica 8: 245-258. Nortcliff, S., andj. B. Thornes. 1981. "Seasonal variations in the hydrology of a small forested catchment near Manaus, Amazonas, and the implications for its management." In: Tropical Agricultural Hydrology, eds. R. Lai, and E. Russell (John Wiley & Sons, Chichester), pp. 37-57. Nortcliff, S., and J. B. Thornes. 1989. "Variations insoil nutrients in relation to soil moisture status in a tropical forested ecosystem." In: Mineral Nutrients in Tropical Forest and Savanna Ecosystems, ed. J. Proctor (Blackwell Scientific Publications, Oxford), pp. 43-54. Nortcliff, S., J. B. Thornes, and M. J. Waylen. 1979- "Tropical forest systems: a hydrological approach." Amazoniana 4: 557-568. Northup, R. R., Z. Yu, R. A. Dahlgren, and K. A. Vogt. 1995. "Polyphenol control of nitrogen release from pine litter." Nature 377: 227-229. Pringle, C. M., and a. s. others. 1988. "Patch dynamics in lotic ecosystems: The stream as a mosaic." Journal of the North American Benthological Society 7: 503-524. Reindl, J. 1903. Die schwarzen Fliisse Stidamerikas. Theodor Ackermann, Miinchen. Rijsdijk, A., and Bruijnzeel, L. A. 1990. Erosion, sediment yield and land-use patterns in the Konto watershed, East Java, Indonesia. Konto River Project Communication no. 18, 2 Volumes. Konto River Project, Malang, Indonesia. Silva, C. P. D. 1992. Influencia das modificagoes ambientais sobre a comunidade de peixes de um igarape da cidade de Manaus (Amazonas). M.Sc. Thesis, Institute Nacional de Pesquisas da Amazonia, Manaus. Silva, E. N. S., and C. P. D. Silva. 1993. "A expansao de Manaus como exemplo do processo de extincao dos igarapes." In: Bases Cientificas para Estrategias de Preservacao e Desenvolvimento da Amazonia Volume 2, eds. E. J. G. Ferreira, G. M. Santos, E. L. Leao, and L. A. Oliveira (Institute Nacional de Pesquisas da Amazonia, Manaus), pp. 25-42. Sioli, H. 1954. "Gewasserchemie und Vorgange in den Boden im Amazonasgebiet." Naturwiss 41: 456-457.
208
Sioli, H. 1955. "Beitrage zur regionalen Limnologie des Amazonasgebietes. III. Uber einige Gewasser des oberen Rio-Negro-Gebietes." Archiv fur Hydrobiologie 50: 1-32. Sioli, H. 1964. General features of the limnology of Amazonia. Verh. Internat. Verein. Limnol. 15: 1053-1058. Sioli, H. 1984. The Amazon: Limnology and Landscape Ecology of a Mighty Tropical River and its Basin. Dr. W. Junk Publishers. Boston. Stallard, R. R, and J. M. Edmond. 1987. "Geochemistry of the Amazon: 3. Weathering chemistry and limits to dissolved inputs." Journal of Geophysical Research 92: 8293-8302. Stark, N. 1971a. "Nutrient cycling II: nutrient distribution in Amazonian vegetation." Tropical Ecology 12: 177-201. Stark, N. 1971b. "Nutrient cycling: I. nutrient distribution in some Amazonian soils." Tropical Ecology 12: 24-50. Sverdrup, H., M. Alveteg, S. Langan, and T. Paces. 1995. "Biogeochemical Modelling of small catchments." In: Solute Modelling in Catchment Systems, ed. S. T. Trudgill (John Wiley & Sons, Chichester), pp. 75-100. Uhl, C., and C. F. Jordan. 1984. "Succession and nutrient dynamics following forest cutting and burning in Amazonia." Ecology 65: 1476-1490. Vannote, R. L, G. W. Minshall, K. W. Cummins, J. R. Sedell, and C. E. Gushing. 1980. "The river continuum concept." Canadian Journal of Fisheries and Aquatic Sciences 37: 130-137. Vasconcelos, F. J., A. Castro de Silva, and J. A. Freitas. 1993. "Deposicao da silica e cristais no xilema de especies tropicais da Familia Caesalpiniaceae." Arvore 17: 369-374. Vegas-Villarubia, T, M. Maass, V. Rull, V. Elias, A. R. Coelho Ovalle, D. Lopez, G. Schneider, P. J. Depetris, and I.
Michael E. McClain and Helmut Elsenbeer Douglas. 1994. "Small catchment studies in tropical zones." In: Biogeochemistry of Small Catchment, eds. B. Moldan, and J. Cerny (Wiley, Chichester), pp. 343-360. Veneklaas, E. J. 1991. "Litterfall and nutrient fluxes in two montane tropical rain forests, Colombia." Journal of Tropical Ecology 7: 319-336. Walker, I. 1990. "Ecologia e biologia dos igapos e igarapes." Ciencia Hoje 11.64: 44-53. Walker, I. 1995. "Amazonian streams and small rivers." In: Limnology in Brazil, eds. J. G. Tundisi, C. E. M. Bicudo, and T. M. Tundisi (ABC/SBL, Rio de Janeiro), pp. 167-193. Went, F. W., and N. Stark. 1968. "Mycorrhiza." BioScience 18: 1035-1039. Wiersum, K. F. 1984. "Surface erosion under various tropical agroforestry systems." In: Proceedings of a Symposium on Effects of Forest Land Use on Erosion and Slope Stability, eds. C. L. O'Loughlin, and A. J. Pearce (IUFRO, Vienna, and East-West Centre, Honolulu, Hawaii), pp. 231-239Williams, M. R., andj. M. Melack. 1997. "Solute export from forested and partially deforested catchments in the central Amazon." Biogeochemistry 38: 67-102. Williams, M. R., T. R. Fisher, andj. M. Melack. 1997. "Solute dynamics in soil water and groundwater in a central Amazon catchment undergoing deforestation." Biogeochemistry 38: 303-335. Zagolski, F., V. Pinel, J. Romier, D. Alcayde, J. Fontanari, J. P. Gastellu-Etchegorry, G. Giordano, G. Marty, E. Mougin, and R. Joffre. 1996. "Forest canopy chemistry with high spectral resolution remote sensing." International Journal of Remote Sensing 17: 1107-1128.
13 Geoecological Controls on Elemental Fluxes in Communities of Higher Plants in Amazonian Floodplains Maria Teresa F. Redade, Martin Worbes, Wolfgang J. Junk
Navigators visiting the Amazon during the fifteenth century provided the earliest descriptions of aquatic systems in the region, but it was not until mid twentieth century that systematic studies of the limnology of Amazon waters began (Sioli 1984). The inclusion of vegetation as an important part of the aquatic biota was only possible after the relatively recent change from traditional potamic limnology to wetlands limnology (Sioli 1975). The first studies of the vegetation of Amazon wetlands consisted mainly of species descriptions, and it is only recently that studies of floodplain vegetation have attained a level of importance equivalent to that of studies dealing with water chemistry, phytoplankton, zooplankton, and fishes. For the past thirty years, fertile floodplain systems along Whitewater rivers (varzea) have been focal areas of colonization. This fertility also supports high rates of primary production within the higher plants, especially of the herbaceous vegetation (Piedade et al. 1991, Junk and Piedade 1993a, 1997). Quickly turning-over pools of nutrients 0unk and Furch 1991, Furch and Junk 1992, 1997a) and direct connections with contiguous terra firme forest and river channels (Alves 1993, Furch and Junk 1997b) are also characteristic of these floodplain systems. As a consequence of the annual floodpulse (Junk et al. 1989), floodplain vegetation is subjected to aquatic and terrestrial phases,
which hold important ecological implications for both the plant populations and related aquatic and terrestrial biota. Life cycles of the species and the time available for growing depend upon the duration of inundation and drought periods and the habit of the species. During the year, pulses of growth and dormancy occur and herbaceous vegetation changes its species composition according to the phase of the hydrological cycle. In this chapter we discuss the distribution and the development of plant communities in floodplain areas, mainly of the big Whitewater rivers, focusing on factors such as diversity, species composition, biomass and primary production. Based upon these factors, we also discuss the annual dynamics of bioelements stocks and their turnover through herbaceous and floodplain forest communities. Finally, we examine the implications of such nutrient dynamics and turnover for the aquatic biota. We do not address carbon and nutrient budgets, as these are thoroughly discussed in chapter 14.
The Floodpulse and the Biota The Amazon River and its big tributaries are characterized by a monomodal flood pattern with an average amplitude of about 10 m at the Manaus harbor on the Rio Negro.
Maria Teresa F. Piedade et al.
210
Considering only three frequencies of flood regimes, the estimated extent of floodplain areas can reach 1,350,000 km 2 (Table 13.1), or about 27% of the total area of the Amazon basin. Although our discussion will concentrate mainly in the big river floodplains, all are of great importance to the aquatic biota. Some floodplain areas may be disconnected from the main river during low water levels, but they are usually connected during extremely high floods. Therefore, we can consider flooded habitats as semiopen systems where the exchange of energy and dissolved and paniculate material occurs in periods of days, months, years, decades, or centuries according to the level of connectivity with the main river and the nature of the material being exchanged (Junk 1997a). Floodplains can be defined as wetlands oscillating between aquatic and terrestrial phases and designated as the Aquatic/ Terrestrial Transition Zone (ATTZ, Junk et al. 1989). Being flooded recurrently, they include permanently lotic and lentic habitats, which make them suitable for organisms ranging from completely aquatic to completely terrestrial species. Both plants and animals make maximal use of available
nutrients during the hydrological phase, which is most favorable for their growth, and somewhat less use during the unfavorable phase. The exchange of energy between the two phases of the hydrological cycle by different groups of organisms accounts for the year-round high productivity of most floodplain systems (Junk 1997b). A difficulty in studying floodplains is distinguishing between terrestrial and aquatic organisms and processes. Processes follow the land-water continuum and many organisms show adaptations for both the aquatic and terrestrial phases. Although some organisms are strictly aquatic or terrestrial, a large group can cope with both phases by simply reducing activities during the less favorable phase. In the case of vegetation, one may encounter difficulties in calculating elemental budgets for specific periods of the hydrological cycle. In order to calculate accurate stocks and fluxes of bioelements, it is necessary to distinguish different plant communities, their distributions, species compositions, biomass, and primary productivity. A very conspicuous characteristic of floodplain vegetation is the variation of species in
Table 13.1 Frequency of flood regimes and the estimated extend of the related floodplains with prevailing associated vegetation. Floodplain Area I. Polymodal unpredictable hydrological regime Along low-order streams within Amazonian lowlands II. Monomodal hydrological regime Along the Amazon River and its big tributaries in Brazil III. Coastal region, tidal regime Amazon River estuary TOTAL AREA Source: Data adopted from Junk 1997.
Extent (km2 103)
Prevailing Vegetation
Ca. 1000
Forest
300
50
1350
Forest, grassland
Mangrove, tidal forest and grasslands
Geoecological Controls on Elemental Fluxes in Communities of Higher Plants in Amazonian Floodplains
relation to the annual duration of inundation, which corresponds in turn to the elevation gradient of each portion of the floodplain. In the Amazon varzea near Manaus, areas greater than 22 meters above the sea level (m.a.s.l) are colonized by a well-established floodplain forest that tolerates less than 230 days of average annual flooding. Less-adapted tree species usually grow on more elevated parts of the plains and, in some cases, also in the neighboring upland forest (Klinge et al. 1990). Flood-tolerant shrubs inundated up to 270 days occur from 20.5 m.a.s.l. upward. Between 20.5 and 19-5 m.a.s.l. herbs and perennial grasses are dominant, while below 19.5 m.a.s.l. annual terrestrial plants colonize land exposed for only a few weeks per year. Even sand bars and mud flats below 16 m.a.s.l., which are exposed for only a few weeks every couple of years, are immediately colonized by terrestrial and semiaquatic herbaceous plants from the seed bank in the drying sediment. Aquatic macrophytes, algae, and periphyton are the dominant forms of plants growing during the aquatic phase of the annual cycle (Putz and Junk 1997). Herbaceous plants and trees in floodplains are distributed according to present and past flood events. Annual species are primarily influenced by the hydrograph of the previous year, perennial herbaceous plants by the most recent interannual flood cycles and trees by flood cycles spanning many decades or even centuries (Junk 1997a). Abnormal floods, exposing or invading areas in extremely low or high positions may play an important role in species distribution patterns and species composition, particularly if they last several years. Seedlings of Salix martiana (S. humboldtiana) may colonize positions about 23 m.a.s.l. during two years of prolonged low water level as in 1991 and 1992, when these low-water periods are followed by less intense floods (Oliveira 1998). In the same way, pronounced high- water
211
levels as observed from 1970 to 1974, when the water did not fall below the 20 m level, may exclude species like Eugenia inundata, Myrcia sp. and Symmeria paniculata which are well adapted to growing at levels from 19.5 to 20.5 m.a.s.l. (Junk and Piedade 1997). The mechanisms regulating the distribution of vegetation along the floodplain are not yet fully understood; however, they are mainly related to differences in tolerance of the vegetation to low oxygen concentrations (Braendle and Crawford 1987, Crawford 1992). Oxygen deficiency is assumed to cause stress that leads to various types of adaptations both in the anatomy and in the metabolism of plants. Herbaceous plants are considered more tolerant to depletion in oxygen in the root systems than trees (Crawford 1992). Trees in the floodplain forest have adapted to the flood through different strategies. They may change their metabolism during the year as a consequence of the inundation. Therefore, many form distinct growth rings which correlate with the extent of the terrestrial phase (Worbes 1985, 1997). Timing of leaf formation, leaf shedding, flowering and fruiting, seed dispersal, and survival strategies of seeds and seedlings are also of special importance (Gottsberger 1978, Revilla 1981, Piedade 1985, Ferreira 1991, Kubitski and Ziburski 1994). Floodpulse-induced zonation of vegetation may be naturally disturbed by sedimentation and erosion which lead both to a permanent setback of plant communities and several stages resulting in highly dynamic and smallscale community development. Early successional stages are considered highly productive with a relatively low biomass. The effects of sedimentation are more hazardous for trees than for annual and perennial herbaceous plants. Although varying along the floodplains and between years, rates of sedimentation can produce layers more than 1 m thick (Junk 1986). Small grain-size
212
sediments reduce aeration and water exchange, burying tree roots. Better adapted trees such as Salix martiana form layers of new roots on top sediment layers; however, even those may suffer massive mortality in years of subsequent high sedimentation events (Oliveira 1998). In the blackwater igapo, trees occupy lower-lying areas, probably because of higher oxygen concentrations in the waters near the bottom (Junk and Piedade 1997).
Herbaceous Plants Distribution and development of herbaceous plant communities Studies of herbaceous plants have been concentrated in the varzea because low pH and nutrient levels characteristic of the blackwater igapo systems limit the presence of terrestrial herbaceous vegetation. On sandy igapo soils drought stress can also be important during the terrestrial phase (Junk and Piedade 1997). Submerged plants are inhibited by poor light conditions in turbid or blackwater, and strong water-level fluctuations. For some aquatic species, such as Eichhornia crassipes, low pH and low electrolyte content of blackwater may be toxic (Berg 1961). Aquatic herbaceous plants may grow in small blackwater streams near cities owing to increasing pollution mainly from domestic refuses. They also may be found in areas of increased sedimentation and nutrient input resulting from large-scale human disturbance of the catchment area. Herbaceous plants are good indicators of the current status of ecological conditions in a specific habitat. Because of their relatively short life cycles and their fast growth they occupy a wide range, from purely aquatic to terrestrial habitats in the floodplains. Establishing patterns of distribution of herbaceous vegetation in floodplains is
Maria Teresa F. Piedade et al.
difficult, since their distribution is regulated by the dynamics of the floodpulse and several interacting factors which make the scenario very complex (Junk and Piedade 1997). Plant distribution and species composition depend upon the duration of the aquatic and terrestrial phases and succession processes linked to the species longevity and the age of the specific habitat in the floodplain. Physical stability of the habitat, related to sedimentation, erosion, wave action, current, and human action, is important. For herbaceous plants such as the grass Echinochloa polystachya that reproduces vegetatively and by seeds, a less intense low water level may be negative, since regrowth and germination will be reduced. Because the species remain attached to the bottom of lakes and mud flats along the rivers, very intense peaks of inundation can remove large stands. This may reduce densities in the following year, with the related reduction in biomass. Species reproducing mainly by seeds like those of the genus Oryza (O. glumaepatula and O. grandiglumis), will not develop in years with a less intense terrestrial phase. Rooted semiaquatic herbaceous plants are well adapted to the dynamics of sediment deposition. During low water levels, perennial grasses forming large monospecific stands like Echinochloa polystachya and Paspalum repens re-grow mainly vegetatively, forming new shoots within the top soil layers. They thereby stabilize the sediments and accelerate sediment deposition. During that period, distribution and maintenance of stands in a given area depend upon the rainfall pattern and capacity of water retention in the soils. A few weeks without rain may lead to water stress even in well-adapted semiaquatic grasses (Piedade et al. 1994). Therefore, species composition and richness in a given area will vary strongly in between years, according to the precipitation pattern.
Geoecological Controls on Elemental Fluxes in Communities of Higher Plants in Amazonian Floodplains
During the aquatic phase plants are detached from the sediments and drift with the current. They contribute to the carbon pool of the main channel of the Amazon River and also accelerate the colonization of newly created open areas formed by the deposition of sediments. Quick colonization creates a range of different stages of allogenic succession in the floodplains. The velocity of the successional process will depend upon the stability of the habitat. Isolated floodplain lakes, where deposition rates of sediments are small, show a slow but gradual colonization process while, processes on sand bars and mud flats in the river channel are more dynamic and often reversible (Junk and Piedade 1997). Natural successional processes especially in the varzea usually overlap with modifications of anthropogenic origin. Floodplain forests are reduced by logging, agriculture, and the establishment of natural and artificial pastures for cattle and buffalo raising. All of these activities favor the growth of herbaceous vegetation. Although areas for agriculture are still relatively small and concentrated near urban centers in higher positions in the floodplains, they are important in allowing the introduction of neophytes which often successfully establish
213
in areas at the upper range of the flood gradient. If forest removal takes place in small areas, for instance by selective logging, it allows the growth of small patches of herbaceous plants and the regrowth of trees in a relatively short period. However, pasture cultivation, usually at bigger scales, complicates forest re-establishment and increases the risk of fires, to which the vegetation of the varzea is not adapted (Junk and Piedade 1997).
Species composition and diversity Numbers of species of herbaceous plants in Amazon floodplains should be related to their classification in aquatic or terrestrial plants, since species richness and composition changes markedly during the year. This is not a simple task, however, because several species show adaptations to both phases and eventually survive the two of them or at least one phase and part of the remaining one. In areas surrounding the city of Manaus in the central Amazon, 388 species of herbaceous plants were found (Junk and Piedade 1993b), belonging to 64 families and 182 genera. When classified according to their mode of existence, 85% of the total
Table 13.2 Site preference of terrestrial herbaceous species in the floodplains of the Central Amazon, per number of species and percentage of occurrence. Site Preference
Number of Species
Percentage of Occurrence*
Disturbed locations in higher positions
273
71%
Areas of fresh sediment deposition
92
24%
Low-lying beds of lakes
26
7%
Inundation forest
25
7%
Bays of the varzea lakes with standing water
42
11%
Floating islands
44
11%
Source: Junk and Piedade 1997. * The total percentage is superior to 100% since several species occur in more than one site. Percentages are calculated in relation to a total number of 388 herbaceous plants collected in the surroundings of the Manaus area.
214
Maria Teresa F. Piedade et al.
number of species were terrestrial. The Lower values are obtained during receding remaining species could be classified as waters and low water levels, when growth aquatic (9%) and of intermediate status (6%). rates are often negative. There is a possible Of the total, 60 were grasses, among which link between nutrient levels and growth some of the most important species were rates of free-floating plants because highest Echinochloa polystachya, Paspalum fascicula- growth values are reached during the begintum and Paspalum repens. These are perenni- ning of the inundation when the plants al C4 species with high densities and receive an input of additional nutrients from biomass values covering large areas in the river (Furch 1984). From low densities, monospecific stands. Biomass incorporation the plants quickly recolonize new habitats in and the fast return of nutrients to the system the floodplain while the waters rise. through a quick decomposition are typical Since free-floating species increase simuland strongly linked to the main period of taneously total biomass and area covered by plants, their biomass and productivity are growth and decaying of the species. Besides mode of existence, several plants estimated through the calculation of the among the terrestrial species showed prefer- relative growth rate. For comparative ence for specific sites, most of them (71%) purposes a useful index is the doubling time preferred disturbed locations at higher (Junk and Howard-Williams 1984). For positions (Table 13-2). Although abundance free-floating plants in Amazonian lakes, was high, near Manaus only 12 species (3%) maximum growth rates are high, although could be considered as very abundant (Junk marked differences between years, lakes, and Piedade 1993b, c), and just 5 species and sites are common. Three of the most (1%) could be considered as dominants, abundant species, Eichhornia crassipes, Pistia forming monospecific stands. Owing to the stratiotes, and Salvinia auriculata were monifrequent introduction of seeds of weeds tored in cage experiments at Lago do together with seeds for cropping, the Castanho by Junk and Howard-Williams number of unreported species in sites of (1984). Results showed a maximum duplicahighest elevations tend to be 10 to 20% tion time of 9-4 days (0.074 g g^day1, dry greater than the total amount of 388 species wt) for E. crassipes. P. stratiotes showed found by Junk and Piedade (1993b, c). a duplication time of 7.9 days (0.088 g g~ 1 day1, dry wt), and S. auriculata showed the Biomass and primary production minimum time, 7.2 days (0.096 g g^day1, dry wt). Biomass production of herbaceous vegeBecause of the horizontal spreading of tation in the varzea floodplain depends on free-floating plants during short time perithe species and its life cycle. Main growth ods, exponential growth can be assumed. S. period is related to the floodpulse, depend- auriculata totally covers the water surface ing on the position of the specific habitat on with a biomass of about 100-150 g rrr2 dry the flood gradient. The conjunction of these weight (Junk and Piedade 1997). Assuming factors will define the maximum biomass for the species an average relative growth and primary production reached by the rate of 0.069 g g^day1 corresponding to a species in a given year. There is also a doubling time of 10 days and starting with a marked difference in production according 1 m2 of the plant, corresponding to lOOg dry to the habit of the species. Free-floating weight biomass, during the main growth plants usually show growth and biomass period (1 of December to 30 of June) peaks related to the peak of inundation. biomass would reach the figure of 200 t dry
Geoecological Controls on Elemental Fluxes in Communities of Higher Plants in Amazonian Floodplains
weight and cover an area of ca. 2 km2. Although this does not happen in the natural systems in the Amazon because of losses by drift and biotic interactions, it may happen in artificial systems like man-made lakes of high nutrient status and stable waterlevel. This shows how quickly S. auriculata, as well as several other floating aquatic species under good nutrient conditions, can recolonize habitats recovering massive population losses during the low-water season. Even higher figures of biomass and net primary production are reported for the rooted species, however, as for the free-floating species, production will depend upon the species and time available for growth. Perennial species, although exhibiting periods of dormancy or collapse in biomass along the year, may grow up to 10 months, e.g. Paspalum fasciculatum (Junk 1986). More conservative values are found in annual communities growing in the ATTZ during the terrestrial phase where time available for growth seldom exceeds 3 months per year. Over a 3-month period, biomass values of communities of annual terrestrial plants
215
reached values ranging from 3.4 (±1.7) t ha'1 in Cyperaceae communities to 8.7 (±5.3) t ha"1 in Aeschynomene sensitiva dominated communities (Table 13-3). Mixed populations, including Aeschynomene sensitiva during the terrestrial phase and Hymenachne amplexicaulis and floating aquatic macrophytes in the aquatic phase, reached a biomass value of 30 t ha"1 during 9.5 months. Considering a monthly loss of biomass by decomposition and consumption between 10% and 25%, the cumulative NPP varies from 37 to 48 t ha"1 (Junk and Piedade 1993a). Mixed populations dominated by Paspalum repens and Oryza perennis (0. glumaepatula) reached, during a growth period of 4 to 5 months, 14.5 to 17.5 t ha"1. The most productive herbaceous species in the Amazon floodplains are the C4 grasses Paspalum fasciculatum and Echinochloa polystachya, both perennial species. E. polystachya grows in mud flats along river margins or lakes. Its life cycle is strongly related to the aquatic phase when the species maximizes its productivity. During the terrestrial phase biomass collapses, intense decomposition takes
Table 13.3 Biomass, including living and dead material of annual herbaceous plant communities in sedimentation areas in the varzea near the city of Manaus, central Amazon, shortly before the beginning of flooding. Day
Locality
Dominating Species
Biomass (t ha'1) Dry Weight*
07.03.74
Costa do Baixio
Cyperaceae Cyperaceae, Gramineae Aeschynomene sensitiva
6.4 (3.4) 3.2 (2.2) 8.7 (5.3)
15.01.82
Marchantaria area
Cyperaceae Gramineae Sphenoclea zeylanica Cyperaceae, Onagraceae Cyperaceae, Gramineae Cyperaceae
5.7 (3.5) 4.8 (2.4) 7.2 (3.3) 6.1 (2.6) 5.5 (2.7) 3.4 (1.7)
Sedimentation
Source: Junk and Piedade 1997. * The harvested area was of 1 m2, n = 10 and the standard deviation is in between parentheses.
216
Maria Teresa F. Piedade et al.
place, and the species regrows from the Bioelements stocks and fluxes nodes of the stems of the old plants Herbaceous plants in Amazon floodplains (Piedade et al. 1992). Conversely, P. fasciculatum has its main growth period related generally contain higher levels of N and P to the terrestrial phase. When the water and major ions (Na , K, Ca, Mg) in their recedes exposing the stems, new tissues than occur in the waters. shoots sprout from the nodes. Maximum Concentrations are comparable to those of biomass and NPP depend on the elevation macrophytes from other regions of the world of the stand in the floodplains (Junk (Junk and Howard-Williams 1984), with the and Piedade 1993a) and the nutritional sta- exception of sodium and calcium concentratus of the system within the varzea tions, which are lower in the aquatic (Conserva 1999). Paspalum fasciculatum macrophytes of the Amazon varzea than reaches a maximum biomass of 53-6 to elsewhere (Furch and Junk 1997a). A more 57.6 t ha'1 dry weight in positions unflood- detailed analysis shows variations among ed during 220 to 230 days per year (Junk terrestrial species and aquatic species, since and Piedade 1993a). E. polystachya may the sources of nutrients that they utilize attain a biomass of 80 t ha'1 in a 1-year differ from sediments to water, respectively. Herbaceous plants growing during the terperiod (Piedade et al. 199D. The NPP of P. fasciculatum may reach about 70 t ha'1 in restrial phase show significant differences in a growth period of about 8 months, while chemical composition between grass E. polystachya shows a figure of 100 t ha*1 (including sedges) and nongrass species in the period of 1 year. Biomass of E. (Furch and Junk 1997a). Nitrogen is higher polystachya starts to collapse when bio- in nongrass species, with a value of 2.88% mass of P. fasciculatum increases strongly, compared to 1.50% in grass species. Grasses maintaining total biomass and NPP of the have, as well, lower concentrations of system at a high level throughout the phosphorus and calcium but higher concenyear (Table 13.4). trations of sulfur, potassium, chloride, and
Table 13.4 Maximum net primary production in dry weight (NPP) of the most important herbaceous plant populations and communities in the varzea of the Central Amazon. Population/Community
Maximum NPP Time for Production* (t ha'1)
Authors
Monospecific stands of Echinochloa polystachya
100.0
1 year
Piedade et al. 1991
Monospecific stands of Paspalum fasciculatum
70.0
7.7 months
Junk and Piedade 1993a
Mixed population dominated by Hymenachne amplexicaulis
48.0
9.5 months
Junk and Piedade 1993a
Monospecific stands of Paspalum repens
33.0
4.0 months
Junk and Piedade 1993a
Monospecific stands of Oryza perennis
27.0
4.0 months
Junk and Piedade 1993a
Mixed population dominated by Oryza perennis
17.5
5.0 months
Junk and Piedade 1993a
'Maximum period of growth for the population or community.
Geoecological Controls on Elemental Fluxes in Communities of Higher Plants in Amazonian Floodplains
silica than nongrass plants. Aquatic macrophytes show similar concentrations of nitrogen, lower concentrations of P, Ca and Mg and higher concentrations of K and Na than those of the plants growing during the terrestrial phase. There are only few fluctuations in nutrient levels during the year, related to the biomass produced by macrophytes. One of the most important species in the varzea, the C4 grass Echinochloa polystachya was monitored and its results were extrapolated on an area basis (Piedade et al. 1997). The content of N, P, and K in leaves of the species showed values of 20, 1.7, and 19 g kg-1, respectively. Comparing these data with concentrations of these elements found in wellfertilized C4 Zea mays (Epstein 1972), levels of N are higher, levels of P are 50% lower, and levels of K are 50% higher. Concentrations of nutrients are within the range of well-established plants and since the species show very high photosynthetic rates all along the year (Piedade et al. 1994), nutrients should be sufficient to allow the species to maximize its potential production in the varzea. E. polystachya grows mainly during the aquatic phase and decomposes during the terrestrial phase, therefore the annual balance of release and uptake of N, P, and K was monitored and then related to the biomass obtained at the same time for the species (Piedade et al. 1997). Results show that N and P uptake follow the rise in water level and are probably derived from the surrounding water, while K appears to be independent of this fluctuation (Fig. 13-1). Total sequestration of nutrients by the stand of vegetation in the peak of the flood (August 1986) was calculated for N, P, and K, respectively, as 377, 51, and 1136 kg ha"1. These values are a lot higher than the 152, 37, and 270 kg rur1 calculated for N, P and K, respectively, by Mengel and Kirkby (1987) in well-fertilized rice-paddy planta-
217
tions. If all the vegetation dies during the following terrestrial phase, September/October, the same amount of nutrients would be released. However, since in the year of this analysis the water did not totally leave the stands, release was calculated in an order of 176, 29, and 343 kg ha-1 of N, P, and K, respectively. If one assumes that the species covers a conservative area of 5000 km2 in the Amazon region (Piedade et al. 199D, the uptake of nutrients sequestered from the waters during the growing period of the species will amount 189, 26, and 568 Mt of N, P, and K, respectively. During decomposition, the nutrients concentrated in the biomass will be released, enriching the floodplain soils, favoring the growth of the terrestrial herbaceous species and, to some extent, the floodplain forest. This illustrates the special importance of perennial grasses to the nutrient status of the varzea (Piedade et al. 1997). Modeling of nutrient fluxes in the water of Lake Camaleao near Manaus showed that annual dynamics of K, N, and P are strongly influenced by the uptake and release of herbaceous plants, and those of Na, Ca, and Mg by the inflowing river water (Weber et al. 1996, Junk and Weber 1996, Weber and Junk 1998, Weber 1997a, b). The model shows that a major factor influencing the contribution of herbaceous plants to the nutrient dynamics of floodplain lakes is the relative size of the ATTZ. During the terrestrial phase, terrestrial and semiaquatic herbaceous plants take up nutrients from the sediments and release them into the water during decomposition. Using nutrient release data from decomposition experiments, measured biomass data and mean bioelement concentrations of the Amazon River, Junk and Weber (1996) calculated that a floodplain lake with 100% ATTZ, that means a lake which dries out completely during low-water period, receives about 5-10% of its P and N, 25% of its K, but 95%
218
Maria Teresa F. Piedade et al.
Fig. 13.1 The net uptake of N, P, and K, calculated from the monthly changes in absolute quantities per unit ground area, by Echinochloa polystachya. A negative uptake indicates a net loss from the plants to the floodplain/water. Water level in m above mean sea level (a.s.l.) at Manaus is given in the upper panel for comparison (Piedade et al. 1997)
Geoecological Controls on Elemental Fluxes in Communities of Higher Plants in Amazonian Floodplains
219
Table 13.5 Comparative structural data of some Amazonian floodplain forests: investigations of studies on floristics in Amazonian inundation forests. Forest Type and Location Seasonal Varzea Forest Rio Napo, Ecuador Rio Xingu, Para Rio Jurua, Acre Lower Solimoes, Amazonas Solimoes, Rio Japura Tidal Varzea Forest Rio Guama, Para Ilha das Oncas
Reference
Plot Size (ha)
Tree Density, N Species Per ha
Per Plot
Per 100 Stems
149 40 20-106 69-127
35.5 9.1 2.6-20.2
5
417 440 420-777 325-1270
Dhh Limit (cm)
Baslev et al. 1987 ca. 1 Campbell et al. 1986 0.5 3x1 Campbell et al. 1992 Revilla 1991, Worbes 17x1 et al. 1992, Klinge et al. 1995, Worbes 1997 2x1 Ayres 1993
10 10 10
10
416-580
109-135
23.3-26.2
Pires and Koury 1959 Anderson et al. 1985
3.8
10
483
53
21.1
0.25
5
1576
51
12.9
of its Ca from the river. The remaining part derives from decomposing higher vegetation of the ATTZ. In lakes with a smaller ATTZ, the amounts of nutrients deriving from higher plants are reduced in relation to the reduction of the ATTZ. The application of the model to Lake Camaleao also showed that bioelement-rich groundwater is an important additional source, mainly for Na, K, Ca, and Mg (Weber et al. 1996, Weber 1997b, Furch in press).
