Advances in Algal Biology: A Commemoration of the Work of Rex Lowe (Developments in Hydrobiology)

Advances in Algal Biology: A Commemoration of the Work of Rex Lowe

Developments in Hydrobiology 185

Series editor

K. Martens

Advances in Algal Biology: A Commemoration of the Work of Rex Lowe

Edited by

R. Jan Stevenson1, Yangdong Pan2, J. Patrick Kociolek3 and John C. Kingston4 1 Department of Zoology, Michigan State University, East Lansing, Michigan 48824, USA Environmental Science and Resources, Portland State University, Portland, Oregon 97207, USA

2

3

California Academy of Sciences, Golden Gate Park, San Francisco, California 94118, USA 4

Natural Resources Research Institute, 1900 E. Camp St., Ely, Minnesota 55731, USA

Reprinted from Hydrobiologia, Volume 561 (2006)

123

Library of Congress Cataloging-in-Publication Data

A C.I.P. Catalogue record for this book is available from the Library of Congress.

ISBN 1-4020-4782-7 Published by Springer, P.O. Box 17, 3300 AA Dordrecht, The Netherlands

Cover illustration: A species of Draparnaldia from the Great Smoky Mountains. Draparnaldia is a morphologically elaborate and beautiful genus of filamentous green algae found mainly in aquatic habitats with low human disturbance. Photo credit: Rex Lowe

Printed on acid-free paper All Rights reserved  2006 Springer No part of this material protected by this copyright notice may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording or by any information storage and retrieval system, without written permission from the copyright owner. Printed in the Netherlands

TABLE OF CONTENTS

Preface Algology and algologists at Bowling Green, a short history R.L. Lowe

vii–viii 1–11

Rexia erecta gen. et sp. nov. and Capsosira lowei sp. nov., two newly described cyanobacterial taxa from the Great Smoky Mountains National Park (USA) D.A. Casamatta, S.R. Gomez, J.R. Johansen

13–26

Large-scale regional variation in diatom-water chemistry relationships: rivers of the eastern United States D.F. Charles, F.W. Acker, D.D. Hart, C.W. Reimer, P.B. Cotter

27–57

Short-term effects of elevated velocity and sediment abrasion on benthic algal communities S.N. Francoeur, B.J.F. Biggs

59–69

The effects of pH on a periphyton community in an acidic wetland, USA J.L. Greenwood, R.L. Lowe

71–82

Food limitation affects algivory and grazer performance for New Zealand stream macroinvertebrates J.R. Holomuzki, B.J.F. Biggs

83–94

Benthic diatom communities in subalpine pools in New Zealand: relationships to environmental variables C. Kilroy, B.J.F. Biggs, W. Vyverman, P.A. Broady

95–110

The relationships among disturbance, substratum size and periphyton community structure M.R. Luttenton, C. Baisden

111–117

Relationships between environmental variables and benthic diatom assemblages in California Central Valley streams (USA) Y. Pan, B.H. Hill, P. Husby, R.K. Hall, P.R. Kaufmann

119–130

Response of periphytic algae to gradients in nitrogen and phosphorus in streamside mesocosms S.T. Rier, R.J. Stevenson

131–147

Comparing effects of nutrients on algal biomass in streams in two regions with different disturbance regimes and with applications for developing nutrient criteria R.J. Stevenson, S.T. Rier, C.M. Riseng, R.E. Schultz, M.J. Wiley

149–165

Differential heterotrophic utilization of organic compounds by diatoms and bacteria under light and dark conditions N.C. Tuchman, M.A. Schollett, S.T. Rier, P. Geddes

167–177

Using diatom assemblages to assess urban stream conditions C.E. Walker, Y. Pan

179–189

vi Developing and testing diatom indicators for wetlands in the Casco Bay watershed, Maine, USA Y.-K. Wang, R.J. Stevenson, P.R. Sweets, J. DiFranco

191–206

Diatom assemblages and their associations with environmental variables in Oregon Coast Range streams, USA C.L. Weilhoefer, Y. Pan

207–219

Algal assemblages in multiple habitats of restored and extant wetlands L. Zheng, R.J. Stevenson

221–238

Ecology and assessment of the benthic diatom communities of four Lake Erie estuaries using Lange-Bertalot tolerance values G.V. Sgro, M.E. Ketterer, J.R. Johansen

239–249

Hydrobiologia (2006) 561:vii–viii  Springer 2006 R.J. Stevenson, Y. Pan, J.P. Kociolek & J.C. Kingston (eds), Advances in Algal Biology: A Commemoration of the Work of Rex Lowe DOI 10.1007/s10750-005-1600-8

Preface The authors of papers in this special issue of Hydrobiologia want to express their respect and gratitude for the inspiration that Dr Rex Lowe has provided. Rex has had extraordinary effects on his students and colleagues with his engaging personality, sometimes dangerous senses of humor and fun, and infectious fascination with the biodiversity and ecology of algae. The occasion for this special issue has been the coincidence of Rex’s 60th birthday, having 60 graduate students, and having taught a field phycology course at the University of Michigan Biological Station for 30 years, where many of us met Rex. The authors include colleagues and graduate students as well as academic grandchildren and great grandchildren from his graduate students that are professors at universities throughout the US. Rex got his Ph.D. from Iowa State University in 1970 studying with Dr John Dodd, who had a great influence on Rex’s commitment to students. Immediately afterward, Rex joined the faculty in the Department of Biological Sciences at Bowling Green State University (BGSU). Throughout his career at BGSU, Rex has maintained an active research program with his graduate students. He has published more than 80 papers, books, and book chapters. He has also been recognized with 4 distinguished teaching awards from 1974 to 1996 at BGSU, which demonstrates that Rex achieves his goal of providing students with ‘‘one of the best and most memorable experiences of their lives.’’ The field of algal biology in the United States has benefited greatly from Rex’s efforts. Through his pioneering leadership in the North American Benthological Society during the 1970s, benthic algal ecology has become a wholly integrated element of research conducted by stream ecologists. His review of diatom ecological tolerances provided a key reference for researchers in environmental assessment. Rex and his students have explored algal flora, described new species, and documented regional biodiversity throughout the US for the last 30 years. This special issue of Hydrobiologia exemplifies these contributions to algal biology. The issue is

introduced by Rex himself, with a description of his career, students, and their research contributions. Papers were contributed in three broad areas of algal biology: aquatic ecology, environmental assessment, and systematics. Effects of disturbance (flow and herbivory), substrates, sediments, light, and organic compounds on benthic algae in streams, lakes, wetlands, and the Great Lakes are explored in aquatic ecology. Algal biodiversity is related to human alterations of streams and wetlands in environmental assessment. Several regional studies document changes in algal species composition and their relation to human disturbances; plus they are used to develop algal indicators and multimetric indices of biological condition. Three ‘‘systematics’’ papers were submitted to a different journal for publication. These included a paper on the development of a diatom flora for freshwater ecosystems in the continental United States and new species of diatoms and cyanobacteria described from Michigan and the Great Smoky National Park. These contributions to algal biology are a tribute to Rex’s inspiration of his students and colleagues. Rex and his wife Sheryn create a special environment for learning and intellectual debate at their home in Bowling Green and their summer retreat at the University of Michigan Biological Station. The warm and friendly atmosphere fosters open dialogue, creativity, and collegiality that attracts fellow scientists and students and produces interactions that have grown into extensive advances in the fields of ecology, biodiversity, systematics, and environmental assessment as well as algal biology. Rex has inspired us to study, teach, and live our lives with more enthusiasm and satisfaction in a way that helps us understand and protect the wonder and diversity in the world around us. Thank you Professor Lowe. R. Jan Stevenson Yangdong Pan Pat Kociolek John Kingston and all other authors of these papers.

viii

List of Reviewers Reviewer

Reviewer institution

Eugene Stoermer Liz Bergey Steve Kohler Christopher Peterson Paul McCormick Jennifer Winter Scott Hagerthey Marina Potapova Robert Sinsabaugh Dean DeNicola Rhonda McDougal Harry Leland Alan Steinman Rex Lowe Walter Dodds Christine Weilhoefer Russel Kreis Don Charles Evelyn Gaiser R. Jan Stevenson Yangdong Pan J. Pat Kociolek

The University of Michigan The University of Oklahoma Western Michigan University Loyola University of Chicago United States Geological Survey Ontario Ministry for the Environment South Florida Water Management District The Academy of Natural Sciences of Philadelphia University of New Mexico Slippery Rock University Ducks Unlimited Canada United States Geological Survey Grand Valley State University Bowling Green State University Kansas State University Portland State University US Environmental Protection Agency The Academy of Natural Sciences of Philadelphia Florida International University Michigan State University Portland State University California Academy of Sciences

Hydrobiologia (2006) 561:1–11  Springer 2006 R.J. Stevenson, Y. Pan, J.P. Kociolek & J.C. Kingston (eds), Advances in Algal Biology: A Commemoration of the Work of Rex Lowe DOI 10.1007/s10750-005-1601-7

Algology and algologists at Bowling Green, a short history Rex L. Lowe1,2,* 1

Department of Biological Sciences, Bowling Green State University, Bowling Green, Ohio, 43403, USA University of Michigan Biological Station, Pellston, Michigan, 49762, USA (*Author for correspondence: E-mail: [email protected]) 2

Key words: algae, periphyton, ecology, environment, assessment, systematics

Abstract This paper summarizes the past 34 years of studies of algae by Rex Lowe and his students and collaborators at Bowling Green State University, Ohio, USA. Sixty-two students have received graduate degrees in this academic program focusing on systematics, ecology and environmental assessment. The taxonomic/ floristic research initially focused on northern Ohio streams but is now continental to international in scope focusing on the algal flora of the Great Smoky Mountains National Park and on the South Island of New Zealand. Ecological research has focused on factors that regulate the structure and function of benthic algae. Variables that have been examined include abiotic resources (nutrients and light), grazers and physical disturbance. Studies on environmental assessment have focused primarily on the impact of pointsource loads of chemicals into water bodies.

Introduction I was surprised, humbled and greatly honored upon learning of my students’ plans to organize a celebration in 2003 to recognize, summarize and reflect on studies of algae undertaken at Bowling Green State University during the past 34 years. When asked to write the introductory chapter to a collection of scientific papers contributed by former academic advisees and research collaborators I decided instead to write about the former students themselves and how their ambitions and interests helped shape my own. Sixty-two graduate students matriculating through the Algal Ecology Laboratory at Bowling Green State University coinciding with my 60th birthday seems an appropriate time for retrospection. The theme of this special volume is ‘‘Benthic algae: Their Roles in Aquatic Ecology, Systematics, and Environmental Assessment.’’ Although these subdisciplines encompass an expansive continuum of questions of scientific interest it has become

increasingly clear that ecology and systematics are inseparable (Kociolek & Stoermer, 2001). One cannot accurately describe species interactions and environmental relationships if the species are unknown. This issue is particularly critical when attempting to extrapolate insights gained from research to other algal communities separated by time (paleolimnolgy) or space (biogeography). In addition, the biological species concept has been difficult to apply to algae given their cryptic sexual behavior. Indeed, most algal ‘‘species’’ have not been observed reproducing sexually. Thus, morphology and increasingly ecology are being employed in algal species concepts. The inseparability of ecology and systematics has been extended to the application of algal communities in environmental assessment (Bahls, 1973, 1974; Descy, 1979; Stevenson & Lowe, 1986; Morgan, 1987; Biggs, 1989; Dixit et al., 1992; Lowe & Pan, 1996; Stevenson & Pan, 1999). Accurate species identification is essential for accurate environmental assessment whether the algal communities are modern or fossil.

2 Students in the algal graduate program over the past three decades pursued a continuum of interests in the three sub disciplines (systematics, ecology and environmental assessment). This research continuum led to close and fertile interstudent collaborations that continue today. Many of the students that matriculated through the program are at the forefront of the field today. I am extremely proud of them. I will divide my summary of these research pursuits into the three sub disciplines, while recognizing their interrelatedness.

Systematics and morphology Our initial interest in the algae laboratory was documenting the local flora of northwest Ohio, USA. Bowling Green rests in an area formerly known as ‘‘The Great Black Swamp,’’ a glacial remnant of Lake Erie (Kaatz, 1955). Streams in northwest Ohio are all low-gradient in nature and strongly influenced by agricultural practices. The Portage River system was studied by McCullough (1971), Jackson (1975) and Rohr (1977). Stevenson (1976), Kline (1975) and Pryfogle (1976) investigated algal communities in the Sandusky River. Acker (1977), Fisher (1980) and Lamb (1983) researched the algae of the Maumee River. The algal flora of these rivers is typical of sluggish, nutrient-rich water. Phytoplankton was typically dominated by small centric diatoms in the genera Cyclotella, Stephanodiscus and Thalassiosira (Lowe & Crang, 1972; Busch, 1974; Lowe, 1975; Lowe & Busch, 1975; Lowe & Kline, 1976; Kline & Lowe, 1976). The benthic algal flora of these sediment-rich rivers was dominated by epipelic species typical of low-gradient streams (Jackson & Lowe, 1978; Kociolek et al., 1983). Following extensive research on northwestern Ohio streams we initiated studied in the southeastern United States in the late 1970s. Camburn (1977) wrote a thesis on the diatom flora of Long Branch Creek in South Carolina that led to his substantial and profusely-illustrated manuscript (Camburn et al., 1978) in which he described eleven new diatom taxa form a flora of 268 total taxa. The appearance of so many undescribed taxa provided incentive for continued taxonomic and floristic research in the southeast. Kociolek (1982) and Keithan (1983) both conducted their graduate research in the

Great Smoky Mountains National Park. While Kociolek focused on taxonomic issues documenting the diatom flora of selected streams including the description of five new taxa (Kociolek & Lowe, 1983; Lowe & Kociolek 1984), Keithan was ecologically focused examining the role of current in structuring benthic algal communities (Keithan & Lowe, 1985). This initial research in the Great Smoky Mountains National Park made our current large-scale algal biodiversity project which is part of a larger all-taxa biodiversity inventory possible (Sharkey, 2001; Gomez et al., 2003; Potapova et al., 2003; Johansen et al., 2004). Taxonomic and floristic research in our laboratory took a more international approach in the 1990s. In collaboration with Barry Biggs at The National Institute for Water and Atmospheric Research while on a sabbatical leave in New Zealand, we observed that diatoms from many local habitats were not easily identified using taxonomic literature from the northern hemisphere. This led to a more intensive investigation into endemic New Zealand diatom species (Sabbe et al., 2001; Kilroy et al., 2003) that is still in progress. In addition to our New Zealand algal floristic work, the arrival of Sophia Passy in Bowling Green enabled us to pursue taxonomic/floristic work on diatoms from Bulgaria (Passy & Lowe, 1994) and from South Africa (Passy-Tolar et al., 1997). Taxonomy and floristics continue to one of the central areas of research interest in the Bowling Green algae laboratory.

Ecology The major focus of research activities of students matriculating through the graduate program in the algology laboratory at Bowling Green has, not surprisingly, centered on ecology. Because algal assemblages are spatially compact and respond to environmental variables relatively quickly, they are ideal subjects for students wanting to pursue a research question addressing community ecology in a limited period of time. Also, algae stand at the interface of the abiotic and biotic components of the ecosystem converting inorganic minerals to organic compounds. Thus, algal community structure, function and dynamics are potentially

3 strongly regulated by abiotic resources and/or consumers and/or disturbance. This complexity is increased when one considers that algal assemblages are composed of a large numbers of species, with species richness often exceeding 100, providing an incredibly complex system for investigation. Initial investigations were descriptive in nature, generating correlative data sets (Lamb & Lowe, 1981; Lowe et al., 1982; Millie & Lowe, 1983; Belanger et al., 1985). Bruno (1978) and Kingston (1980) researching benthic algal assemblages in an Ohio bog lake and Grand Traverse Bay, Lake Michigan were among the early researchers applying multivariate techniques in the analysis of algal communities (Bruno & Lowe, 1980; Kingston et al., 1983). Kingston’s research documented a benthic diatom assemblage living below the maximum penetration of the summer thermocline in Lake Michigan that was highly diverse and structurally stable through seasons. In contrast, he found that shallower benthic assemblages displayed strong seasonal variability in structure. Kingston’s is one of the few data sets detailing this important deep benthic community in Lake Michigan. Passy (1997) applied multivariate analyses to seasonal benthic algal community structure in the Mesta River, Bulgaria. From detailed collections of epilithon, epiphyton, epipelon, epipsammon and plocon across nutrient gradients she was able to define subsets of taxa based on both nutrient and microhabitat preferences (Passy-Tolar et al., 1999). Miller (1983) and Krejci (1985) examined algal distribution patterns at a much finer scale than had been customary in algal ecology. Both students focused on epipsammic diatom communities. Miller investigated the role of micro-topography of sand grains and its influence on diatom distribution (Miller et al., 1987). Her investigation demonstrated habitat partitioning among diatoms on sand grains with prostrate diatoms normally occupying depressions while the ridges were occupied by diatom taxa with relatively short and stout flexible stalks that enabled these forms to better resist crushing during sand drifting events. Krejci examined seasonal phenology of epipsammon in a stable spring-fed brook (Krejci & Lowe, 1987a) documenting, among other findings a preferred temperature range for the spring Meridion bloom

(Krejci & Lowe, 1987b). Krejci also documented the role of sand grain mineralogy using scanning electron microscopy (SEM) and x-ray energy dispersive spectroscopy technology for sand grain elemental analysis as an influence on diatom colonization. Krejci determined that stalked diatoms preferred quartz sand grains, which comprised 65% of the grains he examined. In contrast, motile prostrate diatoms showed no preference between quartz and feldspar sand grains (Krejci & Lowe, 1986). Krejci and Miller’s research was a strong confirmation that algae significantly exploited habitat variability at microscopic scales. Insights into microalgal ecology must focus at the appropriate scale. Greenwood et al. (1999) also employed SEM to examine the distribution and behavior of diatoms moving through sediments (endopelic). This littleexplored microhabitat still holds many interesting mysteries on algal behavior and the interface with endopelic consumers.

Abiotic resources Fairchild’s (Fairchild & Lowe, 1984) development of a means of manipulating nutrients in-situ using clay flower pots stimulated many students to manipulate abiotic variables while investigating interspecific interactions among benthic algal populations (Carrick, 1985; Luttenton, 1989; Marks, 1990; Pillsbury, 1993; Pan, 1993). Nitrogen or phosphorus were found to be a limiting nutrient for periphyton populations in most of the lotic and lentic habitats investigated (Fairchild et al., 1985; Lowe et al., 1986). Although Carrick & Lowe (1988) found some benthic algal populations to be silicon limited when not allowed contact with quartz sand substrate by supplying nitrogen and phosphorus in a silicon-free medium. Carrick’s major contribution resulted from research he conducted in northern Lake Michigan demonstrating that different benthic algal populations are limited by different resources and that it is incorrect to assume that algae are limited by a single resource as if they were a population rather than an assemblage of many populations (Carrick et al., 1988). DeYoe et al. (1992) demonstrated the sensitivity of some taxa to nitrogen/phosphorus ratios in the environment demonstrating that the numbers of endosymbiotic nitrogen-fixing

4 cyanobacteria within the diatom Rhopalodia is partially a function of the external N/P ratio. Fairchild’s nutrient diffusing substrate technique was also employed to understand how pH differences in aquatic ecosystems can impact nutrient limitation. Keithan et al. (1988) investigated periphyton species response to nitrogen and phosphorus manipulation in a culturally acidified stream while Carrick manipulated nutrients along a natural pH gradient in a northern Michigan lake (Carrick & Lowe, 1989). Pillsbury (1993) investigated light resources examining both quantity and quality of light with the application of tannic acid light filters in four acid lakes in northern Michigan, USA (Pillsbury & Lowe, 1999). He found that light accounted for most of the variation in biomass and community structure with high light environment favoring filamentous green algae (Zygnematales) and low light favoring desmids and diatoms. Marks (1990) manipulated light and nutrients simultaneously in a three-way factorial design in oligotrophic Flathead Lake, Montana, USA. While nitrogen and phosphorus together significantly increased algal biomass, light had little effect in this system (Marks & Lowe, 1993). Disturbance The influence of physical flow-mediated disturbance on benthic algal community structure was recently reviewed by Peterson (1996). Although not addressing this topic directly, a few students at Bowling Green did pursue aspects of this phenomenon. Lamb (1983) employed scanning electron microscopy to investigate the role of current in shaping algal community physiognomy. The publication resulting from his thesis (Lamb & Lowe, 1987) was awarded the best paper of the year in the Ohio Journal of Science. Barnese (1989) examined diel patterns of algal drift in the Maple River, Michigan, USA and found that disturbance caused by benthic insect activity partially explained patterns of algal drift (Barnese & Lowe, 1992). Francoeur (1997) studied mechanisms of periphyton disturbance-resistance in a disturbance-prone river in New Zealand. He found that imbricated microform bed clusters of stones served as refugia for disturbance-vulnerable species (Francoeur et al., 1998). These refugia serve

as epicenters for post-disturbance re-colonization of the stream by both periphyton and invertebrates. It is important to remember that nutrient resources and disturbance do not operate in isolation from each other and there may be interactive effects. Resource stress can alter the impact of hydrological disturbance on periphyton communities (Biggs et al., 1999). Biggs et al. (1998b) developed a habitat matrix model that considered the simultaneous impacts of nutrient abundance and disturbance intensity/ frequency. The model predicted responses of several common stream periphyton species and was later successfully tested on three valley segments of a New Zealand grassland river (Biggs et al., 1998a). Grazers Grazers can strongly influence the structure, density and physiology of periphytic algal communities (Steinman, 1996). Grazer-periphyton interactions have been the focus of several research investigations by students and collaborators at Bowling Green. The impact of grazing snails was a focus in of Lowe and Hunter (1988) who examined the impact of Physa integra on periphyton communities in Spring Lake, Michigan and by Barnese et al. (1990) who examined radular ultrastructure and grazing efficiency of six sympatric snails in Douglas Lake, Michigan. Barnese’s investigation demonstrated the capability of the prostrate green alga Coleochaete orbicularis to resist grazing by snails. The thallus often lost erect colorless setae to snails but the prostrate chlorophyll-bearing cells remained largely intact. Further, benthic diatoms associated with the thallus of Coleochaete also often escaped predation. In further research in Douglas Lake, Marks and Lowe (1989) investigated the independent and interactive effects of snail grazing (Elimia livescens) and nutrient enrichment on structuring periphyton communities. Grazing had a more pronounced effect on algal community composition on the nutrientenriched substrates than on the controls. Grazing caused a decrease in periphyton diversity and an increase in the relative proportion of green algae, especially Stigeoclonium tenue. Gresens and Lowe (1994) also working in Douglas Lake manipulated periphyton patches with nutrient-diffusing substrates to examine patch preference by the grazing larva of the chironomid Paratanytarsus dubius. As with Marks & Lowe (1989), addition of nitrogen and

5 phosphorus resulted in a Stigeoclonium-dominated community, which was negatively correlated with Paratanytarsus grazing. Grazing preference was correlated positively, however, with algal diversity. Two stream studies in which nutrients and grazers were manipulated illustrated the impact that grazers can exert on periphyton communities often greatly dampening the expected biomass increase of periphyton from nutrient stimulation. In the Maple River in northern Michigan Pan & Lowe (1995) found that colonization of benthic substrates by hydropsychid caddisflies can have a stronger impact on periphyton accrual than nutrients. Biggs & Lowe (1994) described the same phenomenon in the Kakanui River in New Zealand when the grazing snail Potamopyrgus antipodarum negated the effects of expected biomass accrual from nutrient addition. Environmental assessment The recent literature is rife with examples of the application of algae as tools for monitoring environmental quality both present (Lowe & Gale, 1980; Lowe, 1981; Shubert, 1984; Stevenson & Lowe, 1986; Whitton et al., 1991; Whitton & Rott, 1995; Lowe & Pan, 1996; Stevenson & Pan, 1999) and past (Battarbee et al., 1999; Bradbury, 1999; Fritz et al., 1999). The value of algae as integrators of fluctuating environmental variables is well established and documented. This application has been a continuing interest to students in the algae laboratory at Bowling Green (McCullough, 1971; Rohr, 1977; Maurice, 1982; Karl, 1983; Blake, 1987; Passy, 1997; Gooden, 2002). Algae are excellent integrators of fluctuating abiotic variables but these variables must be measured at correct temporal and spatial scales. Initial studies at Bowling Green were focused on local systems. For example, impact of treated domestic sewage on the Portage River, Ohio was the first research initiative using algae as environmental indicators at Bowling Green (Lowe & McCullough, 1974). Our research group initially focused on aquatic environments near Bowling Green. Pryfogle and Lowe (1979) authored a methodology manuscript based on some of the initial findings and experiences with periphyton monitoring. In the 1970s, Stevenson (1976), Pryfogle (1976), Kline (1975) and others were focused on research projects in the Sandusky River

watershed (Ohio). This was a particularly fertile time in this research group and led to Stevenson developing some of his early thought on environmental monitoring (Stevenson & Pryfogle, 1976; Stevenson, 1984). Stevenson has now become a leading authority on the application of algae for water quality monitoring. In the 1990s the laboratory established research collaboration with Procter and Gamble Experimental Stream Facility near Cincinnati, Ohio. In this controlled environment many synthetic chemicals and particularly surfactants were tested for their potential impact on benthic algal communities (Belanger et al., 1994). This collaboration provided us with the opportunity not for only bioassay research (Belanger et al., 1996), but also facilitated periphyton research not directly related to the bioassays (Lowe et al. 1996; Greenwood et al., 1999). These integrated research activities continue with a new cadre’ of graduate advisees researching and learning the roles of systematics, ecology and environmental assessment.

Acknowledgements and concluding remarks This manuscript is dedicated to Dr. John C. Kingston my first doctoral advisee who succumbed to a brain tumor in 2004. John will forever stand as a shining example of a productive scientist who maintained balance in his life between science, his family, music and nature. I thank all of my graduate advisees, current and future without whom most of this research would never have happened. I thank especially R. Jan Stevenson and Yangdong Pan for supplying the inertia necessary for this event to happen. I thank my collaborators who have enriched my career. Finally, I thank John Dodd, who served as an important role model for teaching and advising students. Dodd wrote an article on this topic that strongly influenced me, ‘‘Science, a Modern Fountain for Youth’’ (Dodd, 1953). This little two-page paper is of a sort that is often undervalued by readers expecting ‘‘hard science’’ from science writers, but the manuscript outlined Dodd’s approach to teaching and training students. I have tried to incorporate Dodd’s ideas in my teaching

6 approach. Dodd asks the following questions in equating teaching of science as a ‘‘fountain for youth.’’ ‘‘1. Is the water in your fountain always clear and sparkling? Do students become thirsty just watching it? 2. Is it free from any taint of bias, boredom or bunk? 3. Is the pressure adjustable, so students can drink their fill without choking or without becoming disgusted because it flows too slowly? 4. Do you shut the fountain off at four o’clock sharp? 5. Can you accept the fact that most of the water goes down the drain, and console yourself with the thought that the little which is imbibed has important metabolic implications? 6. Do you have enough well-springs of information so the fountain never runs dry? 7. Are you patient with the eager ones who always manage to fall into the fountain? 8. Can you recognize the timid ones who are desperately thirsty but are unable to crowd around a busy fountain? 9. And, finally, do you ever take a little drink yourself, just to see if the stuff is as advertised?’’ Thanks again, John.

References Acker, F., 1977. The phytoplankton of the Maumee River between Grand Rapids, Ohio and Maumee, Ohio. Masters Thesis, Bowling Green State University, Bowling Green, Ohio, USA, 90 pp. Bahls, L. L., 1973. Diatom community response to primary wastewater effluent. Journal of the Water Pollution Control Federation 45: 134–144. Bahls, P. A. & L. L. Bahls, 1974. Trophic response to a hatchery effluent. Proceedings of the Montana Academy of Science 34: 5–11. Barnese, L. E., 1989. A survey and experimental study of algal drift in the Maple River, Pellston, Michigan. Doctoral Dissertation, Bowling Green State University, Bowling Green, Ohio, USA, 124 pp. Barnese, L. E. & R. L. Lowe, 1992. Effects of substrate, light and benthic invertebrates on algal drift in small streams. Journal of the North American Benthological Society 11: 49–59. Barnese, L. E., R. L. Lowe & R. D. Hunter, 1990. Comparative grazing efficiency of six species of sympatric snails in Douglas Lake, Michigan. Journal of the North American Benthological Society 9: 35–44. Battarbee, R. W., D. F. Charles, S. S. Dixit & I. Renberg, 1999. Diatoms as indicators of surface water acidity. In Stoermer, E. F. & J. P. Smol (eds), The Diatoms: applications for the Environmental and Earth Sciences. Cambridge University Press, 85–127. Belanger, S. E., J. B. Barnum, D. M. Woltering, J. W. Bowling, R. M. Ventullo, S. D. Schermerhorn & R. L. Lowe, 1994.

Algal periphyton structure and function in response to consumer chemicals in stream mesocosms. In Graney, R. L., J. H. Kennedy & J. H. Rogers (eds), Aquatic Mesocosm Studies in Ecological Risk Assessment. SETAC Special Publication Series Lewis Publishers, Boca Ratan, FL: 535–567. Belanger, S. E., R. L. Lowe & B. H. Rosen, 1985. Epiphytism of Synedra parasitica on Surirella robusta: observations of populations and associations in a Virginia pond. Transactions of the American Microscopical Society 104: 378–386. Belanger, S. E., K. L. Rupe, R. L. Lowe, D. W. Johnson & Y. Pan, 1996. A flow through laboratory microcosm for assessing effects of surfactants on natural periphyton. Environmental Toxicology and Water Quality 11: 65–76. Biggs, B. J. F., 1989. Biomonitoring of organic pollution using periphyton, South Branch, Canterbury, New Zealand. New Zealand Journal of Marine and Freshwater Research 23: 263–274. Biggs, B. J. F. & R. L. Lowe, 1994. Responses of two trophic levels to patch enrichment along a New Zealand stream continuum. New Zealand. New Zealand Journal of Marine and Freshwater Research 28: 119–134. Biggs, B. J. F., C. Kilroy & R. L. Lowe, 1998a. Periphyton development in three valley segments of a New Zealand grassland river: test of a habitat matrix conceptual model within a catchment. Archive fu¨r Hydrobiologie 143: 147–177. Biggs, B. J. F., R. J. Stevenson & R. L. Lowe, 1998b. A habitat matrix conceptual model for stream periphyton. Archive fu¨r Hydrobiologie 143: 21–56. Biggs, B. J. F., N. Tuchman, R. L. Lowe & R. J. Stevenson, 1999. Resource stress alters hydrological disturbance effects in a stream periphyton community. Oikos 85: 95–108. Blake, G., 1987. The effects of the agricultural herbicide alachlor on total biomass and community structure of algal periphyton in artificial streams. Masters Thesis, Bowling Green State University, Bowling Green, Ohio, USA, 97 pp. Bradbury, J. P., 1999. Continental diatoms as indicators of long-term environmental change. In Stoermer, E. F. & J. P. Smol (eds), The Diatoms: applications for the environmental and Earth Sciences. Cambridge University Press, 169–182. Bruno, M. G., 1978. Distribution and periodicity of desmids and diatoms in a Northwestern Ohio bog lake. Masters Thesis, Bowling Green State University, Bowling Green, Ohio, USA, 64 pp. Bruno, M. G. & R. L. Lowe, 1980. Differences in the distribution of some bog diatoms: a cluster analysis. American Midland Naturalist 104: 70–79. Busch, D. E., 1974. Ultrastructure of the filamentous habit in the diatom Navicula confervacea (Ku¨tz.) Grun. Journal of Phycology 10: 241–243. Camburn, K. E., 1977. The haptobenthic diatom flora of Long Branch Creek, South Carolina. Masters Thesis, Bowling Green State University, Bowling Green, Ohio, USA, 266 pp. Camburn, K. E., R. L. Lowe & D. E. Stoneburner, 1978. The haptobenthic diatom flora of Long Branch, South Carolina. Nova Hedwigia 37: 149–279. Carrick, H. J., 1985. The response of Lake Michigan benthic algae to an in situ nutrient manipulation. Masters Thesis,

7 Bowling Green State University, Bowling Green, Ohio, USA, 86 pp. Carrick, H. J. & R. L. Lowe, 1988. Response of Lake Michigan benthic algae to in situ enrichment with Si, N and P. Canadian Journal of Fisheries and Aquatic Science 45: 271–279. Carrick, H. J., R. L. Lowe & J. T. Rotenberry, 1988. Functional associations of benthic algae along experimentally manipulated nutrient-gradients: Relationships with algal community diversity. Journal of the North American Benthological Society 7: 117–128. Carrick, H. J. & R. L. Lowe, 1989. Benthic algal response to N and P enrichment along a pH gradient. Hydrobiologia 179: 119–127. Descy, J. P., 1979. A new approach to water quality estimation using diatoms. Nova Hedwigia Beihefte 64: 305–323. DeYoe, H. R., R. L. Lowe & J. C. Marks, 1992. The effect of nitrogen and phosphorus on the endosymbiont load of Rhopalodia gibba and Epithemia turgida (Bacillariophyceae). Journal of Phycology 28: 773–777. Dixit, S. S., B. F. Cumming, J. P. Smol & J. C. Kingston, 1992. Monitoring environmental changes in lakes using algal microfossils. In McKenzie, D. H., D. E. Hyatt & V. J. MacDonald (eds), Ecological Indicators. Elsevier Applied Sciences, Amsterdam: 1135–1155. Dodd, J. D., 1953. Science, a modern fountain for youth. American Biology Teacher 15: 1–2. Fairchild, F. W. & R. L. Lowe, 1984. Algal substrates which release nutrients: effects on periphyton and invertebrate succession. Hydrobiologia 114: 29–37. Fairchild, G. W., R. L. Lowe & W. B. Richardson, 1985. Nutrient-diffusing substrates as an in situ bioassay using periphyton: Algal growth responses to combinations of N and P. Ecology 66: 465–472. Fisher, D. Z., 1980. Autumn periphyton and phytoplankton diatom communities in relation to depth and current velocity in the Maumee River, Ohio. Masters Thesis, Bowling Green State University, Bowling Green, Ohio, USA, 86 pp. Francoeur, S. N., 1997. Microform bed clusters as refugia for periphyton in a flood-prone headwater stream. Masters Thesis, Bowling Green State University, Bowling Green, Ohio, USA, 34 pp. Francoeur, S. N., B. J. F. Biggs & R. L. Lowe, 1998. Microform bed clusters as refugia for periphyton in a flood-prone headwater stream. New Zealand Journal of Marine and Freshwater Research 32: 363–374. Fritz, S. C., B. F. Cumming, F. Gassee & K. R. Laird, 1999. Diatoms as indicators of hydrologic and climate change in saline lakes. In Stoermer, E. F. & J. P. Smol (eds), The Diatoms: Applications for the Environmental and Earth Sciences. Cambridge University Press, 41–72. Gomez, S. R., J. R. Johansen & R. L. Lowe, 2003. Epilithic aerial algae of Great Smoky Mountains National Park. Biologia Bratislavia 58: 603–615. Gooden, W., 2002. Periphyton responses to surfactants: Community structure and mat architecture. Doctoral Dissertation, Bowling Green State University, Bowling Green, Ohio, USA, 220 pp. Greenwood, J. L., T. A. Clason, R. L. Lowe & S. E. Belanger, 1999. Examination of endopelic and epilithic algal commu-

nity structure employing scanning electron microscopy. Freshwater Biology 41: 821–828. Gresens, S. E. & R. L. Lowe, 1994. Periphyton patch preference in grazing chironomid larvae. Journal of the North American Benthological Society 13: 89–99. Jackson, D. C., 1975. Distribution and morphology of members of the diatom genera Gyrosigma Hassal and Pleurosigma W. Smith in the Portage River Drainage System. Masters Thesis, Bowling Green State University, Bowling Green, Ohio, USA, 75 pp. Jackson, D. C. & R. L. Lowe, 1978. Valve ultrastructure of species of the diatom genera Gyrosigma Hassal and Pleurosigma W. Sm. from the Portage River Drainage system, Ohio. Transactions of the American Microscopical Society 97: 569–581. Johansen, J. R., R. L. Lowe, S. R. Gomez, J. P. Kociolek & S. A. Makosky, 2004. New algal species records for the Great Smoky Mountains National Park, U.S.A., with an annotated checklist of all reported algal species for the park. Algological Studies. Kaatz, M. R., 1955. The black swamp: a study in historical geography. Annals of the Association of American Geographers 45: 1–35. Karl, K. A., 1983. The effects of fly ash extract on periphyton community structure in field enclosures. Masters Thesis, Bowling Green State University, Bowling Green, Ohio, USA, 97 pp. Keithan, E. D., 1983. Primary productivity and structure of phytolithic communities in streams in the Great Smoky Mountains National Park. Doctoral Dissertation, Bowling Green State University, Bowling Green, Ohio, USA, 83 pp. Keithan, E. D. & R. L. Lowe, 1985. Primary productivity and structure of phytolithic communities in streams in the Great Smoky Mountains National Park. Hydrobiologia 123: 59–67. Keithan, E. D., R. L. Lowe & H. DeYoe, 1988. Benthic diatom distribution in a Pennsylvania stream: the role of pH and nutrients. Journal of Phycology 24: 581–585. Kilroy, C., K. Sabbe, E. Bergy, W. Vyverman & R. Lowe, 2003. New species of Fragilariforma (Bacillariophyceae) from New Zealand and Austrailia. New Zealand Journal of Botany 41: 535–554. Kingston, J. C., 1980. Characterization of benthic diatom communities in Grand Traverse Bay, Lake Michigan. Doctoral Dissertation, Bowling Green State University, Bowling Green, Ohio, USA. Kingston, J. C., R. L. Lowe, E. F. Stoermer & T. Ladewski, 1983. Spatial and temporal distribution of benthic diatoms in northern Lake Michigan. Ecology 64: 1566–1580. Kline, P. A., 1975. Survey of the phytoplankton of the Sandusky River at Fremont, Sandusky Co., Ohio. Masters Thesis, Bowling Green State University, Bowling Green, Ohio, USA. Kline, P. E. & R. L. Lowe, 1976. Phytoplankton of the Sandusky River near Fremont, Ohio. In Baker, D. & B. Prater (eds), Proceedings of the Sandusky River Basin Symposium . United States Environmental Protection Agency, 175–208. Kociolek, J. P., 1982. Diatoms from two streams in Great Smoky Mountains National Park. Masters Thesis, Bowling Green State University, Bowling Green, Ohio, USA, 175 pp.

8 Kociolek, J. P., M. A. Lamb & R. L. Lowe, 1983. Notes on the growth and ultrastructure of Biddulphia laevis Ehr. (Bacillariophyceae) in the Maumee River, Ohio. Ohio Journal of Science 8: 125–130. Kociolek, J. P. & R. L. Lowe, 1983. Scanning electron microscopic observations on the frustular morphology and filamentous growth habit of Diatoma hiemale v. mesodon. Transactions of the American Microscopical Society 102: 281–287. Kociolek, J. P. & E. F. Stoermer, 2001. Taxonomy and ecology: a marriage of necessity. Diatom Research 16: 433–442. Krejci, M., 1985. Spatial patterns of epipsammic diatoms in a spring-fed brook with emphasis on the effect of sand grain mineralogy on diatom occurrence. Doctoral Dissertation, Bowling Green State University, Bowling Green, Ohio, USA. Krejci, M. E. & R. L. Lowe, 1986. The importance of sand grain mineralogy and topography in determining microspatial distribution of epipsammic diatoms. Journal of the North American Benthological Society 5: 221–229. Krejci, M. E. & R. L. Lowe, 1987a. Spatial and temporal variation of epipsammic diatoms in a spring-fed brook. Journal of Phycology 23: 585–590. Krejci, M. E. & R. L. Lowe, 1987b. The seasonal occurrence of macroscopic colonies of Meridion circulare (Bacillariophyceae) in a spring-fed brook. Transactions of the American Microscopical Society 106: 173–178. Lamb, M. E., 1983. The effects of current velocity on the structuring of diatom communities. Masters Thesis, Bowling Green State University, Bowling Green, Ohio, USA, 94 pp. Lamb, M. A. & R. L. Lowe, 1987. Effects of current velocity on the physical structuring of diatom (Bacillariophyceae) communities. Ohio Journal of Science 87: 72–78. Lamb, M. A. & R. L. Lowe, 1981. A preliminary investigation of the effect of current speed on periphyton community structure. Micron 12: 211–212. Lowe, R. L., 1975. Comparative ultrastructure of the valves of some species of Cyclotella (Bacillariophyceae). Journal of Phycology 11: 415–424. Lowe, R. L., 1981. The utility of diatoms for hazard assessment of chemicals in ecosystems . Special Publication of the National Academy of Science, Washington, DC, pp 133–136. Lowe, R. L. & D. E. Busch, 1975. Morphological observations on two species of the diatom genus Thalassiosira from freshwater habitats in Ohio. Transactions of the American Microscopical Society 94: 118–123. Lowe, R. L. & R. E. Crang, 1972. The ultrastructure and morphological variability of the frustule of Stephanodiscus invisitatus Hohn and Hellerman. Journal of Phycology 8: 256–259. Lowe, R. L. & W. F. Gale, 1980. Monitoring periphyton with artificial benthic substrates. Hydrobiologia 69: 235–244. Lowe, R. L. & R. D. Hunter, 1988. The effect of grazing by Physa integra on periphyton community structure. Journal of the North American Benthological Society 7: 29–36. Lowe, R. L. & P. E. Kline, 1976. Planktonic centric diatoms of the Sandusky River near Fremont, Ohio. In Baker, D. B. Prater (eds), Proceedings of the Sandusky River Basin Symposium. United States Environmental Protection Agency : 143–152.

Lowe, R. L. & J. P. Kociolek, 1984. New and rare diatoms from Great Smoky Mountains National Park. Nova Hedwigia 39: 465–276. Lowe, R. L., S. Golladay & J. Webster, 1986. Periphyton response to nutrient manipulation in a clear-cut and forested watershed. Journal of the North American Benthological Society 5: 211–220. Lowe, R. L., J. B. Guckert, S. E. Belanger, D. H. Davidson & D. W. Johnson, 1996. An Evaluation of Periphyton Community Structure and Function on Tile and Cobble Substrata in Experimental Stream Mesocosms. Hydrobiologia 328: 135–146. Lowe, R. L. & J. M. McCullough, 1974. The effect of sewage treatment plant effluent on diatom communities in the Portage River, Wood Co., Ohio. Ohio Journal of Science 74: 154–161. Lowe, R. L. & Y. Pan, 1996. Use of Benthic Algae in Water Quality Monitoring. In Stevenson, R. J., M. L. Bothwell & R. L. Lowe (eds), Benthic Algal Ecology in Freshwater Ecosystems. Academic Press, San Diego, CA, USA: 705–739. Lowe, R. L., B. H. Rosen & J. C. Kingston, 1982. A comparison of epiphytes on Bangia atropurpurea (Rhodophyta) and Cladophora glomerata (Chlorophyta) from Northern Michigan. Journal of Great Lakes Research 8: 164–168. Luttenton, M. R., 1989. In situ manipulation of factors affecting periphyton community structure. Doctoral Dissertation, Bowling Green State University, Bowling Green, Ohio, USA. Marks, J. C., 1990. The independent and interactive effects of nitrogen, phosphorus and light on structuring periphyton in Flathead Lake, Montana. Masters Thesis, Bowling Green State University, Bowling Green, Ohio, USA, 92 pp. Marks, J. C. & R. L. Lowe, 1989. The independent and interactive effects of snail grazing and nutrient enrichment on structuring periphyton communities. Hydrobiologia 185: 9–17. Marks, J. C. & R. L. Lowe, 1993. Interactive effects of nutrient availability and light levels on the periphyton composition of a large oligotrophic lake. Canadian Journal of Fisheries and Aquatic Science 50: 1270–1278. Maurice, C. G., 1982. Effects of acidification on the periphyton of an artificial stream. Masters Thesis, Bowling Green State University, Bowling Green, Ohio, USA, 92 pp. McCullough, J. M., 1971. The effect of sewage-treatment-plant effluent on diatom communities in the North Branch of the Portage River, Wood County, Ohio. Masters Thesis, Bowling Green State University, Bowling Green, Ohio, USA, 75 pp. Miller, A. R., 1983. Temporal and spatial relationships in the epipsammic diatom community. Masters Thesis, Bowling Green State University, Bowling Green, Ohio, USA, 87 pp. Miller, A. R., R. L. Lowe & J. T. Rotenberry, 1987. Microsuccession of diatoms on sand grains. Journal of Ecology 75: 693–709. Millie, D. F. & R. L. Lowe, 1983. Studies on Lake Erie’s littoral algae; host specificity and temporal periodicity of epiphytic diatoms. Hydrobiologia 99: 7–18. Morgan, M. D., 1987. Impact of nutrient enrichment and alkalinization on periphyton communities in the New Jersey Pine Barrens. Hydrobiologia 144: 233–241.