8.6-24.3
1986). At the confluence of the Japura and Solimoes rivers, two forest sites of Mamiraua are described by Ayres (1993) with respect to the influence of inundation period on species composition. From the western Amazon there are inventories from the Rio Uyacali (Marmillod 1982), Rio Jurua (Campbell et al. 1992) and Rio Napo (Balsev et al. 1987), and in the eastern Amazon there are inventories from the Rio Guama (Pires and Koury 1959) and the Rio Xingu (Campbell et al. 1986). An overview of samFloodplain Forest ple size, number of species and stems of these inventories is given on Table 13.5. The comparison of the species lists show Floristic composition that many species are -widespread in the The floristic composition of floodplain complete Amazon basin (Table 13-6). forests is described by several forest inven- Ninety-one species are common to several tories covering the complete Amazon basin. studies, 20 species occur in four or more Detailed information is available from cen- locations. Table 13.8 can be divided into sevtral Amazon forests in the vicinity of Manaus eral sections. One species group is obviouswith inventories of about 17 ha of varzea for- ly restricted to the region west of Manaus. est from the Una de Marchantaria, Costa do Another group is distributed from Mamiraua Marrecao, and Costa do Barroso (Worbes to the mouth of the Amazon river (Worbes et 1983, 1986, Revilla 1991, Worbes et al. 1992, al. 1992). Due to the high percentage of Klinge et al. 1995) and from igapo forests unidentified taxa in all references, the (Keel and Prance 1979, Revilla 1981, Worbes similarities of the forests are doubtless even
220
Maria Teresa F. Piedade et al.
Table 13.6 Distribution of tree species from Amazonian floodplain and two adjacent terra firme forests.
Species Pithecellobium cauliflorum Pithecellobium latifolium Leonia glycicarpa Minquartia guianensis Astrocaryum murumuru Vismia cf. baccifera Slonea guianensis Tapura amazonica Sapium marmieri Sterculia elata Cecropia membranaceae Virola elongata Swartzia polyphylla Cordia nodosa Virola surinamensis Matayba arborescens Couepia paraensis Allophyllus amazonicus Maprounea guianensis Ceiba pentandra Hura crepitans Brosimum guianensis Chlorophora tinctoria Caraipa grandifolia Inga cinnamomea Pseudoxandra polyphleba Casearia aculeata Lecointea amazonica Cynometra spruceana Crudia amazonica Pseudobombax munguba Caryocar microcarpum Crataeva benthamii Maytenus guianensis Hevea spruceana Sorocea duckei Pseudolmedia laevigata Iryanthera juruensis Iriartea deltoidea Terminalia amazonica Lacistema aggregatum Brosimum lactescens Protium fimbriatum Protium nodolosum Trichilia maynasiana Margaritaria nobilis Ficus gomelleira
Family O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
0
0
O
O
O
O
O
O O
O
O
O
O
O
O
O
O
O
0
O
O
O O
O
O
O
O
0
O
O
O
O
O
0
0
O
O
O
O
O
O
O
O
O
O
O
O
O
O O
O
O
0
O
O
O
O
O
O
O
O
O
O
O
0
O
O
O O
O
O
0
0
O
O
O
O
O
O
O
O
O
O
O
O
O O O
O
O
O
O
O
?
0
0
O
O
O
O
O
O
O
O
O
0
O
O
O O
0
0
O
0
O
0
O
O
O
O
O
0
O O
Mimosaceae Mimosaceae Violaceae Olacaceae Arecaceae Hypericaceae Elaeocarpaceae Dichapetalaceae Euphorbiaceae Sterculiaeae Moraceae Myristicacaea Caesalpiniaceae Boraginaceae Myristicaceae Sapindaceae Chrysobalanaceae Sapindaceae Euphorbiaceae Bombacaceae Euphorbiaceae Moraceae Moraceae Clusiaceae Mimosaceae Annonaceae Flacourtiaceae Caesalpiniaceae Caesalpiniaceae Caesalpiniaceae Bombacaceae Caryocaraceae Capparidaceae Celastraceae Euphorbiaceae Moraceae Moraceae Myristicacaea Arecaceae Combretaceae Flacourtiaceae Moraceae Boraginaceae Boraginaceae Meliaceae Euphorbiaceae Moraceae
Geoecological Controls on Elemental Fluxes in Communities of Higher Plants in Amazonian Floodplains
Family
Species Pseudolmedia laevis Lecointea amazonica Trichilia micrantha Andira cf. inermis Drypetes variabilis Glycydendron amazonicum Virola cuspidata Ilex inundata Macrolobium acaciifolium Inga coriaceae Moconia poeppigii Platymiscium duckei Cecropia latiloba Hevea brasiliensis Virola surinamensis Astrocaryum jauari Apeiba aspera Trichilia lecointei Tabebuia barbata Licania heteromorpha Unonopsis guatteroides Rheedia brasiliensis Piranhea trifoliata Campsiandra laurifolia Calophyllum brasiliense Rheedia macrophylla Mabea caudata Amanoa oblongifolia Licania cf. parviflora Neoxythece elegans Pterocarpus amazonum Symmeria paniculata Sickingia tinctoria Himatanthus attenuata Malouetia tamaquarina Macrolobium angustifolium Rheedia acuminata Vismia cayennensis Gustavia heoapetala Franchetella anibifolia Malouetia furfuraceae Mabea nitida Quiina rhytidopus Psychotria lupulina Vataira guianensis Gustavia augusta
o o o o o o
o o
o
•>
o o o o
o o 0
0 0
o o o o o o o o
o o o
o
0
o o o o
o o o •>
•> 0
O
0
o o o
o o o o o
o
0
o o o o o o o o o o o o o
O
o o o o
o o o o
0
0
o
o
0
0
o o
o
0
O
o o o
o o O
0 0 0
0
o
0 0 0
o
o
o
O
0
0
0
o
o
0
O
O 0
0
o
o
O
o
o
Moraceae Fabaceae Meliaceae Fabaceae Euphorbiaceae Euphorbiaceae Myristicaceae Aquifoliaceae Caesalpiniaceae Mimosaceae Melastomataceae Fabaceae Moraceae Euphorbiaceae Myristicacaea Arecaceae Tiliaceae Meliaceae Bignoniaceae Chrysobalanaceae Annonaceae Clusiaceae Euphorbiaceae Caesalpiniaceae Clusiaceae Clusiaceae Euphorbiaceae Euphorbiaceae Crysobalanaceae Sapotaceae Fabaceae Polygonaceae Rubiaceae Apocynaceae Apocynaceae Caesalpiniaceae Clusiaceae Clusiaceae Lecythidaceae Sapotaceae Apocynaceae Euphorbiaceae Quiinaceae Rubiaceae Fabaceae Lecythidaceae
Sources: IBalslev et al. (1987), 2Marmillod (1982), 3Campbell et al. (1992), 4Ayres (1995), 5Worbes et al. (1992), 6 Revilla(1981), TWbrbes (1986), SPires and Koury (1959), 9 Campbell et al. (1986).
221
222
greater than Table 13.6 suggests. The similarities in species between forests lying hundreds are even thousands of kilometers apart clearly distinguish the varzea forest as coherent community. Nevertheless there are relations between the flora of the floodplains and the upland areas. A number of species from varzea forests also occur on the terra firme near Manaus, e.g., Mabea caudata, Heisteria spruceana, Minquartia guianensis, Pithecolobium jupumba, and Vatairea guianensis. Two studies of upland and adjacent floodplain forests show concurrence of 18% (Balslev et al. 1987) and 45% (Campbell et al. 1986) on species level. Taxonomic and chorological investigations (Prance 1989, Kubitzki 1989) point out that the flora of the igapo is closely related to the flora of the nutrient-poor savannas. On the other hand, close connections are visible between the flora of the varzea and the flora of more fertile sites on the terra firme. Species of the shorter inundated high-level communities (e.g. Clorophora tinctoria, Hum crepitans) are frequently found in Neotropical dry forests (Frankie et al. 1974). The growth factors which influence species diversity act as differentiating factors in species composition. Tolerance to the intensity of the flood pulse can be explained partially by a change in root metabolism during the submersion phase which varies between species (Schliiter 1989, Worbes 1997). There is also a distinct geological control in distinguishing between long inundated low-lying areas (varzea baixa) and areas of higher elevation (varzea alta). Taken together, these controls produce a clear zonation of the varzea forest (Junk 1989, Ayres 1993). Typical species of the varzea alta include Astrocaryum murumuru, Spondias mombin, Hura crepitans and Ceiba pentandra, which grade into adjacent terra firme forests (Campbell et al. 1986, Balslev et al. 1987). In the western Amazon basin,
Maria Teresa F. Piedade et al.
Piranhea trifoliata often dominates long and high flooded sites. This is a long-living species (up to 400-500 years, Worbes 1989) with a high wood density. Fresh wood of Piranhea trifoliata does not float, causing problems when cut trunks must be skidded during the inundation phase. So Piranhea trifoliata is relatively seldom logged in comparison with light-wood species (Albernaz and Ayres 1999). This is a long-term anthropogenic selection factor considering that the varzea has been settled since pre-Columbian times (Roosevelt 1999). On the other hand, widespread species like Calophyllum brasiliense, Ceiba pentandra, or Hura crepitans do not occur as often as expected. These are valuable timber species and are probably selectively logged. In tropical vegetation analyses, an often neglected differentiating factor is the influence of the dynamic stage of a forest community. The basic problem is the assumed impossibility of dating the exact age of trees and forest stands, primarily due to the assumed absence of tree rings in tropical trees (Whitmore 1990). In the varzea forests of the lower Solimoes river, Worbes et al. (1992) found at different locations different species compositions at similar elevations and, therefore, under equal flood stress. Age dating of the dominant trees by means of tree-ring analysis showed varying age structures for the investigated stands. The young stands were dominated by trees with characteristics typical of pioneers (Swaine and Whitmore 1988), that is, low life expectancy, fast growth, and low wood density (.Salix humboldtiana, Cecropia spp., Pseudobombax munguba, Macrolobium acaciifolium). The oldest stands were dominated by trees with converse features (Piranhea trifoliata, Manilkara sp., Tabebuia barbata). The presence of tree species with opposing, but sequential, life history strategies supported by differences in age structure indicates that the investigated
Geoecological Controls on Elemental Fluxes in Communities of Higher Plants in Amazonian Floodplains
stands may be treated as a chronosequence. The features of the defined successional stages in the varzea also occur in Central American forests (Budowski 1961), as well as in the European floodplain forests (Carbiener et al. 1988). The initial stage of primary succession in floodplain habitats is dominated by fastgrowing grass communities. The following pioneer forest community is dominated by representatives of the genus Salix (Maximum age: < 30 years, Worbes et al. 1992), which is replaced by a species-poor formation dominated by Cecropia spp. This stage soon undergoes a transition to the more species rich Pseudobombax forest type. The Pseudobombax munguba community represents the beginning of a dynamic stage between the "late secondary" and the climax community (Halle et al. 1978). With an age of 80-100 years, the fast-growing dominant species P. munguba reaches its maximum age and is eventually replaced by shadow-tolerant slower-growing species from the lower forest canopy. The Piranhea trifoliata community represents a climax forest type with the highest diversity. This community is
223
dominated by densewood emergents with multisecular rotation and subdominants with a shorter life span. The change in species composition over time incorporates changes in structural elements, i.e. increasing biomass, decreasing population density, and decreasing individual growth rates.
Species composition and diversity Diversity patterns are an important feature in the analysis of tropical forest communities. However, a comparison of existing studies is difficult because size and shape of plots, diameter limit of the included individuals, and criteria of site selection may differ considerably (Table 13-5). A simple but helpful measure is the calculation of species per 100 stems (Klinge et al. 1995) (Table 13-5). Another severe problem is that all individuals can not be identified to species level. In all reports a varying percentage of unidentified taxa is present. This reason impedes a calculation of indices of similarity or diversity. The generally observed trends in changing species diversity of Amazonian floodplain forests are as discussed below.
Fig. 13.2 Diversity expressed in number of species per 100 stems in Varzea forests of different ages and localities. Figures for Jurua forests from Campbell et al. (1992) and for the lower Solimoes forests from Worbes (1997).
Maria Teresa F. Piedade et al.
224
forest at Praia grande (Revilla 1981). According to observations of Tilman (1982) in tropical forests of Malaysia, where diversity is highest under moderate conditions, this type of site can be found in the igapo with clay soils. Diversity decreases either at sites with very poor soils like in white sand regions or on fertile soils as in the varzea. The flood pulse as well as periodic The change in species composition from droughts can be interpreted as climatic stress early pioneer species to species of the factors correlated with a decline of diversity. mature forest includes a general increase in At the lower timber line, where the sites are diversity from monospecific Cecropia- or exposed to more than 250 days of Sa/ix-stands to multilayerd stands with more inundation annually, only species-poor than 100 species per ha (Fig.13-2). shrub communities can persist. This On igapo and varzea sites exposed to an coincides with Grime's model of stress, equal length of inundation, diversity is competition, and disturbance (Grime 1979). higher in the igapo with its lower soil fertili- In Grime's model, species diversity is low for ty. Keel and Prance (1979) and Worbes sites with either high or low degrees of stress (1983, 1986) counted, in small igapo plots of and disturbance. A high degree of distur0.22 and 0.21 ha, 54 and 61 species, respec- bance and stress (flood stress) commonly tively. This is almost twice the number of occurs at river margins of the Amazon. A low species found on plots of comparable size in degree of disturbance occurs in floodplain the varzea (Worbes 1986). In an old varzea forests in the center of great islands and forest in the vicinity of Manaus, a maximum outside the main river channel. On the other of 127 species per ha was counted (Revilla hand, investigations in floodplain forests and 1991, Worbes 1997). In a 15 ha inventory at adjacent terra firme stands along the Napo the Costa de Marrecao, 256 species were river, Ecuador (Balslev et al. 1987), and counted (Revilla 1991) in comparison with along the Xingu river, Brazil (Campbell et al. about 200 species in only 1.8 ha in the igapo 1986), show in both cases a higher diversity Species diversity increases: 1. successional development from young to old stands, 2. from the eutrophic varzea to more oligotrophic stands in the igapo, 3. with decreasing flood stress, and 4. from sites in eastern to sites in western Amazonia.
Table 13.7 Mean content of organic carbon (DOC), total N, total P, exchangeable base cations (Na, K Mg, Ca) cation exchange capacity (CEC), pH (KCl).
Varzea Ah 0-10 10-30 Igapo Ah 0-10 10-30 30-40
pH (KCl) DOC (mg kg-1) Range
N P K Na Ca CEC Mg 1 1 1 1 1 1 (g kg-1) (G kg- ) (Cmol kg: ) (Cmol kg- ) (Cmol kg- ) (Cmol kg- ) (Cmol kg- )
3.7-4.6 3.5^.6 3.6-6.1
691 78 46
7.6 1.4 1.1
0.81 0.61 0.64
0.19 0.18 0.21
0.51 0.34
3.40 2.84
0.29
3.5-3.6 3.6-3.7
612
5.1 3.1 1.1 0.8
0.21
0.10 0.08 0.02 <0.01
0.27 0.18 0.04 0.02
3.7-3.9 3.7
379 148 28
Source: Data adopted from Furch (1997).
0.15 0.08 0.06
17.1
3.63
12.5 12.9
23.3 19.9 20.0
0.31 0.14 0.06 0.04
0.31 0.08 0.07 0.08
5.7 3.8 2.1 2.4
Geoecological Controls on Elemental Fluxes in Communities of Higher Plants in Amazonian Floodplains
225
Table 13.8 Stock of total nutrients in the surface layers (0-10 cm) of soil in six Amazonian forest types in kg ha"1. Location
Forest Type
N
P
K
Mg
Ca
Reference
Varzea
Inundation
19
726
23831
10934
23550
Furch (1997)
Igapo
Inundation
867
412
Furch (1997)
Manaus
Evergreen TF
San Carlos Caatinga Maraca
Evergreen TF
69 25 36 14 21 14 74
143
1156
23
19
211
20-
6
5
Uhl and Jordan (1984)
57
78
36
60
28
Thompson et al .(1992a)
256
98
83
169
Thompson et al . (1992b)
Franken et al. (1979)
o0
Maraca
Semi-evergreen
16
33
in nonflooded areas. However, it is likely that terra firme forests at the margins of the varzea can be flooded unpredictably in certain years. They represent an ecotone with features of both forest types and with a higher diversity than typical forest formations. On nonflooded sites in Amazonia, species richness increases from east to west, from 87 species ha'1 near Belem (Black et al. 1950) to over 179 species ha"1 near Manaus (Prance et al. 1976) and more than 200 species in Ecuador (Balslev et al. 1987). Gentry (1982, 1990) explained species richness generally increases as annual precipitation increases. However in the same direction diversity increases in the floodplain forests from 53 species per ha in Para (Pires and Koury 1959) to over 135 species in Mamiraua (Ayres 1993) to 149 species at the Rio Napo (Balslev et al. 1987). The human influence on species composition in floodplain forests is a poorly investigated factor; however, forests of the eastern Amazon have been exposed to deforestation and severe exploitation for longer times than western forests. In the varzea of the Solimoes river, selective logging nearly caused the extinction of
formerly frequent but valuable species like Ceiba pentandra and others. Therefore, the human impact on species diversity should be taken in account.
Bioelement stocks and fluxes Soils in the varzea have a moderate cation exchange capacity (20-23 C mol kg'1) and a slightly increasing pH with increasing depth (3.5-6.1, Table 13-7). The content of exchangeable base cations and of total phosphorus in the mineral soil of the varzea exceeds that of the igapo. There is a higher concentration of nitrogen and carbon in upper soil horizons in the igapo. In the varzea soils there is a lower C/N ratio (7.4 v. 12.8), indicating higher decomposition rates than in igapo. The low decomposition rates in the igapo lead to a permanent litter layer (Wbrbes 1986) and to the accumulation of carbon and nitrogen in the upper soil layers. The high nutrient concentrations and high nutrient stocks (Table 13.8) in varzea soils can be explained by the high amount of annual sedimentation. Young minerals from the Andes, with a high percentage of chlorite, contribute considerably to the high fertility of varzea soils (Irion et al. 1997). The
226
Maria Teresa F. Piedade et al.
Table 13.9 Nutrient concentrations (%) in leaf litter and wood of an varzea forest.
Element
Leaves Mean
SD
Litter Mean
SD
Wood Mean
SD
N
2,4085
0.25
1.7
0.5
0416
0.03
P
0,1615
0.01
0.09
0.02
0029
0.01
K
1109
0.30
0.46
0.04
0,6965
0.02
Ca
1,6445
0.38
1,7
0.23
1002
0.25
Mg
0341
0.08
0.31
0.01
0,1505
0.02
Na
0013
0.01
0.013
0.06
0,2595
0.26
Sum
5.682
4.26
2.355
Source: Furch and Junk 1997
nutrient stocks of K, Mg, and Ca in the surface layer (0-10 cm depth) of the varzea are 1000 times higher than in terra firme soils in Amazonia (Table 13-8). The clear chemical distinction between Whitewater (Amazon) and blackwater (Rio Negro) is reflected by the soil chemistry as shown above and in the chemical composition of the phytomass in the respective varzea and igapo floodplains (Klinge et al. 1983, Furch and Klinge 1989). A comparison of elemental content of fresh leaves in different environments shows varzea trees to have the highest values in all measured nutrients in comparison with trees from the igapo and terra firme in the vicinity of Manaus (Fig. 13.3). Varzea foliar N, P, K, Ca, and Mg concentrations fall in the range of a moderately fertile soil group from Vitousek and Sanford (1986). The values for igapo, terra firme near Manaus (Klinge 1975), and a seldom flooded site Maraca island in Roraima (Thompson et al. 1992a) fall within the group of infertile Oxisols and Ultisols. The concentrations of N, P, K, and Mg decrease in concentration from fresh leaves to litter to wood of varzea trees (Table 13-9). The difference between nutrient contents in fresh leaves and litter indicates a translocation of these elements
before leaf shedding. Between 10% and 50% of the nutrient concentrations in the leaves are to be found in the wood. The uptake of nutrients per unit of produced biomass is lower in the igapo than in the varzea (Furch 1997). This indicates a higher nutrient use efficiency of the igapo trees in comparison with the varzea trees, which use the nutrients more efficiently than trees from the temperate zone (Vitousek 1982). Nitrogen fluxes in the varzea forest ecosystem can be calculated on the basis of estimations of biomass stock, biomass increment (Worbes 1997) and nutrient stocks in different compartments. The values in Fig. 13.4 and in the following text are calculated for 1 hectare. The data refer to a 40-year-old forest stand at Ilha de Marchantaria, where a biomass of 265 t was measured (Klinge et al. 1995). Total N amounts 5.9 t in the upper 30 cm of the soil, and the exchangeable fraction is 0.11 t. The input from flooding river water, rainwater and sediment to the exchangeable compartment is negligible (0.01 t per year), and the stock of total N increases annually by 0.1 t. The uptake of the phytomass is 0.29 t per year. About 30% of this uptake is stored annually in the wood, 0.2 t is
Geoecological Controls on Elemental Fluxes in Communities of Higher Plants in Amazonian Floodplains
returned with litterfall and to the soil within one vegetation period due to a fast decomposition (Furch and Junk 1997a). The sum of the input from sedimentation, litterfall, and the stock of exchangeable cations exceeds the annual uptake of N by the forest phytomass. There is a loss of nitrogen by denitrification during the terrestrial phase (Korshorreck, personal communication) but also a contribution of N-fixation by mycorrhiza which is not yet estimated. We assume that there is no limitation of forest
227
growth by nitrogen. The same is true for all major nutrients in the varzea and the igapo, •where the amount of exchangeable nutrients in the soil exceeds the uptake of the phytomass (Furch 1997). The lower nutrient supply in the igapo forest probably causes the slower growth, which is also indicated by low rates of annual litterfall of about 5 t, which is significantly less than the 10 t or more measured in the varzea (Adis et al. 1979, Furch and Klinge 1989, Worbes 1997).
Fig. 13.3 Stocks and fluxes of nitrogen in the Varzea forest ecosystem (Furch 1997).
Maria Teresa F. Piedade et al.
228
Table 13.10 Biomass in different successional stages of varzea forests.
Successional Stage (Age)
Dominant Species
t/ha
Pioneer (2 yr) Pioneer (4 yr) Pioneer (12 yr)
Salix Bonafusia Cecropia
3 14 98
Early secondary (44 yr)
Crataeva
258
Late secondary (80 yr, mean)
Pseudobombax
279
Late secondary (80 yr)
Pseudobombax with gaps Most productive plot
118
Late secondary (80 yr)
Pseudobombax
470
Biomass and primary production The biomass of the varzea forest increases along a chronosequence corresponding to forest succession (Worbes 1997). The initial stage of the primary succession formed by a young stand of Salix humboldtiana (Table 13-10) has a very low biomass. Considerably higher is the biomass in a secondary stand, where the dominant species (Bonafusia tetrastarhya) already occurred before clearing and probably germinated by vegetative propagation. In the following stages biomass is increasing rapidly due to high annual biomass production. There are no measurements from a mature forest stand, but it is likely that those stands have an even higher
value than the most productive and dense plot in the measured late secondary stand (470 t). In this stand the dominant species (Pseudobombax mungubd) has a very light wood and is replaced by hard-wood species which may contribute a greater portion to the total biomass of the stands. Proof of the general existence of annual rings in trees from floodplain forests (Worbes and Junk 1989, Worbes 1997) allows the use of tree-ring analysis to estimate the aboveground wood production. Measurement of annual tree-ring width over the complete lifespan of 203 trees show mean figures of 3-5 mm yr1 in the varzea and 1.7 mm yr1 in the igapo. Incremental values from the igapo are in the range of
Table 13.11 Aboveground NPP without losses by herbivory in the investigated varzea stands, in a rain forest in Thailand and in a beech forest in the temperate zone. Site
Varzea (V4, 40 yr) Varzea V5, 80 yr) Rainforest, Thailand** Beech for. Germ. (120 yr)***
Fine Litter* (t ha-1)
Dead Wood (t ha-i)
Wood Increment (t ha-0
NPP
9.9 13.6 12.0 3.6
6.6 11.4 13.3 0.4
7.3
23.8 33.6 28.4 10.7
Sources: * Data for fine litter from Adis et al. (1979). ** The age of the rainforest in Thailand was unkown. (Kira et al. 1967) "* Ellenberg et al. 1986
8.6 3.1 6.7
(t ha-i)
Geoecological Controls on Elemental Fluxes in Communities of Higher Plants in Amazonian Floodplains
229
worldwide estimations from different tropi- 12-year-old Cecropia latiloba stand. In an cal forest communities estimated by repeat- 80-year-old Pseudobombax munguba stand ed diameter measurements. The values from biomass production was calculated at the varzea lie on the upper end of this range. 7.2 t ha'1 yr1. This stand is a patchwork of Rates of radial increment are only relative relatively unproductive gaps (2.2 t ha*1 yr1) figures with no direct relation to biomass and highly productive plots with wellproduction of a specific area. Aboveground structured and differentiated tree stands (8.6 wood production was estimated by dividing t ha-1 yr1) (Worbes 1997). the aboveground biomass by the age of the These figures match closely the world stands (Worbes 1997). Aboveground biomass wide mean of 7.3 t ha"1 yr1 of aboveground was measured by harvesting (Klinge et al. biomass production estimated with various 1995) and allometric estimations of tree methods (Jordan 1983). This result confirms height, diameter at breast height, and densi- earlier findings (Medina and Klinge 1983, ty of the wood. Ten percent of the result was Jordan 1983) that wood production in added for the biomass of the branches and tropical forests, even under good nutrient twigs (Ellenberg et al. 1986). conditions like in the varzea, does not The results show an increase of biomass exceed the wood production of temperate production with increasing age of the stands. forests. The sum of the aboveground wood The figures range from 1.5 t ha'1 yr1 in a production and of the coarse and fine litter 2-year-old monospecific pioneer stand of production leads to aboveground primary Salix humboldtiana to 8.1 t ha"1 yr1 in a production (NPP). Within the varzea the
Fig. 13.4 Leaf chemistry (above) and major nutrient concentrations in the respective soils (below) in Varzea, Igapo and terra firme near Manaus. Data from Klinge (1975), Furch 1997, Furch and Junk (1997a).
230
litterfall varies from 7.8 t ha'1 yr1 in an early successional 12-year-old Cecropia latiloba stand to 13.6 t ha*1 yr1 in a mixed forest stand with an approximate age of 60 years (Adis et al. 1979). The mean of four estimations is 10.3 t ha-1 yr1. The litterfall in the igapo is considerably lower with 6.7 t ha"1 yr1. At both sites leaves contribute about 80 % of the total amount of litterfall. Total NPP was calculated for a 40-year-old stand (V4) and an 80-year-old stand dominated by Pseudobombax munguba (V5). The total aboveground NPP, without considering losses by herbivory, is in stand V4 23.8 t ha'1 yr-1 and in the most productive plot of stand V5 33.6 t ha'1 yr1 (Table 13.11). For the igapo forest no figures on biomass wood production are available. The low figures for litterfall and radial wood growth indicate that this ecosystem is very unproductive. There is no nutrient input from the nutrient- and sediment-poor water of the Negro River.
Conclusions In spite of flood stress, the forest ecosystem of the varzea is more productive than the relatively unproductive forests of the surrounding terra firme area. Varzea forests have an NPP about 16% greater than that of a well-established rain forest on relatively fertile soils in Thailand (Worbes 1997). The growth period of many species is concentrated in the terrestrial phase of only about 200-250 days and often includes a period of drought stress of about 4 weeks after the inundating water retreats and before the rainy season begins. The access of the tree roots to the water table may also contribute to the productivity of the varzea forests. The fertility of varzea soils obviously compensates for the unfavorable growth conditions during the submersion phase. During the flood period metabolism in roots and remaining leaves is observable but energy
Maria Teresa F. Piedade et al.
transfer efficiency is supposed to be generally lower than in the nonflooded period, as indicated by reduced wood increment and tree ring formation. Differences in fertility between igapo and varzea result in different species composition and in different productivity patterns. The igapo forests are more diverse and less productive than the varzea forests. Floodplain forests of the Amazon as a whole can be treated as a distinct vegetation community. This forest type with distinct widespread and abundant species is distributed throughout the catchment areas of the Amazon and Orinoco rivers. Herbaceous C4 plants may exhibit NPP values about three times those of the floodplain forest (Piedade et al. 1991, Junk and Piedade 1993a). Therefore, although restricted to relatively small areas, these herbaceous formations are very important in the carbon and nutrient budgets of the region. During the peak of inundation there is a direct connection of almost all streams and rivers to their respective floodplains. The release of carbon and minerals through the decomposition of herbaceous vegetation in the aquatic or in the terrestrial phases enriches the aquatic system and associated floodplains. Although precise numbers about the area covered by herbaceous vegetation is lacking for the region, Piedade et al. (1994) estimated that about 10% of the varzea is surfaced by the most productive C4 plants, Echinochloa polystachya, Paspalum fasciculatum, and P. repens. This amounts to a total of ca. 20,000 km2 of a total area of 200,000 km2 estimated by Junk (1993) for the floodplains related to the big Whitewater rivers. This is consistent with the high concentrations of nutrients found in floodplain lakes where decomposition processes related mainly to herbaceous vegetation may produce concentrations of nutrients more than 20 times greater than those in the Amazon River itself (Furch 1984, Furch and Junk
Geoecological Controls on Elemental Fluxes in Communities of Higher Plants in Amazonian Floodplains
1985). The size of the ATTZ and high biomass of herbaceous plants per unit area are directly correlated with the amounts of nutrients released to the water from rooted higher vegetation. This becomes evident in the nutrient budgets calculated by Melack and Forsberg in chapter 14 of this volume. Herbaceous vegetation and the floodplain forests are affected by high fluvial dynamics. At the river margins erosion and sedimentation are major factors of natural disturbance. In areas of sedimentation, a primary succession can be observed. In areas of the lower lying varzea, which are far from the main channel, early successional stages may persist for decades or centuries. Annual herbaceous plants react to annual changes in the flood regime, while perennial herbaceous plants have reaction times of several years and the floodplain forest may react on extreme flood events of centuries. Sustainable management systems for Amazon floodplains should consider the importance of higher vegetation in the nutrient budget of the region. Because of the favorable nutrient conditions in the varzea forests, they are frequently replaced by farms and pasture lands for cattle and buffalo ranching. The increasing exploitation of
231
remaining floodplain forests in the middle and upper reaches of the Amazon River and its large Whitewater tributaries leads to a considerable decrease of valuable timber species. In the older stages, less valuable species like Ficus spp. or very hard (and old growing) species like Piranhea trifoliata become more frequent. This leads to a change in species composition of the floodplain forest in the direction of a degraded ecosystem. The substitution of the forest by herbaceous plant communities has little affect on total primary production because natural herbaceous plant communities are often more productive than the forest itself, but it affects strongly biomass. In floodplains of low-fertility status, the large biomass of the forest is important for long term nutrient accumulation and storage. Massive removal of herbaceous plants for the introduction of less-adapted species for agriculture will reduce the fertility of the varzea and will require additional fertilization. Susceptibility of these degraded systems to fire increases during the terrestrial phase, resulting in losses of crops and pastures, and leading to further degradation of the natural plant communities and associated animals of the floodplains, with far-reaching negative consequences for human life.
Literature Cited Adis, J., K. Furch, and U. Irmler. 1979. "Litter production of a Central Amazonian black water inundation forest." Tropical Ecology 20: 236-245. Albernaz, A. L. K. M., and J. M. Ayres. 1999. "Selective logging along the Middle Solimoes river." In: Varzea. Diversity, Development, and Conservation of Amazonia's Whitewater Floodplains, eds. C. Padoch, J. M. Ayres, M. Pinedo-Vasquez, and A. Henderson (The New York Botanical Garden Press, New York), pp. 135-151. Alves, L. F. 1993. The fate of streamwater nitrate entering littoral areas of an Amazonian floodplain lake: the role of plankton, periphyton, inundated soils, and sediments. Ph.D. Thesis, University of Maryland-CEES, College Park. p. 370.
Anderson, A. B., A. Gely, J. Strudwick, G. L. Sobel, and G. C. das Pinto. 1985. "Sistema agroflorestal na varzea do estuario amazonico." Acta Amazonica 15 (suppl.1-2): 195-224. Ayres, J. M. 1993. As matas de varzea do Mamiraua. CNPq/SCM, Brasilia DF. Balslev, H., J. Lutteyn, B. 011gaard, and L. B. Holm-Nielsen. 1987. "Composition and structure of adjacent unflooded and floodplain forest in Amazonian Ecuador." Opera Botanica 92: 37-57. Berg, A. 1961. Role ecologique des eaux de la cuvette congolaise sur la croissance de la jacinthe d'eau [Eichhornia crassipes (Mart.) Solms]. Academic Ryale des Sciences d'Outre-Mer, Brussel. 12(3): 1-120.