9 Pan Y., 1993. The effects of nutrients on periphyton. Doctoral Dissertation, Bowling Green State University, Bowling Green, Ohio, USA, 98 pp. Pan, Y. & R. L. Lowe, 1995. The effects of hydropsychid colonization on algal response to nutrient enrichment in a small Michigan stream, U.S.A. Freshwater Biology 33: 393–400. Passy, S. I., 1997. Ecology and systematics of the periphytic diatoms from the Mesta River system, Bulgaria. Doctoral Dissertation, Bowling Green State University, Bowling Green, Ohio, USA, 241 pp. Passy, S. I. & R. L. Lowe, 1994. Taxonomy and ultrastructure of Gomphoneis mesta sp. nov. (Bacillariophyta), a new epilithic diatom from the Mesta River, Bulgaria. Journal of Phycology 30: 885–891. Passy-Tolar, S. I., J. P. Kociolek & R. L. Lowe, 1997. New Gomphonema species (Bacillariophyta) from South African rivers. Journal of Phycology 33: 455–474. Passy-Tolar, S. I., Y. Pan & R. L. Lowe, 1999. Ecology of the major periphytic diatom communities from the Mesta River, Bulgaria. International Review der Gesempten Hydrobiologie 84: 129–174. Peterson, C. G., 1996. Response of benthic algal communities to natural physical disturbance. In Stevenson, R. J., M. L. Bothwell & R. L. Lowe (eds), Benthic Algal Ecology in Freshwater Ecosystems. Academic Press, San Diego, CA, USA: 375–402. Pillsbury, R. W., 1993. Factors influencing the structure of benthic algal communities in acid lakes. Doctoral Dissertation, Bowling Green State University, Bowling Green, Ohio, USA, 142 pp. Pillsbury, R. W. & R. L. Lowe, 1999. The response of benthic algae to manipulations of light resources in four acidic lakes in northern Michigan. Hydrobiologia 394: 69–81. Potapova, M. G., K. C. Ponader, R. L. Lowe, T. A. Clason & L. A. Bahls, 2003. Small-celled Nupela species from U.S.A. rivers. Diatom Research. 18: 293–306. Pryfogle, P. A., 1976. Seasonal distribution of periphytic diatoms on natural substrates in Tymochtee Creek. Masters Thesis, Bowling Green State University, Bowling Green, Ohio, USA, 79 pp. Pryfogle, P. A. & R. L. Lowe, 1979. Sampling and interpretation of epilithic lotic diatom communities. In Weitzel, R. (ed.), Methods and Measurements of Attached Microcommunities: A Review. American Society for Testing and Materials, Philadelphia, PA, USA: 77–89. Rohr, J. L., 1977. Changes in diatom community structure due to environmental stress. Masters Thesis, Bowling Green State University, Bowling Green, Ohio, USA, 157 pp. Sabbe, K., K. Vanhoutte, R. L. Lowe, E. A. Bergey B. J. F. Biggs, S. Francoeur, D. Hodgson & W. Vyverman, 2001. Six new Actinella (Bacillariophyta) species from Papua New Guinea, Australia and New Zealand: further evidence for widespread diatom endemism in the Australasian region. European Journal of Phycology 36: 321–340. Sharkey, M. J., 2001. The all taxa biological inventory of the Great Smoky Mountains National Park. Florida Entomologist 84: 556–564. Shubert, L. E., 1984. Algae as Ecological Indicators . Academic Press, London, England 434.

Steinman, A. D., 1996. Effects of grazers on freshwater benthic algae. In Stevenson, R. J., M. L. Bothwell & R. L. Lowe (eds), Benthic Algal Ecology in Freshwater Ecosystems. Academic Press, San Diego, CA, USA: 341–373. Stevenson, R. J., 1976. The periphytic diatoms of the Sandusky River. Masters Thesis, Bowling Green State University, Bowling Green, Ohio, USA, 114 pp. Stevenson, R. J., 1984. Epilithic and epipelic diatoms in the Sandusky River, with emphasis on species diversity and water quality. Hydrobiologia 114: 161–175. Stevenson, R. J. & R. L. Lowe, 1986. Sampling and interpretation of algal patterns for water quality assessments. In Isom, B. G. (ed.), Rationale for Sampling and Interpretation of Ecological Data in the Assessment of Freshwater Ecosystems American Society for Testing and Materials. Philadelphia, PA, USA: 118–149. Stevenson, R. J. & Y. Pan, 1999. Assesing environmental conditions in rivers and streams with diatoms. In Stoermer, E. F. & J. P. Smol (eds), The Diatoms: Applications for the Environmental and Earth Sciences. Cambridge University Press, 11–40. Stevenson, R. J. & P. A. Pryfogle, 1976. A comparison of the winter diatom flora of the Sandusky River and Tymochtee Creek. In Baker, D. & B. Prater (eds), Proceedings of the Sandusky River Basin Symposium. United States Environmental Protection Agency, 209–231. Whitton, B. A., E. Rott & G. Friedrich, 1991. Use of Algae for Monitoring Rivers. E. Rott, Publisher, Institut fu¨r Botanik, AG Hydrobotanik, Universita¨t Innsbruck, A-6020 Innsbruck, Austria, 193 pp. Whitton, B. A. & E. Rott, 1995. Use of algae for monitoring rivers II. E. Rott, Publisher, Institut fu¨r Botanik, AG Hydrobotanik, Universita¨t Innsbruck, A-6020 Innsbruck, Austria. 196 pp.

Appendix 1 Graduate theses and dissertations from the Bowling Green State University Algae Laboratory, 1971–2003. 1. J. Michael McCullough, M. A., 1971. The effect of sewage-treatment-plant effluent on diatom communities in the North Branch of the Portage River, Wood County, Ohio. 2. Robert Reitz, M. S., 1973. Phytoplankton periodicity in two Northwestern Ohio ponds. 3. Bill Brower, M. S., 1973. Phytoplankton and periphyton diatom relationships in two highly eutrophic lakes. 4. David E. Busch, M. S., 1974. Vertical and seasonal distribution of the Bacillariophyta in the Miller Blue Hole, Sandusky Co., Ohio.

10 5. Terrance L. Breyman, M. S., 1974. Bangia in Western Lake Erie. 6. David C. Jackson, M. S., 1975. Distribution and morphology of members of the diatom genera Gyrosigma Hassal and Pleurosigma W. Smith in the Portage River Drainage System. 7. Phillip A. Kline, M. S., 1975. Survey of the phytoplankton of the Sandusky River at Fremont, Sandusky Co., Ohio. BGSU. 8. Ronald J. Bockelman, M. S., 1975. The seasonal productivity of zooplankton and benthic macroinvertebrate populations in six northwest Ohio ponds. 9. R. Jan Stevenson, M. S., 1976. The periphytic diatoms of the Sandusky River. 10. P.A. Pryfogle, M. S., 1976. Seasonal distribution of periphytic diatoms on natural substrates in Tymochtee Creek. 11. Keith Camburn, M. S., 1977. The haptobenthic diatom flora of Long Branch Creek, South Carolina. 12. Frank Acker, M. S., 1977. The phytoplankton of the Maumee River between Grand Rapids, Ohio and Maumee, Ohio. 13. J. L. Rohr, M. S., 1977. Changes in diatom community structure due to environmental stress. 14. R. F. Andritsch, M. S., 1977. Seasonal photosynthetic rates of Chara globularis in Steidtmann Pond. 15. Mary Bruno, M. S., 1978. Distribution and periodicity of desmids and diatoms in a Northwestern Ohio bog lake. 16. David F. Millie, M. S., 1979. An analysis of epiphytic diatom assemblages of three species of aquatic vascular plants in three Lake Erie marshes. 17. Robert Foster, M. S., 1980. Selected toxic metal concentrations in several species of western Lake Erie fish with respect to age. 18. Daniel Z. Fisher, M. S., 1980. Autumn periphyton and phytoplankton diatom communities in relation to depth and current velocity in the Maumee River, Ohio. 19. John C. Kingston, Ph.D., 1980. Characterization of benthic diatom communities in Grand Traverse Bay, Lake Michigan. 20. Earl Chilton, M. S., 1982. A comparison of macroscopic invertebrates living in Bangia

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

atropurpurea and Cladophora glomerata beds in Lake Erie. John P. Kociolek, M. S., 1982. Diatoms from two streams in Great Smoky Mountains National Park. Charles G. Maurice, M. S., 1982. Effects of acidification on the periphyton of an artificial stream. David R. Beeson, M. S., 1982. Epiphytic diatom (Bacillariophyceae) community structure in a wetland continuum, Sugar Island, Michigan. Barry H. Rosen, Ph.D., 1982. Physiological and ultrastructural responses to light intensity and nutrient limitation in the planktonic diatom Cyclotella meneghiniana. Mark E. Lamb, M. S., 1983. The effects of current velocity on the structuring of diatom communities. Ann R. Miller, M. S., 1983. Temporal and spatial relationships in the epipsammic diatom community. Robert Genter, M. S., 1983. The effects of different initial colonists on the outcome of periphyton succession in a small stream. Kevin A. Karl, M. S., 1983. The effects of fly ash extract on periphyton community structure in field enclosures. Elaine D. Keithan, Ph.D., 1983. Primary productivity and structure of phytolithic communities in streams in the Great Smoky Mountains National Park. Norlida Anis, M. S., 1985. Effects of water chemistry on the distribution of diatom communities. Mark Krejci, Ph.D., 1985. Spatial patterns of epipsammic diatoms in a spring-fed brook with emphasis on the effect of sand grain mineralogy on diatom occurrence. Hunter J. Carrick, M. S., 1985. The response of Lake Michigan benthic algae to an in situ nutrient manipulation. Gail Blake, M. S., 1987. The effects of the agricultural herbicide alachlor on total biomass and community structure of algal periphyton in artificial streams. Mark R. Luttenton, Ph.D., 1989. In situ manipulation of factors affecting periphyton community structure.

11 35. Lisa E. Barnese, Ph.D., 1989. A survey and experimental study of algal drift in the Maple River, Pellston, Michigan. 36. Douglas Deutschman, M. S., 1990. Response of an algal community to temporal variability of resources. 37. Jane C. Marks, M. S., 1990. The independent and interactive effects of nitrogen, phosphorus and light on structuring periphyton in Flathead Lake, Montana. 38. Craig D. Layne, M. S., 1990. The algal mat of Douglas Lake, Michigan: Its composition, role in lake ecology, and response to chemical perturbations. 39. Hudson DeYoe, Ph.D., 1991. Preliminary characterization of the relationship between Rhopalodia gibba (Bacillariophyceae) and its cyanobacterial endosymbiont. 40. Susan Hardman, Ph.D., 1992. Environmental components influential in epipelic algal community structure. 41. James C. Sferra, M. S., 1992 Potential effects of the zebra mussel, Dreissena polymorpha (Pallas) on the Western Basin of Lake Erie. 42. Carmen Pedraza-Silva, M. S., 1992. A description of the algal floras of Guzmania berteroniana and Vriesia sintenisii (Bromeliaceae) and preliminary investigation of bromeliad-algal interactions. 43. Robert W. Pillsbury, Ph.D., 1993. Factors influencing the structure of benthic algal communities in acid lakes. 44. Diane Longanbach M. A. T., 1993. Survey of aquaria educational curricula across the United States. 45. Yangdong Pan, Ph.D., 1993. The effects of nutrients on periphyton. 46. LouAnne Reich, M. S., 1994. An examination of Douglas Lake, Cheboygan County, Michigan as suitable habitat for the zebra mussel (Dreissena polymorpha): food quality and attachment site preference. 47. Bret Gargasz, M. S., 1994. Non-thesis, plan II. 48. Steven Francoeur, M. S., 1997. The effect of in-stream flow refugia on the recovery of

stream periphyton communities following flooding disturbance. 49. Sophia Passy, Ph.D., 1997. Ecology and systematics of the periphytic diatoms from the Mesta River system, Bulgaria. 50. Joanne Rhoers, M. S., 1997. The Impact of the Crayfish Orconectes propinquis on Benthic Algae and Zebra Mussels. 51. Rebecca Visnyai, M. S., 1997. Wetland restoration: the need to base restoration on function and landscape-level processes. 52. Randy Litteral, 1998. Benthic algal community structure and the compensation point. 53. Jennifer L. Greenwood, M. S., 1998. The effects of pH light on periphyton communities in a Michigan Wetland. 54. Todd A. Clason, M. S., 1999. Diurnal migration and community ultrastructure of benthic algae in Douglas Lake. 55. Timothy Stewart, Ph.D., 1999. Evidence and mechanisms for Dreissena effects on other benthic macroinvertebrates in western Lake Erie. 56. Julianne Heinlein, M. S., 2000. Flood disturbance mechanisms in stream periphyton: individual and interactive effects of shear stress perterbations and suspended sediment concentration. 57. Agnieszka Pinowska, Ph.D., 2001. Indirect effect of sediment nutrient enrichment on epiphytic algal communities. 58. Amy Kireta, M. S., 2001. Benthic algal shifts in response to the round goby. 59. Wanda Gooden, Ph.D., 2002. Periphyton responses to surfactants: Community structure and mat architecture. 60. Jennifer Ress, M. S., 2003. Comparative grazing efficiencies of three aquatic grazers and their impact on periphyton recovery. 61. Jennifer Wearly, M. S., 2004. Changes in algal communities due to zebra mussel invasion of an oligotrophic inland lake. 62. Sarah Zeiler, M. S., 2004. The ratio of periphyton to plankton under variable nutrient regimens in a fen peatland.

Hydrobiologia (2006) 561:13–26  Springer 2006 R.J. Stevenson, Y. Pan, J.P. Kociolek & J.C. Kingston (eds), Advances in Algal Biology: A Commemoration of the Work of Rex Lowe DOI 10.1007/s10750-005-1602-6

Rexia erecta gen. et sp. nov. and Capsosira lowei sp. nov., two newly described cyanobacterial taxa from the Great Smoky Mountains National Park (USA) Dale A. Casamatta, Shannon R. Gomez & Jeffrey R. Johansen* Department of Biology, John Carroll University, University Heights, Ohio, 44118, USA (*Author for correspondence: Tel.: +1-216-397-4487; Fax: +1-216-397-4482; E-mail: [email protected])

Key words: ATBI, Capsosira, Cyanobacteria, endophytic, epilithic, Nostocales, Rexia, Stigonematales

Abstract Two newly discovered taxa of Cyanobacteria from the Great Smoky Mountain National Park (USA) are presented. The first is the newly described species Capsosira lowei (Capsosiraceae), differing from the only other previously described species C. brebissonii Ku¨tz. ex Born. et Flah. in regard to cell size and filament morphology. In addition, C. brebissonii is described as an aquatic or subaerophytic taxon, while our isolate was obtained as a phycobiont from the lichen Hydrothyria venosa J. L. Russell. Capsosira is currently placed in the Capsosiraceae of the Stigonematales due to its ability to have division in two planes. However, molecular evidence gathered in this study indicates closest affinity with Aulosira and Nostoc commune Vaucher, both in the Nostocaceae, Nostocales. Rexia erecta was isolated from concurrently collected aerophytic, epilithic sites. The hormogonia production, near absence of heterocysts and division in two planes are all typical of the Stigonematales, but it fits none of the currently circumscribed families in that order. This genus in other ways appears morphologically similar to members of the Scytonemataceae and Microchaetaceae. Molecular evidence (nearly complete 16S rRNA sequence data and 16S–23S internal transcribed spacer ITS region) places Rexia in the Microchaetaceae. These taxa are both problematic as they indicate that cell division in two planes has likely arisen more than once in the Nostocales, and thus the Stigonematales as currently circumscribed is not a monophyletic group. The Nostocales and Stigonematales are, in our opinion, in need of revision at the family and order level of classification.

Introduction The Great Smoky Mountain National Park (GSMNP) serves as a refuge for one of the largest, richest collections of plants, animals and crytogamic taxa in the world. Comprised of over one million acres, it is the largest contiguous preserve east of the Rocky Mountains (USA). Due to the heterogeneity of habitats afforded by the geological complexity, elevational variety, north-south park orientation and moist climate, this park has been designated an International Biosphere Preserve.

Beginning in 1997, an exhaustive effort was begun to identify every species in every phyla present in the park (Sharkey, 2001). Dubbed the All Taxa Biodiversity Inventory (ATBI), this project includes researchers from myriad academic and governmental positions throughout the Unites States. One of the most poorly characterized groups in the park includes the cryptogamic flora. Given the propensity of aquatic habitats including streams, ponds, perched bogs, drippy cliffs, swamps, and a host of other permanent and ephemeral habitats, the algal community may

14 prove to be quite vast. Park inventories of both diatoms and other algae were undertaken in the 1940s and 1980s, with a complete annotated list and review presented in Johansen et al. (2004). Recently, Gomez et al. (2003) reported on the epilithic, aerial algae from the GSMNP. In their study they noted the potential for 46 potentially new species among all algal groups. Here, we describe two new taxa from drippy walls within the GSMNP system. Both belong to the clade of heterocyte-bearing cyanobacteria, but their higher-level taxonomic status is uncertain. Capsosira lowei clearly fits the genus circumscription. However, its size, filament morphology, and habitat differ markedly from C. brebissonii Ku¨tz. ex Born. et Flah., the only other species in this genus. The newly described Rexia erecta gen. nov. et sp. nov. is problematic in that while undoubtedly a new genus, it does not clearly fit into any currently circumscribed family of cyanobacteria. Both of these taxa will be described and their phylogeny based on molecular data will be presented and discussed.

Materials and methods Sample collection and culturing Aerial algae were collected from GSMNP bedrock formations 19–24 May 2001. Organic matter was scrapped from rocks wetted by springs and splash from waterfalls. Scrapings were placed directly into whirl-paks and kept under refrigeration until culture and preservation. The pH at each site was measured using Whatman pH papers (range 4.0–7.0) while latitude, longitude and altitude were recorded using a Garmin GPS unit. Upon return to John Carroll University, samples were divided into roughly equal parts for preservation in 2% formaldehyde or further culturing. Three culture dilutions were made using Bold’s Basal Medium (BBM) agar (Bold & Wynne, 1978). Cultures were maintained in a growth chamber under 200 lE cm)1 irradiance, on a 16:8 h light:dark cycle, with temperature controlled at 18 C during the light cycle and 10 C during the dark. Cultures were grown for 4 weeks before subsequent analyses. Identifications were made employing fresh, preserved and cultured materials. Both taxa were

examined and photographed using an Olympus BH2 photomicroscope with high resolution Nomarski DIC optics. Line drawings were made from fresh material and represent meticulously measured real specimens. They were chosen in preference to photographs as the plane of focus was consistently flattened in the drawings. Taxonomic references followed Geitler (1932) and Anagnostidis and Koma´rek (1990). Molecular analyses Total genomic DNA was extracted from cultures using the CTAB method as modified by Cullings (1992) for the isolation and purification of DNA from mucilaginous organisms (Doyle & Doyle, 1987). DNA pellets were re-suspended in 50 ll of TE buffer and the resulting genomic DNA was checked using 1% agarose/ethidium bromide gels. Extracted DNA samples were stored at )20 C. PCR primers were modified from Wilmotte et al. (1993) and Nu¨bel et al. (1997) and designated as follows: Primer 1: 5¢ CTC TGT GTG CCT AGG TAT CC 3¢ (after Wilmotte et al., 1993) Primer 2: 5¢ GGG GAA TTT TCC GCA ATG GG 3¢ (after Nu¨bel et al., 1997) Primer 5: 5¢ TGT ACA CAC CGG CCC GTC 3¢ (after Wilmotte et al., 1993) Primer 6: 5¢ GAC GGG CCG GTG TGT ACA 3¢ (after Wilmotte et al., 1993) Initially, DNA samples were amplified using primers 1 and 2, which are cyanobacterial-specific primers. The resulting amplified products were analyzed on 1% agarose/ethidium bromide gels and were determined to be approximately 1600 base pairs in length (‘long PCR’). All PCR reactions were performed in a total volume of 100 ll containing 10.0 ll of 10 Taq polymerase buffer (Promega Corp., Madison, WI); 0.5 ll primer mixture (1.2 ll primer 1 or 6, 1.2 ll primer 2, 7.6 ll dH2O); 0.5 ll of a stock solution of dNTPs [(10 mM in each dNTP); dATP, dCTP, dGTP, and dTTP]; 0.5 ll (Promega) Taq polymerase; 1.0 ll of extracted genomic DNA (50 ng), and the appropriate amount of dH2O to bring the volume to 100 ll. The reactions were overlaid with mineral oil, and thermal cycling was conducted using an Thermolyne’s Amplitron and Temptronic

15 thermalcyclers (Barnstead International, Dubuque, IA) using the following parameters: 94 C for 60 s, 55 C for 45 s, and 72 C for 4 min repeated for 35 cycles (primer pair 1 and 2), and 94 C for 60 s, 55 C for 45 s, and 72 C for 2 min repeated for 20 cycles (primer pair 2 and 6). After amplification, a 7-min/72 C extension step was included for primer pair 1 and 2, whereas primer pair 2 and 6 received no such extension. PCR products were analyzed on 1% agarose/ethidium bromide gels in 1 TBE buffer. Amplified PCR products were cloned into pCR 4-TOPO plasmids containing sites for universal primers M13 forward and reverse using the TOPOTM TA cloning kit (Invitrogen Corp., Carlsbad, CA). After transformation, E. coli cells (Invitrogen) were plated onto Luria Broth plates containing 100 mg l)1 of ampicillin. Plasmids were isolated according to the instructions provided in the QIAprep Mini-prep kit (Quiagen Inc., Valencia, CA). Digests were resolved on 1% agarose/ ethidium bromide gels to detect plasmid inserts. Two replicate plasmid samples were isolated from each cloning plate and sequenced by Cleveland Genomics (Cleveland, OH). Automated sequencing was performed using universal infrared (IR) primers M13IR forward and reverse. Data analysis Forward and reverse sequences were aligned using the CLUSTAL W Multiple Sequence Alignment Program (Thompson et al., 1994). The resulting sequence alignments were checked by eye for ambiguities and PCR errors by the examination of chromatograms, with corrections made where appropriate. Ingroup and outgroup taxa were obtained from Genbank (www.ncbi.nlm.nih.gov) and other sequenced taxa. Parsimony trees were generated using a heuristic search constrained by random sequence addition (1000), steepest descent, and tree-bisection branch swapping using PAUP v.4.02b (Swofford, 1998). Bootstrap values were obtained from 1000 replicates with one random sequence addition to jumble the data using PAUP software. A neighbor joining tree was constructed employing the General Time Reversible model with corrected invariable sites (I) and Gamma distribution shape parameters (G) obtained using Modeltest v3.06 (Posada & Crandall, 1998) and

bootstrap resampled (1000) using PAUP. Maximum likelihood analysis using the HKY85 distance method and assuming a ti:tv ratio of 2 was also performed. Secondary structure of the 16S–23S ITS was determined using Mfold version 3.1 (http:// www.bioinfo.rpi.edu/applications/mfold/old/rna, Zuker, 2003). Structures were determined by folding and identifying each conserved helix separately first, and then constraining the sequence to produce the entire structure. Default conditions were in all cases used.

Results and discussion We observed the phycobiont in Hydrothyria venosa J. R. Russell through the use of epifluorescence microscopy. The specimens resembled Nostoc (kinked, uniserate trichomes with diffluent sheaths), which is reportedly the phycobiont from this unusual and rare aquatic lichen (Fink, 1935). However, it appeared that the cyanobacterium had true branching (Fig. 1f), a feature that would exclude it from Nostoc. We isolated the phycobiont by fragmenting the lichen thallus and culturing on media prepared from nutrient-enriched site water from Hen Wallow Falls, where the lichen was growing. The phycobiont exhibited a very different morphology when grown on agar plates free of H. venosa, an observation also noted by other researchers studying lichen symbioses (e.g. Bubrick, 1988; Ahmadjian, 1989; Davis & Rands, 1993). Examination of the isolated strain revealed all of the morphological features of Capsosira, including a thallus consisting of rigid, vertical, parallel, knobby trichomes with clustered branches and trichomes formed of a series of single-celled vesicles (sensu Ku¨tzing, 1849) (Fig. 1a–e). Geitler (1932) noted intercalary or terminal heterocytes and hormogonia formed at the ends of upright filaments. In particular, the radiating pattern of thallus development due to pseudodichotomous branching (Fig. 1c, d) is very characteristic (Geitler 1932; Anagnostidis & Koma´rek, 1990). This new taxon differs from the only other species in the genus, C. brebissonii, in several key features, most notably vegetative cell dimensions, variability of the filaments and the phycobiont nature (Table 1). While Capsosira has been previously reported

16

Figure 1. Capsosira lowei. (a) Hormogonia; (b) Kinked uniseriate trichome with diffluent sheath; (c–d) Mature filaments radiating in parallel series, which dichotomously branch, and which can become multiseriate when mature, with cell division in two planes evident. Note that this life cycle stage lacked colonial mucilage; (e) Mucilaginous microcolonies; (f) Trichomes as they appear in the intercellular spaces within the lichen thallus. Note the ambiguity in discerning cell division in two planes, and the heterocytes, which were common in the lichen but rare in culture. Scale bars=5 lm.

from New England states by Tilden (1910) and a single stream from North Carolina by Whitford & Shumacher (1984), none of these habitats or drawings corresponded to our taxon.

Our newly described Rexia erecta gen. nov. et sp. nov. is unique in several key features. First, it is isopolar in trichome development (as in the family Scytonemataceae) but filaments taper apically

17 Table 1. Comparison of morphological characters from Capsosira species Taxa

Habitat

C. brebissonii1 Epiphytic on wood

Sheath

Close, thick,

Vegetative cell Vegetative width (lm)

cell length (lm)

4–5

4–6

Cell contents

Heterocytes

Nongranular

Intercalary

or aquatic plants colorless (occasionally in dead cells), or yellow, C. lowei

1

on moistened rock

unlamellated

Endosymbiont in

Facultative,

Hydrotheria venosa

clear, diffluent

or lateral

4.0–6.0 (8.0)

(2.4) 3.0–6.0 (7.2) Nongranular in Facultative, lichen thallus,

commonly

distinctly

intercalary,

granular

can be

in culture

terminal

Description sensu Geitler (1932).

(Fig. 2d), a characteristic of the Microchaetaceae (sensu Anagnostidis and Koma´rek 1990). The evident false branches (both double and single) arise without presence of heterocytes and typically disintegrate into hormogonia (Fig. 2a, e) or at times hormocytes (one celled hormogonia, Fig 2b, c, e). This is found in some families of the Stigonematales, specifically the Loriellaceae and Mastigocladaceae (Angnostidis and Koma´rek 1990). Heterocytes are extremely rare. The cell division in two planes is most similar to pseudodichotomous branching (cf. Koma´rek et al., 2003) (Fig. 2d), but is very rare, occurs only in short hormogonia, and does not precisely fit the description of this type of branching as reported in Angnostidis & Koma´rek (1990). The division in two planes, however, is unequivocal. The branches of the trichomes typically grow erect from the agar. Molecular results Near complete sequence data for a ca. 1150 (90%) base pair (bp) region of the 16S rRNA gene were obtained for the two strains, and deposited in Genbank (assession numbers AY452533 and AY452534 for Rexia erecta and Capsosira lowei, respectively). Representative taxa from all available sequences of Nostocalean and Stigonematalean taxa from Genbank were used as outgroups. In particular, we choose at least one member of every genus of the Stigonematales, which had a nearly complete 16S rDNA gene sequence deposited in Genbank. We also included strains from the Nostocales previously sequenced from our laboratory

for which also had complete ITS data and taxonomic certainty. Maximum likelihood analysis of the 16S rRNA gene sequence data resulted in four equally parsimonious trees with consistency index (CI)=0.601 and retention index (RI)=0.612 (Fig. 3). Overall, there was poor bootstrap support for most major clades and terminal branches. Capsosira lowei clustered with Nostoc commune with modest bootstrap support (80%). Further, these two were associated with a newly sequenced Aulosira sp. in a moderately supported clade (84%). Beyond this cluster, however, little bootstrap support was evident (<50%). The Capsosira did not fall close to a Stigonematalean clade containing Chlorogloeopsis, Fischerella, Mastigocladus, and Haplosiphon. Rexia clustered with the Microchaetacean taxon Coleodesmium wrangelii (C. Agardh) Borzı` (Fig. 3), but with no bootstrap support, and there was no support for any higher-level clades. Distance analysis of the same gene region elicited similar results (Fig. 4). Capsosira again clustered with N. commune, but with a much higher level of bootstrap support (96%). However, the cluster also containing Aulosira was very weakly supported (51%). Rexia was very weakly supported with C. wrangelii (55%), but again higher-level relationships were unsupported. Similarity scores based on the nearly complete 16S rRNA gene region were low between our newly sequenced taxa and all available outgroup taxa (Table 2). Capsosira shared the greatest similarity with Nostoc commune (96.9%) and Aulosira sp. (96.1%), and little similarity with other outgroup taxa. Rexia shared 96.2% similarity with both

18

Figure 2. Rexia erecta. (a) Trichomes cultivated in nitrogen-free Z-8 medium showing intercalary heterocytes and terminal necridial caps; (b–c) Filaments in older cultures showing necridia and hormocyte production; (d) Hormogonia demonstrating cell division in two planes; (e) Copious hormogonial production in young culture. Note the tapering evident in some trichomes. Scale bar=10 lm.

Tolypothrix distorta Ku¨tzing and Spirirestis rafaelensis (Flechtner et al., 2002) both of which were obtained from desert soils. Stackebrandt and

Goebel (1994) proposed that strains sharing <97% sequence similarity are different species, although those strains having >97% similarity are not

19

Figure 3. One of four most parsimonious trees constructed using the 16S rDNA gene for 46 taxa. Numbers above the branches are bootstrap support as a percentage of 1000 replicates resulting from maximum likelihood analysis. Newly sequenced taxa are in larger bold font.

necessarily the same species. They correlated DNA–DNA hybridization results with 16S rRNA gene sequence similarity, and found that taxa with >70% DNA–DNA hybridization typically have >97% 16S sequence similarity. If 16S similarity is <97%, the strains in all cases tested thus far have <70% DNA–DNA hybridization, and are thus separate species by this definition. Our newly sequenced strains fell below the 97% value, and thus are clearly different species. However, the Nostocales in general are very similar in their 16SrRNA sequence, and thus both species and genera can be difficult to separate using this gene alone. The 16S–23S ITS regions were very informative, and generally supported the conclusions based on 16S rRNA sequence data. ITS regions were highly variable, and in no case were identical. Some regions of the ITS, however, have conserved

secondary structure (Iteman et al., 2000). In particular, we have found two domains in the ITS region to be systematically informative: the Dstem, which consistently forms a helix with loops between D1 and D1¢ (Iteman et al., 2000), and the Box B spacer, which in Nostocales generally is located 5–6 bp downstream from the Box A spacer. The ITS is somewhat problematic because it can vary significantly in different ribosomal operons within the same strain (Iteman et al., 2000; Boyer et al., 2001, 2002). However, our observations on numerous ITS regions indicates that the D-stem structure is generally conserved among operons, although exceptions do occur (as in Scytonema hyalinum Gardner, see Boyer et al., 2001). The similarity of D-stem structure in four phylogenetically close taxa that include Capsosira lowei is striking. It is clear that C. lowei shares

20 Table 2. A reduced similarity matrix generated using the 16S rRNA gene sequence. Only the two newly sequenced taxa and their closest outgroup taxa are presented in this table. Stackebrandt and Goebel (1994) proposed that strains sharing <97% sequence similarity are different species, although those strains having >97% similarity are not necessarily the same species Strain

1

2

3

4

5

6

7

95.58

8

9

10

1. Nostoc commune UTEX584 2. Capsosira sp. nov.

96.93

3. Aulosira sp.

95.25

4. Anabaena crassa strain 215

92.21

94.22

93.86

5. Spirirestis rafaelensis

94.81

95.36

95.37

93.92

6. Tolypothrix distorta

94.62

95.16

95.33

94.01

99.02

7. Calothrix brevissima IAM M-249 8. Nodularia sp. PCC73104-1

92.82 92.75

95.36 93.97

96.31 94.56

93.61 94.59

96.08 94.65

95.81 94.74

9. Coleodesmium wrangelii

93.61

93.51

93.91

93.92

96.56

96.11

94.20

94.47

10. Cylindrospermum sp. PCC7417

93.55

94.88

96.67

94.38

95.80

95.54

95.16

95.51

94.01

11. Rexia erecta

93.52

94.37

95.02

94.65

96.07

96.15

95.41

94.94

96.02

96.08

common ancestry with N. commune, having an identical secondary structure and differing in only two bases (Fig. 5). N. punctiforme differs by six bases, which results in a minor change in structure. Aulosira is also similar, being identical in structure to N. punctiforme but differing by 9 bases from Capsosira. Rexia has a D-stem structure very consistent with members of the Microchaetaceae (Fig. 6), but differs strikingly from the Nostocaceae discussed above. It is most similar in structure to Spirirestis, even though the phylogenies based on 16S rRNA sequence placed it closest to Coleodesmium (Figs. 3, 4). The Box B spacer is much more variable than the D-stem, but still reveals similar relationships (Fig. 7). Although none of the structures were identical, it is clear that Capsosira and N. commune share ancestry. The relationship of Rexia to the other Microchaetaceae is less clear, although it shares the longer stems characteristic of the clade (Fig. 8). The combination of molecular evidence we have gathered indicates that Capsosira lowei is in the Nostocaceae, and shares close common ancestry with N. commune. This is problematic, as Capsosira has division in two planes characteristic of the Stigonematales, and has been placed in the Capsosiraceae of the Stigonematales (Anagnostidis & Koma´rek, 1990). This study demonstrates that the traditional placement of Capsosira is not

94.56

phylogenetically correct. We assume that this means that division in two planes has arisen multiple times within the heterocystous cyanobacteria. This is born out in Rexia, which also has division in two planes and cannot be placed in any morphologically circumscribed family within the Stigonematales or Nostocales. It is evidently in the Microchaetaceae, and likely has developed division in two planes independent of all other known taxa, which possess this character. What is even more problematic is the degree of molecular similarity between Capsosira lowei and Nostoc commune. One could support two taxonomic conclusions from our findings: (1) Capsosira lowei is a Nostoc species close in ancestry to the terrestrial N. commune, despite its many morphological autapomorphies; (2) Capsosira lowei belongs in its own genus due to its many derived morphological characters, but shares close common ancestry with at least some Nostoc species. If Capsosira is phylogenetically within the Nostoc clade, as we suspect it is, one must either emend Nostoc to include cell division in multiple planes, pseudodichotomous branching, and absence of colonial mucilage, or one must assume that Nostoc as currently circumscribed is polyphyletic. We suspect the latter is true. Nostoc is defined primarily by an absence of apomorphic characters and the presence of symplesiomorphies such as cell division in one plane only, isopolarity, and copious mucilage production. We are currently examining

21 the genus Nostoc to test this hypothesis. If supported, the genus will need revision.

Species descriptions Capsosira lowei sp. nov. (Fig. 1a–f) Capsosira brebissonio maxime simile, sed intra thallum Hydrothyriae venosae viventi. Coloniae in statibus naturalibus pro filis uniseriatis ramificantibus intra thallum Hydrothyriae venosae viventibus, pro cellulis singularibus vel paribus cellularum ubi thallum primo evadentibus.

Coloniae in cultura in agaro fulginosae, granulatae. Thallus in cultura saepe filis marginalibus in seriebus paralellis crescentibus, aliquando fasciculos compactos cellularum formantes. Fila uniseriata parum descrescentes ad extremia ambo, typice crispa, vel in laqueis angularibus per mutationem distinctam in planities divisionis, interdum dichotome ramificantia, aliquando biseriata subapicaliter per divisionem cellularem parallelam ad axem fili. Vagina facultativa, pellucida, diffluens, nunquam lamellata, absens in contextum lichenis. Cellulae vegitativae in formam maxime variabiles, 4.0–6.0 (8.0) lm latae, (2.4) 3.0–6.0 (7.2) lm longae. Contenta cellularum haud granulata in

Figure 4. Distance analysis tree using the 16S rDNA gene. Numbers above the branches are bootstrap support as a percentage of 1000 replicates. Newly sequenced taxa are in larger bold font.

22 contextum lichenis sed distincte granulata in cultura, fulva ad fusca. Heterocystae facultativae, maxime evidentes in contextum lichenis, saepe intercalares, interdum terminales, straminei, 4.5–6.0 lm latae, 4.0–6.5 lm longae. Hormogonia recta vel leviter flexuosa, sine vagina, uniseriata, destincte constricta ad septa. Cellulae hormogoniorum leviter breviores vel longiores quam latae, 2.8–3.2 lm latae, 2.3–4.4 lm longae, cellulis apicalibus tantum 2.5 lm wide. Most similar to Capsosira brebissonii but living in the thallus of Hydrothyria venosa. Colonies in nature living in the thalli of Hydrothyria venosa as uniseriate, branched filaments (Fig. 1f), when first escaping from the lichen often present as single cells or pairs of cells. Cultured colonies blackish, grainy. Thallus in culture often having marginal filaments growing in parallel rows (Fig. 1c, d), at other times forming compact clusters of cells (Fig. 1b, e). Uniseriate filaments slightly tapered at both ends, typically in

short tight U-shaped loops, or in angular loops formed through a distinct change in division planes (Fig. 1f), sometimes with Y-shaped branching, occasionally becoming biseriate in subapical portions of the filament through cell division parallel to the filament axis (Fig. 1c). Sheath facultative, clear, diffluent, never lamellated, not present within the lichen tissue. Vegetative cells extremely variable in shape, 4.0–6.0 (8.0) lm wide, (2.4) 3.0–6.0 (7.2) lm long. Cell contents nongranular in the lichen tissue, but distinctly granular in culture, yellowish brown to blackish brown. Heterocyte production facultative, most evident in lichen material, commonly intercalary, occasionally terminal, pale yellowish, 4.5–6.0 lm wide, 4.0–6.5 lm wide. Hormogonia straight or slightly flexuous, lacking sheath, uniseriate, distinctly constricted at the crosswalls, with cells slightly shorter to slightly longer than broad. Hormogonial cells 2.8–3.2 lm wide, 2.3–4.4 lm long, with end cells only 2.5 lm wide.

Figure 5. D-stem of the 16S–23S ITS region in Capsosira lowei and taxa that are phylogenetically close. (a) Capsosira lowei; (b) Nostoc commune; (c) Nostoc punctiforme; (d) Aulosira sp.

23 Type collected 5 May 2002 within the lichen, Hydrothyria venosa, growing in the splash zone of Hen Wallow Falls 3545.61¢ N latitude, 8314.27¢ W longitude, Great Smoky Mountains National Park, Cocke County, Tennessee, with water of pH 4.7. Holotype here designated: BRY C 37674, Herbarium of Nonvascular Cryptogams, Brigham Young University, Provo, Utah, USA. Isotype (DNA extraction): BRY C 37675, Herbarium of Nonvascular Cryptogams, Brigham Young University, Provo, Utah, USA. Isotype (in Hydrothyria venosa): BRY C 37676. Reference strain designation HWF3-JJ1: UTEX Culture Collection, University of Texas, Austin, Texas, USA. This species is named in honor of Rex Lowe, who has had significant impact on North American phycology both through his prolific publication record and his legacy of mentorship to undergraduate and graduate students. Rexia gen. nov. A generis totis in Microchaetaceis divisione cellularum in planis duobus differt.

Fila pseudoramosa, trichomis solitariis vel paucis. Trichomae isopolares. Hormogonia paucicellulares, divisibilia in plana duo. Differing from all genera in the Microchaetaceae by cell division in two planes. Filaments falsely branching, with one or more trichomes per filament. Trichomes isopolar. Hormogonia one to few celled, capable of undergoing division in two planes. Rexia erecta sp. nov. (Fig. 2a–e) Fila a substrato erecta, profuse pseudoramosa, hormogonis numerosis exorientibus separatione ramorum brevium, maximam partam trichoma singula, in cultura raro trichomis pluribus, 8–16 lm latae. Trichomae typice ad septa constrictae, raro nonconstrictae, saepe arcte flexae et intricatae, versus apices leviter descrescentes, isopolares, pileo apicali necridiarum saepe evidentes, 5–11 lm latae. Cellulae aeruginosae, saepe granulis grandibus, 2.5–8 lm longae. Hormogonia paucicellulares, raro divisibilia in plana duo. Heterocystae intercalares, infrequentes, semper nodulo polari uno, 8–10 lm latae, 5–9 lm longae.

Figure 6. D-stem of the 16S–23S ITS region in Rexia erectus and taxa that are phylogenetically close. (a) Rexia erectus; (b) Spirirestis rafaelensis; (c) Coleodesmium wrangelii; (d) Tolypothrix distorta.

24 Filaments erect from substrate, profusely false branched, producing numerous Hormogonia, which arise from separation of short branches (Fig 2e), mostly with a single trichome per filament

but in culture occasionally with multiple short trichomes per filament, 8–16 lm wide. Trichomes typically constricted at the crosswalls, infrequently not appearing constricted, often tightly bent and

Figure 7. Box-B of the 16S–23S ITS region for various Nostocales related to the described taxa. (a) C. lowei; (b) N. commune; (c) N. punctiforme; (d) Aulosira sp; (e). R. erectus; (f) S. rafaelensis; (g) C. wrangelii; (h) T. distorta.

25 entangled, tapering toward ends, but never to a hair, isopolar in growth habit, sometimes with an apical cap of one to many necridial cells (Fig. 2a), 5–11 lm wide. Cells bright blue–green, often with large granules, 2.5–8 lm long. Hormogonia a few cells long (Fig 2b–e), capable of undergoing division in two planes, although such division pattern is rare (Fig. 2d). Heterocytes intercalary, very rare, always with a single polar nodule (Fig. 1a), 8–10 lm wide, 5–9 lm long. Type collected 6 May 2002, from a wet rock face facing north northwest, on the side of Cataloochie Road, 3538.46¢ N latitude, 833.60¢ W longitude, elevation 1032 m, Great Smoky Mountains National Park, Haywood County, North Carolina, in water of pH 5.0. Holotype here designated: BRY C 37677, Herbarium of Nonvascular Cryptogams, Brigham Young University, Provo, Utah, USA. Isotype (DNA extraction): BRY C 37678, Herbarium of Nonvascular Cryptogams, Brigham Young University, Provo, Utah, USA. Reference strain designation CAT4-SG4: UTEX Culture Collection, University of Texas, Austin, Texas, USA. This genus is named in honor of Rex Lowe on the occasion of his 60th birthday, in appreciation of his professional accomplishments and personal friendship. The species is named for the false branches, which grow erect from the agar. Acknowledgements This work was supported in part by a grant from Discover Life in America. This material is based upon work supported by the National Science Foundation under grant number 0206360. Any opinions, findings, conclusions, or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation. References Ahmadjian, V., 1989. Studies on the isolation and synthesis of bionts of the cyanolichen Peltigera canina (Peltigeraceae). Plant Systematics and Evolution 165: 29–38. Anagnostidis, K. & J. Koma´rek, 1990. Modern approach to the classification system of Cyanophytes 5-Stigonematales. Algological Studies 59: 1–73.

Bold, H. C. & M. J. Wynne, 1978. Introduction to the Algae. Prentice-Hall, Englewood Cliffs, New Jersey, 706 pp. Boyer, S., V. R. Flechtner & J. R. Johansen, 2001. Is the 16S–23S rRNA internal transcribed spacer region a good tool for use in molecular systematics and population genetics? A case study in cyanobacteria. Molecular Biology and Evolution 18: 1057–1069. Boyer, S., J. R. Johansen & V. R. Flechtner, 2002. Phylogeney and genetic variance in terrestrial Microcoleus (Cyanophyceae) species based on sequence analysis of the 16S rRNA gene and associated 16S–23S ITS region. Journal of Phycology 38: 1222–1235. Bubrick, P., 1988. Effects of symbiosis on the phycobiont. In Galun, M. (ed.), CRC Handbook of Lichenology. CRC Press, Boca Raton, Florida: 133–144. Cullings, K. W., 1992. Simplified Doyle and Doyle extraction procedure. Molecular Ecology 1: 233–240. Davis, J. S. & D. G. Rands, 1993. Observations on lichenized and free-living Physolinum (Chlorophyta, Trentepohliaceae). Journal of Phycology 29: 819–825. Doyle, J. J. & J. L. Doyle, 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin 19: 11–15. Fink, B., 1935. The Lichen Flora of the United States. University of Michigan Press, Ann Arbor. Flechtner, V. R., S. L. Boyer, J. R. Johansen & M. L. DeNoble, 2002. Spirirestis rafaelensis gen et sp nov (Cyanophyceae) a new cyanobacterial genus from arid soils. Nova Hedwigia 74: 1–24. Geitler, L., 1932. Cyanophyceae. In Rabenhorst’s Kryptogamenflora von Deutschland, O¨sterreich und der Schweiz. Reprint by Koeltz Scientific Books, Ko¨nigstein, Germany 14: 1–1196. Gomez, S. R., J. R. Johansen & R. L. Lowe, 2003. Epilithic aerial algae of Great Smoky Mountain National Park. Biologia Bratislavia 58: 603–615. Iteman, I., R. Rippka, N. Tandeau de Marsac & M. Herdman, 2000. Comparison of conserved structural and regulatory domains within divergent 16S rRNA–23S rRNA spacer sequences of cyanobacteria. Microbiology 146: 1275–1286. Johansen, J. R., R. L. Lowe, S. R. Gomez, J. P. Kociolek & S. A. Makosky, 2004. New algal species records for the Great Smoky Mountains National Park, with an annotated checklist of all reported algal species for the park. Algological Studies 111: 17–44. Koma´rek, J., J. Koma´rkova´ & H. Kling, 2003. Filamentous cyanobacteria. In Wehr, J. D. & R. G. Sheath (eds), Freshwater Algae of North America. Academic Press, New York: 117–196. Ku¨tzing, F. T., 1849. Species algarum. Reprint by A. Asher and Co, Amsterdam. Nu¨bel, U., F. Garcia-Pichel & G. Muyzer, 1997. PCR primers to amplify 16S rRNA genes from cyanobacteria. Applied and Environmental Microbiology 63: 3327–3332. Posada, D. & K. A. Crandall, 1998. MODELTEST: testing the model of DNA substitution. Bioinformatics 14: 817–818. Sharkey, M. J., 2001. The all taxa biological inventory of the Great Smoky Mountains National Park. Florida Entomologist 84: 556–564.