232
Black G. A., T. Dobzhansky, and C. Pavan. 1950. "Some attempts to estimate species diversity and populations density of trees in Amazonian forests." Bot Gaz 111: 413-425. Braendle, R., and R. M. M. Crawford. 1987. "Rhizome anoxia tolerance and habitat specialization in wetland plants." In: Plant Life in Aquatic and Amphibious Habitats. Special Publication Number 5 of the British Ecological Society, ed. R. M. M. Crawford (Blackwell Scientific Publications, Oxford), pp. 397-410. Budowski, G. 1961. Studies on forest succession in Costa Rica and Panama. Ph.D. thesis, Yale University, New Haven, p. 189Campbell, D. G., D. C. Daly, G. T. Prance, and U. N. Maciel. 1986. "Quantitative ecological inventory of terra firme and varzea tropical forest on the Rio Xingu, Brazilian Amazon." Brittonia 38(4): 369-393. Campbell, D. G., J. L. Stone, and A. Rosas. 1992. "A comparison of the phytosociology and dynamics of three floodplain (Varzea) forests of known ages, Rio Jurua, western Brazilian Amazon." Botanical Journal of the Limnean Society 108: 213-237. Carbiener, R., A. Schnitzler, and J. M. Walter. 1988. "Problemes de dynamique forestiere et de definition des stations en millieu alluvial." Colloques Phytosociologiques XIV 14: 655-686. Conserva, A. dos S. 1999. Biomassa, ciclo de vida e composicao quimica de duas populacoes de Paspalum fasciculatum Willd. Ex Fluegge (Poaceae) na Amazonia central. M.Sc. Thesis INPA/FUA, Manaus. p. 87. Crawford, R. M. M. 1992. "Oxygen availability as an ecological limit to plant distribution." Advances in Ecological Research. 23: 93-185. Ellenberg, H., R. Mayer, and J. Schauermann, editors. 1986. Okosystemforschung— Ergebnisse des Sollingprojektes 1966-1986. Ulmer, Stuttgart. Epstein, E., editor. 1972. Mineral Nutrition of Plants: Principles and Perspectives. John Wiley & Sons, New York. Ferreira, L. V. 1991- O efeito do periodo de inundacao na zonacao de comunidades, fenologia e regeneracao em uma floresta de igapo na Amazonia Central. M.Sc. Thesis INPA/FUA, Manaus. p. 207. Franken, M., U. Irmler, and H. Klinge. 1979. "Litterfall in inundation, riverine and terra firme forests of Central Amazonia." Tropical Ecology 20: 225-235. Frankie, G. W., H. G. Baker, and P. A. Opler 1974. "Comaparative phenological studies of trees in tropical wet and dry forests in the lowlands of Costa Rica." Journal of Ecology 62: 881-919Furch, K. 1984. "Seasonal variation of the major cation content of the varzea-lake Lago Camaleao, middle Amazon, Brazil, in 1981 and 1982." Verhandlungen International Vereinigen Limnologie. 22: 1288-1293. Furch, K. 1997. "Chemistry of Varzea and Igapo soils and nutrient inventory of their floodplain forests." In: The Amazonian Floodplains: Ecology of a Pulsing System, ed. W. J. Junk (Ecological Studies. Springer-Verlag, Berlin), pp. 47-68. Furch, K. 2000. "Evaluation of ground water input as major source of solutes in an Amazon floodplain lake during the low water period." Verhandlung International Vereinigen Limnologie. 7(1): 412-415
Maria Teresa F. Piedade et al. Furch, K., and W. J. Junk. 1985. "Dissolved carbon in a floodplain lake of the Amazon and in the river channel." In: Transport of Carbon and Minerals in Major World Rivers, eds. E. T. Degens, S. Kempe, and R. Herrera (Geology-Palaentology Institute, University of Hamburg, Hamburg), pp. 285-298. Furch, K., and W. J. Junk. 1992. "Nutrient dynamics of submersed decomposing Amazonian herbaceous plant species Paspalum fasciculatum and Echinochloa polystachya." Revue Hydrobiologie Tropical 25(2): 75-85. Furch, K., and W. J. Junk. 1997a. "Chemical composition, food value and decomposition of herbaceous plants leaves and leaf litter of the floodplain forest." In: The Central Amazon Floodplain, ed. W. J. Junk (Ecological Studies, Springer-Verlag, Berlin), pp. 187-203. Furch, K., and W. J. Junk. 1997b. "Physicochemical conditions in floodplains." In: The Central Amazon Floodplain, ed. W. J. Junk (Ecological Studies, Springer-Verlag, Berlin), pp. 69-108. Furch, K., and H. Klinge 1989: "Chemical relationships between vegetation, soil and water in contrasting inundation areas of Amazonia." In: Mineral nutrients in tropical forests and savanna ecosystems, ed. J. Procter (Blackwell Scient Publ, Oxford), pp. 189-204. Gentry, A. H. 1982. "Patterns of Neotropical plant species diversity." Evolotionary Biology 15: 1-84. Gentry, A. H. 1990. "Floristic Similarities and Differences between Southern Central America and Upper and Central Amazonia." In: Four Neotropical Rainforests, ed. A. H. Gentry (Yale University, New Haven), pp. 141-157. Gottsberger, G. 1978. "Seed dispersal by fish in the inundated regions of Humaita, Amazon." Biotropica 10: 170-183. Grime, J. P., editor. 1979. Plant Strategies and Vegetation Processes. John Wiley & Sons, Chichester. Halle, E, R. A. A. Oldemann, P. B. Tomlinson, editors. 1978. Tropical trees and forests. An architectural analysis. Springer-Verlag, Berlin. Irion, G., W. J. Junk, and J. A. S. N. de Mello 1997. "The large Central Amazonian river floodplains near Manaus: geological, climatological, hydrologgical, and geomorphological aspects." In: The Central Amazon Floodplain: Ecology of a Pulsing System, ed. W. J. Junk (Ecological Studies. Springer-Verlag, Berlin), pp. 23-46. Jordan, C. F. 1983- "Productivity of tropical rain forest ecosystems and the implications for their use as future wood and energy sources." In: Tropical rain forest ecosystems. Ecosystems of the world 14 A., ed. F. B. Golley (Elsevier Publ, Amsterdam, Oxford, New York ), pp. 117-136. Junk, W. J. 1986. "Aquatic plants of the Amazon system." In: The Ecology of River Systems, eds. B. R. Davies and K. F. Walker (Dr. W. Junk, Dordrecht Publishers, The Netherlands), pp. 319-337. Junk, W. J. 1989. "Flood tolerance and tree distribution in central Amazonian floodplains." In: Tropical forests: Botanical dynamics, speciation and diversity, eds. Holm-Nielsen et al. (Academic Press, London San Diego New York Berkely Boston Tokyo Toronto), pp. 47-64. Junk, W. J. 1993. "Wetlands of tropical South America." In: Wetlands of the World, eds. D. F. Whigham, S. Hejny, and D. Dykyjova (Kluwer Publishers, The Netherlands), pp. 679-739-
Geoecological Controls on Elemental Fluxes in Communities of Higher Plants in Amazonian Floodplains Junk, W. J. 1997a. "General aspects of floodplain ecology with special reference to Amazonian floodplains." In: The Central Amazon Floodplain, ed. W. J. Junk (Ecological Studies, Springer-Verlag, Berlin), pp. 3-20. Junk, W. J. 1997b. "Structure and function of the large Central Amazonian river floodplains: synthesis and discussion." In: The Central Amazon Floodplain, ed. W. J. Junk (Ecological Studies, Springer-Verlag, Berlin), pp. 455-472. Junk, W. J., and K. Furch. 1991. "Nutrient dynamics in Amazonian floodplains: decomposition of herbaceous plants in aquatic and terrestrial environments." Verhandlungen International Vereinigen Limnologie 24: 2080-2084. Junk , W. J., and C. Howard-Williams. 1984. "Ecology of aquatic macrophytes in Amazon." In: The Amazon: Limnology and landscape ecology of a mighty tropical river and its basin, ed. H. Sioli (Dr. W. Junk Publishers, Dordrecht, The Netherlands), pp. 269-293. Junk, W. J., and M. T. F. Piedade. 1993a. "Biomass and primary production of herbaceous plants communities in the Amazon floodplain." Hydrobiologia 263: 155-162. Junk, W. J., and M. T. F. Piedade. 1993b. "Herbaceous plants of the Amazon floodplains near Manaus: species diversity and adaptations to the flood pulse." Amazoniana XII(3/4): 467-484. Junk, W. J., and M. T. F. Piedade. 1993c. "Species diversity and distribution of herbaceous plants in the floodplain of the middle Amazon." Verhandlungen International Vereinigen Limnologie 25: 1862-1865. Junk, W. J., and M. T. F. Piedade. 1997. "Plant life in the floodplain with special reference to herbaceous plants." In: The Central Amazon Floodplain, ed. W. J. Junk (Ecological Studies, Springer-Verlag, Berlin), pp. 147-185. Junk, W. J., and G. Weber. 1996. "Amazonian floodplains: a limnological perspective." Verhandlungen International Vereinigen Limnologie 26: 149-157. Junk, W. J., P. B. Bayley, and R. E. Sparks. 1989. "The floodpulse concept in the river-floodplain systems." Canadian Special Publication in Fisheries and Aquatic Sciences. 106: 110-127. Keel, S. H., and G. H. Prance. 1979. "Studies of the vegetation of a white-sand black-water igapo (Rio Negro, Brazil)." Acta Amazonica 9: 645-655. Kira, T., H. Ogawa, K. Yoda, and K. Ogino. 1967. Comparative ecological studies on three main types of forest vegetation in Thailand. Nature Life SE Asia 5: 149-174. Klinge, H. 1975. "Bilanzierung von Hauptnahrstoffen im Okosystem tropischer Regenwald (Manaus). Vorlaufige Daten." Biogeographica 7: 59-77. Klinge, H., K. Furch, E. Harms, andj. Revilla. 1983. "Foliar nutrient levels of native tree species from Central Amazonia. 1. Inundation forests." Amazoniana 8: 19-45. Klinge, H., J. Adis, and M. Worbes. 1995. "The vegetation of a seasonal Varzea forest in the lower Solimoes river, Brazilian Amazonia." Acta amazonica 25: 201-220. Klinge, H., W. J. Junk, and J. C. Revilla. 1990. "Status and distribution of forested wetlands in Tropical South America." Forest Ecology and Management. 33/34: 81-101. Kubitzki, K. 1989. "The ecogeographical differentiation of Amazonian inundation forests." Plant Systematics & Evolution. 163: 285-304.
233
Kubitzki, K., and A. Ziburski. 1994. "Seed dispersal in floodplain forest of Amazon." Biotropica. 26(1): 30-43. Marmillod, D. 1982. Methodik und Ergebnisse von Untersuchungen iiber Zusammensetzung und Aufbau eines Terrassenwaldes im peruanischen Amazonien. Dissertation Gottingen. Medina, E., and H. Klinge. 1983. "Productivity of Tropical Forests and Tropical Woodlands." In: Lange, O.L., Nobel, P.S. Osmond, C. B. & Ziegler, H (eds.): Physiological plant ecology. Encyclopedia of plant physiology, Vol. 12D, Springer Verlad, Berlin, pp. 281-303 Mengel, K., and E. A. Kirkby, editors. 1987. Principles of plant nutrition. International Potash Institute, Bern. Oliveira, A.C. de. 1998. Aspectos da dinamica populacional de Salix martiana (Salicaceae) em areas de varzea da Amazonia Central. M.Sc. Thesis INPA/FUA, Manaus. p. 83. Piedade, M. T. F. 1985. Ecologia e biologia reprodutiva de Astrocaryum jauari Mart. (Palmae) como exemplo de populacao adaptada as areas immdaveis do rio Negro (igapos). M.Sc. Thesis INPA/FUA, Manaus. p. 187. Piedade, M. T. F., W. J. Junk, and S. P. Long. 1991. "The productivity of the C4 grass Echinochloa polystachya on the Amazon floodplain." Ecology 72: 1456-1463. Piedade, M. T. E, W. J. Junk, and J. A. S. N. de Mello. 1992. "A floodplain grassland of the central Amazon." In: Primary Productivity of Grass Ecosystems of the Tropics and Subtropics, eds. S. P. Long, M. B. Jones and M. J. Roberts (Chapman & Hall Publishers, London), pp. 127158. Piedade, M. T. E, S. P. Long, and W. J. Junk. 1994. Leaf and canopy CC>2 uptake of a stand of Echinochloa polystachya on the Central Amazon floodplain. Oecologia. 97: 159-174. Piedade, M. T. F., W. J. Junk, and S. P. Long. 1997. "Nutrient dynamics of the highly productive C4 macrophyte Echinochloa polystachya on the Amazon floodplain." Functional Ecology. 11: 60-65. Pires, J. M., and H. M. Koury. 1959- "Estudo de um trecho de mata de varzea proximo a Belem." Boletim tecnico do Institute) Agronomico do Norte Teen. 36: 3-44. Prance, G. T. 1989. "American Tropical Forests." In: Tropical Rain Forest Ecosystems. Ecosystems of the World 14B, eds. H. Lieth, and M.J.A. Werger (Elsevier Publ, Amsterdam, Oxford, New York, Tokyo), pp. 99-132. Prance, G. T., W. A. Rodrigues, and M. F. da Silva 1976. "Inventario florestal de um hectare de mata de terra firme km 30 da Estrada Manaus-Itacoatiara." Acta Amazonica 6(1): 9-35. Putz, R., and W. J. Junk. 1997. "Phytoplankton and periphyton." In: The Central Amazon Floodplain, ed. W. J. Junk (Ecological Studies, Springer-Verlag, Berlin), pp. 207-222. Revilla, J. D. 1981. Aspectos floristicos e fitossociologicos da floresta inundavel. Praia Grande, Rio Negro, Amazonas. M.Sc. Thesis INPA/FUA, Manaus. p. 129. Revilla, J. D. 1991- Aspectos floristicos e estruturais da floresta inundavel (varzea) do Baixo Solimoes, Amazonas, Brasil. Ph.D. Thesis, INPA/FUA, Manaus. p. 122. Roosevelt, A. C. 1999. "Twelve thousand years of humanenvironment interaction in the Amazon floodplain." Advances in Economic Botany 13: 371-392. Schliiter, U. B. 1989. Morphologische, anatomische und
234
physiologische Untersuchungen zur Uberflutungstoleranz zweier charakteristischer Baumarten (Astrocaryum jauari und Macrolobium acaciifolium) des Weils'- und Schwarzwasseriiberschwemmungswaldes bei Manaus. Diss Kiel. Sioli, H. 1975. "Tropical rivers as expressions of their terrestrial environments." In: Tropical ecological systems. Trends in terrestrial and aquatic research, eds. F. B. Golley, and E. Medina (Springer-Verlag, Berlin), pp. 275-288. Sioli, H. 1984. "The Amazon and its main affluents: hydrography, morphology of the river courses, and river types." In: The Amazon: Limnology and landscape ecology of a mighty tropical river and its basin, ed. H. Sioli (Dr. W. Junk Publishers, Dordrecht, The Netherlands), pp. 127-165. Swaine, M. D., and T. C. Whitmore. 1988. "On the definition of ecological species groups in tropical rain forests." Vegetatio 75: 81-86. Thompson, J., J. Procter, V. Viana, W. Milliken, J. A. Ratter, and D. A. Scott. 1992a. "Ecological studies on a lowland evergree rain forest on Maraca Island, Roraima, Brazil. I. Physical environment, forest structure and leaf chemistry." Journal of Ecology 80: 689-703. Thompson, J., J. Procter, V. Viana, W. Milliken, J. A. Ratter, and D. A. Scott. 1992b. "Ecological studies on a lowland evergree rain forest on Maraca Island, Roreima, Brazil. II. Litter and nutrient cycling." Journal of Ecology 80: 705-717. Tilman, D., editor. 1982. Ressource Competition and Community Structure. Princeton University Press, Princeton. Uhl, C., and C. F. Jordan. 1984. "Vegetation and nutrient dynamics during the first five years of succession following forest cutting and burning in the Rio Negro region of Amazonia." Ecology 65: 1476-1490. Vitousek, P. 1982. "Nutrient cycling and nutrient use efficiency." American Naturalist 1994: 553-572. Vitousek, P. and R. L. Sanford. 1986. "Nutrient cycling in moist tropical forest." Annual Review of Ecology and Systematics. 17: 137-167.
Maria Teresa F. Piedade et al. Weber, G. E. 1997a. Causes of hydrochemical seasonality of major cations in Lago Camaleao, a Central Amazonian floodplain lake. Verhandlungen International Vereinigen Limnologie. 26(2): 408-411. Weber, G. E. 1997b. "Modelling nutrient fluxes." In: The Central Amazon Floodplain, ed. W. J. Junk (Ecological Studies, Springer-Verlag, Berlin), pp. 109-118. Weber, G. E., and W. J. Junk. 1998. "Empirical results on groundwater seepage - a key determinant for hydrochemical seasonality of Lake Camaleao, a Central Amazon floodplain lake." Verhandlungen der Gesellschaft fur Okologie. 28: 485^91. Weber, G. E., K. Furch, and W. J. Junk 1996. "A simple modelling approach towards hydrochemical seasonality of major cations in a Central Amazonian floodplain lake." Ecological Modelling 91: 29-56. Whitmore, T. C., editor 1990. An introduction to tropical rain forests. Oxford University, Oxford. Worbes, M. 1983- "Vegetationskundliche Untersuchungen zweier Uberschwemmungswa'der in Zentralamazonienvorlufige Ergebnisse." Amazoniana 8(1): 47-65. Worbes, M. 1985. "Structural and other adaptations to long term flooding by trees in Central Amazonia." Amazoniana 9(1): 459-484. Worbes, M. 1986. "Lebensbedingungen und Holzwachstum in zentralamazonischen Uberschwemmungswaldern." Scripta Geobotanica 17: 1-112. Worbes, M. 1989. Growth rings, increment and age of trees in inundation forests, savannas and a mountain forest in the Neotropics. IAWA Bulletin, n.s. 10: 109-122. Worbes, M. 1997. "The forest ecosystem of the floodplains." In: The Central Amazon Floodplain, ed. W. J. Junk (Ecological Studies, Springer-Verlag, Berlin), pp. 223-265. Worbes, M., and W. J. Junk. 1989. "Dating tropical trees by means of 14C from bomb tests." Ecology 70: 503-507. Worbes, M., H. Klinge, J. D. Revilla, and C. Martius. 1992. "On the dynamics, floristic subdivision and geographical distribution of varzea forests in Central Amazonia." Journal of Vegetational Science 3: 553-564.
14 Biogeochemistry of Amazon Floodplain Lakes and Associated Wetlands John M. Melack and Bruce R. Forsberg
Floodplains and associated lakes are laboratories in several areas, use of modern important components of the biogeochem- ships capable of regional surveys and istry, ecology, and hydrology of the Amazon equipped for hydrographic studies, and basin. Amazon floodplains contain thou- applications of remote sensing. sands of lakes and associated wetlands Our objective in this chapter is to examine linked to each other and to the many rivers the role of lakes in the hydrology of the of the immense basin. These floodplain floodplain and in the biogeochemistry of lakes modify the passage of flood waves carbon, nitrogen, and phosphorous within (Richey et al. 1989a), increase nutrient the central Amazon basin. Particular retention and recycling (Melack and Fisher emphasis is placed on how inundation 1990), and influence the chemistry of the patterns interplay with carbon balance and rivers (Devol et al. 1995). The mosaic of nutrient limitation. By combining numerous flooded forests, open water, and floating measurements of primary productivity with macrophytes in the central Amazon flood- recent results from studies using isotopes plain makes a significant contribution of of carbon, we will examine the contribution methane to the troposphere (Bartlett et al. of the major plant groups to aquatic 1988, Devol et al. 1990). The fishery poten- foodwebs, and offer a new paradigm for the tial of the large river systems is closely tied processing of organic carbon on the to the area of floodplain and the magnitude Amazon floodplain. The interplay between and duration of inundation (Welcomme the Amazon River and local catchments 1979, Bayley and Petrere 1989). The majori- as sources of nutrients to the floodty of fishes harvested in the Amazon basin plain indicates the potential sensitivity obtain nutrition in flooded forests (Goulding of the lakes to basin-wide and local distur1980) or from organic matter derived from bances. We will focus our analysis on floodplain algae (Araujo-Lima et al. 1986, intensively studied lakes near Manaus (for example, Lake Calado, Melack and Forsberg et al. 1993). Much progress has been made during the Fisher 1990; Lake Camaleao, Junk 1997) last fifty years toward understanding the and on regional surveys conducted with lakes of the Amazon floodplain. Still, the consistent methods. When possible, we vast size of the Amazon basin poses will extrapolate to the basin scale. For challenges to limnologists working in the a review of the ecology of Amazon region. Recent research has been enhanced floodplains, we recommend Sioli (1984) and by the maintenance of functional floating Junk (1997).
236
John M. Melack and Bruce R. Forsberg
plain lakes have provided formal expressions for the temporally and spatially varying conditions. Weber et al. (1996) and Lakes are an integral component of the Weber (1997) present both a box-and-arrow Amazon floodplain, and, as defined here, conceptual picture and a mathematical encompass seasonally flooded areas, other model of cation fluxes in an Amazon than rivers and streams, that include open floodplain lake. Kern and Darwich (1997) water plus vegetated areas. As water levels provide an algorithm for calculation of vary, the proportion of open water to other inputs and outputs of nitrogen at Lake aquatic habitats changes considerably. Camaleao that incorporates temporally At lowest water, floodplain lakes may varying areas of aquatic and terrestrial be reduced to shallow, turbid pools and habitats. Further research is required to occasionally desiccate completely. During extend these efforts to additional elements, rising and high water, the lakes expand in such as phosphorus and carbon, and to area and encompass flooded forests and include additional biogeochemical processseasonal growths of emergent aquatic es and routes of exchange. The open water portion of floodplain macrophytes; as a result, terrestrial areas surrounding the aquatic environments lakes behaves limnologically much as other contract in area. Junk et al. (1989) define shallow tropical lakes (Melack and Fisher this seasonally flooded region as the Aquatic 1990, Melack 1996). Another major aquatic Terrestrial Transition Zone or ATTZ. In habitat within Amazon floodplain lakes is addition, runoff from the local catchment the floating meadow community: extensive and deposition of solutes and aerosols from beds of floating, emergent macrophytes that the atmosphere are components of the develop annually. Seasonally inundated biogeochemical system that influences forests are also ecologically important and extensive throughout the Amazon River floodplain lakes. Seven types of aquatic interfaces, with basin and occupy varying areas within the differing horizontal and vertical gradients lakes (Junk 1986, Junk 1993). The flooded in hydrochemical and physical characteris- plant communities have been classified tics, can occur in floodplain lakes: based on hydroperiod and hydrochemistry river—lake interface, direct rain and through- (Prance 1979), although a continuum fall-lake interface, upland stream-lake inter- of intermediate types occurs. A combination face, lake-lake connections, epilimnion- of flooding, anoxia, sedimentation, and hypolimnion interface, lake-sediment erosion appears to result in fewer species in porewater interface, and lakeshore—ground- inundated forests than in nearby uplands (Dumont et al. 1990). water interface. The strong seasonal variations in depth and extent of inundation and complex Geographical Features hydrologic linkages and interfaces characThe fringing floodplain along a 2600 km teristic of floodplain lakes have important influences on biogeochemical processes and reach of the Amazon River from 52.5°W to fluxes. Much of our discussion in subse- 70.5°W contains about 6500 lakes that vary quent sections will elaborate upon these considerably in shape and size; the lower effects. Recent efforts to model and 400 km of four major tributaries (Japura, calculate seasonal and annual rates of Purus, Negro, and Madeira) contain an biogeochemical processes in Amazon flood- additional 2320 lakes (Melack 1984, Sippel
Floodplain Lakes: A Conceptual Framework
Biogeochemistry of Amazon Floodplain Lakes and Associated Wetlands
et al. 1992; Fig 14.1, reaches 1 through 12). These censuses of lakes and descriptions of morphology are based on side-looking airborne radar (SLAR) imagery (e.g., RADAMBRASIL 1978) and represent the areas of open water excluding river channels and the mouthbays of the Tapajos and Xingu rivers, paranas, and lakes less than about 250 m across, which were too small to reliably distinguish on the 1:250,000 scale images and are not abundant. Most of the imagery was acquired during moderate to low water levels. Hence, the lake areas do not reflect the complete area encompassed by our definition of a floodplain lake or the ATTZ. Area subject to flooding estimated from geomorphological maps derived from the SLAR imagery occupies about 78,000 km2 along the Amazon mainstem and an additional 62,000 km2 along the lower reaches of the four major tributaries (Sippel et al. 1992; 52.5SW to 70.5SW, reaches 1 through 12). Islands comprise a small proportion of the total main-stem floodplain, except in the Amazon delta which is excluded from our
237
treatment. Seasonal changes in stage resulted in variations in inundated area (excluding river channels) from 10,000 km2 to 81,000 km2 along the Amazon River during the period from 1979 to 1987 based on passive microwave measurements (Sippel et al. 1998; Fig. 14.2). The average maximum inundated area (excluding the Amazon River) for the period from 1979 to 1987 was 67,900 km2; this area was computed by averaging the sum of the maximum inundated area for each reach (Fig. 14.1, reaches 1 through 12) in each year. These estimates of regional inundation significantly increase our ability to model biogeochemical processes responsive to changes in oxidation-reduction states or water levels such as exchanges of CH4 or CO2 with the atmosphere. Apportioning the areas of floodplain occupied by one of the three major aquatic habitats, i.e., open water, herbaceous macrophytes, and flooded forests, and determining their temporal changes remains a challenge. As approximations required for calculations of the organic carbon balance
Fig. 14.1 Central and lower Amazon River and its major tributaries; from Sippel et al. (1992). Numbered floodplain reaches indicated along bottom were used in Sippel et al. 1992; arrows mark the upriver limits of measurements along major tributaries.
238
John M. Melack and Bruce R. Forsberg
Figure 14. 2 Monthly estimates of total flooded area (lower line, inundated floodplain plus rivers and lakes) for the mainstem Amazon River (sum of reaches 1-12 in Fig. 14.1), plotted together with river stage at Manacapuru (upper line). Horizontal line indicates open water area. Data from Sippel et al. (1998).
of the floodplain (see below), we have utilized results from recent remote sensing studies of the Amazon basin. Sippel et al. (1992) estimated the open water area of lakes to be 10,370 km2. We based our partition of vegetated habitats on analyses of Landsat thematic mapper imagery (Melack et al. 1994; Mertes et al. 1995; Novo et al. 1997; L. Mertes and E. Novo, personal communication), synthetic aperture radar data (Sippel et al. 1992; Hess et al. 1995; Melack and Hess 1998; L. Hess, personal communication), and on our aerial videography and ground-based surveys. First, we subdivided the Amazon River and its floodplain into two segments because a major shift in lake morphology and forest cover occurs between approximately 58Q and 592W. The western section encompasses about 65% of the floodplain area and extends from 70.5s to 58.5QW (reaches 1 through 8, Fig. 14.1). The eastern section contains about 35% of the floodplain area and extends from 58.5s to 70.5QW (reaches 9 through 12, Fig. 14.1). Within the western
segment, open water covers about 4140 km2 and within the eastern segment about 6230 km2 (Sippel et al. 1992). During periods of moderate to high water levels, we estimate that the vegetated portion of the western section is dominated by herbaceous macrophytes over 30% and flooded forests over 70%. In the eastern section, herbaceous macrophytes are estimated to cover about 90% and flooded forests to cover about 10%. Based on these areas for open water in lakes and proportions of vegetated habitat, and on an average maximum inundated area of 67,900 km2, we estimate that the average maximum area occupied seasonally by herbaceous macrophytes is 29,300 km2 and that the average maximum area occupied seasonally by flooded forest is 28,200 km2. Further basin-wide analyses with new remote sensing data will surely modify and improve our estimates and permit additional refinements in floodplain classification. Following Sippel et al. (1992), we apply a simple morphological classification to Amazon floodplain lakes: ria or blocked-
Biogeochemistry of Amazon Floodplain Lakes and Associated Wetlands
valley lakes and lateral-levee lakes. Ria lakes are formed by flooding of the lower part of a tributary stream or river valley and usually have a levee bordering the main river channel. Lateral-levee lakes are subdivided by a length to width ratio of 5 into dish-shaped (ratio < 5) versus channelshaped basins (ratio > 5). All types are formed largely by fluvial processes. Based on geomorphology, most channel-shaped basins would be considered scroll bar lakes; some would be oxbow lakes, including the world's largest oxbow lake, Lake Aranacu, located between the Solimoes and Japura rivers. Dish-shaped lakes can be called depression lakes to indicate the formation of their basin by incomplete agradation of the post-glacial floodplain (Klammer 1984) or by tectonic subsidence (Mertes et al. 1996). As elaborated below, these differences in morphology lead to contrasting hydrologic and biogeochemical conditions.
239
The frequency of occurrence and area covered by the morphological classes of lakes varies by reach along the Amazon River (Sippel et al. 1992, Fig. 14.3). Much of the limnological research performed on Amazon floodplain lakes has been done in reach 7. Although lake abundances are similar in reaches 3 to 6, reach 5 has much more lake area than do the other three reaches. Reaches 10 and 11 contain the most lake area. Dish lakes are numerically the major class of lakes except in the upper four reaches, where channel lakes are of similar abundance. Dish lakes cover the largest areas, although the uncommon ria lakes are often large and make a significant contribution to total lake area, especially in reach 5. The frequency distribution of individual lake areas has a strong positive skew, fits a hyperbolic probability distribution, and has a statistical property known as self-similarity (Hamilton et al. 1991, Sippel et al. 1992).
Figure 14. 3 Abundance and total area of lakes along Amazon River, divided by lake class; from Sippel et al. (1992).
240
Hence, descriptive statistics such as total abundance per unit area will vary with the scale of the observations. The median open water area of lakes at low to moderate stage ranges from 0.12 to 0.44 km2 among river reaches, a size much smaller than that of most of the lakes studied by limnologists in the Amazon basin. Two major types of hydrologic and morphometric information are lacking for almost all Amazon floodplain lakes: stage data and bathymetric maps. Therefore, it is not possible to determine directly the seasonal and interannual changes in open water area, total flooded area, or volume of water resid-
John M. Melack and Bruce R. Forsberg
ing on the floodplain. Applications of remote sensing with passive microwave sensors (Sippel et al. 1998) and synthetic aperture radar (Hess et al. 1995) help alleviate this data gap, but field surveys will be required to obtain detailed bathymetric and stage data at specific lakes. Transport and distribution of sediments derived from Andean erosion across the cratonic landscape of Brazil have caused development of an upstream to downstream trend in geomorphic and hydrologic character of the channel-floodplain system (Mertes et al. 1996). In upstream reaches, scroll bar topography on the floodplain tends to
Fig. 14.4 Meteorological data measured on a floating platform moored on Lake Calado: air temperature (2-hr averages), 1983, and wind speed (1-hr averages), 1984; a short gap in the wind record occurred around day 275. J. M. Melack, unpublished data.
Biogeochemistry of Amazon Floodplain Lakes and Associated Wetlands
channelize floodwater, decreasing mixing of flooding river water with locally derived water. In middle and downstream reaches, flooding river water flows onto the floodplain as diffuse, nonchannelized overbank flow, and through drainage channels. In both middle and lower reaches considerable mixing of waters from rivers and local drainages occurs. Most limnological studies have been completed near the confluence of the Solimoes and Negro rivers, which combine at Manaus to form the Amazon River. These rivers are biogeochemically distinct systems. The Solimoes River, called a Whitewater river, is rich in dissolved nutrients and suspended sediments and has extensive, fertile floodplains. The Negro River is nutrient poor and contains high concentrations of dissolved organic carbon, hence it is called a blackwater river.
Climate and Physical Environment The central Amazon has a wet and warm climate with a drier period from June through November that corresponds with the northward movement of the intertropical convergence zone. Mean annual rainfall just north of Manaus during the period 1966 to 1983 was 2410 mm; the return period for wet (2800 mm) and dry (2000 mm) years was 10 years (Lesack and Melack 1991). Local differences in the distribution of rainfall have been reported (Ribeiro and Adis 1984), although Lesack and Melack's (1995) transect of rain gauges in the Calado basin did not detect differences. Mean annual temperature is between 26SC and 27QC, with August through November slightly warmer than the mean and January through April slightly cooler than the mean (Irion et al. 1997). Diurnal variations can exceed 10SC. Cool, southern air masses occasionally influence the central Amazon, and minima can fall below 20QC for a few days during the austral winter. Relative humidity is high year
241
round, averaging about 76% (Ribeiro and Adis 1984). Meteorological data are required to calculate evaporation and to model vertical mixing within the water column of lakes, critical aspects of the hydrological cycle and fluxes of nutrients. Meteorological records from lakes in the central Amazon floodplain are short and available from only two sites (Lakes Camaleao and Calado). At Lake Calado, air temperatures typically range from 24s to 32QC year round with occasional, brief cool periods, such as observed on days 164 to 168 in 1983 (Fig. 14.4). Insolation, measured as photosynthetically available radiation (PAR), varies as a function of cloudiness (mean + std. dev. = 33 + 12 Einsteins rrr2 d"1). Wind speeds are generally low (near 0 to 4 m s'1) during morning and around midday, but abruptly increase up to 10 m s'1 as squalls pass (Fig. 14.4). Underwater light attenuation in floodplain lakes is often high because of the high concentrations of suspended particles and dissolved organics (Schmidt 1973, Furch and Otto 1987). Hence, the euphotic zone available to phytoplankton is often shallow. For example, Secchi disk transparency in Lake Calado varied seasonally from ca. 50 to 150 cm (J. M. Melack and T. R. Fisher, unpublished data). Consequently, emergent trees and floating macrophytes with associated periphyton are better positioned to capture sunlight than phytoplankton, and much of the primary productivity in floodplain lakes occurs in nonpelagic habitats.
Hydrological and Hydrodynamic Aspects Regional hydrology The Amazon River has a damped annual hydrograph with a water-stage
242
amplitude of about 10 m in the central basin and a 2-to-3-year cycle of interannual variation coupled to the El Nino Southern Oscillation (Richey et al. 1989b). Linked to the changes in stage are a rise and fall in flooded area, in which lakes are an important component. Large volumes of river water are stored in floodplain lakes during seasonal floods and then return to the main channel when the river level falls. In addition, the discharge of many tributaries passes through lakes, and a net flux of water from the floodplain to the river can occur even when river level is rising (Richey et al. 1989a, Lesack and Melack 1995). Further, lakes and their associated floodplain vegetation are significant sources of water vapor to the atmosphere, especially at the beginning of the dry season when inundated area is largest (Victoria et al. 1991). In a comparison of 51 lakes located along the central Amazon River, Forsberg et al. (1988) used alkalinity as a tracer of sources of water; the Amazon River has relatively high alkalinity, while the streams draining the lowland forests have relatively low alkalinity. During high water from mid-March to mid-September, lake and river alkalinities were similar, indicating that lakes contained primarily river water in most cases. The exceptions were ria lakes with large local drainage basins. During low water from midSeptember to mid-March, the influence of rivers diminished and local inputs to lakes increased in proportion to the ratio between local basin area and lake area (BA:LA). Channel and dish-shaped basins, which typically have small local catchments, appeared to receive small to moderate local runoff and had variable alkalinities at the end of low water. Alkalinities in lakes with BA:LA > 20, such as many ria lakes, approached the average values for lowland runoff by the end of low water. Cor-responding to these different sources of water are differences in concentrations and fluxes of nutrients (see below).
John M. Melack and Bruce R. Forsberg
Hydrological balance of a floodplain lake For any period of time, the water balance of a floodplain lake can be expressed by the equation:
where AS is the change in volume of water stored in the lake, P is rainfall onto the surface of the lake, R is inflowing upland runoff, L is exchange of water in either direction through connections to adjacent lakes, H is exchange of water in either direction through channels to an adjacent river, G is exchange with the surrounding groundwater system, and E is evaporation from the lake; units are volumes (Lesack and Melack 1995). During the course of the year, each of the inflows and outflows in equation 1 changes in response to temporal variations in weather and interactions with changes in stage, surface area, and volume of the lake. Evaluation of the water balance for a floodplain lake requires precise bathymetry of the basin and careful measurements or calculations of all the variables. Bathymetric data are available for very few Amazon lakes, and appropriate meteorological and hydrological measurements in floodplain lakes and their catchments rarely have been made. The first and only detailed analysis of the terms in equation 1 for an Amazon floodplain lake was performed at Lake Calado (Lesack and Melack 1991, 1995; Lesack 1993a, 1995). Lake Calado has a BA:LA ratio ranging from about 7 at high water to about 28 at low water. River water invaded the lake at the start of rising level in the mainstem, but by midrising water, lake water steadily flowed from the lake into the river, while river levels continued to rise. By the end of the water year, Lake Calado had experienced a 10 m range in water level, and local runoff had contributed 57% of the total water input, river inflow 21%, rainfall directly onto the lake
Biogeochemistry of Amazon Floodplain Lakes and Associated Wetlands
243
11%, inflow from an adjacent lake 6% and Flow paths of inflows to floodplain lakes seepage 4%. In contrast to the common view that flooding from river channels is expected to play the major role in the annual flushing Inflow of the turbid, nutrient-rich water of of lakes along floodplains, at Lake Calado, the Amazon River onto the floodplain helps local rainfall and local runoff are the sustain the floodplain's high fertility (Fisher dominant sources of water. Further evidence and Parsley 1979, Setaro and Melack 1984, for seasonal variation in the mixture of Engle and Melack 1993). The inflows follow riverine and upland runoff water in Lake complex flow paths that account, in part, for Calado is provided by an analysis of the differences in productivity among floodplain isotopic composition of the water (Martinelli lakes. For example, Melack et al. (1992) anaet al. 1996). Consider-ably more information lyzed a rare cloud-free Landsat Multispectral about the hydrology of other types of flood- Scanner (MSS) image of the initial stages of plain lakes is needed before generalizations riverine inundation by the Amazon River are possible. which illustrated the complexity of the flow
Fig. 14.5 Limnological gradients in Lake Janauaca; from Fisher and Parsley (1979).
244
John M. Melack and Bruce R. Forsberg
Table 14.1 Limnological conditions and areal coverage in Lake Janauaca, 9 December 1976.
Water type
Area (km2)
Seston (mg Lr1)
Chlorophyll a (jig L'1)
Amazon River Mixing zone Lake water
20 24 22
150 ± 42 41 ± 1 15 ± 2
2.1 ± 0.4 38 ±7 49 ±7
Source: From Melack et al. (1992)
paths. By chance, the date of acquisition of the image coincided with the limnological studies of Fisher and Parsley (1979) at Lake Janauaca. Hence, the gradients of turbidity in the lake could be interpreted in terms of limnological conditions (Fig. 14.5, Table 14.1). In the channel leading from the river to the lake, near the Amazon River, the water is high in suspended solids and dissolved phosphate and nitrate but low in phytoplankton. Further into the lake, transparency increases, as indicated by deeper Secchi depth and lower suspended sediment concentration, and dissolved phosphate and nitrate decline as phytoplankton abundances increase. Mertes et al. (1993) estimated concentrations of suspended sediment in the surface waters of the Amazon River and adjacent floodplain from Landsat images. They documented patterns of decreasing concentration from the main channel onto the floodplain and through channels. Rates of sediment transport into the floodplain were estimated by combining the Landsat-derived sediment concentrations with hydraulic calculations (Mertes 1994). As lake levels rise and inundate upland stream channels, water and solutes from streams often pass through flooded vegetation before reaching open water in lakes. At Lake Calado (Alves 1991; Fig. 14.6), the cool waters (24QC) of Mota Brook flowed beneath the warm lake water (about 28QC). After an initial mixing in the first 50 m, stream water
high in nitrate remained distinct from lake water low in nitrate for about 400 m. Alves (199D attributed consumption of nitrate within the inundated channel to plankton, periphyton, and benthos. Further advances in understanding of how riverine and upland stream inflows interact and influence the biogeochemistry of floodplain lakes will require the merging of remote sensing data and hydraulic and hydrological modeling with information on biogeochemical processes.
Vertical mixing within floodplain lakes Hundreds of vertical profiles of temperature and dissolved oxygen obtained in Lake Calado permitted an analysis of the frequency of mixing (Melack and Fisher 1983, Maclntyre and Melack 1984, 1988), an important factor influencing nutrient supply to the euphotic zone and oxidation-reduction conditions at the sediment-water interface. When water depths •were less than 3 to 4 m, diel mixing from top to bottom was usual. When depths exceeded 5 m, anoxic water developed below a thermocline at about 3 m, and oxycline displacements downward of 1 to 2 m occurred every 3 to 5 days in conjunction with mixing events. When depths exceeded 8 m, mixing to the bottom was rare. Detailed analyses of heat inputs and losses, wind speeds, and thermal structure indicated that evaporative heat losses resulting in convective mixing predominated over direct, wind-induced mixing (S. Maclntyre and J. M. Melack, unpublished data).
Biogeochemistry of Amazon Floodplain Lakes and Associated Wetlands
245
Figure 14. 6 Stratification and horizontal variation from Mota Brook to Lake Calado during high water; SLI is stream-lake interface; from Alves (1991).
246
John M. Melack and Bruce R. Forsberg
Biogeochemical Processes and Fluxes: macrophytes, flooded forest trees, and periphytic algae. The distribution and production Carbon Dynamics Floodplain lakes and their associated wetland habitats play an important role in the organic carbon balance of the Amazon River system. Lakes are the main sites of aquatic plant production, the principal source of organic carbon for aquatic food chains. They are also a major source of methane and other biogenic gases to the troposphere. The temporary storage and transformation of riverine organic matter in lacustrine environments contributes significantly to the downstream spiraling of carbon along the Amazon River. We will describe principal aspects of the organic carbon cycle in Amazon lakes, emphasizing the dominant fluxes, the factors that control them, and their contribution to the organic carbon balance at local, regional and global scales.
Primary production Four groups of plants contribute to autochthonous primary production in floodplain lakes: phytoplankton, herbaceous
dynamics of each of these plant groups varies within and between lakes in response to spatial and temporal variations in lake morphology, geochemistry, and flooding pattern.
Phytoplankton Phytoplankton communities are limited generally to the open waters of lakes where underwater light availability is sufficient. The Amazon River, while rich in nutrients, is too turbid to support significant levels of planktonic production (Wissmar et al. 1981). As the river flows onto the floodplain, sediments settle from the water column and light availability and phytoplankton production can increase (Schmidt 1973, Fisher and Parsley 1979, Forsberg et al. 1988; Fig. 14.5). For example, more than 90% of the suspended material entering Lake Calado from the Solimoes River was lost from the upper water within 2 km of the river during periods of channelized and overbank flow (Engle and Melack 1993). Phytoplankton populations in Lake Calado had moderate abundance (1 to 10 pg chlorophyll L'1) and
Table 14.2 Average values of phytoplankton production parameters for 36 Amazon floodplain lakes. Lakes Associated With:
n
Pmax
Ed
TP
TN
TSS
TN/TP
River type Whitewater Black- and clearwater
959 366
114 48
3.0 2.9
37 18
460 472
33 19
17 36
River stage High water Low water Low water
894 684
82 124 124
2.2 2.2 4.1 4.1
25 42
468 473
19 43
25 19
Lake average
818
98
2.9 2.9
32
470
28
30
Source: B. R. Forsberg (unpublished). '11 = daily integral photosynthesis (mg C nr2 d'1), Pmax = light saturated rate of photosynthesis (mg C irr3 d'1), Ed = vertical attenuation coefficient of PAR (nv1), TP = total P (mg nr3), TN = total N (mg nr3), TSS = total suspended sediments (mg L-1) and TN/TP = total N to total P ratio by weight.
Biogeochemistry of Amazon Floodplain Lakes and Associated Wetlands
247
modest gross photosynthesis (0.4 to 1.2 g C This variability makes it difficult to calculate rrr2 cH) (Melack and Fisher 1983, Fisher et the total annual production of individual al. 1988a, Smith-Morrill 1987). Seasonal min- lakes and to estimate the combined contriima in phytoplankton biomass and produc- bution of phytoplankton in all lakes to total tion tended to occur during low water when floodplain production. A reasonable the water was turbid from resuspension approach is to use the total lake area of of sediments. Annual gross planktonic 10,370 km2 determined by Sippel et al. production was estimated to be about 300 g (1992) for the central Amazon floodplain. C rrr2 y1 in Lake Calado. A similar rate of Since images used by Sippel et al. (1992) were annual production was reported for Lake acquired at different times during the year, Castanho by Schmidt (1973). their value is assumed to approximate average Daily integral rates of phytoplankton open water area. The average areal producproduction, (11) are known to vary inversely tion value of 300 t C krrr2 yr1 determined by with the vertical light attenuation coefficient B. R. Forsberg reflects the proportional area (Ed) and as a direct positive function of the of different lake types along the central floodlight saturated rate of photosynthesis (Pmax) plain. Using these two values, we estimate the (Tailing 1957, Vollenweider 1965), reflecting total gross production of phytoplankton on the effects of light limitation and nutrient or the central Amazon floodplain to be 3.1 Tg C temperature limitation (Smith 1979), respec- yr1, where Tg indicates teragrams or 1012 g. tively. B. R. Forsberg (unpublished data) has Converting this value to net production for investigated the influence of variation in Ed better comparison to production estimates for and Pmax on levels of primary production in herbaceous macrophytes, flooded forests and 36 lakes associated with the Amazon River periphtyon (see below), is difficult. Although and its major tributaries in the central Amazon approximate conversions are possible, they floodplain (Table 14.2). Daily integral gross are confounded by assumptions about rates of production, averaged over all lakes, was 0.82 respiration in the dark versus light, bottle g C rrr2 d'1, which corresponds to an annual effects on measurements of photosynthetic production rate of about 300 g C rrr2 yr1. rates which usually reduce rates, and difficulDaily integral production was higher in flood- ties calculating integrated daily rates. We judge plain lakes associated with Whitewater rivers our estimate of gross production by phytodue primarily to higher Pmax levels. The plankton to be not more than 50% above net average light saturated rate of photosynthesis production, that is, net production is about in Whitewater lakes was more than twice 2 Tg C yr1. that encountered in lakes fed by clear and blackwater rivers. Daily integral production Herbaceous macrophytes tended to be lower during low water, and this Large expanses of floating macrophytes was attributed to a lower light availability (i.e., higher average Ed) associated with develop and decay each year on the floodresuspension of sediments. The average H plain of the central Amazon basin (Junk 1970 value determined for Whitewater lakes (0.96 g 1997). Herbaceous macrophytes are especialC rrr2 d'1) corresponded to an annual rate of ly abundant in lakes associated with 350 t C km-2 yr1. Whitewater rivers, where nutrient-rich Total daily phytoplankton production can sediments and waters stimulate their developvary considerably in Amazon floodplain ment. A seasonal succession of herbaceous lakes due to large seasonal changes in plant communities occurs in lakes as the surface area linked to water-level variations. ATTZ expands and shrinks in synchrony with
248
the annual flooding cycle (Junk and Piedade 1997). At lowest water, the surface area of many lakes is reduced and much of the lake bed is dry mud flats. These dry areas are colonized by a variety of perennial and annual macrophytes. Some of these plants (e.g., Alternanthera pilosa, A. brasiliana, Paspalum conjugation, Ludwigia densiflora and Sorghum arundinaceum) are exclusively terrestrial, while others (e.g., Echinochloa polystachya, Hymenachne amplexicaulis, Leersia hexandra, Oryza perennis, Paspalum repens, and Montrichardia arborescens) have both terrestrial and aquatic growth phases. All of these species are rooted in the exposed lake sediment during the terrestrial phase. When flooded, the obligate terrestrial species die and decompose while transitional species continue growing as emergent aquatic macrophytes and dominate the herbaceous plant community during the early rising water period. As the water level continues to rise, some of these species (for example, P. repens and L. hexandra) become free-floating and, together with other obligate floating species (e.g., Pistia stratiotes, Scirpus cubensis, Eichhornia crassipes, and Salvinia auriculata), form an extensive floating meadow community Qunk 1970). Rooted, submerged macrophytes are generally absent and submerged unrooted species (e.g., Utriculariafoliosd) are restricted by the shallow euphotic zones encountered in floodplain lakes. When the water level falls, the aquatic macrophyte community begins to decompose; annual species generally die and perennial species persist as rhizomes. The cumulative, sequential net production of terrestrial, semiaquatic, and aquatic macrophyte communities is all potentially available to the lacustrine ecosystem. The seasonally changing biomass of floating macrophytes was determined at Lake Calado with a combination of sampling and airborne videography (Fisher and Moline 1992; T.R. Fisher, personal communication; Fig. 14.7). Further analysis of these data
John M. Melack and Bruce R. Forsberg
indicate that the biomass of the macrophytes turned over 2 to 3 times per year. Casual observations of the area covered by floating plants at Lake Calado during the 1980s suggest large variations occur with low coverage in years with high floods and possibly increased human activities reducing abundance, except where deforestation of the floodplain has increased habitat for macrophytes. While many species of macrophytes in Amazon floodplain lakes are C-3 species, a
Figure 14.7 Periphyton and macrophyte biomass and areal coverage at Lake Calado; dominant genera are listed in bottom panel; T. R. Fisher, D. Engle, and R. Doyle, unpublished data.
Biogeochemistry of Amazon Floodplain Lakes and Associated Wetlands
few C-4 grasses (especially Echinochloa polystachya and Paspalum repens) frequently dominate community production and biomass (Junk and Piedade 1997). Growth in C-4 plants is efficient under drought stress conditions (Teeri and Stowe 1976), and this characteristic gives C-4 plants a competitive advantage over C-3 plants during the terrestrial growth phase. C-4 plants also tend to have a higher photosynthetic efficiency than C-3 plants under the high light conditions which prevail in the tropics. Piedade et al. (1991, 1994) measured an average photosynthetic efficiency of 2.3 g dry wt MJ"1 and an average energy conversion of 4% in natural stands of Echinochloa polystachya growing in Lake Camaleao, one of the highest efficiencies measured for a vascular plant. The high rates of photosynthesis in C-4 plants allow them to grow faster during the rising water period and colonize deeper water than C-3 plants. The combination of high growth rates and high drought efficiency thus provides a competitive advantage to C-4 plants through the year and appears to explain their dominance in the herbaceous plant community of Amazon floodplain lakes. The complex structure and seasonal dynamics of herbaceous macrophyte communities make it difficult to estimate their total annual contribution to floodplain lake production. Annual production estimates must incorporate the cumulative, sequential production of terrestrial, semiaquatic and aquatic plant communities and the spatial and temporal variation in their distributions. To date all measurements of macrophyte production have been made in a limited area on the central Amazon floodplain near Manaus. Only a few of these estimates have included contributions of more than one species. Junk and Piedade (1993) estimated the cumulative biomass increase of three successive macrophyte communities (terrestrial, semiaquatic, and aquatic) growing under
249
favorable conditions on the central Amazon floodplain to be 3000 t dry wt. km"2 yr1. Assuming a monthly biomass loss of 10-25% during the growing season, they estimated net annual primary production to be 5000 t dry wt km~2 yr1. Assuming a dry weight carbon content of 50%, this represents a net annual production of 2500 t C km"2 yr1. Using 29,300 km2 as the area of floodplain potentially covered by herbaceous macrophytes (see section on geographical features) together with the cumulative, net areal production estimated above (2500 t C km"2 yr1), we estimate the total contribution of herbaceous macrophyte communities to floodplain net production along the central Amazon to be about 73 Tg C yr1.
Flooded forest trees Floodplain forests often occupy a significant portion of the littoral region of Amazon floodplain lakes. Their distributions are generally restricted to the shallower portions of lakes where the annual flooding period is less than 230 - 270 days (Junk and Piedade 1997). The biomass and species diversity of floodplain forests vary spatially as a function of geochemical and flooding characteristics. Inundation forests associated with Whitewater and blackwater rivers are taxonomically distinct (Kubitzki 1989, Filoso 1996) and have been classified as varzea and igapo forests, respectively (Prance 1979). Varzea forests tend to be more productive and have a lower species diversity than igapo forests due to the higher nutrient inputs from Whitewater rivers (Worbes 1997). Species diversity and biomass also tend to diminish within a given forest with increasing annual inundation period, reflecting the physiological stress of long-term root inundation (Ferreira 199DLocal variations in floodplain forest community structure have been identified as successional sequences related to riverine
250
John M. Melack and Bruce R. Forsberg
depositional and erosional processes Periphyton (Worbes et al. 1992, Salo et al. 1986). Net production estimates for Amazonian Periphytic algae require solid substrata floodplain forests have been based primari- and adequate light levels for growth. They ly on litterfall data collected at sites near are generally found near the water surface Manaus (Junk 1985, Bayley 1989). Litter in the littoral regions of lakes attached to production rates of 670 and 1030 t dry wt. the submerged portions of emergent knr2 yr1 have been estimated for igapo macrophytes and flooded forest trees. The (Adis et al. 1979) and varzea forests (J. Adis, highest biomass is usually encountered at unpublished data cited in Worbes 1997), the interface between flooded vegetation respectively. The higher litterfall in varzea and open water areas where light availabiliforests has been attributed to the higher lev- ty is greatest. However, significant densities els of phosphorus and potassium present in can also occur in the middle of macrophyte soils (Worbes 1997). In addition to fine and forest stands when light penetration is litterfall, large woody detritus and root pro- sufficient for growth. Periphyton communiduction can contribute significant amounts ty structure has been found to vary among of carbon to floodplain lakes. Worbes (1997) lakes associated with geochemically distinct estimated the combined production of fine rivers. In a study of periphytic algae grown litter and large woody detritus to vary from during in situ incubations on roughened 1600-2500 t dry wt km-2 yr1 in well devel- cellulose-acetate, Putz and Junk (1997) oped varzea forests. He estimated root found differences in species composition production to be 30% of the live wood between white- and blackwater rivers. The increment (that is, 730 and 860 t dry wt knr2 periphyton communites in blackwater yr1) or 219 to 258 t dry wt km-2 yr1. We environments were generally dominated by assume that fine root production and some diatoms throughout the year while those in coarse litter derived from roots is available to Whitewater contained predominantly green benthic organisms. However, we did not algae during high water with diatoms and include Worbes1 measurements of increment cyanophytes increasing in importance at in wood or his estimates of aerial grazing lower water levels. The increased imporlosses in our production values because we tance of nitrogen fixing cyanophytes at low considered that production unavailable to water provided evidence for the existence aquatic organisms. The total contribution of of nitrogen-limiting conditions in whitewavarzea forests to floodplain lake production ter lakes at low water. The underwater stems and roots of the thus varied between about 1820 - 2760 t dry wt km-2 yr1. Taking the average of these two floating plants, so common in lakes such as values and assuming a dry weight carbon Calado, are colonized by periphyton. These content of 50% results in an average net epiphytic algae, whose biomass per unit production rate of 1150 t C km'2 yr1. This area can exceed that of phytoplankton production, which includes inputs during (Engle and Melack 1990), are responsible both flooded and dry conditions, is all poten- for high rates of primary productivity (Doyle tially available to the lacustrine ecosystem. 1991) and can be an important food for fishApplying this value to a maximum high- es (Forsberg et al. 1993). Doyle (199D water flooded forest area of 28,200 km2, measured using closed, recirculating chamwe estimate the total contribution of flooded bers in situ photosynthesis and respiration forests to production along the central of intact periphyton growing on macrophyte Amazon floodplain to be about 32 Tg yr1. roots and stems. Biomass-specific photosyn-
Biogeochemistry of Amazon Floodplain Lakes and Associated Wetlands
thesis rates were high, and biomass-specific respiration rates averaged 37% of biomassspecific photosynthesis rates, indicating considerable heterotrophic activity in the periphyton. Depth- and time-integrated annual, gross primary productivity for the periphyton attached to floating macrophytes in Lake Calado for 1989 was 1.2 t C km'2 d'1 (Doyle 199D. Converting this value to net primary productivity is confounded by the difficulty of distinguishing the considerable heterotrophic respiration from autotrophic respiration. Based on Doyle's (199D measurements of community respiration and biomass increments of periphyton, we estimate net primary productivity to be at least half of gross productivity. Periphyton accumulations on leaves and branches within flooded forests may be lower than those in floating meadows. For example, Doyle (199D reported a mean of about 40 mg chlorophyll nr2 in the periphyton of floating meadows of Lake Calado, while Alves (1991) measured an average of 10 mg chlorophyll rrr2 in the periphyton of the flooded forest near an upland stream in Lake Calado. However, more data on periphyton biomass from a variety of sites are needed before generalizations are warranted. Putz (1997) estimated production as uptake of 14C by periphyton that grew on artificial cellulose-acetate substrata suspended in several habitats near Manaus. Her results could not be used to compute an areal value for periphyton associated with floating macrophytes due to the complex surface of the natural substrata. To extrapolate her results to areal rates in floodplain forests requires an estimate of leaf area in the euphotic zone per unit area of water. Alves (199D reported between about 0.5 and 1.5 m2 of leaves in the euphotic zone of flooded forest per m2 of water; for simplicity we used 1.0. Based on net productivities measured by Putz (1997) at a site dominated
251
by Solimoes River water in Lake Camaleao and at a site receiving a mixture of Solimoes and Negro river water in Lake Catalao, we estimate average areal periphyton production in varzea forests to be 0.76 t C krrr2 d'1. The contribution of periphytic algal communities associated with herbaceous macrophytes and floodplain forests to lacustrine production along the central Amazon floodplain was determined by multipying the maximum area of each habitat (28,000 and 29,300 km2 for floodplain forest and herbaceous macrophytes, respectively) by the corresponding daily areal production rate and the average estimated inundation period. Since most lakes along the central Amazon floodplain are associated with the main or side channels of the Amazon River, the areal production values of Doyle (1991) for macrophyte habitat (1.2 t C km'2 d"1) and Putz (1997) for varzea forest (0.76 t C km'2 d'1) were considered appropriate. The maximum inundation period for the most flood-tolerant trees in flooded forests along the central Amazon floodplain has been estimated as 270 days (Junk and Piedade 1997). Assuming an average inundation period of half this value (135 days), the annual areal and total regional, net production of periphyton communities associated with floodplain forests were estimated to be 100 t C km'2 yr1 and about 3 Tg C yr1, respectively. Aquatic macrophyte communities can be encountered throughout the year in Amazon floodplain lakes Qunk and Piedade 1997). Assuming an average inundation period of half a year (182.5 days), the annual areal and total regional, gross production of periphyton associated with herbaceous macrophytes were estimated at 220 t C km'2 yr1 and about 6 Tg C yr1, respectively; total regional, net production of periphyton associated with herbaceous macrophytes is about 3 Tg C yr-1. Hence, the regional, net production of periphytic
John M. Melack and Bruce R. Forsberg
252
Table 14.3 Regional net production estimates for four plant groups contributing to primary production on floodplain of the Amazon River between 70.5QW and 52.5SW.
Plant Group
Total area (km2)
Phytoplankton Herbaceous macrophytes Floodplain forest Periphytic algae: flooded forest Periphytic algae: aquatic macrophytes
10,370 29,300 28,200 28,200 29,300
Total production
Regional Net Production Tgyr-i(%) 2 (2%) 73 (65%) 32 (28%) 3 3 (5%)
113
* See text for information regarding individual estimates.
algae in both habitats was estimated to be about 6 Tg C yr-1.
Combined primary production A summary of the regional, net production rates estimated for the four main plant groups encountered in Amazon floodplain lakes is given in Table 14.3. Considering the different methods used to estimate production in the field, the limited number of measurements available, and the various assumptions used to determine these values they can only be considered rough estimates. They do incorporate new data on plant production and habitat distributions which were unavailable in earlier analyses (e.g., Junk 1985, Bayley 1989) and provide an improved basis for evaluating the relative contribution of each plant group to lacustrine primary production. However, estimates for periphyton, aquatic macrophytes, and flooded forests, in particular, are based on a small number of measurements in a restricted geographical area and may not represent basin-wide averages. For the entire floodplain herbaceous macrophytes accounted for the largest share of total primary production, 65%, followed by floodplain forests, periphyton and phytoplankton which accounted for 28%, 5% and
2%, respectively. The order of contribution (macrophytes, flooded forest, algae) is in general agreement with that reported by Junk (1985) and Bayley (1989) in earlier analyses of floodplain carbon dynamics. However, the proportional contributions of the plant groups differ. The relative contribution of floodplain forests is higher in the present analysis due to the inclusion of root increment and dead wood fall in the production estimate. The proportional input from periphyton communities is also higher due to the higher areal production rates and larger habitat areas used in our estimates. Our estimate of the proportional contribution from phytoplankton production is less than the values derived by Junk (1985) and Bayley (1989).
Allochthonous inputs In Amazon floodplain lakes, daily water column respiration usually exceeds planktonic gross photosynthesis, and dissolved oxygen is usually undersaturated (Melack and Fisher 1983). Wissmar et al. (1981) and Richey et al. (1988) found Amazon floodplain lakes to be consistently undersaturated in O2 and supersaturated in CO2. The deficit of oxygen in these systems is a consequence of high respiration to production ratios.
Biogeochemistry of Amazon Floodplain Lakes and Associated Wetlands
In an analysis of planktonic production and respiration rates measured by incubating samples in bottles from 30 Amazon floodplain lakes, B. R. Forsberg (unpublished data) found an average respiration to production ratio of 5.4 for oxygen and 6.5 for carbon. The predominance of respiration in these systems is due, in part, to the input of significant quantities of allochthonous carbon from riverine and upland sources. The main cause for the imbalance is the predominance of aerial photosynthesis and aquatic respiration in the plant communities. Emergent macrophytes and flooded forest trees, which account for the bulk of lacustrine production, both have aerial photosynthetic parts and respire predominantly through submerged roots (Hamilton et al. 1995). Hence, these plants contribute organic carbon to the aquatic environment and consume dissolved oxygen, but they return almost no oxygen to the water column, which results in a major deficit of O2 and excess of CO2. Quay et al. (1995) reported lower respiration to production ratios (1 to 1.7) in seven Amazon floodplain lakes. These ratios were derived from stable isotope ratios of dissolved oxygen in surface waters where light availability and photosynthesis are often high. The higher ratios calculated by B. R. Forsberg were based on depthintegrated rates and thus more accurately reflect the total pelagic carbon balance. Inputs of organic carbon to floodplain lakes from the Amazon River are largest during rising water and reach a cumulative maximum at high water. Using alkalinity as a hydrological tracer, Forsberg et al. (1988) calculated that many lakes along the central Amazon floodplain contain predominantly river water at high water, but the larger lakes often are ria lakes with a considerable contribution of water from uplands (Lesack and Melack 1995). Based on Sippel et al. (1992), ria lakes cover about 2400 km2 and other lakes the remaining 7970 km2. Based
253
on Forsberg et al. (1988) and Lesack and Melack (1991), we estimate that 90% and 30% of the water in nonria and ria lakes, respectively, at high water, is derived from the Amazon River. Using an average highwater lake depth of about 7 m (n = 86, B. R. Forsberg, unpublished data) and the appropriate areas and percent contributions for ria and nonria lakes, we estimate a cumulative annual input of 55 x 109 m3 of river water to open water in lakes. In the remaining 57,530 km 2 of floodplain occupied by flooded forests and floating macrophytes at high water, after approximating the average depth to be 4 m based on numerous sampling transects into these habitats and estimating that 80% of the water is derived from the Amazon River in these habitats based on examination of optical remote sensing imagery, we calculate an annual input of 184 x 109 m3 of river water. The sum of riverine inputs to the whole floodplain is 239 x 109 m3. The particulate and dissolved organic carbon in the Amazon River have been shown to be predominantly terrestrial in origin and highly refractory (Hedges et al. 1986, Ertel et al. 1986). The average concentrations of coarse particulate organic carbon (CPOC > 63 pni), fine particulate organic carbon (FPOC < 63 urn), and dissolved organic carbon (DOC) in the Amazon River were determined by Richey et al. (1990) to be 0.5, 2.7, and 3.6 g C nr3, respectively. Using this average composition and the cumulative water storage on the floodplain, we estimate the total riverine, organic carbon input to the floodplain to be 1.6 Tg C yi. The distribution on the floodplain of particulate carbon derived from riverine inputs depends on its route of entry. In an investigation of sedimentation patterns during diffuse overbank flow onto the central Amazon floodplain, Mertes (1994) estimated that > 80% of the riverine particu-
254
lates entering the floodplain were deposited within a few hundred meters of the river-floodplain boundary. Riverine sediments entering lakes through discrete channels tend to penetrate further into floodplain lakes. For example, during an early rising water period, Fisher and Parsley (1979) found evidence of turbid river waters penetrating over 30 km into Lake Janauaca. Riverine DOC is expected to behave conservatively and to be distributed throughout the floodplain. In a stable isotope study of autotrophic carbon sources for heterotrophic bacteria in Lake Calado, Waichman (1996) found that DOC in all lacustrine habitats was derived predominantly from C-3 plants and was refractory, characteristics similar to those of riverine DOC. Inputs of water to lakes from upland drainage basins occurs year around (Lesack and Melack 1995) but makes a major contribution to the water occupying lakes during falling and low water (Forsberg et al. 1988). Richey et al. (1986, 1989a) estimated that up to 30% of the discharge of the Amazon River passes through the central floodplain and that part of this flow represents diffuse upland runoff and discharge from small tributaries passing through floodplain lakes. The magnitude of inputs from local upland runoff and direct precipitation to the floodplain was estimated by Richey at al. (1989) and plotted for upriver, midriver, and downriver reaches. We integrated the flow volumes in their plot, subtracted an estimate of direct precipitation and scaled the integrated sum to the entire floodplain. Assuming that all of upland runoff flowed across the floodplain, we calculated the annual input of upland runoff to the floodplain to be 400 x 109 m3 yr1. The organic carbon in upland runoff is expected to be refractory and predominantly terrestrial in origin (Hedges et al. 1986). The local drainage basins of Amazon floodplain lakes are generally forested and local runoff is
John M. Melack and Bruce R. Forsberg
low in suspended particulates. The concentrations of dissolved and particulate organic carbon in these waters can be estimated from values for black- and clearwater tributaries of the Amazon River which have predominantly forested basins. The average concentration of suspended sediments in these black- and clearwater tributaries was determined by Forsberg et al. (1988) to be 12 g m-3. The suspended sediments in black and clear water tributaries are predominantly fine (< 63 pm, Richey et al. 1986) and have an average carbon content of 5.45% (Hedges et al. 1986). The average particulate organic carbon concentrations in these rivers is estimated to be 0.65 g C rrr3. The corresponding average concentration of DOC, determined from measurements in seven different black and clear water tributaries (J. E. Richey, unpublished data, n = 29), is 6.08 g nr3. Using this average carbon composition and the estimated regional input of local runoff, we calculate the total organic carbon input from local drainage basins to floodplain lakes to be 2.7 Tg C yri. Allochthonous inputs from the Amazon River and local drainage thus contribute 4.3 Tg yr1 of organic carbon to floodplain. This represents approximately 4% of total plant production within the floodplain. However, the majority of the allochthonous inputs are refractory carbon, while most of the aquatic plant production is available to other trophic levels.
Heterotrophic activity and foodwebs Once organic carbon is produced in or transported into floodplain lakes, it is potentially available to heterotrophic organisms. The point at which the organic carbon enters the foodweb, and the rate at which it is assimilated depend on a variety of factors including its physical form, its nutritional quality, and the spatial and temporal distributions of the carbon source and its potential consumers.