26 Stackebrandt, E. & B. M. Goebel, 1994. Taxonomic note: a place for DNA–DNA reassociation and 16S rRNA sequence analysis in the present species definition in bacteriology. International Journal of Systematic Bacteriology 44: 846– 849. Swofford, D. L., 1998. PAUP – Phylogenetic Analysis Using Parsimony, Version 4.02. Sinaur Associates, Sunderland, Massachusetts. Thompson, J. D., D. G. Higgins & T. J. Gibson, 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting positions-specific gap penalties and weight matrix choice. Nucleic Acids Research 22: 4673–4680. Tilden, J., 1910. Minnesota algae. 1. The Myxophyceae of North America and Adjacent Regions. University of Minnesota, Minneapolis, Minnesota, 302 pp.

Whitford, L. A. & G. J. Schumacher, 1984. A Manual of FreshWater Algae, revised ed. Sparks Press, Raleigh, North Carolina 337 pp. Wilmotte, A., G. Van der Auwera & R. De Wachter, 1993. Structure of the 16S ribosomal RNA of the thermophilic cyanobacterium Chlorogloeopsis HTF (‘Mastigocladus laminosus HTF’) strain PCC7518, and phylogenetic analysis. FEBS Microbiology Letters 317: 96–100. Zucker, M., D. H. Mathews & D. H. Turner, 1999. Algorithms and thermodynamics for RNA secondary structure prediction: a practical guide. In Barciszewski, J. & B. F. C. Clark (eds), RNA Biochemistry and Biotechnology. NATO ASI Series Kluwer Academic Publishing, Dordrecht: 11–43. Zuker, M., 2003. Mfold web server for nucleic acid folding and hybridization prediction. Nucleic Acids Research 31: 3406– 3415.

Hydrobiologia (2006) 561:27–57  Springer 2006 R.J. Stevenson, Y. Pan, J.P. Kociolek & J.C. Kingston (eds), Advances in Algal Biology: A Commemoration of the Work of Rex Lowe DOI 10.1007/s10750-005-1603-5

Large-scale regional variation in diatom-water chemistry relationships: rivers of the eastern United States Donald F. Charles1,*, Frank W. Acker1, David D. Hart1, Charles W. Reimer2 & Patrick B. Cotter1,3 1

Patrick Center for Environmental Research, The Academy of Natural Sciences, 1900 Benjamin Franklin Parkway, Philadelphia, PA 19103, USA 2 Diatom Herbarium, The Academy of Natural Sciences, 1900 Benjamin Franklin Parkway, Philadelphia, PA 19103, USA 3 12 Llanfair Lane Ewing, NJ 08618, USA (*Author for correspondence: E-mail: [email protected])

Key words: diatoms, ecology, water quality, indicators, multivariate analysis, distribution, scale, Academy of Natural Sciences

Abstract We analyzed diatom and water chemistry data collected by The Academy of Natural Sciences from 47 rivers throughout the eastern United States to address several ecological questions. How does the composition of diatom assemblages vary over large regional scales? What are the most important environmental factors affecting assemblage composition and how does their influence vary among regions and with spatial scale? How do distributions and autecological characteristics of individual taxa vary spatially? What are the implications of answers to these questions for use of diatoms as water quality indicators? Data for 186 samples at 116 sites were collected from 1951 to 1991 on moderate- to large-sized rivers ranging from Maine to Texas as part of Academy monitoring and survey programs, most initiated and implemented by Dr. Ruth Patrick. Several sites were highly impaired by point and non-point source pollution. Diatom assemblages grouped into four main categories, based on multivariate analyses. Group membership correlated equally well with intermediate-scale geographic regions and water chemistry: (1) Northeastern US rivers with lower alkalinity and hardness, and pH 6.5–7.8; (2) Primarily dilute coastal plain rivers in the southeastern United States with the lowest average pH (5.5–7.3) of all sites and some with high DOC; (3) Rivers within and west of the Appalachian Mountains, generally having higher pH (>7.5) than those in other regions, but with relatively low chloride concentrations; and (4) Gulf Coast rivers with the highest chloride (>100 mg l)1), hardness (>250 mg l)1), and pH of rivers in all the groups. Hardness, pH, alkalinity, and Cl explained most of the variation among diatom assemblages, based on ordination analysis. Factors related to water quality problems, such as BOD, P, NH4, and turbidity explained much less variability at the eastern US scale, but were more important in the four intermediate-scale regions. Diatom taxa abundance-weighted mean values for water chemistry characteristics varied among the four intermediate-scale regions, often greatly, and in proportion to the average measured values for each region. Design of calibration data sets for development of water quality indicators should account for spatial scale in relation to species dispersal, regional geochemistry and habitat types, and human-influenced water chemistry characteristics.

Introduction Diatoms are excellent indicators of the ecological condition of rivers and streams, and have been

used for water quality assessment for many decades. Many useful approaches are reviewed by Cholnoky (1968), Patrick (1973), Lowe (1974), Descy (1979), Lange-Bertalot (1979), Whitton &

28 Kelly (1995), Lowe & Pan (1996), and Stevenson & Pan (1999). Increasingly, diatoms are being used in monitoring programs in the United States (Charles, 1996; Barbour et al., 1999) and throughout the world (Whitton & Friedrich, 1991; Whitton & Rott, 1996; Prygiel et al., 1999). The geographic scale of the monitoring is expanding. Examples include the USGS National Water-Quality Assessment program (NAWQA) (Gurtz, 1994) and the US EPA Environmental Monitoring and Assessment Program (EMAP) (Pan et al., 1996) in the United States. One of the greatest limitations to more effective use of diatoms as water quality indicators is the lack of detailed information on the autecology of individual taxa and how this varies with spatial scale and along major environmental gradients (e.g., pH, conductivity, nutrients). This is especially important when diatom indicators are being developed for use on a large scale (e.g., national). Because of this limitation, existing metrics often do not work as well as we would like. Few large-scale studies have been designed to provide this autecological information. The objective of this study was to gain a better understanding of diatom ecology, especially in those areas necessary to make advances in use of diatoms as water quality indicators. We addressed several questions. How do diatom assemblages vary among study sites in the eastern United States? Which water chemistry characteristics have greatest influence on diatom assemblage composition overall, and how does their importance vary with spatial scale and geographic region? What is the relative importance of natural vs. ‘‘pollutionrelated’’ factors? How important are geographic factors that are unrelated to water chemistry in influencing diatom distributions (e.g., historical dispersal events)? What are the autecological characteristics of diatom taxa common in larger rivers and streams? How and by what amounts do they vary throughout the eastern US? What are the implications of our study results for designing large-scale diatom monitoring programs? Our analytical approach to address these questions involved several steps. We first used a clustering technique to examine underlying patterns in diatom distributions and to identify groups of samples with similar species composition.

We then examined how the sample groups related to geographic location and environmental characteristics. We used the cluster analysis groups to define intermediate-scale geographic regions for subsequent examination of spatial variability. We then used ordination analysis to examine relationships between diatom assemblages and water chemistry, within the entire dataset, and within the sets of samples in the geographic regions. Next, we calculated diatom taxa abundance-weighted mean (AWM) values of several water chemistry characteristics using the entire dataset and the sets of samples in the regions defined by the cluster analysis. We examined the spatial variability in ecological characteristics of individual taxa among the regions. We then interpreted these results, addressing questions concerning interaction of scale, spatial pattern, and environmental influences posed in the introduction. Finally, we examined the implications of the results for the design of diatom-based water quality monitoring programs. Water chemistry data and the full set of AWM calculations are available at http://diatom.acnat sci.org/autecology/.

Study area and database We addressed our objectives using a unique set of data created at The Academy of Natural Sciences (ANS) during the past half century. In 1948, under the direction of Dr. Ruth Patrick, Academy scientists began conducting integrated biological– chemical–physical studies at sites on rivers and streams throughout the country. These studies were undertaken at the request of corporations, citizens groups and government agencies wanting information on the condition or ‘‘health’’ of rivers. Many of these studies were repeated over a span of many years, and some are ongoing. The studies usually involved collecting and identifying diatoms and other algae, invertebrates, and fish, and measuring a standard set of water quality variables. Assessments of ‘‘health’’ were based largely on the diversity of biological groups measured (Patrick, 1950, 1951). Standardized techniques were used to collect diatom assemblages, and to evaluate spatial and temporal variation in water quality and biological integrity. Results of these studies were published in Academy reports prepared for the project sponsors, and many have been summarized

29 and synthesized by Patrick (1994, 1995, 1996, 1998a, b, 2000, 2003) in a series of books on rivers of the United States. From the large amount of data available, we selected 186 samples for this study. Each had welldocumented diatom counts and corresponding water chemistry data. We selected samples to represent a wide geographic area, a wide range in physical and chemical characteristics, and both good and poor water quality conditions. The samples were collected between 1951 and 1991 from 116 sites on 47 rivers from Maine to Texas (Fig. 1). Maximum separation of sites was about 2800 km. The rivers are generally larger in size than in most river diatom studies, ranging from 4th to 10th order (lower Mississippi R.). Most samples were collected in the vicinity of industrial facilities, typically one at a site upstream from an effluent source and two at sites downstream. Most differences between upstream and downstream assemblages were minor; however, there were some major differences. Some sites were sampled at intervals of about 10 years; water quality changed during these periods. Most sites were sampled between July and October. Fewer than 10 samples, all from southern rivers, were collected in March, April, May and November. The rivers varied considerably in size and hydrology. Types of water quality problems addressed by the river studies included: nutrient enrichment and biochemical oxygen demand (BOD) associated with wastewater treatment plants, industrial effluent discharges, urban development, dams and impoundments, oil well brine, high turbidity due to various causes, and acid mine drainage. Some sites were influenced by estuarine conditions. All samples were collected from natural substrates, and were usually a composite of collections from all the microhabitats in the sampling area. Water quality data (Table 1) to accompany the diatom counts were sometimes limited. A complete set of measurements was available for only 104 samples, and included pH, total alkalinity, total hardness, conductivity, temperature, Ca, Mg, Na, K, SO4, Cl, dissolved oxygen (DO), BOD, total PO4, soluble reactive phosphorus (SRP), NO3, NH4, turbidity, total Si, and Fe. Correlations among variables are shown in Table 2. This set was used to investigate the relative importance of relationships of chemical characteristics with

diatom assemblages. The water chemistry data for the main set of 186 samples (Table 1) had between 1 and 30 missing values for each variable. This data set was therefore used primarily to calculate AWMs of chemistry variables for diatom taxa, and to assess the geographic variation in ecological characteristics of the taxa. The ‘‘health’’ of a site was categorized based on assignments made or implied in the ANS reports: healthy=1; semi-healthy=2; polluted=3; very polluted=4. Ruth Patrick made the majority of the health assessments, based primarily on the diversity of all biological groups (e.g., Patrick, 1950, 1951), including diatoms, and factors such as water chemistry and local pollution sources. In many cases, determinations were based on quantitative comparisons of biological diversity with sites in the same region that were minimally influenced by human activity. When sites were not specifically categorized in reports, we made assignments in a manner consistent with the approaches used in these reports. Thirty-two samples were identified as originating from sites designated as ‘‘polluted’’ or ‘‘very polluted.’’ Other sites were designated ‘‘healthy’’ (34) or ‘‘semihealthy’’ (104); 16 samples were not classified because information was insufficient. The ‘‘semihealthy’’ category often included situations where human influence was obvious, but biological diversity was still relatively high.

Methods Diatom sample material was digested with nitric acid at boiling temperatures. After settling, liquid was decanted. Distilled water was then added, the samples were allowed to settle again, and the liquid overlying the diatom material removed. This process was repeated several times. Subsamples of the cleaned diatom suspension were distributed on cover slips and affixed to microscope slides with Hyrax mounting medium. All counts were made using research quality microscopes at 1000 or 1250 magnification. Diatom counts were made over a period of 40 years by 8 diatomists (Raymond Cummins, Roger Daum, Robert Grant, Luzern Livingston, Katherine Pearson, Christine Parker Smith, Noma Ann Roberts, and John Wallace). Dr. Charles

30

Figure 1. Location of 116 ANS study sites on 47 rivers, each studied during 1 to 3 years. Symbols represent eight groups of samples based on multivariate analysis (TWINSPAN) (see text for explanation). Lines separate the four major TWINSPAN Groups: NE=Northeast; SE=Southeast; WA=West of Appalachians; GC=Gulf Coast.

Reimer supervised the identifications and enumerations during most of this period, and required the analysts to document new taxa by making drawings in special notebooks and by adding a microscope slide with a circled specimen to the ANS Diatom Herbarium. Dr. Reimer kept a record of all taxonomic name changes during the years of study. Diatom counts were entered into the North American Diatom Ecological Database (NADED) at the ANS. Original data were recorded on paper forms or stored electronically on computer floppy disks. The taxa included in each count were reviewed carefully, and were changed when necessary to make them consistent with taxonomic nomenclature used at the ANS during the mid1980s. In many cases this required the additional effort of examining original slides and making partial recounts of some slides. The number of diatom valves or cells counted varied among samples, ranging relatively evenly

from about 100 to 2500. Sample sets for some rivers had both high and low counts. Types of analyses included standard counts of 400–800 frustules, multiple 100 frustule counts of samples from different microhabitats, and detailed and semi-detailed (Hohn, 1961) counts. Only the ‘‘first row’’ portion of detailed and semi-detailed counts were entered; full counts can total up to 40,000 frustules. The ‘‘first row’’ counts were typically the first 500–1000 frustules counted, all of which were identified and counted. In subsequent rows, common taxa were not counted. We recognize that the relatively large range in number of frustules counted could have influenced the results of statistical and multivariate analyses. Output from the various analyses were sufficiently robust, however, to indicate that the large range in count size did not substantially influence the results (see Results). For multivariate analyses we used the 377 taxa, of the total of 818, that occurred with a minimum abundance of 0.5% in 2 or more samples.

31 Table 1. Summary statistics for water chemistry characteristics of sample locations Parameter

Units

Mean

Median

Minimum

Maximum

Temperature

C

24.4

25.1

24.9

24.8

7.8

37.5

8.7

32.7

DO 6.5

mg/l 6.5

6.7 0.5

6.6 11.2

0.5

11.4

pH

Units

7.2

7.3

7.2

7.3

3.4

8.8

3.4

8.8 )41.5

238

4

1600

Total alkalinity

mg/l

65

53

71

54

)41.5

220

Hardness

mg/l

155

95

129

100

4

740

Conductivity 693

l S/cm 284

NC 30

NC 1200

NC

NC

Turbidity

Units

64

33

0.3

381

77

44

0

660

Total Solids

mg/l

602

251

46

9506

307

252

46

2616

BOD

mg/l

2.7

1.8

0.1

29

2.6

2.0

0.1

29

Ca 29

mg/l 25

30 1.4

25 95

1

135

Mg

mg/l

21

9

1

308

14

8

1

144

NH-4

mg/l

0.28

0.06

0.002

4.9

0.34

0.12

0.001

4.9 2

519

SO4

mg/l

59

21

49

22

1

295

NO3 0.85

mg/l 0.32

1.7 0.002

0.3 5.9

0.002

100

Cl

mg/l

212

18

1

4597

100

18

1

2090

SRP

l g/l

103

35

1

1836

NC

NC

NC

NC

Total phosphorus

l g/l

156

77

3

796

93

36

0.3

1836

Fe 308

l g/l 100

308 1

100 8750

0.5

8750

Top level of numbers are for the 186-sample set. Numbers in bottom level are for the 104-sample set. Values were missing for many of the 186-sample set. NC=not calculated, because too few data were available.

Water chemistry data were obtained from the ANS river survey reports and entered into the NADED database. Typically, reports contained data for a standard set of characteristics that had been measured one to five times during a 1-week period. If only a summary value was reported, we used that value. If not, we calculated and entered

the mean value, or median (in cases where one or two values were very different from the others). The several chemists involved in the ANS studies used different analytical methods. All were stateof-the-art methods for the time. Most are described in earlier editions of Standard Methods for the Examination of Water and Wastewater

32 Table 2. Pearson correlation matrix of all water chemistry characteristics in the 104-sample dataset SRP

0.10

Cl

0.12

0.04

Hardness

0.59

0.10

0.64

Temp

0.11

0.03

0.20

0.05

Cond

0.30

)0.05 0.60

0.74

%DO sat

0.23

)0.17 )0.29 )0.06 )0.13 )0.21

Total alk

0.10

0.11

0.19

0.64

0.32 0.17

0.17 0.58

)0.02 0.30 0.02 0.67 )0.03 0.59

)0.17 0.02 )0.15 0.29

0.01

BOD

0.04

0.40

0.27

0.23

0.19

)0.48 0.14

0.26

Si

)0.27 )0.10 )0.07 )0.09 )0.05 0.16

)0.16 0.00

)0.09 )0.14 )0.08

Ca

0.67

0.15

0.47

0.95

0.02

0.63

)0.04 0.69

)0.06 0.57

0.23

)0.10

Mg

0.42

0.06

0.73

0.92

0.08

0.79

)0.10 0.51

0.06

0.70

0.24

0.02

0.79

NH4

)0.04 0.21

0.02

0.07

0.05

0.20

)0.37 0.05

0.19

0.11

0.36

0.01

0.09

0.06

SO4

0.41

0.59

0.76

0.05

0.62

)0.08 0.36

0.18

0.66

0.35

)0.23 0.69

0.77

NO3 Fe

0.12 0.25 0.06 0.36 )0.29 0.23 )0.09 0.30 )0.07 0.37 0.28 0.10 )0.37 )0.01 )0.13 )0.42 )0.14 )0.18 0.01 )0.48 0.00 )0.20 )0.03 0.02

0.40 0.28 0.20 0.30 )0.43 )0.41 0.08 )0.38 )0.30

pH

Ca

0.22

SRP

Cl

0.18

0.15

0.37

)0.06

0.76

Turbidity 0.01 Total solids 0.23

0.28

Hard Temp Cond %DO T Alk Turb T Sol BOD Si

(published by the American Public Health Association). We recognize that the use of different methods is a source of uncertainty, and that some chemical data are more accurate and precise than others. However, we believe the uncertainty introduced is acceptable given the geographic scale at which we are working, and the magnitude of differences among the sites. If there was significant doubt about some aspect of chemistry data, (e.g., units, method, sample location), we excluded those data from our analyses. We used TWINSPAN (Hill, 1979) to identify groups of samples with similar taxonomic composition. To determine robustness of these groupings, we ran several analyses, each time varying the dataset in some way, or selecting a different set of program options. For example, we ran individual analyses (1) with all taxa included, (2) with only the 377 taxa that occurred in abundance of at least 0.5% in 2 or more samples, (3) excluding the 12 most common taxa (those that occurred in 60 or more samples), and (4) removing the 32 sites with health categorized as ‘‘polluted’’ or ‘‘very polluted’’. Percent abundance counts and log10+1 of percentages were both used. We varied the five cut levels used to define pseudospecies to assess the relative influence of rare and common

Mg

0.12

NH4 SO4

NO3

taxa, but in most cases used the default values of 0, 2, 5, 10 and 20%, or their logarithmic equivalents. Default values were used for other program options. Final selection of TWINSPAN groups was based on the first three divisions of samples. To explore relationships between diatom assemblages and water chemistry, we used several options within the CANOCO program (ter Braak, 1986; ter Braak et al., 1995; ter Braak & Sˇmilauer, 1998). We performed principal components analysis (PCA) of all water chemistry variables in the 104 sample dataset to determine the main chemistry gradients within the dataset. To select a subset of chemistry variables to use in ordination analysis, we used stepwise forward selection in canonical correspondence analysis (CCA) to choose chemistry variables that represented the major gradients, but did not correlate strongly with each other in explaining variability among diatom assemblages. We used Detrended CCA (DCCA; detrending by segments) to quantify the variation in diatom assemblages explained by the selected chemistry characteristics within the entire dataset, and in the sets of samples in the intermediate-scale geographic regions. We chose DCCA, which is based on unimodal responses of species to environmental gradients, because lengths of

33 DCCA axes varied from 3 to 4, and the data matrix contained a large proportion of zeros (Lepsˇ & Sˇmilauer, 2003). Also, DCCA eliminated the ‘‘horseshoe’’ effect apparent when CCA was used. We eliminated variables from the analyses if the variance inflation factor (VIF) was greater than 20. For all analyses, we transformed (log10+1) both species and water chemistry data (except pH). We used CCA in the CANOCO program to partition the variation in diatom assemblage composition due to environmental and spatial factors. We used data for the 104 samples with complete water chemistry, but included only the variables pH, alkalinity, hardness, Cl, BOD, SRP, and turbidity. To develop spatial data, we centered latitude and longitude on their means, and then calculated the cubic trend surface regression for x, y, x2, x3, y2, y3, xy, xy2, and yx2, where x=latitude and y=longitude. We then used forward selection in CCA to determine which combination of these variables contributed most to explaining the variation in diatom distributions. All except xy and yx2 made a statistically significant contribution (p<0.05) based on Monte Carlo simulations, and were included in the partitioning analysis. All variables explained about the same amount of variance; Lambda 1 for marginal effects ranged from 0.23 to 0.28. We ran the variance partitioning analyses following the steps outlined in Borcard et al. (1992): (1) CCA with both environmental and spatial variables to determine the variance due to these two factors together, (2) CCAs with environmental and spatial data separately to determine the variance due to each individually, (3) partial CCA with environmental data as explanatory variables and spatial data as co-variables to determine the amount of variability not explained by spatial data that could be explained by environmental data, (4) the reverse, partial CCA using spatial data as explanatory variables and environmental data as covariables, and (5) calculation of the amount of variability that could be attributed to environmental data alone, to spatial data alone, and to an inseparable combination of environmental and spatial factors. AWM values for diatom taxa were calculated using WACALIB 3.3 (Line et al., 1994) and an in-house software program. We included only taxa that occurred in at least five sites in abundance >0.5%. When calculating the ‘‘health’’ AWM

values, we used counts for taxa in a sample only if the relative abundance was greater than 0.5% and maximum abundance was >2% in at least one sample; data for sites influenced by acid mine drainage were not used.