Biogeochemistry of Amazon Floodplain Lakes and Associated Wetlands
Measurements made in Lake Calado indicate that bacteria are responsible for 60 to 90% of the planktonic respiration (E. Peele and T. R. Fisher, personal communication). While bacterial abundances were intermediate (2 to 5 x 109 cells per liter), rapid division rates produced bacterial doubling times of about 15 hours. Grazing measurements with tritium-labeled bacteria indicated that bacterial growth was about balanced by bacterivory by flagellates (0.2 to 25 x 106 per liter), ciliates (102 to 103 per liter), rotifers (900 per liter) plus cladocera and copepod nauplii (E. Lessard and E. Peele, personal communication). Abundances of adult calanoid and cyclopoid copepods and cladocera were low (1 to 100 per liter) (Fisher et al. 1983, Lenz et al. 1986, Silva 199D, and this may reflect predation by fish larvae. Macrophytes are apparently the principal source of labile DOC for bacterial communities due to their high production rates and small losses from grazing. The DOC leached from decomposing macrophytes is nutritious and rapidly consumed by heterotrophic bacteria. In enclosure experiments with freshly cut Echinochloa polystachya, the dominant C-4 plant in many macrophyte communities, Furch and Junk (1992) observed an initial weight loss accompanied by a rapid decline in dissolved oxygen concentrations and increase in CO2 associated with rapid bacterial metabolism of labile DOC. Based on analyses of stable carbon isotope ratios, DOC derived from C-4 plants was found to be the principal carbon source for bacteria, accounting for 89% of bacterially respired CO2 (Waichman 1996). The DOC present in lake water, in contrast, was found to be predominantly derived from C-3 plants. These results suggest that there are two distinct pools of DOC in lakes: a small pool of labile DOC dominated by C-4 macrophyte inputs with a fast turnover time, and a large pool of more refractory DOC dominated by inputs from C-3 plants with slower turnover times. The
255
refractory DOC component may include degraded terrestrial carbon derived from upland and riverine inputs or older carbon leached from flooded forest soils or sediments. The contribution of different plant groups to carbon flow in aquatic foodwebs on the central Amazon floodplain has been investigated in a series of studies utilizing natural variations in carbon stable isotope ratios (Araujo-Lima et al. 1986, Padovani 1992, Fernandez 1993, Forsberg et al. 1993, and Waichman 1996). In an investigation of carbon sources for 35 species of adult fish, Forsberg et al. (1993) found that a minimum of 82% of carbon in fishes was derived from C-3 plants (phytoplankton, periphyton, C-3 macrophytes, and floodplain forest trees), of which 36% was derived from algae. Eleven fish species, including 8 characiform detritivores, derived all of their carbon from algae. Recent results indicate that the delta 13 C of periphyton is variable and can be quite negative as are the values from some detritivores (B. R. Forsberg, unpublished data). C-4 plants, which are responsible for the bulk of herbaceous macrophyte production and more than half of all aquatic primary production, accounted for less than 18% of carbon in fishes (Forsberg et al. 1993). In a similar study of carbon sources for larval fish, Fernandez (1993) found a higher contribution of C-4 plant carbon, especially in smaller sizes. In larvae of Semaprochilodus insignis and Prochilodus nigricans smaller than 55 mm, C-4 plants accounted for 43% and 36% of their carbon, respectively. Percent C-4 carbon diminished above lengths of 55 mm, indicating a shift to C-3 carbon sources in both species as they became fingerlings. In a study of carbon sources for detritivorous shrimp species, Padovani (1992) obtained results similar to those for adult fish. While living in close association with C-4 macrophyte beds, the shrimp derived more than 85% of their carbon from C-3 plants, predominantly periphytic algae associated with macrophyte roots.
256
The relative lack of carbon derived from C-4 plants in higher organisms is surprising considering the large contribution of C-4 macrophytes to floodplain primary production and suggests that these plants are selectively avoided by aquatic herbivores. Forsberg et al. (1993) presented data on the nutritional quality of plant materials available in floodplain lakes which indicated C-4 grasses to have low food value. The occurrence of C-4 carbon in larval fish can be explained by their spatial distribution and available food sources. Since parental care is rare in Amazonian fish, larvae are vulnerable to predators and tend to seek refuge under the abundant floating macrophytes present in most lakes. These fish are forced to consume the less nutritious C-4 plant detritus which is plentiful under these beds. The predominance of C-3 plant carbon in most higher organisms reflects the selective consumption of C-3 plants by herbivores and detritivores. Forsberg et al. (1993) found tree fruits and seeds and C-3 macrophytes and algae to have high nutritional value. The high percentage of fruits and seeds encountered in the diets of herbivorous and omnivorous fish tends to support this conclusion (Goulding 1980, Goulding et al. 1988). The predominance of phytoplanktonderived carbon in diets of many fish species, despite its small contribution to floodplain production, can be explained by the selective consumption of algae. However, Bayley (1989) argued that phytoplankton production was too low to contribute significantly to regional fish production. He based his argument on a food chain with three trophic levels (phytoplankton—zooplankton—fish), and assumed a 10% transfer efficiency between trophic levels. This model is inappropriate for detritivorous and herbivorous fish which consume plant materials directly. A model with two trophic levels would be more appropriate and would indicate a higher potential contribution to
John M. Melack and Bruce R. Forsberg
fish from phytoplankton. Further, measurements of carbon isotopic ratios for periphyton (B. R. Forsberg, unpublished data) suggest that periphyton may also be a major carbon source for detritivores and other higher aquatic organisms. Overall, these results suggest a general pattern of carbon flow in lakes on the central Amazon floodplain. Nutritious plant materials, primarily derived from C-3 plants (phytoplankton, periphyton, C-3 macrophyte leaves, tree fruits, and seeds) are selectively consumed by aquatic herbivores and detritivores and dominate the organic carbon flow to higher trophic levels. Some C-3 plant materials and the bulk of C-4 plants decompose and release their organic carbon, predominantly in the dissolved form, to the water column. The labile component of this DOC, dominated by C-4 plant carbon, is rapidly consumed by heterotrophic bacterial communities and released as either CO2 or CH4 to the atmosphere. The more refractory DOC component has a much slower turnover rate and tends to persist in the system where it is eventually exported to the main river channel.
Net carbon losses Organic carbon is lost from Amazon floodplain lakes by permanent burial, emission as CO2 and CH4 to the atmosphere or export to the river. Sufficient data are now available to estimate most of these losses. Methane is produced predominantly in anoxic environments associated with flooded habitats. Methane emission rates have been estimated in a variety of habitats and sites along the central Amazon floodplain (Devol et al. 1988, 1990, 1994, Crill et al. 1988, Bartlett et al. 1988, 1990, Wassmann et al. 1992, Wassmann and Thein 1994, Engle and Melack 2000). The average emission rates encountered in aquatic macrophyte beds, flooded forest and open water, including
Biogeochemistry of Amazon Floodplain Lakes and Associated Wetlands
data from both seasonal and regional studies, were 0.233, 0.103, and 0.050 t C krrr2 d'1, respectively. Using these values, together with average flooding periods of 182.5, 135, and 365 days and maximum flooded areas of 29,300, 28,200 and 10,370 km2, we estimated the total methane emissions from aquatic macrophyte beds, flooded forests, and open water areas in lakes to be 1.25, 0.39 and 0.19 Tg C yr1, respectively, and total regional methane emission to be about 1.8 Tg C yr-1. This represents 1 to 2% of the total global methane flux from wetlands to the troposphere (Bartlett and Harriss 1993). Carbon dioxide emissions from lacustrine habitats include respiratory losses from living plants and animals but are dominated by the metabolic losses from bacterial communities consuming dead organic matter. While these losses occur throughout the year under both flooded and dry conditions, few measurements have been made of CO2 emissions in flooded environments. A single regional survey was conducted along the central Amazon floodplain by Devol et al. (1988) during high water. They reported average CO2 emission rates of 2.18, 0.94, and 1.98 t C krrr2 d'1 for aquatic macrophyte beds, flooded forests, and open water, respectively. These measurements were made in open water areas within each habitat and include only below-water respiratory losses. No data are available on CO 2 emissions from floodplain forests and macrophyte communities during their terrestrial growth phase. It has been reported that decomposition of forest litter and macrophyte leaves is about 3 times slower in terrestrial environments than in aquatic habitats (Furch and Junk 1997). If we assume a similar reduction in other metabolic losses, CO2 emissions in both habitats would be approximately three times lower during the terrestrial phase. Applying this relation and using average flooding
257
periods of 182.5, 135 and 365 days and maximum flooded areas of 29,300, 28,200, and 10,370 km2, we estimated the total annual CO2 emissions from macrophyte beds, floodplain forests and open water areas in lakes to be 11.7, 3.6, and 7.5 Tg C yr1, respectively, resulting in a total floodplain CO2 loss of 22.8 Tg C yr1. This value does not include active respiration of above ground and above water vegetation, and, therefore, underestimates the total regional CO2 flux. Particulate organic carbon which is incorporated into lake sediments and not oxidized by detritivore and bacterial communities is lost through permanent burial. In an analysis of carbon diagenesis in the pelagic sediments of Lake Jacaretinga, Devol et al. (1984a) estimated the permanent burial rate of particulate organic carbon to be 43 t C km'2 yr1. Lake Jacaretinga is a small (about 5 ha) floodplain lake connected to the Amazon River by a 500 m long channel during 6 months of the year. Smith-Morrill (1987) determined burial of organic carbon based on cores collected from three sites in Lake Calado, a moderately sized (2 to 8 km2) ria lake connected year round to the Amazon River. Lake-wide, burial rates were estimated to be 42 t C km"2 yr1. Assuming that burial in these two lakes is representative of the entire floodplain area of 67,900 km2, we estimated the total regional loss of particulate organic carbon to permanent burial to be about 2.9 Tg C yr-1. As part of an analysis of exchanges of sediment between the floodplain and channel of the Amazon River in Brazil, Dunne et al. (1998) calculated the amount of suspended sediments accumulating on a decadal time scale on the floodplain along the reach of the Amazon River from about 52Q to 68SW. They used two different approaches and estimated a net transfer of 209 Mt yr1 and 500 Mt yr1 of sediments from the river to the floodplain.
258
John M. Melack and Bruce R. Forsberg
Hedges et al. (1986) reported organic carbon permanent burial were estimated to be 27.6 to be about 1% of the coarse, sand-size frac- Tg C yr1. The residual loss term of 89.7 Tg tion (> 63 pm) of suspended sediments and C yr1 is high, but the majority of the residual about 1.2% of the fine, silt-clay fraction (< 63 probably corresponds to organic carbon urn) of suspended sediments. Using the export to the Amazon River. average of these values (about 1.1%), we Richey et al. (1990) used flow-weighted calculate organic carbon deposited on the estimates of organic carbon flux and microfloodplain from the river to be between 2.3 bial respiration measurements to evaluate Tg C yr1 and 5.5 Tg C yr1. The estimate we the organic carbon balance along 1800 km made of burial in the floodplain based on of the Amazon River. Organic carbon inputs cores from lakes should incorporate inputs of estimated from tributaries were half those suspended sediments from the river, produc- required to support in situ oxidation. They tion occurring on the floodplain and inputs hypothesized that the residual oxidation from the local catchments. Hence, the value was sustained by diffuse inputs of labile of 2.9 Tg Cyr1 may be low. Given the range organic carbon from floodplains with a total of values and the uncertainties in each, we flux at least as large as the final TOG flux of will estimate the burial of organic carbon in the mainstem, 36.1 Tg C yr1. The large the floodplain to be 3 Tg C yr1. Organic organic carbon residual that we calculated carbon losses due to export to the river were suggests that the flux from the floodplain estimated by the difference in the overall may be significantly greater. carbon balance (see next section).
Total organic carbon balance
Biogeochemical Processes and Fluxes: Nitrogen and Phosphorus Dynamics
The total organic carbon balance for lakes on the central Amazon floodplain can be examined by comparing total inputs due to primary production and external loading with total losses (Table 14.4). The combined input of organic carbon due to primary production, river import and local runoff was estimated at 117.3 Tg C yr1. Combined losses due to biogenic gas emission and
Understanding the biogeochemistry of nitrogen and phosphorus in floodplain lakes requires measurements of numerous processes, pools, and fluxes. Key processes include nitrogen fixation, denitrification, nitrification, and nitrogen and phosphorus uptake and regeneration. Fluxes within lakes include sedimentation, sediment-water exchange and burial. Further, measuring
Table 14.4 Organic carbon balance for floodplain of Amazon River between 52.5s and 70.5SW. Input or Output Aquatic net primary productivity Amazon River input Local runoff input Methane emission CO2 emission Burial Export to river and residual errors * See text for derivation of values.
Regional Total (Tg C y1)
113.0 1.6 2.7 - 1.8 - 22.8 -3.0 -89.7
Biogeochemistry of Amazon Floodplain Lakes and Associated Wetlands
fluxes between lakes and the atmosphere and neighboring rivers requires merging frequent measurements of solutes or gases with hydrological, hydrodynamic, and meteorological analyses. We emphasize the importance of multiyear, multisite data on these processes and fluxes to determine adequately their spatial and temporal differences and variations. Because most of these processes and fluxes have been examined at Lake Calado, we focus our discussion on it. Lake Calado (3Q15'S, 60Q34'W) varies from about 2 to 8 km2 in area and 1 to 12 m in depth depending on the stage of the Solimoes River. This dendritic, ria lake is connected year round to the river via a welldefined channel. All major aquatic habitats and interfaces (as described in our conceptual framework of a floodplain lake) are represented in the lake. As is typical of floodplain lakes, Lake Calado receives solutes and particulates from the Solimoes River and local sources that include direct rainfall and dry deposition, upland runoff, groundwater seepage, exchanges with neighboring lakes, and nitrogen fixation within the lake. All these fluxes have been measured at Lake Calado (Table 14.5; Lesack 1988, Lesack 1993a, Lesack and Melack 1991, Doyle and Fisher 1994). While the river was the major supplier of P, inputs of N predominantly were from local
259
sources. Nitrogen fixation provided about 8% of the total annual N inputs (Doyle and Fisher 1994). Measurements of atmospheric deposition spanning at least one full year, based on event collections with timely assays and including inorganic and organic fractions of N and P are very rare for the Amazon basin. The only two studies available (Lesack and Melack 1991, Williams et al. 1997a) were done at Lake Calado in 1983-1984 and 1988-1990, respectively. Solute concentrations were not statistically different between these two studies, and both studies attributed most of the solutes to biogenic rather than anthropogenic sources. To detect temporal trends and spatial differences in atmospheric deposition to the Amazon as development proceeds will require multiyear monitoring programs at several sites. Throughfall and litterfall from flooded forests represent a complex aspect of nutrient dynamics in floodplain lakes that has received little attention. Filoso (1996) measured wet deposition and throughfall from igapo forest of the Anavilhanas archipelago in the lower Negro River, and calculated that about two thirds of the rain that fell on floodplain lakes in the Anavilhanas was influenced by flooded forest canopy. The deposition of dissolved P increased about seven times after rain was
Table 14.5 Inputs of total nitrogen (TN) and total phosphorus (TP) to Lake Calado in 1984-1985. Source
TN
%
TP
%
Direct rainfall Surface runoff Groundwater Adjacent lakes Amazon River
0.24 1.29 0.14 0.24 1.03
9 43 5 8 35
0.006 0.011 0.018 0.013 0.049
11 19 13 51
Source: Data are from Lesack (1988), and calculations modified from Fisher et al. (1991). * Units are 106 moles of N or P yr1, and % is the percentage of the total contributed by each source.
6
John M. Melack and Bruce R. Forsberg
260
intercepted by the canopy, and the N to P to the river exceeded annual inputs from the ratio in throughfall was 3. In contrast, the N river, interannual variations in these fluxes to P ratio in Negro River waters was 20, and are to be expected, and it is premature to these waters supplied only about half of the state that the floodplain is a net source of N or P to the river. P to the lakes. Conditions conducive to denitrification While nutrients supplied by throughfall are immediately available to the algae, seem to occur only intermittently in selected forest litter must first decompose to supply sites in Amazon floodplain lakes. Melack nutrients to the overlying water. For exam- and Fisher (1988) were unable to detect ple, based on measurements of decomposi- denitrification in water or sediments of Lake tion of Pseudobombax munguba leaves in a Calado during periods of rising and falling stagnant tank and a litterfall of 4 tons ha-1 water. However, undersaturated concentrayr1 decomposing underwater, Furch et al. tions of nitrous oxide observed in lake (1984) estimated that after 4 weeks of water could indicate denitrification is occurdecomposition concentrations of P would ring. In Lake Camaleao, Kern et al. (1996) increase about 50% in a 4 m water column. measured denitrification, without substrate However, given that flooded forests can be amendment, during low water when nitrate composed of 60 to 80 tree species per ha, in overlying water reached about 5 pM, but each with different foliar composition not at other times during their seven month (Klinge et al. 1983) and decay rates, and that study. Denitrification was highest in sedidecomposition varies as a function of the in ments exposed to air at low water, and N2O situ physicochemical conditions, consider- was evolved from these exposed sediments, ably more study (e.g., Furch et al. 1989) is but not from flooded sediments. Sediment required to determine the quantitative role slurries amended with 100 pM nitrate effectof decomposition of litterfall in floodplain ed detectable denitrification, a result indibiogeochemistry. cating limitation by nitrate of denitrifiers. Recycling of N and P occurs in the water Floodplain lakes lose N and P via outflow to rivers and neighboring lakes, burial in column and at interfaces between the water sediments and groundwater seepage. and substrata such as profundal sediments. Nitrogen can also be lost via denitrification. In Lake Calado, regeneration of ammonium During the period of study at Lake Calado, and phosphorus is dominated by planktonic outflow to the river was the major loss heterotrophs less than 53 pm in size (Table (Table 14.6; Lesack 1988). Burial in lacus- 14.7, Lenz et al. 1986, Fisher et al. 1988a, trine sediments was of secondary impor- Fisher et al. 1988b, Morrissey and Fisher tance (Smith-Morrill 1987), and groundwater 1988). Sediment-water exchange is smaller seepage was small. Although annual losses than planktonic processes, but is substantial
Table 14.6 Losses of total nitrogen (TN) and total phosphorus (TP) from Lake Calado. Term
TN
%
TP
%
Groundwater Burial Outflow
0.18 1.36 3.26
4 28
0.012 0.072 0.152
5 31 64
68
Source: Data are from Lesack (1988) and Smith-Morrill (1987), and calculations are from Fisher et al. (1991). * Units are 106 moles of N or P yr1, and % is the percentage of the total contributed by each term.
Biogeochemistry of Amazon Floodplain Lakes and Associated Wetlands
(Smith and Fisher 1985, Figuereido 1986, Smith-Morrill 1987). Because the magnitude of internal recycling exceeds external supply, in-lake processes can exert major control over nutrient concentrations in floodplain lakes, a finding corroborated by Forsberg et al.'s (1988) comparative study. In Lake Calado, mass balance calculations indicate that inputs are recycled about 100 times before burial or export (Fisher et al. 1991). Hence, it is essential to make direct measurements of recycling to determine correctly the total supply of nutrients within a floodplain lake. An analysis of nitrogen turnover in Lake Camaleao, a floodplain lake located on an island, provides comparative information on nitrogen dynamics in a lake without a large upland catchment but with a large ATTZ (Kern and Darwich 1997). Mass balance calculations were done over a nine month period (June to March) and excluded the period of overbank flow (April to June). With respect to the Amazon River, the lake basin was a sink for nitrate, a slight source for ammonium, about in balance for dissolved organic N and stored particulate N. Decomposition led to an accumulation of ammonium within interstitial waters of the sediments, which led to efluxes to the overlying water. Nitrogen fixation was highest in the periphyton, modest on exposed sediments, not detected in the plankton, and possibly high in some
261
leguminous plants. Denitrification within unforested, exposed sediments and in periphyton was an important route of N loss in contrast to very small outputs from aquatic sediments. While the uptake of N from soils, sediments, and water by the floodplain forests and herbaceous plants roughly balanced decay with annual plant growth, the magnitude of this internal recycling was several times larger than inputs or outputs. In contrast to Lake Calado, in Lake Camaleao riverine N contributed the majority to the total input (70%, Fig. 14.8). When not inundated, the ATTZ of Lake Camaleao was the location of significant gaseous N output, a finding likely important in most floodplain lakes. Periphyton played a major role in denitrification and nitrogen fixation, a result further supporting the evidence from other lakes that periphyton is a key component in the biogeochemistry and foodwebs of floodplain lakes.
Nutrient Availability and Limitation Seasonal and regional differences in the relative importance of N or P limitation occur in the Amazon basin. Most research has concerned phytoplankton, and some recent studies have examined periphyton and floating meadows. Comparable experimental assessments of flooded forests are lacking. Enrichment bioassays in Lake Calado during rising and high water consistently
Table 14.7 Summary of nitrogen (as NH^ and phosphorus (as PO^ regeneration in Lake Calado. Term Zooplankton (> 53 um) Total plankton Sediments
N
P
N:P
1.0 (5%) 13.8 (65%) 6.5 (30%)
0.11 (2%)
9
3.36 (78%) 0.89 (20%)
4 7
Source: Calculations are from Fisher et al. (199D, and data are from references cited in text. * Total plankton and zooplankton data are integrated over a 3 m epilimnion. Units are mmol nr2 d'1. Percentages (in parenthesis) are the percent of total recycling represented by each process. N:P is the atomic ratios of the rates.
262
indicated limitation by phosphorus; physiological assays implied that the phytoplankton were impoverished in both N and P (Setaro and Melack 1984). In contrast, during falling and low water, N and P or nitrogen alone were required to stimulate algal growth; physiological measurements detected nitrogen deficiency or N or P sufficiency. Bioassays done in another floodplain lake adjacent to the Amazon River during low water also indicated responses to N or to N
John M. Melack and Bruce R. Forsberg
plus P (Zaret et al. 1981, Devol et al. 1984b). Measurements of sestonic N to P ratios and bioassay experiments done in a small lake bordering the Negro River indicated phosphorus limitation (Forsberg 1984). Setaro and Melack (1984) hypothesized that the seasonal changes in nutrient limitation in Lake Calado were related to differences in stratification and mixing. At high water, when the lake was deep and stratified, phosphorus-rich particulates
Figure 14.8 Input and output of nitrogen (tons per 650 ha) at Lake Camaleao during the period from 2 June, 1992 to 14 March, 1993; from Kern and Darwich (1997).
Biogeochemistry of Amazon Floodplain Lakes and Associated Wetlands
derived from the river tended to sink from the water column, increasing the potential for P-limitation. At low water, these sediments were resuspended, increasing P availability and driving the system toward N-limitation. However, the presence of an oxic-anoxic interface in the water column when floodplain lakes are stratified, as commonly observed in Lake Calado (Maclntyre and Melack 1988) and other floodplain lakes, could create a sink for dissolved inorganic nitrogen because of enhanced denitrification at the interface but have little influence on the upward flux of P. B. R. Forsberg (unpublished data) provides further evidence for nutrient limitation (Table 14.2). Light saturated rates of photosynthesis (Pmax) were found to vary as a linear function of either total nitrogen (TN) or total phosphorus (TN) concentration. TP was the main source of variation in Pmax in black- and clearwater lakes at both high and low water. This indicates a consistent pattern of P-limitation in these systems, which was linked to the high TN to TP ratios. In Whitewater lakes, Pmax varied as a function of TP at high water and TN at low water. This reflects a shift from P-limitation to N-limitation and was accompanied by a decrease in average TN to TP ratios. The N and P adsorbed to suspended sediments in Amazon floodplain lakes are available to phytoplankton (Grobbelaar 1983). Up to 30% of available P may be adsorbed to clay particles (Engle and Sarnelle 1990). Further, rapid exchange can occur among the pools of particulate, dissolved inorganic P (DIP) and dissolved organic P (DOP). Carrier-free 32P (as PO4) additions to surface waters of Lake Camaleao, collected during low stage, were fractionated over 12 hours into four compartments: particulate ( > 0.2 pm, 83%), dissolved inorganic P (9.4%), high molecular weight DOP (about 66,000 Da, 7.3%) and very high molecular weight DOP ( >200,000
263
Da, 0.3%); turnover times of DIP were 8 to 11 hours (Pinheiro 1996). In addition, enzymatic hydrolysis of DOP occurs in white- and blackwaters with low molecular weight fractions ( < 1,000 Da) being the most susceptible (Pinheiro 1996). Because attached algae retain their position close to the water surface, they may be better placed to exploit nutrients in turbid riverine floodwaters. Further, because particulate-bound nutrients are retained among the floating macrophytes, epiphytic algae have access to these nutrients, as well. Engle and Melack (1993) investigated the interactions between suspended sediments, dissolved nutrients, and epiphytic algal growth in floating meadows in a series of in situ experiments in Lake Calado. During rising and high water, chlorophyll per gram of macrophyte roots was less than controls in experimental treatments bathed in river water. Addition of N and P without sediments increased chlorophyll above controls. Further, when epiphytic algae were treated with dilutions of river water (0%, 10%, 25%, and 100%), only the 10% treatment resulted in significantly more algal chlorophyll. Hence, the influence of nutrient-rich, turbid water on epiphytic algal abundance will depend on subtle differences in suspended sediment concentrations. Nutrients for macrophyte growth in lakes can be derived from sediments, soils, rivers, and local streams. Nutrient availability in lake sediments or soils tends to be higher on Whitewater floodplains and lower in blackand clearwater environments (Furch 1997). Nutrient availability for floating species tends to be highest during low and rising water and lowest during falling water when river inputs are reduced. Junk and Piedade (1997) found little evidence for short-term nutrient limitation in macrophyte communities growing in Whitewater environments. Nutrient limitation was more common in acidic blackwaters and appeared to restrict the growth
264
and occurrence of aquatic and semiaquatic macrophytes in these environments. Bioassays with Salvinia auriculata, Eichhornia crassipes and Pistia stratiotes transferred from Whitewaters to acidic blackwaters indicated short-term growth limitation (Junk and Piedade 1997). The N to P ratio in the sources and sinks of N and P provides an ecosystem level indication of the likelihood of limitation by N or P (Howarth 1988). Based on the measurements at Lake Calado summarized above, it is possible to determine the N to P ratios for inputs, losses and recycling within the lake. Overall, inputs to Lake Calado from rain, upland runoff, groundwater seepage, adjacent lakes, and riverine inflows had an aggregate molar N to P ratio of 32; inclusion of nitrogen fixation would increase the ratio. Losses via burial as sediments and outflow to the river, but not including gaseous losses of N, had an N to P ratio of 20; denitrification would decrease the ratio. Internal recycling by bacteria and zooplankton within the water column and by regeneration from sediments had molar N to P ratios of 4 and 7, respectively. Typically, molar N to P ratios below 10 are indicative of N-limitation, and molar N to P ratios above 20 are indicative of P-limitation in primary producers. Therefore, the likelihood of N or P-limitation will vary as a function of spatial and temporal differences in the relative importance of inputs or recycling.
John M. Melack and Bruce R. Forsberg
and expansion of cities proceeds and as more hydroelectric dams are constructed, associated impacts will surely increase.
Impacts of bauxite tailings on an Amazon floodplain lake
Mining activities occur at several sites within the Amazon basin, but their impacts on floodplain lakes have seldom been examined. One well-studied example is provided by Esteves and his students for Lake Batata, a lake, separated from the Trombetas River by a lateral levee, that received discharges of bauxite tailings totaling 18 million m3 per year from 1979 to 1989 (Esteves et al. 1990). The tailings now cover about 30% of the lake's bottom. These sediments have a much higher proportion of very fine silts and clays with lower caloric content than do those not impacted by tailings (Callisto and Esteves 1996). Phosphorus, carbon, and nitrogen concentrations in the sediments were much lower in areas affected by tailings than in unaffected parts of the lake (Roland and Esteves 1993). Concentrations of dissolved inorganic nitrogen differed between the area of the lake with tailings and that without. Ammonium varied from 565 ug L'1 to 33 pg L"1, most likely as a result of anoxia under the clay-rich tailings. Nitrate "was higher above the tailings. In contrast, phosphorus fractions were similar at all the stations. During periods of low water, resuspension Impacts of Human Disturbances of the clays associated with the tailings impeded light penetration and probably Floodplain lakes can be perturbed by reduced photosynthesis by phytoplankton. direct alterations to the lakes and fringing Chlorophyll concentrations were about ten wetlands, by changes to the river with which times higher at stations not impacted by they are associated, and by modifications to tailings than at those with tailings. the uplands surrounding their local catchments and to their airshed. To date, human- Impacts of hydroelectric dams induced impacts on floodplain lakes appear only in limited areas within the Amazon Several large hydroelectric reservoirs have basin. As development of agriculture, mining been built within tropical forested catch-
Biogeochemistry of Amazon Floodplain Lakes and Associated Wetlands
ments of the Amazon, including Curua-Una, Samuel, Tucurui, and Balbina in Brazil (Junk et al. 1981, Fearnside 1989, MatsumuraTundisi et al. 1991, Petrere and Ribeiro 1994). Many more are planned. Melack (2001) reviewed published studies of conditions within the reservoirs, a topic outside the scope of this chapter. Alterations of flows downstream of dams have major implications for floodplain ecosystems adapted to and dependent on a natural flood regime. While detailed studies are not available for any of the major Amazonian dams, numerous interactions between the biota and the flood pulse documented elsewhere in the Amazon (Junk 1997) are fair warning that myriad impacts are likely. The regulated flows will reduce flood levels and maintain higher low flows with a net effect of decreasing the area of floodplain enriched with riverine sediments each year and hence
265
available for agriculture and use by fishes and other organisms (Barrow 1988).
Mercury pollution Mercury contamination is a significant problem in many Amazon lakes. While commonly linked to the indiscriminant use of azougue (liquid mercury) in gold mining areas, a portion of the mercury in lakes appears to be derived from natural sources. The relative importance of these sources and their influence on the biota and human inhabitants of lakes can only be understood in context of the larger regional and global mercury cycles. The primary source of mercury in most lacustrine environments is the atmosphere (Watras et al. 1995). Atmospheric mercury concentrations have tripled globally during the last hundred years due to increased emissions from
Figure 14. 9 Historic increase in mercury concentration in sediments of Lake Cristalino, located on the shore of the lower Negro River. B. R. Forsberg and E. Campos, unpublished data.
266
John M. Melack and Bruce R. Forsberg
anthropogenic sources (Mason et al. 1994). results suggest that natural inputs of mercuEmissions of mercury vapor from Amazon ry from terrestrial sources and river chemgold fields have contributed to this rise, istry may be the principal factors controlling accelerating significantly during the last 25 mercury levels Amazonian lakes. years (Forsberg 1992). Natural sources (i.e., volcanism, weathering, and air-ocean Impacts of deforestation on Amazon exchange) also contribute significantly to floodplain lakes the atmospheric mercury load and have contributed to atmospheric deposition. The deforested area in the Amazon basin Atmospheric mercury reaches lakes directly increased through the 1980s (Skole and through atmospheric deposition and indi- Tucker 1993) and has continued to increase rectly through inputs of atmospherically in the 1990s. While accessible areas borderderived mercury from the terrestrial catch- ing navigable rivers and roads are particument (Swain et al. 1992). B. R. Forsberg larly prone to deforestation, effects of defor(unpublished data) has found evidence of a estation on the ecology of adjacent streams, gradual increase in mercury inputs to Lago rivers and lakes have received limited attenCristalino, a floodplain lake associated with tion (Bruijnzeel 1991). Lake Calado was the Negro River, during the last 60 years studied during the 1980s when its catchment (Fig. 14.9). This finding suggests that atmos- was undergoing rapid development (Melack pheric deposition from global and regional and Fisher 1990, Melack 1996). Downstream anthropogenic sources have had a signifi- lacustrine waters were influenced by cant impact in central Amazon basin. increased inputs of water and solutes resultExceptionally high levels of mercury, ing from deforestation. apparently derived from natural sources, The hydrologic characteristics of runoff in have been encountered in Amazonian soils the Amazon determine the magnitude of the (Roulet et al. 1998, Silva-Gorsberg et al. various hydrologic pathways that transport 1998). These soils represent an important nutrients and other solutes (see McClain and natural source of mercury to lacustrine Elsenbeer, this volume). Given the high ecosystems. Mercury is released from soils frequency of storms in wet equatorial by both physical and chemical weathering, climates, it is important to understand how and generally complexed to paniculate or export of nutrients and solutes is influenced dissolved organic matter (Roulet et al. 1998). by storm size and frequency, and the volume The highest levels of mercury are conse- of generated runoff. However, most hydrologquently encountered in blackwater rivers ical studies in the Amazon basin have been and lakes which receive large inputs of done in areas of intact rain forest (Nortcliff terrestrially derived organic matter. The low and Thornes 1981, Lesack 1993a), and do not pH and high DOC in these systems also address the effects of deforestation. Intrasystem processes and their effects on favors the methylation and bioaccumulations of mercury resulting in high levels of solute transport can be evaluated by meascontamination in predatory fish and fish- uring solution chemistry at different stages eating human populations (Forsberg et al. of a hydrologic pathway. Studies of nitrogen 1995). Strong, basin-wide correlations have transformations along hydrologic pathways been found between human mercury con- in Amazon catchments have been conducttamination and both river pH and DOC ed by McClain et al. (1994) and Brandes et (Silva-Forsberg et al. 1998), which appear to al. (1996). Only Williams et al. (1997b) have be independent of mining activities. These addressed the effects of land conversion on
Biogeochemistry of Amazon Floodplain Lakes and Associated Wetlands
solution chemistry for a suite of solutes at different stages of a hydrologic pathway through an upland catchment. At Lake Calado, Lesack (1993a) measured all the water balance components for a firstorder catchment draining a 23 ha stand of undisturbed upland rain forest adjacent to the Amazon floodplain. In parallel studies, Lesack (1993b) measured fluxes of solutes exported by streamflow and subsurface outflow, and Lesack and Melack (1996) calculated mass balances based on input and outputs. Williams and Melack (1997) and Williams et al. (1997b) measured water and solute fluxes from the same catchment after partial deforestation had occurred. Large increases in solute mobilization from the upper soil horizons to groundwater were observed from a 2 ha plot after cutting and burning in a partially deforested catchment. The first water collected from lysimeters in the cut-and-burned plot contained very high nitrate concentrations (about 1800 pM). In groundwaters, nitrate increased by a factor of 5 to about 100 uM after cutting and burning. However, mean volume-weighted concentrations of nitrate in the stream water were similar before and after partial deforestation. The high nitrate concentrations
267
evident in groundwaters after deforestation may have been denitrified prior to reaching the stream or sequested in riparian vegetation or the stream bed. Solute export via streamwater increased and nutrient ratios were altered subsequent to deforestation (Table 14.8). Rainfall was similar during the two periods, but runoff was higher in the catchment after disturbance. While inputs of total N and total P via rain were smaller in 1989-1990 than in 1984-1985, export via streamflow was greater in 1989-1990. In fact, after partial deforestation, more N and P were exported than added via rain. The N to P ratio of fluvial flux from the intact forest was 120:1; following deforestation, the ratio decreased to 33:1. Although the total N yield doubled after disturbance, the total P yield increased by a factor of 7, primarily due to increased particulate P.