Results Diatom taxa In all, 818 diatom taxa were identified within the 186 samples selected. Of these, 191 taxa occurred in only 1 sample. There were 377 taxa that occurred with a minimum abundance of 0.5% in 2 or more samples. Fifteen taxa occurred in onehalf or more of the samples; 56 occurred in onequarter or more. The most frequently occurring taxa are considered generalists, and most are often associated with poorer water quality conditions. Large-scale patterns in assemblage distributions We used results of the several TWINSPAN analyses of the 186 sample dataset to define four major Groups and four minor Groups of samples. This set of groups is described below and represents a synthesis of results from all TWINSPAN analyses. Most samples always occurred in the same group. Those that sometimes fell into different groups were assigned to groups in which they occurred most frequently. Final assignments of the least stable samples, those few that tended to shift among groups more than others, were made with the objectives of keeping samples from the same river together and of maintaining similar water chemistry characteristics among groups. The four major TWINSPAN Groups corresponded with geographic regions (Fig. 1; river names are listed in legends for Fig. 4). Each of these four groups is henceforth referred to by a two-letter code indicating its geographic location: 3 (NE), 4 (SE), 6 (WA), and 8 (GC). There were substantial differences in water chemistry among the eight TWINSPAN Groups (Fig. 2). Group NE (Northeast, Group 3) includes all samples from that region except those from three of the minor TWINSPAN Groups (Group 1, Pennsylvania sites significantly affected by acid mine drainage (pH<5.4); Group 2, the Kennebec R., Maine,

34 20

NE SE

800

WA

GC

15 600

NE SE

400

WA

GC

BOD (mg/L)

Total Hardness (mg/L)

1000

200

10

5

0

0 1

2

3

4

5

6

7

8

1

9

3

4

5

6

7

8

1000

8

6

5

NE SE

WA

GC

Phosphate (μg/L)

800

7

pH

2

NE SE

WA

GC

600

400

200

4 0

3 1

2

3

4

5

6

7

8

TWINSPAN Groups

1

2

3

4

5

6

7

8

TWINSPAN Groups

Figure 2. Hardness, BOD, pH, and SRP concentrations in the 8 TWINSPAN Groups based on the data set of 104 samples with complete water chemistry. Box and whisker diagrams show mean, median, range, and 25th and 75th percentiles of values for samples in each group.

which is separated by considerable geographic distance from all other Northeast sites; and Group 5, the Nanticoke R. sites in Maryland, which had high Cl concentrations due to estuarine influence). In some TWINSPAN runs the Nanticoke sites were included with samples in Groups 3 or 4. Group NE sites had somewhat lower levels of alkalinity and hardness than samples from other groups, and pH ranged from 6.5 to 7.8. Group SE (Southeast, Group 4) includes samples primarily from coastal plain rivers. Group SE sites were generally more dilute than those in other regions, and had the lowest pH values (5.5–7.3). Group WA (West of Appalachians, Group 6) includes a diverse set of river types. In general, the sites had higher pH (>7.5) than sites in other groups. Group

7, a minor group, consists of sites on the high hardness Ottawa/Auglaize rivers in a limestone area of northwestern Ohio. Group GC (Gulf Coast, Group 8) sites are all relatively near the Gulf of Mexico. They had the highest chloride (>100 mg l)1) and hardness (>250 mg l)1) of all groups. In some TWINSPAN runs, the furthest upstream site on the Guadalupe R. was included with the samples in Group WA (it was the Group GC site closest geographically to Group WA). Also, the Loosehatchie R., Tennessee, sites were sometimes included in Group GC, though they were usually included in Group WA. These sites were geographically near Group GC sites, and have a high turbidity level, in common with many of its members.

35 Relationships between diatom assemblages and water chemistry characteristics Water chemistry characteristics explain a considerable amount of the variation among diatom assemblages. We examined diatom – water chemistry relationships at both large and intermediate regional scales. We defined ‘‘large scale’’ as the eastern US, and ‘‘intermediate scale’’ as the four regions based on the TWINSPAN analysis (Groups NE, SE, WA, and GC; Fig. 1). Large scale PCA of the 18 chemistry variables in the 104 sample data set extracted two major water quality gradients. The first PCA axis correlates strongly with major-ion chemistry variables: hardness, Mg, Ca, conductivity, SO4, total solids, and Cl, in order of decreasing correlation coefficient (Table 3). Total alkalinity, pH and Fe are less highly expressed on this axis. The second axis correlates with variables often associated with human influences on water quality: percent DOsat, pH, BOD, NH4, SRP, and turbidity (Table 3). The stepwise forward selection in CCA, with all water chemistry variables in the 104 sample set,

provided a statistical basis for excluding only two variables from ordination analyses. Of the 18 variables, 16 are significant components of a multiple regression model based on Monte Carlo permutation tests ( p<0.05; p of the first 15 was<0.006). In order of decreasing importance they are: total hardness, Cl, SRP, total alkalinity, conductivity, total solids, BOD, pH, total Si, turbidity, SO4, % DO saturation, Ca, Mg, temperature and Fe. Only NH4 and NO3 were not significant additions and therefore were not included in subsequent ordination analyses. DCCA of a 96-sample dataset (the 104 sample dataset minus the eight samples not in the four major TWINSPAN Groups, NE, SE, WA, GC) (Fig. 3), shows that hardness, pH, alkalinity and chloride explain more of the variability among samples than any other factors at the scale of the eastern US (Table 4). These variables correlate most strongly with the first axis. SRP, turbidity and Si correlate most closely with the second axis. Temperature and turbidity are most closely associated with the third axis. The chemistry variables included in the analysis were the 16 significant variables in the forward selection analysis described above, but excluding Ca and Mg because they had a high VIF (>20). In this

Table 3. Correlation of water chemistry variables in the 104-sample set with PCA axes 1–4 Axis 1

Axis 2

Axis 3

Axis 4

0.48

)0.59

)0.44

0.06

SRP Cl

0.21 0.71

0.38 0.20

)0.33 0.32

0.45 )0.30

Hard

0.94

)0.16

0.19

0.05

Temp

0.18

)0.02

)0.42

)0.57 )0.24

pH

Cond

0.80

0.07

0.29

%DO sat

)0.26

)0.68

0.29

0.18

Total alk

0.51

)0.44

)0.61

)0.11

Turbidity

0.09

0.30

)0.35

)0.12

Tot solids BOD

0.77 0.37

0.10 0.57

0.30 )0.39

0.13 0.25 )0.33

)0.02

0.18

0.07

Ca

0.86

)0.22

0.09

0.18

Mg

0.92

)0.05

0.25

)0.08 0.15

Si

NH4

0.21

0.55

)0.35

SO4

0.81

0.10

0.23

0.25

NO3

0.37

0.07

)0.16

0.54

)0.49

0.35

0.38

)0.08

Fe

36 Northeast

SRP

Southeast Turbid

West of Appalach.

Temp

Gulf Coast SO4

BOD

Cl

NH4-N

TotSol Hard

Fe % DO Sat

pH

Cond Si

0

8

Figure 3. Relationships among diatom assemblages and water chemistry characteristics in the 104 sample data set. Detrended CCA includes all environmental variables with a statistically significant contribution in automatic forward selection. Symbols distinguish the four main TWINSPAN Groups. The arrow for total alkalinity is not shown; the arrowhead would be located between those for pH and hardness.

analysis, the 16 water chemistry variables explain 26% of the total variance in diatom assemblage composition. The variables related to the first axis (Fig. 3) are clearly most important in distinguishing among the four major groups of samples, especially the Gulf Coast sites (right side of

the plot) and the much more dilute Southeast sites (left side). Most characteristics related to anthropogenic pollution (e.g., total P, NO3, BOD) are considerably less important at this large scale, and are most closely related to the second axis.

Table 4. Correlations between water quality variables and sample scores on DCCA axes 1-4 for the 104-sample dataset Axis 1

Axis 2

Axis 3

Axis 4 0.21

0.70

)0.06

)0.14

SRP

0.02

0.60

)0.15

0.19

Cl

0.64

0.21

0.28

)0.03

Hard

0.88

)0.05

)0.12

0.11

Temp

0.24

0.33

0.48

)0.26

Cond %_DO_sat

0.62 0.09

)0.28 )0.13

0.02 )0.06

0.18 0.08

Total alk

0.75

)0.09

)0.17

)0.03

Turbidity

0.04

0.35

0.60

0.20

Total solids

0.52

0.03

)0.22

0.45

PH

BOD

)0.02

0.10

0.12

0.27

Si

)0.12

)0.37

0.04

)0.10

0.67

0.26

)0.07

0.43

)0.33

)0.10

)0.12

)0.26

SO4 Fe

37 West of Appalachians Cumberland Tennessee Green N.F. Holston Loosahatchie Holston White Clinch

Hard

Northeast

+2.0

+3.5

Turb

Cl

Kennebec Potomac Brandywine Schuylkill Susqehanna Assunpink Clarion Slippery Rock

SRPhos BOD

Cl Temp

SRPhos

-2.5

Temp

Wolf

TotAlk

BOD

+2.5

-2.5

PCTDO

+2.5

Hard

DO

Turb

TotAlk

pH

BOD

-2.0

Gulf Coast Neches Guadalupe Sabine Mississippi

Temp

DO

+1.5

+2.5

-2.5

PCTDO pH

TotAlk DO Hard PCTDO pH -2.5

TotAlk Turb -2.0 SRPhos Cl

+3.0

PCTDO

Hard

-1.5

DO

+2.5

Southeast Coastal Cape Fear Wateree Savannah Flint Congaree Black Edisto Escambia North Anna Pocalla

SRPhos Turb

BOD

Temp Cl

-2.5

pH

Figure 4. Relationships among diatom assemblages and water chemistry characteristics of samples in each of the four major TWINSPAN Groups. (a) West of the Appalanchians (WA), (b) Northeast (NE), (c) Gulf Cost (GC), and (d) Southeast (SE). Detrended CCA includes all environmental variables with a statistically significant contribution in automatic forward selection. Symbols indicate in each of the major TWINSPAN Groups.

Intermediate scale Northeast (NE) In general, there is little overlap in sets of samples from different rivers (Fig. 4). The exception is that all Susquehanna sites group within the Potomac cluster. The Susquehanna samples were all collected within 1 year, whereas the Potomac samples were collected during 3 years; this probably explains the greater dispersion of Potomac samples. The water chemistry factors explaining the highest proportion of variation in the eastern US

96-sample data set also explain most of the variation at the scale of the Northeast (Fig. 4); water quality variables such as SRP and BOD are of less importance. The 1st axis has highest intraset correlations with alkalinity, pH, hardness, Cl and temperature. Also, going from left to right along the horizontal axis, rivers tend to decrease in size, increase in gradient, and vary in location from south to north. Location, habitat and size factors vary consistently with major-ion chemistry, and may be responsible for part of the variation attributed to these variables. The 2nd axis

38 corresponds with variations in SRP and DO. The higher SRP and BOD, and lower DO, in the Schuylkill and Assunpink rivers are clearly indicated. At the time samples were collected, both these rivers had significant point source inputs affecting water quality. Southeast (SE) There is more similarity and overlap in diatom assemblages from SE rivers than for any other major group (Figs. 3 and 4). Samples with greatest diversity and those from sites located nearest each other tend to group at the center of the plot. The three sets of samples that appear separately on the right side of the biplot are from rivers with the lowest pH and high color (Black, Edisto and Pocalla). The river represented at the top of the biplot, the North Anna, is located further north than all other samples. The most important water chemistry factors correlating with the first axis are those related to pH and DO. The second axis represents a water quality gradient, defined by the variables SRP, turbidity and BOD. Interpretation of the biplot is complicated somewhat by the presence of samples from the black water rivers, which tend to have higher natural levels of DOC that may have an influence on DO concentrations. Nonetheless, samples from sites that had phosphorus and BOD-related water quality problems plotted in the lower part of the graph (e.g., Cape Fear, Wateree). West of Appalachians (WA) Diatom assemblages from individual WA rivers group separately from each other; there is little overlap (Fig. 4). Some groups lie near each other, however. The WA sites generally had higher pH (Fig. 2) and lower Cl than those in other regions. Chemistry differences among sites were less than among groups of sites in the other three regions, and physical habitat differences are probably more important. There is a general trend for higher elevation, steeper gradient sites to be on the right side of the biplot, and this trend is probably also reflected in the strong relationship with temperature. The relationship with temperature is stronger in this region than in any others. Turbidity levels were much higher for the Loosahatchie than other sites, and are responsible for the length of the turbidity arrow. Another important difference in

physical characteristics is that sites on the Tennessee and Cumberland rivers were above or immediately below impoundments that had deep, slowly moving water. The vertical axis has a weaker relationship with SRP and BOD than for other regions, probably at least in part because there is less of a range in these variables for this group of rivers. There is a stronger relationship with SRP on the third axis. This may be because the temperature/elevation gradient displaces pH related factors as being most closely associated with the first axis; pH-related factors then become more important on the second axis, and perhaps SRP and BOD/DO factors are displaced to the third axis, compared with other regions.

Gulf Coast (GC) Assemblages from the Neches, Sabine and Mississippi are relatively similar and overlap (Fig. 4). Samples from the Guadalupe are separate, probably because sites from which they were collected have a higher alkalinity than sites on the other rivers. There is no clear explanation for the distribution of samples based on geography. Hardness and Cl concentration are considerably higher, on average, than in other regions/groups. Levels of turbidity and SRP are also higher, due to natural conditions, and correlate with hardness, Cl and pH. A primary reason for these conditions is that the sites are located near the coast and marine influences cause higher Cl, and that some are in a drier climate resulting in higher dissolved solids and turbidity. These differences set most samples from this region apart from the others. Unlike the case for some other regions, SRP correlates more strongly with the first than the second axis. The second axis correlates much more strongly with BOD than any other factor, and represents pollution-related water quality conditions. It also explains the spread along the second axis of samples within river clusters. The Guadalupe, Mississippi/Reserve Canal, and Sabine data sets each include samples collected during three different years, each sampling interval 5–10 years apart. There were significant improvements in water quality during that time, particularly related to BOD. Samples collected in the earlier years at sites where higher BOD prevailed are located above

Achnanthes biporoma Hohn et Hellermann Achnanthes clevei Grunow Achnanthes exigua var. heterovalva Krasske Achnanthes hauckiana Grunow Achnanthes lanceolata (Bre´bisson in Ku¨tzing) Grunow Achnanthes lanceolata var. dubia Grunow Achnanthes linearis (Smith) Grunow Achnanthes minutissima Ku¨tzing Amphipleura pellucida (Ku¨tzing) Ku¨tzing Amphora ovalis (Ku¨tzing) Ku¨tzing Amphora ovalis var. pediculus (Ku¨tzing) Van Heurck ex DeTony Amphora submontana Hustedt Anomoeoneis vitrea (Grunow) Ross Bacillaria paradoxa Gmelin

Taxon

Health AWM

pH AWM

SRP AWM

BOD

5

32

6

42

73

20

25

10

27

14

8

18

27 13

14

12

5

36 32

125 33

5

5

29 19

12

69

16

20 14

27

81

16

5

6

15

12

19

27

26

5

8

5

9

14

8

10

10

14

1.8

2.0

2.0 1.4

2.3

1.3 1.6

1.6

1.7

2.0

1.9 1.3

2.0

1.8 2.2 2.7 1.3

1.8

1.3

2.0 2.1 1.8 2.1

1.7

1.7 2.1 1.9 1.4

1.9 2.0 1.8 1.7

2.2

2.0

1.9 2.2 2.9

1.6

6.8

6.8 7.7

7.0

7.0 7.7

1.9 7.2

1.8 6.8

2.1 7.8

6.8 7.6

7.7

1.0 7.8 8.2 6.9 7.9

2.0 7.4

7.2

1.4 7.1 7.2 7.0 7.6

6.9

1.5 7.2 8.2 6.9 7.5

2.0 7.6 8.0 6.9 7.8

2.2 7.1

2.0 7.7 8.3 6.9

7.6 8.2 6.9

6.9

7.4 70

6.9 26

8.0 53

7.7 47

7.6 30

18

7.4 47

118

7.5 66

7.1 51

7.1 35

7.8 19

61

57

55

33

41

72

28

26

9

74

18

57

42

22

302 54

90

34

34

65

113 35

26

34

24

20

45

97

27

71

5

16

28

51

73

36

18

1.1

4.9 1.6

2.7

1.6 1.4

2.0

1.7

3.1

2.6 0.8

1.4

2.3 3.3 1.0 2.2

1.5

1.6

1.7 3.1 1.0 1.9

1.7

2.2 3.2 2.6 1.6

2.0 2.0 2.5 2.1

1.9

2.0 5.1 1.4

2.4 4.4 1.6

1.0

2.2 228

2.6 374

4.0 56

1.4 59

1.6 536

11

1.4 20

9

0.9 55

1.4 47

2.0 1251

2.0 50

11

14

9

9

16

7

18

24

13 17

8

8

4

6

11 11

9

9

9

9

854

732

84

18

902

224

2004

294

1602

55

Continued on p. 40

21

26

22

23

21

21

11

NE SE WA GC

AWM

Cl

All NE SE WA GC All NE SE WA GC All NE SE WA GC All NE SE WA GC All NE SE WA GC All

samples

Number of

Table 5. Occurrence and AWM values of health, pH, SRP, BOD and Cl for 177 diatom taxa found in 15 or more samples

39

Biddulphia laevis Ehrenberg Caloneis bacillum (Grunow) Cleve Caloneis hyalina Hustedt Capartogramma crucicula (Grunow ex Cleve) Ross Cocconeis fluviatilis Wallace Cocconeis pediculus Ehrenberg Cocconeis placentula var. euglypta (Ehrenberg) Cleve Cocconeis placentula var. lineata (Ehrenberg) Van Heurck Coscinodiscus pygmaeus var. micropunctatus Brun et Peragallo Cyclotella aliquantula Hohn et Hellermann Cyclotella atomus Hustedt Cyclotella florida Voight Cyclotella meneghiniana Ku¨tzing Cyclotella pseudostelligera Hustedt

Taxon

Table 5. (Continued) Health AWM

pH AWM

SRP AWM

BOD AWM

Cl

31

21 20

25 19

13

8

21

117 31

17

64

83

6

6

29

14 13

5

5

13

19

16

13

5

28

22 8

8

47

5

16 16

27

74

10 12

13

23

12

16 6

28

5

17

8

27

14 6

7

29

9

43

7

9

24

1.5

1.9 1.7

1.9 1.0

2.1 2.4

2.1 2.2 1.5 2.0

2.2 2.3 2.2 1.2

2.1 2.2

2.1 2.3 2.0 1.7

2.1 2.2 1.7

2.1 1.9 1.3

1.8 2.0 1.6 1.8

1.4 1.6 2.0 1.5

1.5 1.9

1.8

1.5

2.0

2.0 2.0 1.7 1.2

1.9

7.1

6.7 7.9

6.7 7.6

6.6 7.5

2.7 7.9 8.3 6.9 7.5

2.4 7.7 7.9 7.1 7.6

2.8 7.8 8.5

2.1 7.7 8.5 6.9 7.4

2.0 7.7 8.5 7.0

2.4 7.3 8.6 7.0

7.6 8.7 6.9 7.8

1.0 7.5 7.5 6.6 7.7

7.9 8.2

1.7 6.9

7.1

1.8 6.9

2.2 7.4 7.9 6.8 7.7

2.0 7.3

7.6 50

45

7.7 116 71

46

7.4 152 52 6.9 91

57

57

26

153 49

261 63

72

38

26

45

31

186 50

101 37

99

45

139 17

58

80

60 32

74

42

7.3 113 29

7.5 48

65

7.7 41

30

7.1 35

110

7.6 88

7.3 93

7.7 41

2.1

4.0

1.8 1.1

3.0 1.3

2.6 4.6

1.2 3.6 1.2

2.4 3.9 1.8 1.7

1.5 2.5 0.7 2.3

1.2 3.6

1.8

2.3

2.3

3.2 3.7 1.6 1.5

121 2.9 4.0 1.5 1.6

81

678 4.0 2.1

133 3.7 4.9 2.5 1.8

184 1.9 1.7 2.4

70

5

10

14

113 2.4 2.5 0.8 1.5

26

41 18

15 15

19

1.4

20 22

2.3 108 26

3.9 190 24

2.0 87

1.8 767 26

0.8 912 18

1.1

18 18

1.8 206

241

1.4 144

2.8 642 25

1.3 251 8

18

9

18

29

15 8

9

15 22

15

15

9

6

11

11 12

12 141

9

12

125

294

198

1316

1240

897

18

415

29

1043

67

All NE SE WA GC All NE SE WA GC All NE SE WA GC All NE SE WA GC All NE SE WA GC All NE SE WA GC

samples

Number of

40

Diploneis oblongella (Naegeli ex Ku¨tzing) Ross Diploneis puella (Schumann) Cleve Eunotia curvata (Ku¨tzing) Lagerstedt Eunotia exigua (Bre´bisson ex Ku¨tzing) Rabenhorst Eunotia incisa Smith ex Gregory Eunotia maior (Smith) Rabenhorst Eunotia pectinalis (Mu¨ller) Rabenhorst Eunotia pectinalis var. minor (Ku¨tzing) Rabenhorst Eunotia tenella (Grunow) Cleve

Cyclotella stelligera (Cleve et Grunow) Van Heurck Cyclotella striata (Ku¨tzing) Grunow Cymbella affinis Ku¨tzing Cymbella lunata Smith Cymbella minuta Hilse ex Rabenhorst Cymbella minuta var. silesiaca (Bleisch ex Rabenhorst) Reimer Cymbella prostrata (Berkeley) Cleve Cymbella sinuata Gregory Cymbella tumida (Bre´bisson ex Ku¨tzing) Van Heurck Diatoma vulgare Bory 7.4 41