Research Frontiers Regionalization of biogeochemical fluxes Characterization of the areal extent and temporal changes of the tropical floodplains would extend our understanding of trace gas exchange within these ecosystems and
Table 14.8 Comparison of precipitation, runoff and input and export of total nitrogen (TN) and total phosphorus (TP) in the Mota Brook catchment within the Lake Calado basin during a period without deforestation (1984-1985) and a period after about 80% clearing (1989-1990). Parameter
Units
Rain quantity RainTN RainTP Runoff Runoff TN Runoff TP Runoff TN:TP
cm yr"1 kg N ha"1 yr"1 kg P ha"1 yr"1 cm yr"1 kg N ha"1 yr"1 kg P ha"1 yr'1 atomic ratio
1984-1985
Source: Modified from Williams and Melack (1997) and Lesack (1993 a and b).
280.0 6.7 0.34 160.0 4.3 0.082 120:1
1989-1990 275.0 4.5 0.21 208.0 9.1 0.61 33:1
268
of their significance to regional hydrological and biogeochemical processes. The recent availability of optical and synthetic aperture radar data from satellites can be used to study these processes by providing information on the type and distribution of aquatic plants and temporal distribution of inundation (Melack et al. 1994, Melack and Hess 1998, Mertes et al. 1995). For example, delineation of both flooding status and vegetation, with accuracies greater than 90%, has been demonstrated using multifrequency, polarimetric synthetic aperture radar data for floodplains in the central Amazon (Hess et al. 1995). Application of spectral mixture analysis to Landsat multispectral scanner and thematic mapper imagery has permitted calculation of concentrations of suspended sediments in surface water of the Amazon (Mertes et al. 1993). When combined with models of ecological processes, the synoptic information offered by remote sensing will permit calculation of rates on a larger scale and provide much improved data for resource management.
Comparative limnological studies As is evident throughout our review, a much wider diversity of lakes and floodplain environments must receive the attention of limnologists. Additional measurements of periphyton productivity, of burial of C, N, and P in sediments, of nitrogen fixation and denitrification, and of carbon dioxide emissions, especially, are needed. A further challenge is to investigate biogeochemical processes in the ATTZ, and to link these processes with the whole floodplain ecosystem.
Hydrological and hydrodynamical analyses Basic hydrological information is lacking from all but a few floodplain lakes. Given the intimate connection between hydrology
John M. Melack and Bruce R. Forsberg
and biogeochemistry on floodplains, it is essential to examine a representative set of lakes in a manner similar to that employed at Lake Calado (e.g., Lesack and Melack 1995). Complementary approaches using naturally occurring and introduced chemical and isotopic tracers should also be tried. Recent advances in measuring and modeling mixing and advective processes in lakes (Maclntyre and Melack 1995) must be applied to floodplain lakes to increase our understanding of processes causing nutrient fluxes. Improved equipment is now available to measure currents (e.g., acoustic Doppler current meters) and mixing (e.g., microstructure profilers). Highly accurate locations can be determined with portable global positioning units.
Experimental approaches Experimental manipulations of whole lakes or of enclosures within lakes has proven a valuable approach to decipher processes controlling ecological and biogeochemical activity in lakes. The large number of small to moderate-sized lakes on the Amazon floodplain offers an excellent opportunity to employ replicated, wholesystem manipulations. Of particular importance would be experimental alterations of flooding periodicity and depth, and of nutrients inputs.
Management issues The continued construction of hydroelectric dams and expansion of agricultural and urban areas adds urgency to increasing our ability to forecast biogeochemical consequences of such developments in the Amazon basin. Applications of limnological understanding to timber management in flooded forests, to fisheries, to conservation of aquatic biodiversity, to control of waterborne disease vectors and to amelioration of
Biogeochemistry of Amazon Floodplain Lakes and Associated Wetlands
chemical pollution offer important challenges. To meet such challenges will require improved understanding of the structure and functioning of flood-plain ecosystems. Acknowledgments: Our research and perspectives on Amazon floodplain lakes have benefitted from collaborations with many individuals in Brazil and the United States, and we especially acknowledge T. R. Fisher, L. F. V.
269
Lesack, M. R. Williams, S. Maclntyre, S. K. Hamilton, D. Engle, L. Smith, L. Alves, J. R. Richey, R. L. Victoria, and L. Martinelli. We thank M. McClain, G. Kling, S. Hamilton, D. Engle. T. Hovanec, V. LaCapra, J. Sickman, and T. Dunne for thoughtful comments on earlier drafts. Funding from the U.S. National Science Foundation and National Aeronautical and Space Administration has supported much of our research in the Amazon.
Literature Cited Adis, J., K. Furch, and U. Irmler. 1979. "Litter production of a Central Amazonian inundation forest." Tropical Ecology 20: 236-245. Alves, L. F. 1991. The fate of streamwater nitrate entering littoral areas of an Amazon floodplain lake: The role of plankton, periphyton, inundated soils, and sediments. Ph.D. Thesis, University of Maryland, p. 370. Araujo-Lima, C. A. R. M., B. R. Forsberg, R. Victoria, and L. A. Martinelli. 1986. "Energy sources for detrivorous fishes in the Amazon." Science 234: 1256-1258. Barrow, C. 1988. "The impact of hydroelectric development on the Amazonian environment: with particular reference to the Tucurui Project." Journal of Biogeography 15: 67—78. Bartlett, K.B., P. M. Grill, D. I. Sebacher, R. C. Harriss, J. O. Wilson, and J. M. Melack. 1988. "Methane flux from the central Amazonian floodplain." Journal of Geophysical Research 93: 1574-1582. Bartlett, K. B., P. M. Grill, J. A. Bonassi, J. E. Richey, and R. C. Harris. 1990. "Methane flux from Central Amazonian floodplains: Emission during rising water." Journal of Geophysical Research 95: 16,733-16,788. Bartlett, K. B., and R. C. Harriss. 1993. "Review and assessment of methane emissions from wetlands." Chemosphere 26: 261-320. Bayley, P. B. 1989. "Aquatic environments in the Amazon basin, with an analysis of carbon sources, fish production and yield." Canadian Fisheries and Aquatic Science Special Publication 106: 399-408. Bayley, P. B., and M. Petrere Jr. 1989. "Amazon fisheries: Assessment methods, current status and management options." Canadian Fisheries and Aquatic Science Special Publication 106: 385-398. Brandes, J. A,. M. E. McClain, and T. P. Pimentel. 1996. "15N evidence for the origin and cycling of inorganic nitrogen in a small Amazonian catchment." Biogeochemistry, 34: 45-56. Bruijnzeel, L. A. 1991. "Nutrient input-output budgets of tropical forest ecosystems: a review." Journal of Tropical Ecology 7: 1-24. Callisto, M., and F. A. Esteves. 1996. "Composicao granulometrica do sedimento de um lago Amazonico impacta-
do por rejeito de bauxita em lago natural (Para, Brasil)." Acta Limnologia Brasiliensia 8: 115-126. Grill, P. M., K. B. Bartlett, J. O. Wilson, D. I. Seabacher, R. C. Harriss, J. M. Melack, S. Maclntyre, L. Lesack, and L. SrnithMorrill. 1988. "Tropospheric methane from an Amazonian lake." Journal of Geophysical Research 93: 1564-1570. Devol, A. H., T. M. Zaret, and B. R. Forsberg. 1984a. "Sedimentary organic matter diagenesis and its relation to the carbon budget of tropical Amazon floodplain lakes." Verhandlungen International Vereinigen Limnologie 22: 1299-1304. Devol, A. H., H. dos Santos, B. R. Forsberg, and T. M. Zaret. 1984b. "Nutrient addition experiments in Lago Jacaretinga, Central Amazon, Brazil: 2. The effect of humic and fulvic acids." Hydrobiologia 109: 97-103. Devol, A. H., J. E. Richey, W. A. Clark, and S. L. King. 1988. "Methane emissions to the troposphere from the Amazon floodplain." Journal of Geophysical Research 93: 1583-1492. Devol, A. H., J. E. Richey, B. R. Forsberg, and L. A. Martinelli. 1990. "Seasonal dynamics in methane emissions from the Amazon River floodplain to the troposphere." Journal of Geophysical Research 95: 16,417-16,426. Devol, A. H., J. E. Richey, B. R. Forsberg, and L. A. Martinelli. 1994. "Environmental methane in the Amazon river floodplain." In: Global Wetlands, ed. W. J. Mitsch (Elsevier, Amsterdam), pp. 155-166. Devol, A. H., B. R. Forsberg, J. E. Richey, and T. P. Pimentel. 1995. "Seasonal variation in chemical distributions in the Amazon (Solimoes) River: A multiyear time series." Global Biogeochemical Cycles 9: 307-328. Doyle, R. D. 1991. Primary production and nitrogen cycling within the periphyton community associated with emergent aquatic macrophytes in an Amazon floodplain lake. Ph.D. Thesis, University of Maryland, p. 269. Doyle, R. D., and T. R. Fisher. 1994. "Nitrogen fixation by periphyton and plankton on the Amazon floodplain at Lake Calado." Biogeochemistry 26: 41-66. Dumont, J. F., S. Lamotte, and F. Kahn. 1990. "Wetland and upland forest ecosystems in Peruvian Amazonia: Plant species diversity in the light of some geological and
270
botanical evidence." Forest Ecology and Management 33/35: 125-139. Dunne, T., L. A. K. Mertes, R. H. Meade, J. E. Richey, and B. R. Forsberg. 1998. "Exchanges of sediment between the flood plain and channel of the Amazon River in Brazil." Geological Sociely of America Bulletin 110: 450-467. Engle, D. L., and J. M. Melack. 1990. "Floating meadow epiphyton: Biological and chemical features of epiphytic material in an Amazon floodplain lake." Freshwater Biology 23: 479-494. Engle, D. L., and O. Sarnelle. 1990. "Algal use of sedimentary phosphorus from an Amazon floodplain lake: Implications for total phosphorus analysis in turbid waters." Limnology and Oceanography 35: 487-490. Engle, D. L., andj. M. Melack. 1993. "Consequences of riverine flooding for seston and periphyton of floating meadows in an Amazon floodplain lake." Limnology and Oceanography 38: 1500-1520. Engle, D., and J. M. Melack. 2000. "Methane emissions from an Amazon floodplain lake: Enhanced release during episoded mixing and during falling water." Biogeochem-istry 51: 71-90. Ertel, J. R., J. I. Hedges, A. H. Devol, J. E. Richey, and M. Ribeiro. 1986. "Dissolved humic substances of the Amazon River system." Limnology and Oceanography 31: 739-754. Esteves, F. A., R. L. Bozelli, and F. Roland. 1990. "Lago Batata: Um laboratorio de limnologia tropical." Ciencia Hoje 11: 26-33Fearnside, P. M. 1989. "Brazil's Balbina dam: Environment versus the legacy of the Pharaohs in Amazonia." Environmental Management 13: 401-423Fernandez, J. M. 1993. Fontes autotroficas de energia em alevinos de jaraqui Semaprochilodus insignis (Schomburgki, 1841) e curimata Prochilodus nigricans (Agassiz, 1829) (Pisces: Prochilodontidae) na Amazonia Central. M.S. Thesis. Instituto Nacional de Pesquisas da Amazonia/Universidade do Amazonas. Ferreira, L. V. 1991. O efeito do periodo de inundacao, fenologia e regeneracao de clareiras em uma floresta de igapo na Amazonia Central. M.S. Thesis. Instituto Nacional de Pesquisas da Amazonia/Universidade do Amazonas. Figuereido, A. C. 1986. O papel dos sedimentos em dois lagos da Amazonia. M.S. Thesis. Instituto Nacional de Pesquisas da Amazonia/Universidade do Amazonas. Filoso, S. 1996. Throughfall and aquatic biogeochemistry in the Anavilhanas archipelago, Negro River, Brazil. Ph.D. Thesis. University of California, Santa Barbara, p. 187. Filoso, S., M. R. Williams and Jim Malade. 1999"Composition and deposition of throughfall in a flooded forest archipelago (Anavilhanas, Negro River, Brazil)." Biogeochemistry 45: 169-195. Fisher, T. R., and P. E. Parsley. 1979- "Amazon lakes: Water storage and nutrient stripping." Limnology and Oceanography 24: 547-553Fisher, T. R., J. M. Melack, B. Robertson, E. Hardy, and L. Fernando. 1983. "Vertical distribution of zooplankton and physico-chemical conditions during a 24-hour period in an Amazon floodplain lake (Lago Calado, Brazil)." Acta Amazonica 13: 475-487. Fisher, T. R., K. M. Morrissey, P. R. Carlson, L. F. Alves, and J. M. Melack. 1988a. "Nitrate and ammonium uptake by plankton in an Amazon River floodplain lake." Journal of Plankton Research 10: 7-29.
John M. Melack and Bruce R. Forsberg Fisher, T. R., R. D. Doyle, and E. R. Peele. 1988b. "Size-fractionated uptake and regeneration of ammonium and phosphate in a tropical lake." Verhandlungen International Vereinigen Limnologie 23: 637-641. Fisher, T. R., L. F. W. Lesack, and L. K. Smith. 1991. "Input, recycling, and export of N and P on the Amazon floodplain at Lake Calado." In: Phosphorus Cycles in Terrestrial and Aquatic Ecosystems. Regional Workshop 3: South and Central America, eds. H. Tiessen, D. Lopez-Hernandez, and I. H. Salcedo (Saskatchewan Inst. of Pedology, Univ. Saskatchewan, Saskatoon, Canada), pp. 34-53. Fisher, T. R., and M. Moline. 1992. "Seasonal plant cover on the Amazon River floodplain determined with aerial videography and image analysis." In: Color Aerial Photography in the Plant Sciences and Related Fields, ed. C. H. Blazquez (American Society for Photogrammetry and Remote Sensing, Bethesda, MD), pp. 207-216. Forsberg, B. R. 1984. "Nutrient processing in Amazon floodplain lakes." Verhandlungen International Vereinigen Limnologie 22: 1294-1298. Forsberg, B. R. 1992. "Mercury contamination in the Amazon: Another Minamata?" Water Report 2: 6-8. Forsberg, B. R., A. H. Devol, J. E. Richey, L. A. Martinelli, and H. Santos. 1988. "Factors controlling nutrient concentrations in Amazon floodplain lakes." Limnology and Oceanography 33: 41-56. Forsberg, B. R., C. A. R. M. Araujo-Lima, L. A. Martinelli, R. L. Victoria, and J. A. Bonassi. 1993. "Autotrophic carbon sources for fish of the central Amazon." Ecology 74: 643-652. Forsberg, B. R., M. C. Silva-Forsberg, C. R. Padovani, E. Sargentini, and O. Malm. 1995. "High levels of mercury in fish and human hair in the Rio Negro basin (Brazilian Amazon)." In: Proceedings of the International Workshop on Environmental Mercury Pollution and its Health Effects in the Amazon River Basin, eds. H. Kato, and W.C. Pfeiffer (National Institute of Minamata Disease, Federal University of Rio de Janeiro), pp. 33-40. Furch, B., and K. Otto. 1987. "Characterization of light regime changes (PAR) by irradiance reflectance in two Amazonian water bodies with different physico-chemical properties." Archiv fur Hydrobiologie 110: 579-587. Furch, K. 1997. "Chemistry of varzea and igapo soils and nutrient inventory of their floodplain forests." In: The Central Amazon Floodplain. Ecological Studies, ed. W. J. Junk (Springer-Verlag, Berlin), pp. 47-68. Furch, K., and W. J. Junk. 1992. "Nutrient dynamics of submersed decomposing Amazonian herbaceous plant species Paspalum fasciculatum and Echinochloa polystachya." Revue Hydrobiologie Tropical 25: 75-85. Furch, K., and W. J. Junk. 1997. "Chemical composition, food value and decomposition of herbaceous plant leaves and leaf litter of the floodplain forests." In: The Central Amazon Floodplain. Ecological Studies, ed. W. J. Junk (Springer-Verlag, Berlin), pp. 187-203. Furch, K., W. J. Junk, and Z. E. S. Campos. 1984. "Release of major ions and nutrients by decomposing leaves of Pseudobombax munguba, a common tree in the Amazonian flooplain." Verhandlungen International Vereinigen Limnologie 23: 642-646. Furch, K., W. J. Junk, and Z. E. S. Campos. 1989. "Nutrient
Biogeochemistry of Amazon Floodplain Lakes and Associated Wetlands dynamics of decomposing leaves from Amazon floodplain forest species in water." Amazoniana 11: 91-116. Goulding, M. 1980. The Fishes and the Forest. University of California Press, Los Angeles. Goulding, M., M. L. Carvalho and E. G. Ferreira. 1988. Rio Negro: rich life in poor water. SPB Academic Publishing, the Hague, the Netherlands. Grobbelaar, J. U. 1983. "Availability to algae of N and P adsorbed on suspended solids in turbid waters of the Amazon River." Archiv fur Hydrobiologie 96: 302-316. Hamilton, S. K., J. M. Melack, M. F. Goodchild, and W. M. Lewis, Jr. 1991. "Estimation of the fractual dimension of terrain from lake size distributions." In: Lowland Floodplain Rivers: Geomorphological Perspectives, eds. P. A. Carling, and G. E. Petts (John Wiley & Sons, New York), pp. 145-163. Hamilton, S., S. Sippel, and J. M. Melack. 1995. "Oxygen depletion and carbon dioxide production in waters of the Pantanal wetland of Brazil." Biogeochemistry 30: 115-141. Hedges, J. I., W. A. Clark, P. D. Quay, J. E. Richey, A. H. Devol, and H M dos Santos. 1986. "Compositions and fluxes of particulate organic material in the Amazon River." Limnology and Oceanography 31: 717-738. Hess, L. L., J. M. Melack, S. Filoso, and Y. Wang. 1995. "Delineation of inundated area and vegetation along the Amazon floodplain with the SIR-C synthetic aperture radar." IEEE Transactions in Geoscience and Remote Sensing 33: 896-904. Howarth, R. W. 1988. "Nutrient limitation of net primary productivity in marine ecosystems." Annual Review of Ecology and Systematics 19: 89-110. Irion, G., W. J. Junk, and J. A. S. N. de Mello. 1997. "The large central Amazonian River floodplains near Manaus: Geological, climatological, hydrological, geomorphological aspects." In: The Central Amazon Floodplain. Ecological Studies, ed. W. J. Junk (Springer-Verlag, Berlin), pp. 23-46. Junk, W. J. 1970. "Investigations on the ecology and production biology of the "floating meadows" PaspaloEchinochloetum on the Middle Amazon. 1. The floating vegetation and its ecology." Amazoniana 2: 449-^95. Junk, W. J. 1985. "The Amazon floodplain: A sink or source of organic carbon?" In: Transport of Carbon and Minerals in Major World Rivers, part 3, eds. E. T. Degens, S. Kemp, and R. Herrera (Mitteilungen aus dem GeologischPalaontologischen Institut der Universitat Hamburg 58), pp. 267-283. Junk, W. J. 1986. "Aquatic plants of the Amazon system." In: The Ecology of River Systems, eds. B. R. Davies, and K. F. Walker (Dr. W. Junk Publishers, Dordrecht, The Netherlands), pp. 319-337. Junk, W. J. 1993. "Wetlands of tropical South America." In: Wetlands of the World, eds. D. F. Whigham, S. Hejny, and D. Dykyjova (Kluwer Publishers, Dordrecht, The Netherlands), pp. 679-739. Junk, W. J., editor. 1997. The Central Amazon Floodplain. Ecological Studies, Springer-Verlag-Verlag, Berlin, p. 525 Junk, W. J., B. A. Robertson, A. J. Darwich, and I. Vieira. 1981. "Investigafoes limnologicas e ictiologicas em CuruaUna, a primeira represa hidrelectrica na Amazonica Central." Acta Amazonica 11: 689-716. Junk, W. J., P. B. Bayley, and R. E. Sparks. 1989. The flood-
271
pulse concept in river-floodplain systems. Canadian Fisheries and Aquatic Science Special Publication 106: 110-127. Junk, W. J., and M. T. F. Piedade. 1993. "Biomass and primary-production of herbaceous plant communities in the Amazon floodplain." Hydrobiologia 263: 155—162. Junk, W. J., and M. T. F. Piedade. 1997. "Plant life in the floodplain with special reference to herbaceous plants." In: The Central Amazon Floodplain. Ecological Studies, ed. W. J. Junk (Springer-Verlag, Berlin), pp. 147-185. Kern, J., and A. Darwich. 1997. "Nitrogen turnover in the varzea." In: The Central Amazon Floodplain. Ecological Studies, ed. W. J. Junk (Springer-Verlag, Berlin), pp. 119-135. Kern, J., A. Darwich, and W. J. Junk. 1996. "Seasonal denitrification in flooded and exposed sediments from the Amazon floodplain at Lago Camaleao." Microbial Ecology 32: 47-57. Klammer, G. 1984. "The relief of the extra-Andean Amazon basin." In: The Amazon, ed. H. Sioli (Dr. W. Junk Publishers, Dordrecht, The Netherlands), pp. 47-83. Klinge, H., K. Furch, E. Harms, and J. Revilla. 1983. "Foliar nutrient levels of native tree species from Central Amazonia. 1. Inundation forests." Amazoniana 9: 19-45. Kubitzki, K. 1989. "The ecogegraphical differentiation of Amazonian inundation forests." Plant Systematics and Evolution 163: 285-304. Lenz, P. H., J. M. Melack, B. Robertson, and E. A. Hardy. 1986. "Ammonium and phosphate regeneration by the zooplankton of an Amazon floodplain lake." Freshwater Biology 16: 821-630. Lesack, L. F. W. 1988. Mass balance of nutrients, major solutes, and water in an Amazon floodplain lake and biogeochemical implications for the Amazon Basin. Ph.D. Thesis, University of California, Santa Barbara, p. 494. Lesack, L. F. W. 1993a. "Water balance and hydrologic characteristics of a rain forest catchment in the central Amazon basin." Water Resources Research 29: 759-773Lesack, L. F. W. 1993b. "Export of nutrients and major ionic solutes from a rainforest catchment in the central Amazon basin." Water Resources Research 29: 743-758. Lesack, L. F. W. 1995. "Seepage exchange in an Amazon floodplain lake." Limnology and Oceanography 40: 598-609. Lesack, L. F. W., and J. M. Melack. 1991. "The deposition, composition, and potential sources of major solutes in rain of the central Amazon basin." Water Resources Research 27: 2953-2978. Lesack, L. F. W., and J. M. Melack. 1995. "Flooding hydrology and mixture dynamics of lake water derived from multiple sources in an Amazon floodplain lake." Water Resources Research 31: 329-345. Lesack L. F. W., and J. M. Melack. 1996. "Mass balance of major solutes in a rainforest catchment in the central Amazon: Implications for nutrient budgets in tropical rainforests." Biogeochemistry 32: 115-142. Maclntyre, S., and J. M. Melack. 1984. "Vertical mixing in Amazon floodplain lakes." Verhandlungen International Vereinigen Limnologie 22: 1283-1287. Maclntyre, S., and J. M. Melack. 1988. "Frequency and depth of vertical mixing in an Amazon floodplain lake (L. Calado, Brazil)." Verhandlungen International Vereinigen Limnologie 23: 80-85.
272
Maclntyre, S., and J. M. Melack. 1995. "Vertical and horizontal mixing in lakes: linking littoral, benthic and pelagic habitats." Journal of the North American Benthological Society 14: 599-615. Martinelli, L A., R. L. Victoria, L. S. L Steinberg, A. Ribeiro, and M. Z. Moreira. 1996. "Using stable isotopes to determine sources of evaporated water to the atmosphere in the Amazon basin." Journal of Hydrology 183: 191-204. Mason, R. P., W. F. Fitzgerald, and F. M. M. Morel. 1994. "The biogeochemical cycling of elemental mercury: Anthropogenic influences." Geochimica Cosmochimica Acta 58: 3191-3198. Matsumura-Tundisi, T., J. G. Tundisi, A. Saggio, A. L. Oliveira Neto, and E. G. Espindola. 1991. "Limnology of Samuel Reservoir (Brazil, Rondonia) in a filling phase." Verhandlungen International Vereinigen Limnologie 24: 1482—1488. McClain, M. E., J. E. Richey, and T. P. Pimentel. 1994. "Groundwater nitrogen dynamics at the terrestriallotic interface of a small catchment in the central Amazon basin." Biogeochemistry 27: 113—127. Melack, J. M. 1984. "Amazon floodplain lakes: Shape, fetch and stratification." Verhandlungen International Vereinigen Limnologie 22: 1278-1282. Melack, J. M. 1996. "Recent developments in tropical limnology." Verhandlungen International Vereinigen Limnologie 26: 211-217. Melack, J. M. 2001. "Floodplain lakes and reservoirs in tropical and subtropical South America: Limnology and human impacts." In: The Lakes Handbook, eds. P. O'Sullivan, and C. Reynolds (Blackwell Science Ltd., Oxford, U.K.), in press. Melack, J. M., and T. R. Fisher. 1983. "Diel oxygen variations and their ecological implications in Amazon floodplain lakes." Archiv fur Hydrobiologie 98: 422-442. Melack, J. M., and T. R. Fisher. 1988. "Denitrification and nitrogen fixation in an Amazon floodplain lake." Verhandlungen International Vereinigen Limnologie 23: 2232-2236. Melack, J. M., and T. R. Fisher. 1990. "Comparative limnology of tropical floodplain lakes with an emphasis on the central Amazon." Acta Limnologia Brasiliensia 3: 1-48. Melack, J. M., S. J. Sippel, D. M. Valeriano, and T. R. Fisher. 1992. Environmental conditions and change on the Amazon floodplain: analysis with remotely sensed imagery, pp 377-387. 24th International Symposium on Remote Sensing of the Environment. ERIM, Ann Arbor, Michigan. Melack, J. M., L. L. Hess, and S. Sippel. 1994. "Remote sensing of lakes and floodplains in the Amazon Basin." Remote Sensing Reviews 10: 127-142. Melack, J. M., and L. L. Hess. 1998. "Recent advances in remote sensing of wetlands." In: Modern Trends in Ecology and Environment, ed. R. S. Ambasht (Backhuys Publisher, The Netherlands), pp. 155-170. Mertes, L. A. K. 1994. "Rates of floodplain sedimentation on the central Amazon River." Geology 22: 171-174. Mertes, L. A. K., M. O. Smith, and J. B. Adams. 1993. "Estimating suspended sediment concentrations in surface waters of the Amazon River wetlands from Landsat images." Remote Sensing of Environment 43: 281-301. Mertes, A. K. L., D. L. Daniel, J. M. Melack, B. Nelson, L. A. Martinelli, and B. R. Forsberg. 1995. "Spatial patterns of hydrology, geomorphology and vegetation on the flood-
John M. Melack and Bruce R. Forsberg plain of the Amazon River in Brazil from a remote sensing perspective." Geomorphology 13: 215—232. Mertes, L. A. K., T. Dunne, and L. A. Martinelli. 1996. "Channel-floodplain geomorphology along the SolimoesAmazon River, Brazil." Geological Society of America Bulletin 108: 1-18. Morrissey, K. M., and T. R. Fisher. 1988. "Regeneration and uptake of ammonium by plankton in an Amazon floodplain lake." Journal of Plankton Research 10: 31-48. Nortcliff, S., and J. B. Thornes. 1981. "Seasonal variations in the hydrology of a small forested catchment near Manaus, Amazonas, and the implications for its management." In: Agricultural Hydrology, eds. R. Lai, and E. W. Russel Qohn Wiley & Sons, Chichester, U.K.), pp. 37-57. Novo, E. M. L. M., F. A. Leite, J. Avila, V. Ballester, and J. M. Melack. "Assessment of Amazon floodplain habitats using TM/Landsat data." Ciencia e Cultura 49: 280-284. Padovani, C. R. 1992. Determinacao das fontes autotroficas de carbono para camaroes em um lago de varzea da Amazonia Central. M.S. Thesis. Institute Nacional de Pesquisas da Amazonia/Universidade do Amazonas, p. 72 Petrere, M. Jr., and M. C. L. B. Ribeiro. 1994. "The impact of a large tropical hydroelectric dam: the case of Tucurui in the middle River Tocantins." Acta Limnologia Brasiliensia 5: 123-133. Piedade, M. T. F., W. J. Junk, and S. P. Long. 1991. "The productivity of the C-4 grass Echinochloa polystachya on the Amazon floodplain." Ecology 72: 1456-1463. Piedade, M. T. F., S. P. Long, and W. J Junk. 1994. "Leaf and canopy photosynthetic CO2 uptake of a stand of Echinochloa polystachya on the Central Amazon floodplain." Oecologia 97: 193-201. Pinherio, P. R. C. 1996. Fracionamento, caracterizacao e dinamica do fosforo dissolvido em alguns ambientes aquaticos da Amazonia. Ph.D. Thesis. Institute Nacional de Pesquisas da Amazonia/Universidade do Amazonas, p. 120. Prance, G. T. 1979. "Notes on the vegetation of Amazonia III. The terminology of Amazonian forest types subject to inundation." Brittonia 31: 26-38. Putz, R. 1997. "Periphyton communities in Amazonian blackand Whitewater habitats: Community structure, biomass and productivity." Aquatic Sciences 59: 74-93. Putz, R., and W. J. Junk. 1997. "Phytoplankton and periphyton." In: The Central Amazon Floodplain, ed. W. J. Junk (Springer-Verlag, Berlin), pp. 207-222. Quay, P. D., D. O. Wilbur, J. E. Richey, A. H. Devol, R. Benner, and B. R. Forsberg. 1995. "The 18Q:16O of dissolved oxygen in rivers and lakes in the Amazon Basin: determining the ratio of respiration to photosynthesis in freshwaters." Limnology and Oceanography 40: 718-729. RADAMBRASIL. 1978. Levantamento de recursos naturais, geologia, geomorfologia, pedologia, vegetacao, e uso potencial da terra. Folha SA.20 Manaus. Departamento Nacional da Producao Mineral, Rio de Janeiro. Ribeiro, M. N. G., and J. Adis. 1984. "Local rainfall variability—A potential bias for bioecological studies in the central Amazon." Acta Amazonica 14: 159-174. Richey, J. E. , R. H. Meade, E. Salati, A. H. Devol, C. F. Nordin Jr., and H. M. dos Santos. 1986. "Water discharge and suspended sediment concentrations in the Amazon River, 1982-1984." Water Resources Research 22: 756-764. Richey, J. E., A. H. Devol, S. C. Wofsy, R. Victoria,
Biogeochemistry of Amazon Floodplain Lakes and Associated Wetlands and M. N. G. Ribeiro. 1988. "Biogenic gases and the oxidation and reduction of carbon in Amazon River and floodplain waters." Limnology and Oceanography 33: 551-561. Richey, J. E., L. A. K. Mertes, T. Dunne, R. L. Victoria, B. R. Forsberg, A. C. N. S. Tancredi, and E. Oliveira. 1989a. "Sources and routing of the Amazon River flood wave." Global Biogeochemical Cycles 3: 191-204. Richey, J. E., C. Nobre, and C. Deser. 1989b. "Amazon River discharge and climatic variability: 1903-1985." Science 246: 101-103. Richey, J. E., J. I. Hedges, A. H. Devol, P. D. Quay, R. Victoria, L. Martinelli, and B. R. Forsberg. 1990. "Biogeochemistry of carbon in the Amazon River." Limnology and Oceanography 35: 352-371. Roland, F., and F. A. Esteves. 1993. "Dynamics of phosphorus, carbon and nitrogen in Amazonian lake impacted by bauxite tailing (Batata Lake, Para, Brasil)." Verhandlungen International Vereinigen Limnologie 25: 925-930. Roulet, M., M. Lucotte, N. Farella, G. Serique, H. Coelho, C. J. Sousa Passes, E. Jesus da Silva, P. Scavone de Andrade, D. Mergler, J. R. D. Guimaraes, and M. Amorim 1999. "Effect of recent human colonization on the presence of mercury in Amazonian ecosystems." Water Air and Soil Pollution, 112: 297-313. Salo, J., R. Kalliola, I. Hakkinen, Y. Makinen, P. Niemela, M. Puhakka, and P. D. Coley. 1986. "River dynamics and the diversity of Amazon lowland forest." Nature 322: 254-258. Schmidt, G. W. 1973. "Primary production of phytoplankton in the three types of Amazonian waters. 3. Primary productivity of phytoplankton in a tropical floodplain lake of Central Amazonia: Lago Castanho, Amazonas, Brazil." Amazoniana 4: 379-404. Setaro, F. V., and J. M. Melack. 1984. "Responses of phytoplankton to experimental nutrient enrichment in an Amazon lake." Limnology and Oceanography 28: 972-984. Silva-Forsberg, M. C., B. R. Forsberg, and V. K. Zeidemann. 1999. "Mercury contamination in humans linked to river chemistry in the Amazon Basin." Ambio, 28: 519-521. Sioli, H. 1984. editor. The Amazon. Dr. W. Junk Publishers, Dordrecht, The Netherlands, p. 763. Silva, E. N. S. 1991. Variacao sazonal e abundancia de copepoda calanoida em um lago de varzea, Lago Calado. M.S. Thesis. Institute Nacional de Pesquisas da Amazonia/ Universidade do Amazonas. Sippel, S. J., S. K. Hamilton, and J. M. Melack. 1992. "Inundation area and morphometry of lakes on the Amazon River floodplain, Brazil." Archiv fur Hydrobiologie 123: 385-400. Sippel, S. J., S. K. Hamilton, J. M. Melack, and E. M. M. Novo. 1998. "Passive microwave observations of inundation area and the area/stage relation in the Amazon River floodplain." International Journal of Remote Sensing, 19: 3055-3074. Skole, D., and C. Tucker. 1993. "Tropical deforestation and habitat fragmentation in the Amazon: Satellite data from 1978 to 1988." Science 260: 1905-1910. Smith, L. K., and T. R. Fisher. 1985. "Nutrient fluxes and sediment oxygen demand associated with the sediment-water interface of two aquatic environments." In: Sediment Oxygen Demand: Processes, Modeling, and Measurement,
273
ed. K. J. Hatcher (American Water Pollution Control Federation), pp. 1—24. Smith, V. H. 1979. "Nutrient dependence of primary productivity in lakes." Limnology and Oceanography 24: 1051-1064. Smith-Morrill, L. 1987. The exchange of carbon, nitrogen, and phosphorus between the sediments and watercolumn of an Amazon floodplain lake. Ph.D. Thesis. University of Maryland, p. 209. Swain, E. B., D. R. Engstrom, M. E. Brighan, T. A. Henning, and P. L. Brezonik. 1992. "Increasing rates of atmospheric mercury deposition in midcontinental North America." Science 257: 784-787. Tailing, J. F. 1957. "The phytoplankton population as a compound photosynthetic system." The New Phytologist 56: 133-149. Teeri, J. A., and L. G. Stowe. 1976. "Climatic patterns and the distribution of C—4 grasses in North America." Oecologia 23: 1-12. Victoria, R. L., L. A. Martinelli, J. Mortatti, andj. Richey. 1991. "Mechanisms of water recycling in the Amazon Basin: Isotopic insights." Ambio 20: 384-387. Vollenweider, R. A. 1965. "Calculation models of photosynthesis-depth curves and some implications regarding day rate estimates in primary production measurements." Memorie dell1 Istituto italiano di Idrobiologia 18: 427-457. Waichman, A. V. 1996. "Autotrophic carbon sources for heterotrophic bacterioplankton in a floodplain lake of Central Amazon." Hydrobiologia 341: 27-36. Wassmann, R., U. G. Thein, M. J. Whiticar, H. Rennenberg, W. Seiler, and W. J. Junk. 1992. "Methane emissions from the Amazon floodplain: characterization of production and transport." Global Biogeochemical Cycles 6: 3-13. Wassmann, R., and U. G. Thein. 1994. "Spatial and seasonal variation of methane emission in an Amazonian floodplain lake." Mitteilungen Internationale Vereinigung Limnologie 25: 179-185. Watras, C. J., K. A. Morrison, J. S. Host, and N. S. Bloom. 1995. "Concentrations of mercury species in relationship to other site-specific factors in the surface waters of northern Wisconsin Lakes." Limnology and Oceanography 40: 556-565. Weber, G. E. 1997. Modelling nutrient fluxes in floodplain lakes. In: The Central Amazon Floodplain. Ecological Studies, ed. W. JJunk (Springer-Verlag, Berlin) pp. 109-117. Weber, G. E., K. Furch, and W. J. Junk. 1996. "A simple modeling approach towards hydrochemical seasonally of major cations in a central Amazonian floodplain lake." Ecological Modelling 91: 39-56. Welcomme, R. L. 1979- Fisheries Ecology of Floodplain Rivers. Longman, London. Williams, M. R., and J. M. Melack. 1997. "Solute export from forested and partially deforested catchments in the central Amazon." Biogeochemistry 38: 67-102. Williams, M. R., T. R. Fisher, and J. M. Melack. 1997a. "Chemical composition and deposition of rain in the central Amazon, Brazil." Atmospheric Environment 31: 207-217. Williams, M. R., T. R. Fisher, andj. M. Melack. 1997b. "Solute dynamics in soil water and groundwater in a central Amazon catchment undergoing deforestation." Biogeochemistry 38: 303-335. Wissmar, R. C., J. E. Richey, R. F. Stallard, and J. M.
274 Edmond. 1981. "Plankton metabolism and carbon processes in the Amazon River, its tributaries and floodplain waters." Peru-Brazil, May-June 1977. Ecology 62: 1622-1633. Worbes, M., H. Klinge, J. D. Revilla, and C. Martius. 1992. "On the dynamics, floristic subdivision and geographic distribution of varzea forests in Central Amazonia." Journal of Vegetation Science 3: 553-564.
John M. Melack and Bruce R. Forsberg Worbes, M. 1997. "The forest ecosystem of the floodplains." In: The Central Amazon Floodplain. Ecological Studies, ed. W. J. Junk (Springer-Verlag, Berlin) pp. 223— 260. Zaret, T. M., A. H. Devol, and A. dos Santos. 1981. "Nutrient addition experiments in Lago Jacaretinga, central Amazon Basin, Brazil." Verhandlungen International Vereinigen Limnologie 21: 721-724.
15 Organic Matter and Nutrients in the Mainstem Amazon River Alan H. Devol and John I. Hedges
The Amazon, like smaller rivers, is the daughter of its drainage basin. Local climate and interactions over time with the template of topography, geology, and vegetation determine the size and flow of rivers. Likewise, the compositions of the particulate and dissolved materials carried by rivers result from initially similar rainwaters that have been uniquely imprinted by contact with almost every plant, animal, and mineral in the catchment. Rivers thus provide a continuously flowing signal, recorded by isotopes, ions and molecules, of the cumulative effects of drainage basin processes such as weathering, oxidation/ reduction, gas exchange, photosynthesis, biodegradation, and partitioning. This recording is complementary to more classical methods of remote sensing based on electromagnetic radiation, but is composited over a wider range of time and space scales and includes effects of subcanopy and subsurface processes. The Amazon River is similar to other rivers in this regard, but is unusual in the size and extent of different environments its waters touch. The Amazon River is the world's largest river and drains the world's largest single catchment (-6,000,000 km2). It discharges an average of about 200,000 m3 of water per second to the Atlantic Ocean. This volume is about 5 times more than the Congo, the second largest river. The Amazon has 1100
major tributaries, three of which are nearly as large as the Congo. From its origins at about 5200 m in the Andes about 200 km from the Pacific Ocean, the Amazon goes through at least 10 name changes as it snakes its way 6500 km eastward to the Atlantic Ocean (Schreider and Schreider 1970). The flooded areas along the lower mainstem are important sources of greenhouse gases such as methane (Bartlett and Harriss 1993, Devol et al. 1994) and the latent heat release from convective precipitation in the basin is sufficient to influence global climate. The Amazon drainage basin contains 40% of the world's tropical rain forest (dos Santos, 1987) and is home to countless species of plants and animals. The river itself contains some 2000 described species of fish. Nevertheless, the Amazon is a classic river basin, with a vast central plain bordered by highlands and an extensive floodplain (varzea). Even now, the Amazon is largely undisturbed by anthropogenic activity and includes a series of hydrological and chemical regimes that are representative of world rivers. The diverse natural ecosystem of Amazonia supports a large resource-based economy. Deforestation for lumber and agriculture has made perhaps the greatest impact on Amazon ecosystems. Extensive road-building during 1960-1980 converted
276
large areas of forest to crop land and cattle pasture. Between 1978 and 1988 the amount of deforested land in Amazonia increased from 78,000 to 230,000 km2 (Skole and Tucker 1993). Mining and extraction of gold and aluminum have severely impacted the chemistry of certain Amazon River subcatchments. Development has resulted in elevated concentrations of mercury in the river, sediments, fish, and some humans. This input is associated with both gold mining (Martinelli et al. 1988) and increased erosion of mercury-rich surface soils due to land clearing (Roulet et al. 1998). Another developmental activity that has directly impacted the Amazon River system is dam building. The main purpose of these dams is
Alan H. Devol and John I. Hedges
to produce inexpensive electricity to attract investment (Bunyard 1987). However, the reservoirs they create also produce methane gas that is eventually emitted to the troposphere (Martinelli et al. 1993). Because of its immense size, the Amazon River basin plays a key role in global biogeochemical cycles. In order to predict the consequences of land use change, both locally and globally, it is necessary to understand how the Amazon interacts with its geological environment, component ecosystems, and the atmosphere, as well as how these interactions might change as a consequence of anthropogenic impacts. This is the central question that has guided Amazon River research for the past several decades.
Figure 15.1 Map of the Amazon River drainage basin. Major tributaries are denoted by open circles with numbers. Locations refered to in the text are referenced by filled circles with letters. Major geologic provinces are also noted.
Organic Matter and Nutrients in the Mainstem Amazon River
277
In this chapter we present an overview of Rainfall in the basin is seasonally and current knowledge of the geochemistry of geographically variable due to the movebiologically active elements, (e.g. C, N, P, ment of the intertropical convergence zone and O) in the Amazon River mainstem. The (Salati et al. 1979). Maximal rainfall (-7000 data and concepts upon which this summa- mm yr1) occurs on the eastern flanks of the ry is based result primarily from the CAM- Andes. Rainfall decreases to about 3500 mm REX project (Richey et al. 1990, Richey and yr1 in the northwest lowlands and about Victoria 1993). This ongoing study has 2000 mm yr1 in the northeast. There is a included a series of 13 seasonal cruises distinct local dry season that extends from between 1982 and 1991 which covered the June-August in the South and from Amazon mainstem between Vargem Grande January-March in the North. As a result, the and Obidos (1800 km), as well as a seasonal flood cycles of the north and south time-series record collected at Machanteria tributaries are not in phase (Richey et al. near Manaus (Devol et al. 1995). Sampling 1990). The inflows of lagged tributaries from and analytical methods will not be reiterat- the south and north produce a mainstem ed here, but are discussed in detail in the hydrograph that becomes progressively various cited CAMREX references. more damped downstream (Fig. 15.2). Within Brazil, the hydrograph is reasonably Geologic and Hydrologic Setting sinusoidal, with no sharp spikes. Upstream, near the border with Peru and Colombia, Several major geomorphological zones discharge varies from a low of about 20,000 distinguish the Amazon drainage basin m3 s'1 in September to a high of about (Fig. 15.1), each with a distinctive geology, 50,000 m3 s'1 in April. The minimal and peak vegetation, and local climate. The basin is discharges increase dramatically downstream flanked to the north and the south by two and at the city of Obidos (-800 km from the Precambrian shields, the Planalto das sea, Fig. 15.1) vary annually from about Guianas and the Planalto Brasileiro. The 100,000 m3 s-1 to 225,000 m3 s'1. Andean Cordillera and the Subandean The seasonal cycle of discharge in the Trough constitute the western boundary. Amazon mainstem is accompanied by a In between lies the nearly flat central plain dramatic change of roughly ten meters in of the Amazon trough where elevation river height. At high water the river overchanges only about 120 m in 3400 km. flows its banks and enters the surrounding These different source regions imprint floodplain through a complex network of distinct chemical signals on the various channels (paranas). Within Brazil the floodtributaries to the Amazon mainstem (Fig. 15.1). plain is typically 10—50 km wide (Mertes et West of the Brazilian border, the Rios Ucayali al. 1996) and is occupied by numerous and Maranon come together to form the Rio permanent and seasonal lakes. Extensive Solimoes, as the Amazon mainstem is called flooded forests and beds of floating grasses above its confluence with the Rio Negro. plus other macrophytes also cover the mainThese rivers drain primarily Andean regions. stem floodplain (varzea). The expanse of The two largest of the lower tributaries are the floodplain along the Amazon mainstem has Rio Negro, which drains highly weathered been estimated at 94,000 km2 (Sippel et al. lowland terrain, and the Rio Madeira, with 1992), but accounts for less than one third headwaters in the Bolivian Andes. The Rios of the total floodplain area within the Trombetas, Xingu, and Tapajos in the lower Amazon basin (Junk and Firch 1993). The course of the river drain primarily shield areas. Amazon floodplain has been compared to a
278
capacitor, which stores water during rising water and releases it during falling water (Fischer and Parsley 1979). At high water as much as 20,000 m3 s~l of water enter the floodplain, as compared to a net downstream discharge of about 170,000 m3 s*1 at Obidos (Richey et al. 1989)- As the river drops, floodplain water drains back into the mainstem. This seasonal flooding results in expansive areas of inundated and waterlogged soils, as well as extensive exchanges of water and chemical signals between the mainstem and its floodplain.
Elemental Biogeochemistry of the Amazon Mainstem Early on, the linkage between the geology, biology and hydrology of the Amazon River mainstem was recognized from visual-
Alan H. Devol arid John I. Hedges
ly observable features of the waters and was formalized into an Amazon river "typology" by Sioli (1950, 1951, 1956). This classification consists of visually distinct Whitewater, clearwater, and blackwater rivers. Whitewater rivers are ochre colored due to their high suspended sediment content, a prime example being the Rio Madeira. These rivers have a major portion of their drainage in the Andes mountains where physical and chemical erosion of the basement rock are sufficiently intense to produce high turbidity and dissolved ion contents. At the opposite end of the spectrum are the clearwater rivers, such as the Tapajos, that drain highly weathered, low-gradient shield areas. Clearwater rivers are generally low in suspended and dissolved solids and have high transparencies due to low concentrations of suspend-
Figure 15. 2 Amazon River discharge at three locations within Brazil between 1988 and 1992: (1) OBI = Obidos, (2) MAN = Manacapuru, and (3) VG = Vargem Grande. The horizontal dotted line represents the overbank condition at Manacapuru.
Organic Matter and Nutrients in the Mainstem Amazon River
279
Figure 15.3 Distributions of carbon, nitrogen, and phosphorus between dissolved inorganic, paniculate, and dissolved organic fractions in the Amazon River mainstem, its major tributaries and floodplain waters. Inorganic carbon in the mainstem is divided into its free CO2 and alkalinity components. The total concentration of each species is given below each individual pie diagram and the percentages of the total attributable to the different forms are given next to each pie segment. For carbon the major dissolved inorganic species are alkalinity, which is almost exclusively bicarbonate ion, and dissolved CO2 (Devol et al. 1987). The dissolved inorganic nitrogen is virtually all NO3" and phosphorus is mainly PO4~3.
280
ed sediment and DOC. Blackwater rivers, such as the Rio Negro, have low concentrations of suspended solids and are typically the color of tea due to staining by high concentrations of dissolved organic matter. These rivers drain in low-gradient areas of the central basin that contain podsols rich in bleached sands. These highly weathered soils are typically covered by a characteristic caatinga vegetation and frequently lack a distinct B-horizon. Although this classification was derived from visual characteristics of the rivers, subsequent analysis of dissolved and particulate constituents has shown that the three water types are distinctly different in patterns that reflect upstream processes. Biogeochemical reactions transform elements between particulate, dissolved organic, and dissolved inorganic phases. Therefore, it is instructive to look at the relative sizes and distributions of these various pools in the three major water types and their influence on elemental cycling in the mainstem, its tributaries and floodplain. The highest total levels of carbon, nitrogen, and phosphorus are found in the mainstem, with progressively lower levels in tributary and varzea waters (Fig. 15.3). Dissolved inorganic carbon (DIG) in Amazonian waters takes primarily two forms, carbonate alkalinity, derived mostly from Whitewater Andean drainages (Devol et al. 1995, Stallard and Edmond 1983, Gibbs 1972) and dissolved CO2 gas. In the mainstem (pH -6.2-7.2) the dominant DIG species are dissolved CO2 and bicarbonate, with virtually no carbonate (Devol et al. 1987). Consequently, particulate forms of inorganic carbon are virtually absent. Alkalinity, defined as the total equivalents of dissolved weak bases, is essentially equivalent to bicarbonate concentration and composes about 82% of the DIG in the Amazon mainstem. However, bicarbonate (alkalinity) percentages drop to an average of 57% in tributaries and 50% in
Alan H. Devol and John I. Hedges
the varzea due to generally decreasing pH in these waters. The CO2 gas component of the DIG is directly involved in biogeochemical cycling through the processes of respiration and photosynthesis. Due to the lack of carbonate or carbonate ion to buffer respiratory CO2 production, respiration drives the dissolved CO2 up and the pH down. Therefore, the high CO2 concentrations in varzea waters reflect active respiration. The second largest carbon reservoir in all environments is dissolved organic carbon (DOC), which increases as a percentage of total carbon from the mainstem, to tributaries and the varzea. Conversely, particulate carbon, which is essentially all organic, decreases in the same sequence, comprising less than 2% of the total carbon in varzea waters. Nitrogen and phosphorus both decrease continuously from the mainstem to the tributaries to the varzea. All three forms of nitrogen are present in significant quantities in all environments. The dissolved inorganic nitrogen is primarily NO3 and the particulate nitrogen is exclusively organic (Hedges et al. 1986a). In contrast, most phosphorus is present in particulate phases in all environments. The particulate P shown in Fig. 15.3 is equivalent to a concentration 7.8 pM and is similar to the total particulate P values reported by Berner and Rao (1994) of about 22 ug/g of sediment given an average sediment load of about 250 g/1 (Devol et al. 1995). According to Berner and Rao, about one quarter to one third of the particulate P is organic, with the remainder composed of several inorganic phases. The distributions of C, N, and P in the mainstem, tributaries, and varzea are related to the white-, black-, and clearwater types. The high concentrations of alkalinity in the mainstem indicate that much of this white water originated in the Andes where fresh, mineral-rich rock is being weathered. This Andean contribution to the mainstem also is seen in the higher fractions of C, N, and P
Organic Matter and Nutrients in the Mainstem Amazon River
281
The previous distribution patterns reflect in the paniculate phase and the generally greater total concentrations of all elements. variations in both sources and transport of The lower amounts of almost all con- particulate materials during different stages stituents in the tributaries reflect increased of the hydrograph. To some extent, FSS and influences of black- and clearwater rivers CSS concentrations simply reflect water that drain weathered ion-poor basins. Due sources, because a greater fraction of water to settling, particulate phase concentrations flowing through the lower Amazon during decrease even further in quiescent waters rising water is derived from turbid Andean on the floodplain. Carbonate alkalinity, tributaries (Devol et al. 1995). Physical however, is a conservative property in the processes, however, modulate this pulsed lowland Amazon system (Devol et al. 1987) input by differential particle storage and and remains relatively concentrated. This mobilization within the mainstem. The persistence indicates that the fraction of power of a river to suspend and transport mainstem water stored on the floodplain is sediments increases with turbulent energy, large relative to local contributions from and hence with river surface slope (Richards black- and clear water streams. 1982). The surface slope of the Amazon Downstream distributions of suspended mainstem is greater during rising versus particulate material and associated bioactive falling water periods (Devol et al. 1995), constituents are shown in Figs. 15.4, creating higher particle concentrations and 15.5 and 15.6, which compare falling and transport at this time (Dunne et al. 1998). rising water distributions. These distribuIn most cases, distributions of particulate tions are subdivided into coarse and bioactive elements within the mainstem corfine-size fractions, where coarse fraction is respond directly with changing abundances that material that was retained on a 65 pm of particles in the two size fractions. This mesh sieve and the fine fraction is that relationship indicates that most particulate which passed through the sieve but was species bioactive elements are closely captured on a 0.5 pm filter. The rising associated with mineral particles whose and falling portions of the hydrograph are compositions are relatively uniform characterized by distinctly different sus- throughout the hydrograph (Keil et al. 1997, pended sediment distributions. During Richey et al. 1993). Thus, fine particulate falling water, concentrations of both fine organic carbon (FPOC) concentrations vary suspended sediment (FSS, 0.5—65 pm) and upstream from about 200 pM (falling water) coarse suspended sediment (CSS, > 65 pm) to 400 pM (rising water). Coarse particulate are relatively uniform throughout the organic carbon (CPOC) concentrations mainstem. FSS varies between about 100 reach about 60 pM (falling water) and 120 and 200 mg H, while CSS is about 25% as mM (rising water) in the upper half of the abundant. During rising water, however, reach. During rising water there is a concentrations of both particle size fractions pronounced decrease to values as low as 10 are higher and generally decrease down- pM at the downstream end. Distributions of stream. At the upstream stations, FSS and fine and coarse particulate nitrogen parallel CSS concentrations are nearly twice as high those of carbon, both between rising and as during low water. CSS concentrations falling water and among cruises within each decrease proportionally more than FSS period. Particulate phosphorus concentraconcentrations and actually fall below their tions (PP, Fig. 15.6) follow the trends of FSS, corresponding falling water values at the likely because most particulate phosphorus downstream end of the reach. is in the fine-size fraction.
282
Alan H. Devol and John I. Hedges
Figure 15.4 Downstream distributions of various particulate species between Vargem Grande and Obidos (-1800 km) in the Brazilian Amazon. Distributions for each species are given for both rising-water and falling-water periods, with the particular concentration scale applicable to both periods. Species defined as follows: FSS = fine (< 65 pm) suspended sediment; CSS = coarse (> 65 pm) suspended sediment; FPOC and FPON = fine particulate carbon and nitrogen, respectively, and FPON and CPON = fine and coarse particulate nitrogen, respectively. For falling water, the solid black line is CAMREX cruise 5, dotted line is cruise 8 and thick gray line is cruise 2. For rising water, the solid black line is CAMREX cruise 4, dotted line is cruise 7, and thick gray line is cruise 3 (see Richey et al. 1990). All variables are mM except FSS and CSS, which are mg I"1.
Organic Matter and Nutrients in the Mainstem Amazon River
Figure 15.5 Downstream distributions of dissolved species between Vargem Grande and Obidos the Brazilian Amazon. Distributions for each species are given for both rising-water and falling-water periods, with the particular concentration scale applicable to both periods. Alk = alkalinity, pCO2 = dissolved CO2 gas, O2 = dissolved oxygen, PO4 = phosphate ion, DOC = dissolved organic carbon and NO3 = nitrate ion. Solid dashed and thick gray lines are as defined in Fig. 15.4. All units are uM except alkalinity, which is ueq H.
283
284
Covariations among concentrations of suspended sediment, POC, PN, and total participate phosphorus suggest that the mechanisms controlling their downstream distributions are closely linked. Indeed, plots of paniculate carbon and nitrogen for the different fractions (including the major tributaries) all show a clear dependence on suspended sediment concentration (Fig. 15.7). However, there are differences between C and N and P. Phosphorus shows the strongest relationship with an intercept of zero P at zero FSS. In contrast, both the C and N regressions have significant positive intercepts (C and N are present at zero FSS). Although this difference might be due to a small component of organic debris that is not sediment associated, it is more likely a result of the way the organic matter is associated with the sediments. Keil et al. (1997) have shown that much of the particulate organic matter in the Amazon mainstem exists in association with mineral grains and is thus a function of the total surface area of suspended sediment
Alan H. Devol and John I. Hedges
particles (Fig. 15.8). Consequently, low overall FSS concentrations in more quiescent rivers may correspond to smaller remnant particles with proportionately more surface area than an equivalent mass of larger particles. This relationship would generate downward curvature to the FPOC v. FSS relationship, thereby causing a higher intercept on the carbon axis. As has been noted previously for other rivers (Meybeck 1982), weight % total POC decreases sharply with TSS concentration in the Amazon River system (Fig. 15.9)Although all plots of this form (A/B versus B) have a boomerang shape (Berges 1997), Amazon basin waters have only about half the carbon content per unit weight compared to the rivers in the Meybeck compilation (noted by a line in Fig. 15.9). One reason for this offset may be that the Amazon samples are depth-integrated composites (Richey et al. 1986), as opposed to more conventional surface grab samples of the type largely compiled by Meybeck. Because larger particles are concentrated
Figure 15.6 Downstream distributions of total particulate phosphorus (PP) between Vargem Grande and Obidos in the Brazilian Amazon. Distributions are given for both rising- and falling-water periods. Solid dashed and thick gray lines are as defined in Fig. 15.4. All concentrations are uM.
Organic Matter and Nutrients in the Mainstem Amazon River
285
Figure 15.1 Fine (< 65 urn) and coarse (> 65 um) particulate organic carbon (POC, top), particulate organic nitrogen (PON, middle), and particulate phosphorus (PP, bottom) versus their respective fine or coarse sediment concentration. The symbol key in the lower panel applies to the entire figure.
286
lower in the river channel (Meade et al. 1985) and typically are poorer in organic carbon than fine particles (Hedges et al. 1994, Keil et al. 1997), depth-integrated samples will tend to exhibit lower percentages of organic carbon. Although there is a good correlation between POC and PN for the entire data suite, the fine and coarse fractions have distinctly different relationships (Fig. 15.10). The slopes of the two regression lines correspond to atomic N:C ratios of 0.093 (correlation coefficient = 0.96, n = 165) for the fine fraction and 0.037 (correlation coefficient = 0.84, n = 133) for the coarse fraction (see also Hedges et al. 1986a, 1994). Neither regression has a statistically significant intercept, indicating that PN is primarily organic rather than inorganic. Dissolved organic matter isolated by ultrafiltration (UDOM) has a N:C ratio of 0.028 (Hedges et al. 1994), and hence is more
Alan H. Devol and John I. Hedges
nitrogen depleted than organic matter in either particulate fraction. Phosphorus shows no relationship with organic carbon. The good relationship between C and N within the individual coarse and fine fractions indicates that their compositions are internally uniform throughout the lower mainstem. The lack of a corresponding relationship for particulate phosphorus indicates that the phosphorus fraction is primarily inorganic (Berner and Rao 1994, Devol et al. 1991, Devol et al. 1995). The downstream distributions of the dissolved species of the bioactive elements are distinctly different from those of the particulate forms (see Figs. 15.4 and 15-5). Alkalinity is high in the upper mainstem ££J313and decreases continuously downstream regardless of flood stage, although the decrease is greater for falling versus rising water periods. This trend reflects the high percentage of Andean white water
Figure 15.8 Weight percent organic carbon (%OC) versus mineral surface area (SA). Capital letters represent bulk samples from Obidos (O), Vargen Grande (V), and Manacapuru (M); lowercase letters represent data from SPLITT-fractionated samples from Vargem Grande (v) and Manacapuru (m).
Organic Matter and Nutrients in the Mainstem Amazon River
287
Figure 15.9 Weight-percent organic carbon in the participate phase versus total suspended sediment concentration (TSS, mg I-1) for the Amazon River mainstem and its major tributaries within Brazil. The solid line gives the Meybeck (1982) relationship derived from a global compilation of river data.
upstream and continuous dilution downstream by lower ionic strength clear- and blackwater tributaries (Stallard 1980, Gibbs 1972). In contrast, CO2 generally increases during falling water and remains relatively constant during rising water. The mainstem Amazon is supersaturated in CO2 gas (150-250 pM) by more than an order of magnitude (atmospheric equilibrium CO2 concentrations -10 uM). Dissolved oxygen concentrations are the mirror image of the CO2 distributions. During falling water, O2 concentrations decrease downstream. During rising water they are constant, with concentrations below atmospheric equilibrium (-250 pM) at all times. Patterns for dissolved nitrate and phosphate are similar, with downstream decreases during falling
water and relatively constant concentrations during rising water. Three processes primarily control the forms and fluxes of paniculate and dissolved materials in the Amazon mainstem: input, mixing, and within-reach processing. With respect to the data presented, the CAMREX program has taken the upstream input to the Brazilian Amazon as the measurements made at Vargem Grande (Fig. 15.1). With the notable exception of the Rio Madeira, all the lower tributaries are more dilute in most chemical species than the upstream input. Consequently, the downstream decreases seen in most variables are due primarily to mixing of Andean white water with dilute, less turbid lowland water (Gibbs 1972, Richey and Victoria 1993,
288
Alan H. Devol and John I. Hedges
Figure 15.10 Paniculate organic carbon (POC) versus paniculate nitrogen (PN) for the fine and coarse fractions of sediment in the Amazon mainstem and its tributaries.
Stallard and Edmond 1983). However, other processes must come into play. For example, distributions of most particulate components (and dissolved nitrate and oxygen) are relatively uniform during falling water. This constancy demonstrates the importance of physical processes, such as particle resuspension and in-channel chemical reactions. While it is possible to see seasonal signals in the mainstem data, e.g. rising versus falling water trends, a clearer seasonal picture is available from time series data (Fig. 15.11) obtained at the Marchanteria station (Fig. 15.1) ~50 km downstream of Manacapuru (Devol et al. 1995). At this location the rising stage of the hydrograph extends from about the beginning of November through mid-June, after which the river falls through October. At this site near the midpoint of the Brazilian Amazon, water discharge varies from about 130,000
m3s"1 at high water to about 70,000 mV1 at low water. Among the dissolved chemical species, oxygen, carbon dioxide and nitrate all show seasonal patterns that are either directly or inversely related to discharge. Dissolved phosphorous is the only bioactive element that does not show this pattern. Concentrations of particulate nitrogen, carbon and phosphorus are directly proportional to total suspended sediment (see Fig. 15.7). Alkalinity has a distribution intermediate between other particulate and dissolved constituents. To understand the mechanisms governing the seasonal cycles of bioactive elements it is helpful to know the geographic sources of mainstem water as a function of time. Devol et al. (1995) used water discharge data (from the Brazilian Deparatmento Nacional de Aguas e Energia Electrica) to calculate the fractions of water in the
Organic Matter and Nutrients in the Mainstem Amazon River
289
Figure 15.11 Average discharge (Q) and "composite hydrographs" for fine (< 65 pm) suspended sediment (FSS), coarse (< 65 pm) suspended sediment (CSS), dissolved oxygen (O2) fine particulate organic matter (FPOM), coarse particulate organic matter (CPOM), free dissolved CO2 gas (pCO2), alkalinity (Alk), phosphate, nitrate, and respiration (resp) at the Marchanteria time series station. The composite hydrographs were constructed from the time series (-10 year) data by plotting all data for a given variable by day and month as though they were all collected during a single year. The mean trend is given by the solid line and the standard deviation around that mean is the dashed line. The lower right graph shows the seasonal variation in fraction of the total mainstem discharged at Marchanteria derived from Andean drainages, major tributaries, and local lowland drainages throughout the year (redrawn from Devol et al. 1995).
290
mainstem derived from the Andes versus major tributaries and local drainages (Fig. 15.11, bottom row). The slope of the river surface at the time series station over the course of the year was also determined (not shown). Multiple regression analyses of these compositions against hydrologic parameters (individual water sources, the total discharge, and river slope) showed that FSS and alkalinity were positively correlated with the percentage of Andean water in the mainstem. Oxygen, carbon dioxide, and nitrate distributions were correlated with overall discharge. Surprisingly, CSS, FPOC, CPOC, FPN, CPN, and PP were best correlated with river slope. Phosphate and respiration rates were not significantly correlated with any hydrologic parameter.
Alan H. Devol and John I. Hedges
The relationship between alkalinity and FSS concentration and Andean source waters was expected because > 80% of suspended materials and dissolved solids in transport in the Amazon are derived from Andean regions (Gibbs 1972, Stallard 1980, Meade et al. 1985). Although the Andes are also the dominant source of CSS, its seasonal cycle is tied more to the river surface slope than the percent Andean water. In fact, the seasonal pattern of river slope has a nearly identical shape to CSS concentration (Devol et al. 1995). Deposition occurs during periods of low slope (low turbulence), whereas resuspension predominates during high slope (Meade et al. 1985). This changing balance would also explain the relationships between coarse paniculate
Figure 15.12 Oxygen concentration versus time at Marchanteria. The solid line is the solution to the numerical model discussed in the text, while the filled circles are the time series average data from Fig. 15.5 (redrawn from Devol et al. 1995).