6 14 11 13 15 22 27 23

27

25

17

17

19

31

36

27

17

32

13

14 23 22

67

6

10

17

33

17

12

10

16

21

2.0

1.7

1.3

1.9

2.0

2.7

1.8

1.5

1.7

2.1

20 1.8

10 1.7

1.8 1.8

1.9

2.0

2.8

1.7

1.5

2.3

2.2

1.4

1.3

1.9

1.8 1.7 1.9 1.7

1.7 1.7

1.4

2.0

6.7

6.6

7.5

7.8

6.3

5.8

6.3

6.7

5.9

4.8

6.5

1.8 7.5

1.8 7.2

7.8 7.4

6.3

5.8

6.3

6.7

5.8

6.2

6.5

7.0

6.9

7.9

7.5 7.9 7.0 7.8

7.7 7.9

7.8

7.7

6.5

7.6

7.3

97

28

29

153

94

232

154

94

252

100

21

33

22

13

126

102

43

59

10

47

50

16 45

15 12

18

2.1

6

13

1.7

3.4

1.8

3.1

3.1

6.7

2.2

2.1

0.6

1.2

2.4

2.0

0.7 3.0

3.1

3.1

7.0

2.2

2.1

1.7

1.8

1.0

1.8

0.3

2.2 4.0 2.7 1.6

3.4 3.5

1.9

0.8

48

9

10

8

7

21

8

722

11

8

11

6

7

6

7

7

5

10

17

13

7

21

10

752

1733

81

24

512

15 167

Continued on p. 42

12 15 2.0 1449 2.5

15

10

15 7

5

13 4

24 9

17 17 8

14

8

27

1.8 2.1 1.9 1.6 1.9 19

0.9

31

1.9 648

1.6 2.8 0.7 1.5 1.8

1.8

42 102 1.6 3.0 1.4 1.4 2.0

55 35

149

94

46

44

14 52

41 70

64

60

29

49 52

1.2

111 27 40 23 13 1.9 1.7 2.0 1.9 1.9 7.1 7.3 6.3 7.8 7.4 37

1.8

65 11

30 54

41

6

2.3 6.9

1.6 1.6 1.8 1.6 1.6 7.7 7.4 6.8 7.8 7.8 15

2.1

1.6 2.2 2.0 1.3 2.0 7.5 8.4 6.9 7.7 7.9 50

59

16

7

13 5

9

35

16 15 6

5

7

18

47

41

Fragilaria pinnata Ehrenberg Fragilaria vaucheriae (Ku¨tzing) Petersen Frustulia rhomboides (Ehrenberg) De Toni Frustulia rhomboides var. amphipleuroides (Grun.) DeT. Frustulia rhomboides var. crassinervia (Bre´b. ex W. Sm.) Ross Frustulia vulgaris (Thwaites) DeT. Frustulia weinholdii Hustedt Gomphonema affine Ku¨tzing Gomphonema clevei Fricke Gomphonema dichotomum Ku¨tzing Gomphonema gracile Ehr. emend. V. H. Gomphonema grunowii Patr. Gomphonema intricatum Ku¨tzing Gomphonema olivaceum (Lyngb.) Ku¨tz. Gomphonema parvulum (Ku¨tz.) Ku¨tz. Gomphonema sphaerophorum Ehrenberg

Taxon

Table 5. (Continued) Health

pH AWM

AWM

SRP AWM

BOD

16 7

5

5

30

32

6

36

17

161 31

44 36

11

6

16

31

18 6

29

9

14 17

16

59 15

16

18

16

16

22

20

8

18

5

24

20 19

25

82

30

17 9

11

60

10

35

5

5

8

18

2.1 1.6

1.8

2.1 1.1

1.7

2.0

2.0 1.3

2.8 1.1

2.5

2.0 2.4 2.1 1.8

1.7 2.0

1.8

2.1

1.7 1.9 1.3

1.9 2.0 1.8 1.8

2.1

2.0

1.6 2.0 1.6

1.9

1.7

1.7

1.7

1.7 2.0 1.6 1.7

2.1 2.5 1.8 2.2

6.4

6.6 8.0

6.1

6.6

6.8 7.8

6.7 7.6

2.6 7.6

1.9 7.0 8.0 6.5 7.4

7.9 8.4

1.5 7.3

7.0

7.2 7.7 6.8

7.4 8.1 6.8 7.4

6.6

2.0 7.7

7.0 7.6 6.8

7.0 7.4 6.9 6.9

6.6

7.1

6.1

1.4 7.2 7.3 6.8 7.7

2.1 7.1 7.3 6.9 7.3

53

54

85

133

170

95

79

7.6 109

7.3 97

55

132 103

7.7 57

137

79

76

88

97

97

60 60

58 62

7.8 20

7.6

7.3

32

40 121 47

135

214 95

19

111 39

88

51

55

122

199 74

171

87 36

21 4

1.2

4.5 2.0

1.8

1.6

1.4 1.5

5.2 1.6

2.5 2.1 2.8 3.3

2.6 4.0

1.6

3.8

1.9 1.2 1.5

2.7 2.2 2.3 3.7

1.6

1.4

2.0 2.9 1.8

2.8 2.9 3.3 1.3

1.2

3.2

1.8

2.6 3.5 2.3 2.6

120 1.8

72

26

25

11

153 2.3 3.4 1.7 2.4

37

22

18

1.6 140

2.0 42

22

1.2 37

19

27

20

13

11

9

23

18

13

18

21

28

104 20

317 31

2.0 33

1.3

1.5

10

13 11 65

13

14 27

6

12 19

13

6

10 19

5

14 5

7

12 41

9

165

169

86

46

1497

1268

NE SE WA GC

AWM

Cl

All NE SE WA GC All NE SE WA GC All NE SE WA GC All NE SE WA GC All NE SE WA GC All

samples

Number of

42

Meridion circulare (Grev.) Ag. Navicula aikenensis Patr. Navicula arvensis Hustedt Navicula atomus (Ku¨tz.) Grun. Navicula bicephala Hustedt Navicula biconica Patr. Navicula canalis Patr. Navicula capitata Ehrenberg Navicula capitata var. hungarica (Grun.) Ross Navicula cincta (Ehr.) Ralfs Navicula cohnii (Hilse) Lange-Bert.

Gyrosigma acuminatum (Ku¨tz.) Rabh. Gyrosigma nodiferum (Grun.) Reim. Gyrosigma scalproides (Rabh.) Cl. Gyrosigma spencerii (Quek.) Griff. & Henfr. Melosira ambigua (Grun.) O. Mu¨ll. Melosira distans var. alpigena Grunow Melosira granulata (Ehr.) Ralfs Melosira granulata var. angustissima O. Mu¨ll. Melosira varians Ag. 90 28

14 11 19 12 10 1.6 1.3 2.0 1.2 2.0 7.2 7.3 6.7 7.2 7.5 11 8 6 6 15

16

24 5

48 6

32

30

24

29

6

6

10

17 5

1.7

1.8

11 2.0

9

1.9

2.3 7.0

2.0 2.1 1.7 7.8

1.8 1.0 2.0 7.2

1.9

6.4 7.1

73 38

70 61

58

46 45

92

42

6.7

7.4

63

8.4 6.9 7.6 147

6.9 7.6 7.2

6.8 7.6 7.2 131

2.0 7.0 8.4 6.8

6.4

2.3 1.1 2.4 7.2

2.1 2.3 1.9

1.9

6.8 7.2 7.3

7.1 7.7 6.8

2.3 1.4 1.9 6.9

2.0 2.2 2.0

2.1

10 1.6

6

5

6

24 6

6.9

2.4 3.1 2.6 1.3 1.3

2.4

248

1.5

247 1.6

20 20 195 1.1

19 73 42

15

44

75

27

32

6

11

73

1.3

1.8 608

2.3 0.9 0.6 1907

2.7 1.1 1.4 755

2.1 1.2 2.6

1.0 4.2 1.2 0.8 1.0 215 82 46 405 1.7

49

2.5 1.8

3.9 3.0 1.5

2.7

1.6 3.2 1.1

3.6

2.7

2.6 4.2 2.5

134 2.0 4.9 0.8 27 59 76

63

57

47

101 44 81

37

39

6.9

39

14

15

125 37 92

85 88

1.9

0.7 1.9

147 37 158 1.7 4.1 1.6 1.5 1.3 151

161 103 179

1.9

1.6

36 48 121 1.7 3.9 1.6 1.7 0.8 168

8

6.5 8.0 6.3

1.1 1.9 180

2.1 0.7 2.0 266

2.3 104

2.1 2.9 1.8 1.2 2.9 19

1.5

2.0

1.9

18 32 104 1.3 3.6 1.3 1.3 1.7 26

21 49

21 7

2.3 1.9 2.4

39 21

21 12

356 1.5

133 34 39

26

73

112 28 34 31 10 1.9 1.9 2.0 1.7 2.5 7.3 7.5 6.8 7.4 7.5

46 47

33 101

35

20 10 12 1.8 1.8 1.7 1.6 2.7 7.2 7.4 6.7 7.5 7.5

6.7 7.6

52 6

7.1

1.5 2.0 1.4 1.4 2.8 7.3 7.5 6.6 7.6 7.5

1.6 1.0

15 18 9

1.6

53 6

70

21 19 19 1.7 2.0 1.4 1.7 2.3 7.4 8.4 6.8 7.6 7.3

7.0 7.7 97

67 5

1.5 1.9 7.4

1.8 2.4 1.6 1.5 2.0 7.4 8.4 6.7 6.8 7.9 65

1.7

7.0 99

7.0 7.7 7.6 15

6.7

13 11

6

2.4 7.1

1.3 1.9 1.6 7.5

2.3

33

25

16 1.6

1.7

14 18 10 8

34

5

50

12 5 8

6

18

11

3

3

11

9

14

714

1445

593

445

99

734

850

362

61

1072

690

Continued on p. 44

10

26 12 2619

10 14 1850

10 13

15 5

19 10

8

21 9

8

6

18 10

17 12 9

18 13 27

25 9

470 28 350

18 11 6

9

21 9

6

14

43

Health

19 10

55

8 15 6 28 7

13

20

16

42

66

5

7

15

5

20

39

12

19 5

26

11

35 24

119 32

6

5

18

12 8

17

14

55

9 19

12

22

53

7

6

5

12

21

16

5 2.1

2.0 1.1

2.0 2.9

2.0 1.5

1.9 2.0

2.3 2.1 2.5 1.4

2.5 2.0 2.3 2.5

1.9

2.0

1.8 1.5 1.8

2.6

1.7 1.4 1.6 2.5

2.1 2.1 2.3 2.0

1.4

2.4 2.0 2.0 2.8

2.0

1.8 2.2 2.4 1.6

1.9 2.0 2.2

1.7 1.2 2.1 1.2

9

9

41

Navicula gregaria Donk. 18

18 9

1.5 1.9 2.2 1.3

9

16

21 20

1.6 2.7

2.1 1.9 2.4 1.9

1.7

2.0 2.0 2.1

71

13

10

13

2.0 2.1

39 22

112 28

26 27 5

13

47

54

10

Navicula lanceolata (Ag.) Ku¨tz. Navicula lateropunctata Wallace Navicula luzonensis Hustedt Navicula menisculus Schum. Navicula minima Grunow Navicula mutica Ku¨tzing Navicula mutica var. stigma Patr. Navicula notha Wallace Navicula paratunkae Peters. Navicula paucivisitata Patr. Navicula pelliculosa Hilse Navicula pupula Ku¨tzing

AWM

pH AWM

SRP AWM

BOD AWM

Cl

6.7 7.2

6.7

6.7 7.6

6.9 7.1

6.7 7.6

7.0 7.8

2.7 7.0 7.8 6.6 7.6

3.0 7.5 7.6 6.8 7.1

6.8

7.6

6.5 7.5 6.3

7.0

1.8 7.0 7.2 6.4 7.4

1.9 7.2 7.5 6.7 7.4

7.3

2.1 7.7 7.7 6.8 7.6

7.0

1.4 7.4 8.1 6.8 7.2

7.5 7.6 7.4

7.0 7.0 6.8 7.6

1.3 7.5 7.8 6.6 7.7

7.8 7.8

2.0 7.0 7.7 6.7 7.3

1.8 6.9

2.0 7.0 7.6 6.7

49 68

37 57

55 60

126 94

90 63

41

92

57

18 68

150

93 56

7.6 183 74

7.5 159 81

7.3

7.1 122 78

136

7.8 127 44

51

7.6 115 64

7.7

7.1

7.3

7.4 131 68

48

16

42

0.9

1.1

0.9 1.2

2.4 3.3 2.3 1.1

1.8

2.1 2.9 1.8 2.1

1.4 3.1 0.7

1.5 2.8 1.4 1.6

2.1 2.9 2.3 1.5

2.7 2.9

196 2.8 3.4 2.0 4.5

231 40

5.5 5.4

1.6 1.3

1.8 0.9

195 4.0 3.1 4.6 2.4

2.9 3.4 1.4 4.7

1.6

1.2

2.8 2.8 2.8

5.4

214 4.0 4.2 3.5 6.4

61 116 84

99 47

61 46

14

304 25

126 29

115 122 154 2.7 3.3 3.0 1.2

156 40

151 75

51

97 42

34

34 53

88 46

119 28

1.4 2.4

100 2.4 2.8 2.2

33 130 104 1.5

159

34 26

20

5

1.8 14

19

22

10 13

10

90 14

35 25

23

33 23

28

65 20

16 21

31 9

42 19

24 19

86 18

1.2 40

2.3

1.7

2.1

1.5

2.5

2.9

1.8 111

3.0

103

46

37

3

9

11 9

67

93

63 506

17 469

40

25

53 251

12

7 240

181 313

10 773

11 26

18

6

8

11

10

12

11

12

8

17

10

4

10

9

9

15

All NE SE WA GC All NE SE WA GC All NE SE WA GC All NE SE WA GC All NE SE WA GC All NE SE WA GC

samples

Number of

16

Navicula confervacea (Ku¨tz.) Grun. Navicula contenta var. biceps (Arnott) V. H. Navicula cryptocephala Ku¨tzing Navicula cryptocephala var. exilis Grunow Navicula cryptocephala var. veneta (Ku¨tz.) Rabh. Navicula decussis Østr.

Taxon

Table 5. (Continued)

44

Navicula tripunctata (O. F. Mu¨ll.) Bory Navicula tripunctata var. schizonemoides (V. H.) Patr. Navicula viridula (Ku¨tz.) Ku¨tz. emend. V. H. Navicula viridula var. linearis Hustedt Nitzschia accommodata Hustedt Nitzschia acicularis (Ku¨tzing) Smith Nitzschia amphibia Grunow Nitzschia bacata Hustedt

Navicula radiosa Ku¨tzing Navicula radiosa var. parva Wallace Navicula radiosa var. tenella (Bre´b. ex Ku¨tz.) Grun. Navicula rhynchocephala Ku¨tzing Navicula rhynchocephala var. germainii (Wallace) Patr. Navicula salinarum var. intermedia (Grun.) Cl. Navicula schroeteri var. escambia Patr. Navicula secreta var. apiculata Patr. Navicula seminulum Grunow Navicula symmetrica Patr.

Navicula pupula var. capitata Skv. & Meyer Navicula pupula var. mutata (Krass.) Hust. Navicula pupula var. rectangularis (Greg.) Grun. Navicula pygmaea Ku¨tzing 14 11

18

23

37 36 80 77 124 38 38 43 84

68 94

14 30 13 8 2.0 1.5 2.3 1.8 1.9 7.2 7.4 6.9 7.3 7.4 115 60 14 29 25 19 1.9 2.3 1.9 1.8 2.0 7.3 7.9 6.8 7.4 7.3 105 52 53 85

12 10 20 35 2.3 2.0 1.9 1.5 2.5 7.6 8.3 7.0 7.6 7.5 127 40

54

10

11 7 10 17 6 9

28 27 21 18 2.0 2.0 2.1 1.6 2.0 7.5 7.5 6.8 7.5 7.6 122 91 10 20 7

68 95

35

84

18

24

41

64

99

47

1.5

7.6

7.2 7.3 6.7 7.8

7.6 8.6 6.8 7.8

7.5 8.0

7.5 8.0

110 43

99 59

57 43

4.1

1.4

3.5 2.2 4.2

2.4 3.2 2.3 1.0

1.7

3.1

3.1

2.2

1.5

5.7

13

20

14

10

3.6 733

2.0 60

9

55

16

3.1 2.8 1.6 3.9 3.7 39

2.5 2.2 3.3 1.4 2.2 284

2.1

1.6

200

88

39

2.6 3.2 2.2 1.6 2.3

1.8 2.6 1.5 2.0 1.6

2.7 4.1 2.9 0.8

2.0 4.1 0.8 1.8

2.6 2.9

7

9

11

15

9 104

7 4

14

8 13

25

18

5

9

7

4

5

133

127

76

777

243 472

127

626

202

1308

67

Continued on p. 46

64 20 10 22

59 25 11 24

40 15

154 34 16

8 17

9 20

9

13 10 17 19 12 14

13

24

16

15

24 10

24 11 13

17 22

113 1.8 2.8 2.0 1.0 1.8 555

1.6 3.6

195 2.8 2.3 2.8 2.7 3.4 23 123 1.9 3.3 2.1 1.1 2.6 68

66

8

2.5 3.2

105 1.7 3.7 1.3 1.1 2.6 30

60

62

1.5

4.3

130 117 101 1.8 2.7 2.5 0.9 2.0

72

49

100 111 174 52

65

66

1.8 1.6 2.1 1.6 2.0 7.3 7.4 6.8 7.7 7.1 69

2.1 2.3 2.0 2.0

1.8 1.9 2.2 1.0

2.4 2.5

1.6 2.1

10 1.5 2.4 2.1 1.0 2.1 7.4 8.3 6.4 7.4 7.8 109 52

22 16 6

11 6

12

10 1.8 1.6 2.2 2.1 2.0 7.3 7.2 6.9 7.6 7.9

34 51

31

54 60

7.7

17 14 6

7.8 7.9

10 1.6 2.0 1.6 1.7 1.5 7.2 8.5 7.0 7.9 7.2

1.6

48

13 6

1.7 1.8

5

24

35

54

54 49 58 70

114 25

20

143 71

41

59

235

50

364

53 85

7.3 8.2 7.0

76

7.3 174

7.9

45

141

24

2.3 2.3 2.5

7.7

7.3 8.0 6.5 7.8

7.3

6.7

6.9

6.5

54

12

1.8 2.0 1.7 2.0

1.8

6.7

3.4 7.6

2.0 7.8

7.0

7.6

70 79

7

21

23 19

2.4

2.1

2.5

125 29 34 30 30 1.8 1.7 2.1 1.2 2.1 7.2 7.7 6.8 7.4 7.7

6

62

1.8

9

17

2.6

2.2

1.6

7

5

2.0

2.1

15

20

9

16

45

12

SRP AWM

1.6 7.0

7.1

7.2 16

17

1.3

1.4 7.4

6.9

7.6 35

6.8 7.1 7.2 136 13

26 30

2.0 7.2

7.1

7.7 37

53

8

18

6

Nitzschia palea (Ku¨tzing) Smith 156 31 40 33 Nitzschia sigma (Ku¨tz.) W. Sm. 35 10

26

21

15

9

BOD AWM

1.2

2.6 0.9

0.7

1.7

136 2.8

72 2.2

1.8

3.0

2.8

2.0

110 2.2 2.4 1.9 1.8 2.5

93 2.9 3.7 2.1 1.9 2.6

9.8 3.2 2.3

96 1.8

65 2.0 3.8 1.3 0.4 1.9

107 1.7 3.2 0.6

96 1.7 3.3 2.0 1.1 1.6

127 1.8 3.5 3.0 1.3 2.1

35 1.9 3.7

64 1.8 3.7 2.6 0.8 1.8

1.5

1.7

3.1 7.2

6.9

7.3 41

119

3.4

494

27 20

492

10

9

12 9

9

3 9 16

7

9 10

8 23

6

9 14

7

3.1 1291

15

91 21 10 41

803

255

57

20 22

66 27

39 23

767

28 29

854 15

303 22 11 19

537 13

632 21

720 28 11 17

79

23 19

10

9 14

16

1540

277

815

266

56

451

1895

84

1746

431

1028

1455

1028

129

133

2940

810

39

297

NE SE WA GC

Cl AWM

0.8 1667

31 2.9

19

1.1

1.9 1.7 2.7

1.3 3.5 1.7 1.3 1.1

20 1.4

5

48 1.1

309 2.3

15 2.8

7.2 134

7.7 76

33

2.7

117 5.6 3.5 3.1 7.4 2.0

2.5 7.2

7.0

0.8

29 2.9 4.3 1.5 0.7 1.5

125 1.0

35 2.5 2.1 2.0 2.7 2.6 7.5 8.1 6.7 7.4 7.6 261 60 66 34

25 2.5

6.9 2.4 7.7

1.2

15 2.5

1.5

2.2 2.1 2.4 2.0 2.0 7.5 8.0 6.8 7.6 7.7 69 46 90 54

30 2.0 2.1 2.2 1.6 2.0 7.3 7.6 6.8 7.7 7.4 99 62 100 48

550 50 110

7.5 225

5

7.6 7.6 6.6

1.4 7.6

36 15 6

2.3 2.0 1.7

1.7

1.8 2.2 2.3 1.1 2.0 7.4 7.1 6.8 7.7 7.7 96 48 44 42

136 29 32 35

30 13 6

6

7.5 69 41 32

51 11 13 10

1.8 7.5 7.2 6.7

15 1.8 1.6 2.0

40 10 8

74

39

16 2.1 2.2 1.3 1.5 2.3 7.2 7.8 6.8 7.6 7.2 82 77 17 33

8.1 7.4 69 60

78 22 18 16

2.0 1.8 7.6 7.7

27 1.6 1.5 1.1 1.5 1.9 7.2 7.1 6.7 7.6 7.3 74 74 7

97 16 18 25

9

11 2.0 2.2

2.0

39 11

1.8

1.6 1.9 2.0 1.6 1.0 7.7 7.8 7.2 7.7 7.7 39 50 63 38

1.4

1.1 1.8 2.5 7.1

37 2.3 2.4 1.8 2.2 2.4 7.3 7.0 6.7 7.7 7.3 65 75 103 51

5

1.1

113 17 33 21

12

7

28

8

60 10 9

25

9

24

Nitzschia dissipata (Ku¨tzing) Grunow Nitzschia dissipata var. media (Hantz.) Grun. Nitzschia filiformis (W. Sm.) V. H. Nitzschia fonticola Grunow

Nitzschia linearis (Ag. ex W. Sm.) W. Sm. Nitzschia lorenziana var. subtilis Grun. in Cl. et Grun. Nitzschia obtusa var. scalpelliformis Grun. in Cl. et Mo¨ller Nitzschia obtusa W. Sm.

pH AWM

2.2 2.4 2.6 1.2 2.0 7.8 8.2 6.9 7.6 7.8 133 43 31 35

27 1.9

5

12 1.5

Nitzschia confinis Hustedt

24 16

41 14 13 5

29

74

Nitzschia frustulum (Ku¨tzing) Grunow Nitzschia frustulum var. perminuta Grunow Nitzschia frustulum var. subsalina Hustedt Nitzschia gracilis Hantz. ex Rabh. Nitzschia hungarica Grunow Nitzschia intermedia Hantz. ex Cl. et Grun. Nitzschia kuetzingiana Hilse

Health

All NE SE WA GC All NE SE WA GC All NE SE WA GC All NE SE WA GC All NE SE WA GC All

Number of samples

Nitzschia clausii Hantz.

Nitzschia brevissima Grun. in V. H. Nitzschia capitellata Hustedt

Taxon

Table 5. (Continued)

46

9 13 7

22

7

27

24

28

52 7

Nitzschia tropica Hustedt

10

Nitzschia tarda Hustedt

12 15

23 6

20

Pinnularia biceps Greg.