Organic Matter and Nutrients in the Mainstem Amazon River
carbon and nitrogen and river slope. The corresponding relationship between fine paniculate carbon and nitrogen and river slope is not so clear. Although there was a partial regression association with the percent Andean water, the primary regression parameter was river slope. This pattern also may result from deposition of organic poor larger particles during periods of low slope, and resuspension at high river surface slope. The seasonal distributions of O2 and CO2 are related to respiration rate and gas exchange across the river surface. In situ photosynthesis within the turbid mainstem is virtually zero (Wissmar et al. 1981, Richey et al. 1990). Both Quay et al. (1992) and Devol et al. (1987) suggested that the Amazon mainstem is in quasi steady state with respect to dissolved O2 and CO2. If the respiration rate per unit volume is approximately constant, as it appears to be at the time series site (Fig. 15.11), then the total respiratory consumption of oxygen in the water column is almost exactly balanced by invasion of O2 across the air-water interface. Consequently, as areal respiration increases with increasing river depth, oxygen concentration is drawn down and the saturation deficit (and the gas-exchange rate) increases until the two processes balance. Thus, the seasonal cycle of dissolved oxygen is driven primarily by river depth. Devol et al. (1995) used the time series data to test the quasi steady state hypothesis in a model where change in oxygen concentration with time was equal to the integrated respiration rate plus air-water-gas exchange and a residual term (dO2/dt = respiration + gas exchange + residual). The model was then solved numerically, given the initial condition that dissolved O2 was equal to the observed value during early January, respiration rate was 0.5 uM h'1, and the residual was zero (Fig. 15-12). The best fit to the data was
291
obtained with a gas exchange boundary layer of 78 urn, a realistic value for a turbulent river. The model clearly reproduced the shape and timing of the observed distributions, but was only able to produce about 60% of the amplitude. The remaining variation is the residual term and could be due to seasonal fluctuations in respiration, gas exchange across the river surface or lateral exchange with the floodplain (Devol et al. 1995). Lateral exchange with the floodplain seems likely and also would help explain seasonal NO 3 distributions. Further evidence for mainstem floodplain exchange is seen in the stable carbon isotope data presented by Quay et al. (1992). They show that 13C depleted FPOC of Andean origin is gradually replaced with more ^C-enriched material as the river flows across Brazil to the Atlantic Ocean. PO4 concentrations exhibited little seasonality possibly due to buffering by sorption reactions (Berner and Rao 1994, Froelich 1988, Fox 1991). These data suggest that the cycling of biogeochemically important elements in the Amazon mainstem is influenced by three primary factors: source, physical processing, and biogeochemical reaction (Fig. 15-13). The ultimate source for most of the material in transport in the Amazon is the Andes. Consequently, as distance from the Andes increases, concentrations of most constituents decrease due to dilution by sediment- and ion-poor lowland tributaries. Superimposed on the source imprint, however, are the effects of physical and biochemical alteration. Particulate materials undergo a series of physical depositionremobilization cycles as they move downstream, age, and undergo chemical alteration (see also next section). DOM enters the river from the adjacent floodplain and contributes significantly to the total organic carbon pool (Richey and Victoria 1996). While in the river, the different forms of carbon are subject to oxidation and
292
modification. Carbon-driven respiration consumes dissolved oxygen from the river and produces CO2 and NO3. Consumption of O2 and the production of CO2 sets up concentration gradients across the air—river interface that drive gas exchange. In the mainstem it appears that gas exchange is counterbalanced by in situ respiration such that the river is in quasi steady state with respect to dissolved O2 and CO2. The lowland mainstem contains virtually no mineral carbonates, thus at typical pH values (6.4-7.2) nearly all of the titration alkzlinity occurs as bicar-
Alan H. Devol and John I. Hedges
bonate. Thus, it is the relative concentrations of dissolved CO2 gas and alkalinity that poise the pH of the mainstem, which is highly sensitive to respiratory CO2 injection. Respiration also liberates organic N and P. Remineralized nitrogen is quickly nitrified to nitrate, which produces a NO3 flux at Obidos that is about 30% greater than the accumulated upstream inputs (Richey and Victoria 1996). To gain further insight into the nature of the physical and chemical control one can examine Amazon mainstem biogeochemistry as reflected by organic constituents.
Figure 15.13 Schematic representation of processes affecting the cycling of biogeochemically important elements in the Amazon River. The ultimate source of most particulates is the Andes mountains. Mineral material and fine paniculate carbon are transported downstream but are deposited and eroded from local floodplains many times before being discharged to the ocean. Coarse paniculate material (CPOM) enters and dissolved organic matter (DOM) enters from the floodplain. Within the river some fraction of these carbon sources are respired consuming oxygen and producing carbon dioxide. A byproduct of the organic matter oxidation is nitrate. The dissolved oxygen and carbon dioxide levels in the river come to quasi steady state concentrations that are maintained by gas exchange with the atmosphere.
Organic Matter and Nutrients in the Mainstem Amazon River
293
Figure 15.14 Ratio of fine paniculate organic carbon (FPOC) to dissolved organic carbon (DOC) versus fine suspended sediment (FSS). Capital letters denote the mainstem stations at Obidos (O), Manacapuru (M), and Vargem Grande (V) and lowercase letters are tributaries: Negro (n), Madeira (m), Jurua (j), Japura (r), Purus (p), and lea (i).
Amazon Biogeochemistry from an Organic Perspective
patterns are consistent with a model in which organic matter is progressively degraded to smaller units, which upon The isotopic, elemental, and molecular becoming submicron in size distribute seleccomposition of organic matter carried by the tively between water and the surfaces of Amazon River provide a continuous record- minerals. The interplay of microbial degraing of the sources and reactions of biogenic dation and dynamic partitioning between matter in the drainage basin. Although only dissolved and paniculate phases modulates a tiny fraction of the organic recording flow- the forms, compositions and reactions of ing through the Amazon system has as yet organic substances in the Amazon River and been viewed, it is clear that the data stream sets the stage for their fates following disis extremely detailed and provides informa- charge into the ocean (Keil et al. 1997). tion complementary to that obtained from nutrients, gases, and other inorganic sub- Forms stances. The observed organic compositions indicate that fine and coarse paniculate The amount of organic matter carried in organic matter (FPOM and CPOM) have paniculate form by the Amazon depends contrasting histories and dynamics through- largely on concentrations of total suspended out the Brazilian mainstem. In contrast, particles (Fig. 15.7). The partitioning of organic compositions within individual size riverine organic matter between fine panicclasses are relatively uniform throughout ulate and dissolved forms as a function of different seasons and stages of the hydro- fine suspended solids is illustrated in Fig. graph (Hedges et al. 1986a, 1994). These 15.14. The direct relationship of FPOM/DOC
294
with FSS (by far the predominant particle form) indicates that an increasing fraction of transported organic matter occurs in fine particles as the overall concentration of suspended fine solids increases. There are at least two possible explanations for this relationship. One is that FPOM occur as discrete organic particles whose concentrations vary in proportion to the total amount of physically separate mineral particles in suspension. This parallel relationship, however, would have to hold over a large range of hydrodynamic conditions and for discrete organic and mineral particles of sharply contrasting origins and densities. Such close coupling is not the case for CPOM, which includes a component of discrete plant debris that is concentrated versus sand when water velocities drop in the mainstem (Richey et al. 1990). A second possibility is that the organic matter in the 0.5-63 pm size range is physically associated with minerals, so that a direct covariation of FPOM with FSS is "locked in." This relationship would explain the remarkably uniform %OC within the fine-size fraction (1.2 ± 0.2%; n = 50), versus coarse suspended solids (1.0 ± 0.5%), throughout the mainstem and year. Keil et al. (1997) have recently found that the %OC of size fractions of fine suspended solids from throughout the lower Amazon mainstem vary directly with the surface area (SA) of the component mineral particles at a slope in the range of 0.5-1.0 mg OC/m2 (Fig. 15.8). Such a direct OC/SA relationship is the hallmark of organic matter associated with mineral surfaces (Mayer 1994a,b). The observed organic concentration of 0.5-1.0 mg OC/m2 (Fig. 15.8) is typical of mineral particles suspended in many rivers (Keil et al. 1997) and coastal marine sediments and may correspond to an environmentally stable loading (Mayer 1994 a,b, Keil et al. 1994). Sorption of organic molecules from natural waters onto mineral surfaces might
Alan H. Devol and John I. Hedges
also contribute toward the observation that the FPOC/FSS relationship for Amazon system rivers (Fig. 15.8) is more linear when the particle loading is normalized to ambient DOC.
Sources Organic substances in river systems have many potential sources and are subject to extensive alteration as they pass from their origins to the lower river. Because of these challenges and the physical complexity of drainage basins, the biological and geographic provinces of riverine organic substances are best assessed on the basis of multiple characterizations at different chemical scales (nuclear, atomic, and molecular). Source inferences are most likely to be accurate when drawn from concordant information provided by multiple tracers with contrasting sensitivities to natural processes and their rates. The sources of organic materials in the Brazilian Amazon have now been assessed based on the elemental (N/C), stable carbon isotopic, radiocarbon, and major biochemical (lignin, carbohydrates, and amino acids) compositions of dissolved organic matter isolated by ultrafiltration (UDOM) and fine and coarse particulate organic matter (CPOM and FPOM: Ertel et al. 1986, Hedges et al. 1986a, b, 1994). These characterizations indicate that each of the three organic forms is uniform in its composition throughout the Brazilian mainstem, although each is compositionally distinct from the other two. Stable carbon isotopic compositions are among the most persistent characteristics of organic substances (Fry and Sherr 1984). The averages and ranges (± 1 standard deviation) of the delta 13C values of dissolved, coarse, and fine particulate organic matter from the Amazon mainstem (Fig. 15.15a) are in the range of -26 to -30%o, with UDOM falling toward the negative
Organic Matter and Nutrients in the Mainstem Amazon River
extreme (Fig. 15.15a). These values are within the delta 13C range of woods and leaves from local angiosperm trees (n = 15), all of which fix carbon by the conventional C-3 pathway (Fry and Sherr 1984). In contrast, the two C-4 grasses, Paspalum repens and Echinochloa polystachya, which account for 80-90% of the floating grasses predominating vegetation of the Amazon varzea, have a delta^C range of 12.0 to -12.4%o. At the "light" carbon isotopic extreme are phytoplankton recovered from varzea lakes
Figure 15.15 (A) Total lignin phenol yield per 100 mg of organic carbon (L) versus delta 13C for Amazon mainstem samples (C=CPOM, F=FPOM, and D=UDOM) and potential source materials, and (B) atomic carbon:nitrogen atom ratio, C/N)a, versus cinnamyl:vanillyl phenol ratio (C/V) of the lignin oxidation products from the same samples.
295
that exchange water with the mainstem during high water. These algae incorporate 13Cdepleted CO2 of respiratory origin and resultantly have delta 13C values in the range of -30 to -36%o (Araujo-Lima et al. 1986). Based on these comparisons (Fig. 15.15a), organic matter carried by the Amazon mainstem is not derived predominantly from either floodplain grasses or plankton en-trained from surrounding lakes. Although a grass/plankton mixture cannot be ruled out isotopically, it seems unlikely that these two disparate sources would be so well balanced throughout 1800 km of the mainstem. In addition, organic matter has similar 13C compositions in the major intervening tributaries (Quay et al. 1992), including those of the blackwater type (for example, Rios Negro and Jutai) that do not have varzea C-4 grasses. Thus, stable carbon isotope measurements point toward trees, which cover about 80% of the drainage basin, as the most likely ultimate source of organic remains carried by the lower Amazon. All three forms of organic matter in the Amazon mainstem include appreciable amounts of vascular plant remains, as indicated by substantial yields of total lignin-derived phenols per 100 mg of total organic carbon (A= 1-9 mg ). This result is definitive because lignin polymers only occur in vascular plants, which are the sole source of vanillyl, syringyl, and cinnamyl phenols (Hedges and Mann 1979, Goni and Hedges 1992). It is evident from the vertical ranges in Fig. 15.15a that coarse suspended sediments include appreciably more vascular plant remains than fine sediments and UDOM. Comparison of these ranges to those for potential plant sources indicates that neither plankton nor angiosperm woods can be the predominant organic matter source. Tree leaves appear to be a major source of all three forms of riverine organic matter, although some addition of
296
wood, or preferential loss of nonlignin components (for example, polysaccharides) from leaves, would be necessary to explain the higher lignin content of CPOM. Additional resolution of these potential sources is possible via a combination of elemental and lignin phenol analysis. For example, the four potential organic matter sources vary characteristically in their average atomic C/N ratios (Fig. 15.15b). Plankton from varzea lakes are nitrogenrich, as reflected by low C/N values of about 5-8. Microorganisms such as bacteria and fungi, are also nitrogen-rich (C/N < 10). Tree leaves and varzea grasses have much higher C/N ratios averaging about 25 and 70, respectively. Woods contain high concentrations of N-free biopolymers such as polysaccharides and hence exhibit extremely high (> 100) C/N ratios. As with the previous figure, the compositional ranges of the three forms of riverine organic matter are remarkably small within the elemental spread of the different likely sources. Again, coarse and dissolved organic matter plot near the C/N range of tree leaves. FPOM, however, is considerably more nitrogen-rich, with an average C/N near 11. This high nitrogen content could result either from the presence of a major fraction of plankton or microbial remains, or from some process that has selectively added nitrogen to fine particulate material. The other axis of this property/property plot is the C/V lignin ratio (Fig. 15.15b), which is the weight ratio of total cinnamyl phenols to total vanillyl phenols. This tracer is based on the general relationship that cinnamyl phenols are obtained from the CuO oxidation of nonwoody plant tissues, whereas vanillyl phenols are produced by all vascular plant tissues, including woods. Phytoplankton do not produce either cinnamyl or vanillyl phenols, and hence have no effect on C/V ratios. Woods produce only trace amounts of cinnamyl
Alan H. Devol and John I. Hedges
phenols and have C/V ratios of effectively zero. In contrast, leaves of trees from the Amazon basin and the two floodplain grasses yield C/V ratios on the order of 0.2 and 1.3, respectively. All three forms of riverine organic matter produce measurable amounts of cinnamyl phenols (C/V > 0.1), and hence contain remains of nonwoody vascular plant tissues. The measured values are within the range of local tree leaves and over an order of magnitude lower than for floodplain grasses. Although not shown, the syringyl/vanillyl phenol ratios (S/V) of these samples are also similar to those of leaves. Thus all the measured isotopic, elemental, and lignin parameters indicate that leaves of locally predominant angiosperm trees are the major source of dissolved and particulate organic materials in the lower Amazon mainstem. Lack of evidence for input from the C-4 grasses that predominate in the varzea, point toward upland (terra firme) forests as the geographic province of these leaf remains.
Degradation In spite of similar sources, the coarse, fine, and dissolved organic components of Amazon River water exhibit consistent biochemical patterns that reflect sharply contrasting degradation histories. For example, CPOM, FPOM, and UDOM yield markedly smaller percentages (18, 15, and < 5%) of total sugars (aldoses) than their angiosperm leaf source. This sequence of depletion is accompanied by distinct compositional trends among individual sugars. One of the most outstanding differences is a consistent decrease with particle size in the relative concentration of glucose, and a corresponding increase (on a glucosefree basis) in the percent yields of the two deoxy sugars, rhamnose, and fucose (Fig. 15.l6a). The four potential sources of riverine organic matter yield 30-70% percent of
Organic Matter and Nutrients in the Mainstem Amazon River
their total aldoses as glucose. Glucose levels among total sugar mixtures from CPOM fall near the center of this range. Glucose yields from FPOM (~35%) lie at the bottom extreme of this interval and UDOM produces less glucose (20-30%) than any source. In comparison, percentage yields of deoxy sugars increase from the middle of the source field for CPOM, to the upper extreme for FPOM, to well beyond all the measured sources for UDOM. Since glucose percentages characteristically decrease, and
Figure 15.16 (A) Percent glucose versus percent rhamnose plus fucose on a glucose-free basis (%(Rha+Fuc)b) for Amazon mainstem samples and sources, and (B) percent b-alanine plus g-aminobutyric acid (%(BALA+GABA)) versus the vanillic acid:vanillin ratio (Ad/Al)v. Symbols are as in Fig. 15.13.
297
deoxy sugar percentages increase, with advancing degradation of carbohydratecontaining materials (Cowie and Hedges 1994, Hedges et al. 1999), the pattern in Fig. 15.l6a suggests that coarse, fine, and dissolved riverine organic matter become increasingly degraded. Because this sizerelated trend extends well beyond the limits of the measured sources in the case of UDOM, it appears to be caused by selective alteration of leaf remains. However, an unmeasured additional source rich in deoxy sugar, such as bacteria (Cowie and Hedges 1986), cannot be ruled out on the basis of these data alone. Complementary information on the diagenetic histories of organic fractions in the Amazon mainstem can be drawn from comparisons of their lignin and amino acid compositions (Fig. 15.l6b). This approach is particularly useful because all the measured sources give low yields (< 2 mole %) of the two nonprotein amino acids, b-alanine and g-aminobutyric acid, and low ratios (0.1-0.3) of vanillic acid to vanillin. As before (Fig. 15-l6a), CPOM exhibits a composition that is in the range of the natural sources. FPOM and UDOM plot at increasing offsets along a trajectory that again leads away from all measured potential sources. Although not illustrated, the same size fractions from major Amazon tributaries exhibit similar patterns (Hedges 1994). The elevated nonprotein amino acid and vanillic acid yields of Amazon UDOM are characteristic of heavily degraded organic matter in soils (Hedges et al. 1986a) and deep marine sediments (Cowie and Hedges 1994, Cowie et al. 1995). In the case of these two parameters (Fig. 15.l6b), there are no known examples of fresh biological materials having such high values. Thus, organic matter in the Amazon mainstem becomes increasingly degraded as the size of the associated particles becomes smaller. Amon and Benner (1996) report a similar
298
situation for DOC. They isolated high molecular weight DOC ( > 1 kDa) and low molecular weight DOC ( < 1 kDa) and found that the bacterial availability of high molecular weight DOC was significantly greater. The observation that the large diagenetic offsets among the size fractions are observed in the major tributaries and do not increase downstream within the mainstem, indicates that degradation occurs predominantly on the landscape or in much lower order streams.
Partitioning Consistent patterns among the nitrogen contents of organic matter in the Amazon mainstem suggest that the previous source and degradation signatures are overprinted by a third process, selective fractionation of organic substances between water and solid phases within the catchment. In addition to the previously presented concentration patterns (Fig. 15.3), compositional evidence for selective partitioning occurs at three levels of nitrogenous organic materials. Elementally, the atomic N/C ratios of UDOM (0.2-0.25), CPOM (0.30-0.45) and FPOM (0.45-0.55) become progressively greater. The carbon-normalized yields of total hydrolyzable amino acids (THAA) also increase in the same order (Fig. 15.17). This sequence (FPOM > CPOM > UDOM) is seen even at the molecular level (Fig. 15.17), where nitrogen-rich basic amino acids (for example, lysine and arginine) are concentrated in FPOM at increased ratios to the sum of basic plus acidic amino acids (aspartic and glutamic acids). These patterns are consistent throughout all the measured tributaries (Hedges et al. 1994), even though higher relative yields of amino acids from FPOM than CPOM goes against the general trend that advanced degradation results in lowered biochemical concentrations (Cowie and Hedges 1994). Because all other diage-
Alan H. Devol and John I. Hedges
netic parameters indicate that smaller organic materials are more degraded (Fig. 15.16), the pattern in Fig. 15.17 suggests a fractionation of the degradation products of CPOM, such that nitrogenous substances are concentrated in FPOM at the expense of UDOM. The previous patterns in nitrogen distribution may result from preferential uptake of dissolved nitrogenous organic substances from natural waters onto minerals. It is known from experiments with pure biochemicals (Theng 1979, Hedges and Hare 1987), artificial polymers (Hedges 1978, Letey 1994) and natural organic matter (Henrichs and Sugai 1993, Wang and Lee 1994) that clay minerals sorb nitrogenous organic substances in preference to nitrogen-free counterparts, as well as basic amino acids over acidic amino acids. The primary explanation for this affinity appears to be that most organic amines carry locally positive charges at the pHs of environmental waters, whereas clay minerals are negatively charged. Opposite charges lead to long-range attractive forces and tend to
Figure 15.17 Ratio of basic amino acids (B), (lysine and arginine) to the sum of basic plus acidic amino acids, (B+A), (aspartic and glutamic acids) versus carbon-normalized yields of total hydrolyzable amino acids (THAA). Symbols are as in Fig. 15.13-
Organic Matter and Nutrients in the Mainstem Amazon River
299
guide and anchor nitrogen-bearing mole- collected in the early 1980s from about 400 cules to clay mineral surfaces (Hedges et al. km upstream of Manaus (Fig. 15.1) had a 1994). Once tethered, large molecules can A14C (+227%o) that was indistinguishable then spread onto mineral surfaces, leading from contemporary atmospheric CO2 (Fig. to multiple-site binding that is entropy 15.18). This similarity indicates material driven and difficult to overcome because all exhibiting a short (< 5 yr) cycling time attachment must be broken simultaneously between biosynthesis and degradation in before release. This mechanism, however, the basin (Hedges et al. 1986b). FPOM from applies only for organic molecules that are the same water had a A14C near +20%o. This sufficiently small and polar to dissolve in lower activity reflects a large fraction of water. Insoluble organic particles will be older (prebomb) carbon, and hence a affected primarily by gravity, which is not slower average turnover time in the basin. charge specific. Such radiocarbon contents are often exhibited by soil organic matter, which also Dynamics has C/N and lignin compositions typical of riverine FPOM (Hedges et al. 1994). Although only a small number of One striking result of this study was that radiocarbon analyses have been published dissolved total humic substances isolated for organic materials from the Amazon from two sites in the mainstem (Ertel et al. mainstem (Fig. 15.18), the observed 1986) exhibited A14C values of +265 to compositions reflect dynamics which are as +285%o (Hedges et al. 1986b). Although different as the previously discussed chemi- dissolved humic substances (isolated by cal compositions. A mainstem CPOM sample hydrophobic sorption onto a resin column)
Figure 15.18 The A14C of atmospheric CO2 versus year, plotted with 1981 values for the FPOC (F), CPOC (C), and UDOM (D) from the Amazon mainstem.
Alan H. Devol and John I. Hedges
300
and ultrafiltered DOM are not necessarily the same material (Hedges et al. 1994), both fractions represent more than two-thirds of Amazon DOC and are likely to have similar overall radiocarbon contents. The measured A14C of the two DOM samples are substantially higher than for atmospheric CO2 at the time of collection (Fig. 15.18), meaning that a major portion of mainstem DOM had to be synthesized between I960 and 1980, when the radiocarbon content of local atmospheric CO2 was higher. A specific "age" cannot be assigned to these samples. However, division of the maximal time since photosynthesis (~30 years), by the minimal fraction of radiocarbon that could have been fixed in this time (for example, 265%o divided by the highest possible 14C input at +750%o), yields a ratio of -307(1/3), or about 100 years (Hedges et al. 1986b). This is the maximal estimated residence time of dissolved organic matter in the Amazon basin prior to export by the river. Much shorter (decadal) lags between photosynthetic uptake of DOM carbon and its discharge to the ocean are more likely. Even though CPOM, DOM, and FPOM from the same liter of mainstem water all appear to share a predominant leaf source, their average residence times between formation and export are dramatically different. The presence of contemporary CPOM indicates the potential for rapid organic matter export with minimal attending degradation. Since the time required to transport water directly down the mainstem into the ocean is on the order of a month, the older carbon in the FPOM and UDOM fractions must age somewhere on land, or in pooled groundwaters. Storage may also provide time for the extensive degradation evident from the compositions of these smaller components. The mean residence times and extents of degradation of organic matter forms in the Amazon basin, however, are not always strictly parallel. For example, FPOM moves
through the basin more slowly than UDOM, and yet is less degraded.
Model One challenge for formulating a conceptual model for organic matter processing in the lower Amazon is to explain how leaf-derived remains of contrasting size from throughout the basin might exhibit such consistent differences in their compositions and rates of export. The model must also explain how nitrogenous organic matter, which is often considered to be relatively reactive, might be concentrated in the "oldest" FPOM fraction, possibly at the expense of a more degraded, but faster moving, dissolved fraction. Because the Amazon mainstream discharges precipitation too fast to explain the long residence times of the fine and dissolved organic materials it carries, the mechanisms responsible for the observed dynamics (and presumably the corresponding compositions) must temporary storage at upstream sites within the basin. Our current attempt to explain how the previously discussed patterns might be generated is illustrated in Fig. 15.19 (see also Hedges et al. 1986a, 1994). This hypothetical scenario involves leaves from trees of the upland forest as the predominant organic matter source. Although clearly an oversimplification, this starting point is consistent with the previously discussed compositional information and the observation that the Amazon basin is 80% covered with hardwood forest where leaves constitute a major fraction of long-term productivity. This assumption also aligns with the lack of evidence for a strong influence of varzea vegetation. Upon falling to the forest floor, tree leaves are subject to rapid and extensive degradation. Advancing breakdown of leaves (and woods) by animals and microorganisms is attended by stepwise
Organic Matter and Nutrients in the Mainstem Amazon River
reduction in the size of the tissue remains, and an increase in their extent of chemical alteration, thereby imprinting the diagenetic relationships with size seen throughout the river (Fig. 15.8). Although parallel and 99% complete overall (Richey et al. 1990), the processes of physical and biochemical breakdown are not necessarily uniform or stepwise on all scales. A scheme for selective storage and transport of surviving organic substances is now needed to carry the previously imprinted degradation signal to the river and impose the additional dynamic and nitrogen signatures. The process we envision involves selective transport based on the size and nitrogen content of the degradation products. Being physically intact at the micron to millimeter scale, CPOM is the most likely of
301
the size fractions to retain the biochemical composition of its parent plant tissues. Even this fraction, however, has suffered extensive polysaccharide loss. Since organic particles in the sand-size range are too large to infiltrate the soil, they must be transported by throughfall, wind, or slopewash into the river system. Because the half-life of physically intact leaf tissue is short in wet tropical settings, only relatively fresh debris will be available for export to streams. More complete degradation of vascular plant remains to dissolved intermediates is the rule, because this is the only pathway to uptake through the cell walls of microorganisms. Since respiration is the fate of 99% of the organic production in the Amazon basin, at least that fraction of all biomass must at one time or another become
Figure 15.19 A conceptual model of organic matter processing in the Amazon basin.
302
dissolved. The possible fates of degradation byproducts dissolved by leaching or microbial breakdown of forest litter are more varied than for coarse debris because DOM can percolate into the soil and interact with minerals. Many nitrogen-rich organic molecules are preferentially sorbed onto minerals and will be preferentially retained in soils until mobilized into streams by colloid infiltration (Kaplan et al. 1993) or erosion (Hedges et al. 1994). In contrast, highly soluble organic substances with neutral or net negative charges should pass rapidly through the soil (if not degraded) via groundwater and into local streams and rivers. Supporting this "regional chromatography" model is growing evidence that organic substances sorbed to soil and sedimentary minerals are physically protected from microbial degradation (Nelson et al. 1993, Keil et al. 1994). Such shielding would greatly improve the odds that nitrogenous organic matter might persist unusually long in association with soil minerals on the landscape and after erosion into streams and rivers. The elevated concentration of amino acids in the older FPOM fraction might be thus explained. The direct relationship of organic carbon concentration with surface area within fine particulate material suspended in the lower mainstem (Fig. 15.8) attests to a partitioning history. It is not clear, however, whether this pattern is predominantly remnant from the soil or might be reinforced by active partitioning in the river system. Evidence for sorption of dissolved humic acid below the confluence of the particle-free Rio Negro with the mainstem Amazon (total suspended solids -250 mg/L) suggests some withinriver partitioning. Although internally consistent, the previous paradigm is still largely hypothetical. The notion that nitrogenous organic materials within the Amazon system are preferentially sorbed onto the surfaces of soil miner-
Alan H. Devol and John I. Hedges
als is yet to be tested with local samples. The alternative explanation that organic nitrogen may be immobilized to soil minerals in the form of remnant "microbial metabolites" that were never dissolved (Hedges and Oades 1997) remains to be tested. The question of the fate of DOM in groundwater is also open. Evidence has been presented that DOM undergoes major changes in concentration and composition at the interface with small Amazon streams (McClain et al. 1997), a process which is not specifically considered here. Bordering wetlands also may contribute DOM directly and efficiently to rivers with little or no effect by exposure to mineral particles. In addition, very labile molecules that might disproportionately fuel respiration in the mainstem could exhibit different pathways and dynamics than the more slowly cycling organic matter forms which predominate among the fractions we have thus far characterized. The present paradigm addresses the major sources and reactions of only a small component of the remarkably complex assemblage of organic chemicals derived from 6 million square kilometers of Amazon basin. The 600 micromoles of organic carbon in every liter of Amazon water (Fig. 15.3), is equivalent to 10 large (-2000 molecular weight) organic molecules from every square centimeter in the entire drainage basin. Given that every molecule of this size itself carries thousands of "bites" of structural and isotopic data, the immense wealth of environmental information provided by the river remains almost completely untapped.
Conclusions The emerging picture of the Amazon is an enigmatic one of a metabolically active mainstem transporting predominantly refractory organic materials (Benner et al. 1995, Amon and Benner 1996). The studied
Organic Matter and Nutrients in the Mainstem Amazon River
303
reach of the Amazon mainstem (Fig. 15.1) carbon in transit (Hedges et al. 1986a, 1994). receives on average about 26 Tg yr1 of Another potential source of labile substrates DOC 'and 13 Tg yr1 of total POC from its may groundwaters directly entering the tributaries and varzea. The corresponding mainstem. This pathway is indicated for outputs at Obidos are 22 Tg yr1 for DOC methane, which must come from a reducing and 12.5 Tg yr1 for total POC. subenvironment outside the river and Independently measured rates of respiration expresses sharp concentration gradients and CO2 evasion in the Brazilian reach are from the banks toward the center of the both order of 18 Tg yr1 (Richey et al. 1990), mainstem (Bartlett et al. 1990). Suboxic suggesting the lower Amazon mainstem is in groundwaters would be a particularly quasi steady state. The observed respiration effective route of introduction of methane, rates of 1.0-0.5 pM C tr1 (Benner et al. 1995, ammonia, and low molecular weight Devol et al. 1995) are sufficient if unabated organic acids whose reaction rates might be to completely oxidize all transported sufficiently slow in the absence of dissolved organic matter within a period of 25—50 O2 to allow large-scale storage and longer days, which is shorter than the mean water distance transport. travel time through the -1800 km reach. The contrasting distributions and dynamCorresponding rates on in situ organic ics of the different types of biogenic nitrogen oxidation are sufficient to generate materials borne by the Amazon mainstem a nitrate flux past Obidos that is 30% greater reflect a large range of sensitivities to differthan the sum of all the tributary and varzea ent processes at work within the basin and inputs (Richey and Victoria 1996). Given their anthropogenic perturbations. For that most POM and DOM in transport example, dissolved O2 and CO2 maintain a appears relatively refractory (Hedges et al. delicate balance between rapid rates of 1986a, 1994), the half-life of the small substrate input from the landscape, respiraorganic fraction that is reactive must be tion within the turbulent mainstem, and gas much shorter, on the order of hours to a few exchange across the river surface. Because days. This model of a small pool of these small pools of dissolved gases and highly reactive substrate is in agreement labile organic molecules turnover on the with experimental evidence that bacteria in scale of hours to days, they reflect predomthe river are limited by low concentrations inantly active processes occurring locally. of oxidizable organic carbon (Benner et Such chemical species could therefore indicate perturbations such as loading with al. 1995). Such fast cycling substrates must either sewerage and agricultural wastes or changes come from a nearby source on the flood- in biological activity due to introduction of plain, or express a much greater reactivity toxic materials. Nitrate and methane can be within the river than during transport to its similarly responsive, whereas dissolved waters. One proximate source of potentially phosphate is buffered by exchange with a reactive organic matter is the C-4 grasses huge pool of mineral-bound ion and that occupy much of the varzea. Based on responds little if any to local change. The the 13C-rich composition of CO2 evading the amounts of suspended particulate materials lower mainstem, Quay et al. (1992) have are so strongly swayed by water turbulence, calculated that as much as 40% of the and hence hydrographic stage, that environrespired carbon may be derived from C-4 mental effects of human activities such as varzea grasses, which compose a minor deforestation, mining, and cultivation are (< 10%) component of the total organic difficult to sort out based concentrations
304
alone. Compositions of paniculate materials, especially within specific size classes, are much more uniform, establishing a background of relatively diagnostic ratios against which human alterations can be more sensitively detected. The 14C compositions of organic matter in the coarse and fine-size fractions indicate, however, that characteristic response times will vary with particle size. Coarse plant debris in the sand fraction is exported on an annual time scale—slow enough to represent major regions of the landscape, yet sufficiently rapid to denote recent activities such as forest clearing. In
Alan H. Devol and John I. Hedges
contrast, organic matter associated with fine particles will change in composition on a decadal to century time scale. Such cycle periods are more apace with processes related to extensive alterations in land and water usage, as well as local climate change. Dissolved organic constituents probably have the greatest diagnostic potential because of the information richness of the complex molecular blend, the spectrum of residence times represented, and the sensitivity of the chemical recording to myriad subsurface processes that are invisible to electromagnetic radiation.
Literature Cited Amon, R. M. W., and R. Benner. 1996. "Bacterial utilization of different size classes of dissolved organic matter." Limnology and Oceanography 41: 41-51. Araujo-Lima, C. A. R. M., B. R. Forsberg, R. Victoria, and L. Martinelli. 1986. "Energy sources for detritivorous fishes in the Amazon." Science 234: 1256-1258. Bartlet,t K. B., and R. C Harriss. 1993. "Review and assessment of methane emissions from wetlands." Chemosphere 26: 261-320. Bartlett, K. B., P. M. Grill, J. A. Banasi, J. E. Richey, and R. C. Harriss. 1990. "Methane flux from the Amazon River floodplain: emissions during rising water." Journal of Geophysical Research 95: 16733-16738. Benner, R., S. Osphal, G. Chin-Leo, J. E. Richey, and B. R. Forsberg. 1995. "Bacterial carbon metabolism in the Amazon River system." Limnology and Oceanography 40: 1252-1270. Berges, J. A. 1997. "Ratios, regression statistics, and 'spurious1 correlations." Limnology and Oceanography 42: 1006-1007. Berner, R. A., and J-L Rao. 1994. "Phosphorus in sediments of the Amazon River and estuary: Implications for the global flux of phosphorus to the sea." Geochimica Cosmochimica Acta 58: 2333-2339. Bunyard, P. 1987. "Dam building in the tropics: Some environmental and social consequences." In: The Geophysiology of Amazonia, ed. R. E. Dickinson (John Wiley & Sons), pp. 63-68. Cowie, G. L., and J. I. Hedges. 1994. "Biochemical indicators of diagenetic alteration in natural organic matter mixtures." Nature 369: 304-307. Cowie, G. L., J. I. Hedges, F. G. Prahl, and G. J. De Lange. 1995. "Elemental and major biochemical changes across an oxidation front in a relict turbidite: An oxygen effect." Geochimica Cosmochimica Acta 59: 33-46.
Devol, A. H., P. D. Quay, J. E. Richey, and L. A. Martinelli. 1987. "The role of gas exchange in the inorganic carbon, ox