Pinnularia microstauron (Ehr.) Cl. Pinnularia subcapitata Greg.

Synedra radians Ku¨tzing Synedra rumpens Ku¨tzing Synedra rumpens var. familiaris (Ku¨tz.) Hust.

5

7

7

11

23

5 10

2.4

2.1

2.2

1.5

2.1

6.9

6.9

6.9 61

71

38 25

24 24

178

64 87 49 19

175

50 58 130

7.1 141 46 138

1.8

1.4 1.2

1.4

1.6 1.9

7.1 7.1

7.0

7.5 7.5

7.0

1.3 1.4 7.5 1.9 6.6 2.6

6.2

6.8

6.9

7.5

7.6 7.7

1.6 1.2 1.7

6.9 7.0 6.9

32

27

19 13

33

170 68

36 87 34

1.9

1.7

1.9

2.5

3.3

4.4

4.1

3.6

0.5 0.4

1.4

1.6 3.2

1.5

2.2

0.4 2.5

1.7 4.6 1.3

9

6

5

10 12

9

9 22

14

8

9 13

4

8 7

11

19 23

12 20

16

19

10

6

28

77

1772

967

239

857

1731

926

490

90

269

544

Continued on p. 48

25 10

18 21 14 54

13

27 190

4

366 19 11 45

8

19 27

55

9 21

60 28 16 5

18 18

14

571 33 13

191 32 13

8 14

244 22 11

231

234

26

0.5 1.4 1378

1.8

2.0

1.5 2.8 1.9 5.4 1.1

2.7

1.2

2.3 2.9 2.6 2.6

3.4

3.4 4.1 1.4 2.8

13 1.7

2.1

116 2.1 3.9 1.2 2.1 1.4

2.4 2.9 2.9 2.7

2.7 4.3

12 52 17

2.1 2.0 1.6 1236

3.6 3.6 3.8 4.3

3.0

0.7 1.5 2.8

311 3.3 4.1 2.4

20 1.8

99 5

17

1.1

34 1.7 3.7 2.1

254 2.6

88 2.1

328 1.1

14 125 1.2

17

54

1.8 363 2.3 2.8 1.3

38 18 1.0 156 1.4 53 84 41

30

27 25

21

134 82 90

29

123 66 40 179

7.4 364

35 42

7.7 7.4 100

7.5

7.5 7.8 7.0 7.5

6.2

1.8 2.1 7.4

2.2

7.9

7.1 7.0 7.2 7.1

2.0 7.9

8.0 8.4

1.8 2.1 1.6 2.6 1.0 7.0 7.4 7.0 7.2 7.7

2.2

1.3 2.0

1.7

2.1 2.2 2.4 2.2

1.8

6

9

2.0

2.5 2.3 2.2 2.8

1.6

14

6

2.1 2.4

13 1.8 2.0 1.4 1.8 2.9 7.7 7.9 7.0 7.9 7.8 43 83 24 39

7.7 7.8 6.9 7.5

6.7

6.7

2.3

2.3

7.2 6.9 6.6

70

7.5 43 38 63

7.4 216

7.8 62

7.0 7.7 7.2 21

3.2 6.7 6.5 6.7

2.7 2.7 1.9

1.8 2.0 1.6 2.0

73 7.1 110 78 139

6.7 7.5 6.9 56

1.8 7.0 8.3 6.9

2.3 7.4

2.0 7.2

2.5 2.0 2.0 7.2

2.4 2.7 1.9

14 2.1 5

7.2 2.4 7.5 8.1 6.6

2.4 1.7 2.3 7.0

16 1.7 2.4 1.8

17 2.0

7

2.5

2.1 2.0 2.2

13 1.9

7

5

6

9

9

8

11

18

22

11

8

8

67 16 21 13

16

Synedra fasciculata (Ag.) Ku¨tz. 31 Synedra goulardi Bre´b. 22 Synedra pulchella Ralfs ex Ku¨tz. 22

27

19

18

32 10 7

Surirella ovata Ku¨tzing Surirella ovata var. pinnata (W. Sm.) Brun Surirella tenera var. nervosa A. S. Synedra acus Ku¨tzing

13

18

Surirella delicatissima Lewis

Rhoicosphenia curvata 49 25 6 (Ku¨tz.) Grun. ex Rabh. Stephanodiscus astraea var. 60 8 11 minutula (Ku¨tz.) Grun. in V. H. Stephanodiscus invisitatus 35 12 Hohn & Hellerm. Stephanodiscus minutus 15 H. L. Sm. Surirella angusta Ku¨tzing 57 21 22

19

33

30 6

Nitzschia tryblionella var. debilis (Arnott) Hust. Nitzschia tryblionella var. levidensis (W. Sm.) Grun. in Cl. et Grun. Nitzschia tryblionella var. victoriae (Grun.) Grun. in Cl. et Mo¨ller Opephora martyi He´rib.

15

15

29 8

Nitzschia subtilis Grunow

47

Health

pH AWM

SRP AWM

BOD AWM

1.9

9 2.4 2.3

1.6 1.6

1.5

1.9 2.0 2.0 1.5 1.4 1.1

8.0 8.4

6.6

6.2

2.4 7.1 7.1

6.0

6.2

2.0 7.1 7.5 6.9 7.7

6.7 7.3 6.8 7.6

7.2

29 35

50

52

72 41

7.0 120 62

61 63

28

48

167

47

3.1 4.0

0.8

2.7 3.2

3.2

2.4 2.1 2.8 2.0

1.7 3.2 1.4 6.4

189 3.7 4.6

41 81

18

1.7

2.0

29 22

6

6

40 13

28 18

16 15

8

7

11 25

10 51 176 136

Values are given for the full data set of 186 samples and for the four major TWINSPAN Groups (NE, SE, WA, GC). The number of samples in which the taxa found in 30 or more samples occurred is highlighted in bold. An electronic version of this table that includes all taxa is available at http://diatom.acnatsci.org/autecology/.

7

14 7

1.7 1.5 1.6 2.5

14

37 21

98 19

6

22 6

48 13

NE SE WA GC

Cl AWM

All NE SE WA GC All NE SE WA GC All NE SE WA GC All NE SE WA GC All NE SE WA GC All

Number of samples

Synedra ulna var. 17 oxyrhynchus (Ku¨tz.) V. H. Tabellaria fenestrata 18 (Lyngb.) Ku¨tz. Tabellaria flocculosa 20 (Roth) Ku¨tz. Thalassiosira fluviatilis 15 Hustedt

Synedra rumpens var. meneghiniana Grunow Synedra ulna (Nitz.) Ehr.

Taxon

Table 5. (Continued)

48

49 the horizontal axis; those collected more recently are located below.

Relative roles of water chemistry and spatial location factors in explaining variation in diatom assemblage distributions The TWINSPAN and DCCA analyses, described above, show that water chemistry and geography (spatial location) can both explain patterns in distributions of algal assemblages. The variance partitioning analysis quantified their importance. It attributed 14% of the variation in diatom assemblage composition to chemistry variables alone, 1% to spatially structured environmental (chemistry) variables, and 11% to spatial (geographical) variability alone. The amount of unexplained variation is high (74%), but is typical for data matrices such as those analyzed in this study that contain a large number of zero occurrences of taxa (Borcard et al., 1992).

Diatom taxa distributions and AWMs for water chemistry characteristics Most of the taxa occurring in 20 or more samples were widely distributed and found in more than 1 of the major TWINSPAN Groups (NE, SE, WA, GC) (Table 5). The 21 most common taxa (in 74 or more samples) occurred in five or more sites in all four major groups. Many of the less common taxa were in only one sample, one river, or within one region. About one half of the taxa (177 of 356) occurred in five or more sites in only one of the four major groups. Many taxa were limited to, or most common in, only one or two of the four main geographic regions (Table 5). These patterns suggest that many of the less common taxa may have local or regional distributions. Regional floras, defined specifically as groups of consistently co-occurring species, were, however, not distinguishable by casual observation. In general, taxa seemed to be distributed independently of each other. The clearest patterns of taxonomic distributions were related to major ion water chemistry characteristics of the intermediate-scale regions. For example, Eunotia and other acidophilic taxa were more common in the SE Group, and taxa

often associated with high hardness and Cl were more common in the GC Group (Table 5). Planktonic taxa occurred more frequently in the larger rivers, and there were geographic patterns in their distributions. We calculated AWM values, both because these values are useful for scientists and managers using diatoms for water quality assessments, and to better understand how these measures of species autecology varied among regions. Values for ‘‘health’’, pH, SRP, BOD and Cl for the 177 taxa that occurred in five or more samples are presented in Table 5. Values are given for data available from all 186 sites and for each of the four main TWINSPAN Groups. We chose to present values for pH, SRP and Cl because they represent major environmental gradients relevant to this study and there are AWM data from other studies with which they can be compared. We present results here for SRP because there were many more measurements of SRP than total P. The BOD values are of interest because there are few diatom calibration data sets with this measurement, and because many European water quality indices often used in the United States are based on response to BOD. AWM values for other chemical constituents (including total P), plus tolerances and other statistical measures are available at http://diatom.acnatsci.org/autecology/. We decided to include a ‘‘health’’ AWM value for taxa to take advantage of the ‘‘health’’ ranking that was made for the study sites in the study reports, and because such values are useful for making environmental assessments using diatoms. The AWM ‘‘health’’ values for each taxon could potentially range from 1(‘‘healthy’’) to 4 (‘‘very polluted’’). The index values are useful, but should be interpreted carefully. They are calculated mostly for larger rivers. The human factors influencing river health varied widely. For all water chemistry parameters, there are substantial differences in AWM values for many taxa among the four regions. To illustrate how one characteristic varies spatially, we show how AWM values for SRP vary among the four main TWINSPAN Groups (NE, SE, WA and GC; Fig. 5). In general, they increase from the Southeast to the Northeast to W of Appalachians to the Gulf Coast. The differences are greatest for the taxa with the highest AWM values for SRP.

50

377 Diatom Taxa

WA - West of Appalachian Mountains (41)

NE - Northeast (34)

GC - Gulf Coast (41)

0

200

400

SE - Southeast (46)

600 0

200

400

600

AWM SRP - μg/L Figure 5. SRP AWM for 377 taxa in the 4 TWINSPAN Groups. All taxa occurred in five or more samples. Order of taxa is the order of SRP AWMs calculated using all 186 samples, and is the same in all four plots.

Discussion Large-scale patterns in diatom assemblage composition At the scale of the eastern United States, and using data from 116 sites on 47 rivers, we defined four major regions based on similarities in diatom assemblage composition. The groups of samples are defined equally well by unique combinations of geochemical parameters (pH, hardness, alkalinity and Cl), and by geographic location, both of which correspond with one another. The physical characteristics of river systems or level of human influence were not as important. Definition of the four main TWINSPAN Groups is robust, in that the same basic groups emerge using different multivariate analyses and various modifications of the procedures, and despite limitations in the consistency of the data. The four minor TWINSPAN Groups all had diatom assemblages

reflecting water chemistry that was different from sites in geographic regions represented by the four major TWINSPAN Groups. The geographic locations of the four major groups (NE, SE, WA, GC) generally correspond to the USGS Water Resource regions (Seaber et al., 1987) and combinations of Omernik’s Level III ecoregions (Omernik, 1987, 1995) (Table 6). Data are insufficient, however, to assess quantitatively how well they agree or whether one scheme matches any better than another. The Omernik Level II tier of ecoregions (Commission for Environmental Cooperation, 1997) does not agree well with the four groups defined here, mainly because the Southeast Plains ecoregion covers substantial areas of all four of the groups. Differences in assemblage composition are consistent with those found in analysis of USGS NAWQA samples collected from throughout the United States (Potapova & Charles, 2002), though possibilities for comparisons are limited because of

51 Table 6. Correspondence between geographic regions defined based on the four main TWINSPAN Groups of diatom assemblages and USGS Water Resource regions and Level II ecoregions TWINSPAN region

USGS Water Resource region

Level II ecoregion

NE

North Atlantic

Northeastern Highlands (58), North Central Appalachians (62), Northern Piedmont (64), Central Appalachian Region Ridges and Valleys (Northern half; 67), Western Allegheny Plateau (70)

SE

South Atlantic Gulf

Middle Atlantic Coastal Plain (63),

WA

Ohio and Arkansas-White

Ozark Highlands (39), the Central Irregular

Southeastern Plains (65) Plains (40), Central Appalachian Region Ridges and Valleys – (Southern half; 67), GC

Texas Gulf and Lower Mississippi

Interior Plateau (71), Mississippi Valley Loess Plains (74) Texas Blackland Prairies (32), the Western Gulf Coastal Plain (34), the Southern Coastal Plain (75), Mississippi Alluvial Plain (73)

Ecoregion numbers are in parentheses following name.

differences in nature and composition of the data sets (e.g., taxonomy). One clear area of agreement, however, is the major distinction between sites along the southeastern coast and in Texas. The general correspondence of spatial groupings of diatom assemblages and ecoregions or water resource regions suggests that boundaries of these regions may serve to delineate areas within which the majority of diatom taxa will exhibit relatively similar ecological characteristics. Any of the schemes for defining regions may be appropriate for developing calibration data sets. The pattern also suggests that development and application of diatom indicators would be more successful if done within each of the four geographic regions than within a larger region, up to and including the scale of the eastern United States. Data from this project are insufficient to assess the level of improvement that could be achieved by working at smaller scales. Influences of water chemistry on diatom assemblages DCCA showed that at the scale of the eastern United States, major-ion chemistry (pH, alkalinity, Cl and hardness) was the most important factor explaining variation in assemblage composition. At the intermediate scale, physical characteristics

(e.g., river size, temperature/elevation) and humaninfluenced water quality gradients were more important (e.g., nutrients, BOD, turbidity). Chemistry constituents important at the large scale are likely to be important in most rivers in the four regions studied. The variables found to be important at the intermediate scale are probably more influenced by the particular set of rivers and conditions represented within those groups (number, location, human influences). These results are generally consistent with the hierarchy of determinants (ultimate, intermediate, and proximate, and associated scales) proposed by Stevenson (1997) and Biggs (1996) for algal species composition. Results are also similar, in terms of relative importance of water chemistry variables, to those of other regional studies, for example the MidAtlantic Highlands streams (Pan et al., 1999; 2000), the Upper Illinois R. basin (Leland & Porter, 2000), the San Joaquin R. (Leland et al., 2001), the Yakima R. basin (Leland, 1995) and the Central Columbia Plateau (Munn et al., 2002). They are difficult to compare in other respects, however, because of differences in study design, geographic scale, size of rivers or streams studied, and environmental data used in the analyses (e.g., inclusion of geology and land-use data). William’s (1964) study of planktonic diatoms in larger rivers throughout the United States and his

52 assessment of the most important physical and water chemistry factors influencing species distributions (streamflow, turbidity, temperature, alkalinity, salinity, impoundments, irrigation) were consistent with our findings. Potapova & Charles (2002) describe three major ecological gradients explaining variation in diatom assemblages at the scale of the conterminous United States: (1) a complex ‘‘downstream’’ gradient from fast-flowing, mostly oligotrophic rivers, to predominantly lowland, more eutrophic rivers, (2) a pH/alkalinity/mineral content gradient, and (3) a latitudinal and elevational variation in temperature. Multivariate analysis of the ANS data set clearly demonstrates the importance of the second two gradients, but the ‘‘downstream’’ gradient is not obvious. This is probably because there are relatively few high-gradient rivers. Most are moderately to gently sloping and at lower elevation. This gradient is apparent in the major TWINSPAN Groups that contain higher elevation, faster-flowing rivers (NE and WA). Relative importance of factors influencing spatial distributions other than water chemistry Two main theories that may explain patterns of diatom distributions are: (1) taxa are adapted to specific environmental conditions, and because diatoms can be widely dispersed, they are likely to exist where those conditions occur, and (2) factors related to historical events, disturbance, chance introductions, and dispersal ability determine distributions and are more important than environmental conditions. To the extent that the first theory is true, distributions of taxa can be predicted based on knowledge of environmental conditions at particular sites. To the extent that distributions are determined by chance and dispersal factors, occurrence of taxa can be predicted based, in part, on knowledge of spatial location. The variance partitioning analysis was performed to assess the relative importance of these two theoretical explanations. Results show that 14% of variation could be explained by water chemistry factors alone, nearly equal to the 11% that could be explained by geographic factors alone. These represent 54 and 42% of the total variance explained (TVE) (26%), respectively. These percentages are comparable to those found in analysis

of a USGS NAWQA dataset of 582 samples (Potapova & Charles, 2002), though the percent explained by environmental parameters is less and the percent for spatial parameters is higher. In that analysis, the TVE was 11.9%, the % TVE due to environmental parameters was 57.7, and the % TVE due to spatial parameters was 28.3. For the 306 samples in the Eastern Temperate Forest ecoregion, which overlaps substantially with ANS sites in this study, the corresponding values are 21.1, 56.7 and 25.9. These results show that water chemistry data are important for understanding some large-scale patterns, but that spatial-dispersal factors are key to understanding others. The underlying causes for the spatial relationships could be chance factors affecting the distribution of taxa or unmeasured ecological factors that vary geographically. They could also be a function of small sample size and some taxa not being well represented. Intermediate-scale regional variation in ecological characteristics of individual taxa Knowing how ecological characteristics of diatom taxa vary geographically is an important aspect of understanding their distribution patterns and using them as environmental indicators. In particular, if measured ecological characteristics are consistent over wide geographic areas, then information developed in one area can be applied in many other areas. If, however, ecological characteristics vary significantly from one area to another, precise ecological data would need to be calculated for each area. We found considerable geographic variation in ecological characteristics of many taxa, but patterns were complex and difficult to quantify and interpret. AWM values are useful measures of ecological characteristics of taxa, and are helpful for understanding how ecological characteristics of individual taxa vary from region to region and at different scales. Diatom taxa AWM values of water chemistry parameters varied among the four intermediate-scale regions (Table 5). Differences were small for some taxa and substantial for others. The extent to which these differences reflect true variation in ecological characteristics is not clear. Some variation could be due to genetic variability, or because of cryptic species and

53 lumping of taxa, or because some extraneous environmental variables restrict the distribution of taxa differently in different areas. A portion is probably due to the small number and diversity of samples in which many taxa occurred. For example, some values are based on fewer than 10 samples, several of which might have been from 1 river. Also, some values may be influenced by high relative abundance values in a few streams. The size and composition of the calibration data sets used to calculate the AWM values for each of the four major groups (NE, SE, WA, GC) are limited, and there is considerable diversity in the type of systems represented and the range of human influences to which they are exposed. Until ecological characteristics of taxa are more clearly quantified and presented in other studies, the AWM values in Table 5 should be used carefully and in conjunction with other sources of information. The ‘‘health’’ AWMs of many taxa are consistent with assignments of taxa in related classification systems (e.g., ‘‘pollution tolerance’’) (Lange-Bertalot, 1979; Van Dam et al., 1994). For example, Nitzschia palea is an indicator of poor health according to these systems, and has a high AWM (2.5); Cymbella species, considered indicative of good health, have low AWM values (1.4–1.8). Taxa that occurred only or primarily in one TWINSPAN Group should only be used for assessments in similar geographic regions. For example, high numbers of Eunotia pectinalis may be a good indicator of poor health in Southeast rivers, but not elsewhere. The AWM of many taxa are similar among regions. The average health values for the sites in each major region were also relatively similar: (NE=2.0, SE=1.9, WA=1.7, GC=2.0; average for all samples=1.3). In many cases these interregional differences in taxa AWM values are less than differences in the corresponding water chemistry values. The 89 taxa with widest occurrence (in 30 to 159 samples) all have ‘‘health’’ AWMs near the mean for all taxa (1.92). Taxa with the highest and lowest AWMs occurred in lower numbers of samples (5–20). The causes and implications of this distribution are unclear. On one hand, it suggests that widespread taxa may be generalists and their occurrence is not influenced by ecosystem ‘‘health’’. On the other

hand, it may mean that these taxa reach their greatest abundance in ‘‘semi-healthy’’ to ‘‘polluted’’ systems because they can be most competitive there, and that these taxa are indeed good indicators of these conditions. Both factors are likely involved. The variation in SRP AWMs (Fig. 5) shows that many taxa at the low range of AWM values had low values in all four intermediate-scale regions, but taxa at the higher AWM ranges had values that differed considerably among the four regions. The AWM values vary as a function of the average values of SRP for all samples in each region, especially at the higher ranges. The same pattern exists for other variables as well (Table 3). This point further emphasizes the influence of calibration data set composition on AWM values. Some taxa were most common in only the highest SRP sites within each of the regions, and therefore were good indicators of high SRP, even though their SRP AWM values for individual regions were very different. Thus, in many cases the relative rank of a taxon’s AWM value might be more useful and reliable for making water quality interpretations than the AWM value per se. To assess how well the AWM values calculated for all sites in the full 186-sample data set agreed with those for the much larger national data set of USGS NAWQA samples (Potapova & Charles, 2003), we compared AWM Cl values. This was one of the few chemistry variables for which there was a large number of values in both data sets. There were 103 taxa common to both data sets that were found in 10 or more samples. Agreement was reasonable for many of the taxa that had an AWM of 10 mg l)1 Cl or less. Above this value, there was rapid divergence, with values based on the ANS river data set being much higher. The range of NAWQA AWM values was 3.2–26.5 mg l)1 Cl. For the ANS river data set the range was 6.1–2029 mg l)1. This difference reflects the much greater proportional representation of high Cl sites in the ANS data set, primarily because of the dominating influence of the high Cl Gulf Coast sites. The correlation between the two sets of AWM values was low (r=0.46), suggesting that even the relative rankings of taxa AWMs along the Cl gradient were not the same. To eliminate the effect of the high Cl ANS sites, we reran the analysis excluding the GC sites so that the ranges

54 of both sets of AWM values were comparable. Correlation of the two sets of AWM was still not significant. The lack of correlation may be because distribution of most diatoms is not appreciably influenced by low values of Cl concentrations. Also, although some taxa with high Cl AWM seemed restricted to high Cl sites, many of the taxa with the highest Cl AWM are widely distributed and would be considered ecological generalists. We compared AWM values for SRP and total P to see how well they agreed both in values and in rank. The correlations for sets of taxa occurring in 5 or more and 10 or more sites, respectively, were 0.34 and 0.38. This low level could be due to several factors, including the low correlation of measured SRP and total P values and the much lower number of sites in the total P data set used to calculate AWMs than the SRP data set. The AWM values presented here are most useful for showing general differences among taxa and to demonstrate potential variability in values calculated for relatively large regions. If used for environmental assessments, they should be used cautiously. For example, AWMs for taxa with the highest values tend to vary the most, geographically. Careful attention should be paid to these. It is clear that the AWM values based on the largest data set are often not appropriate to apply to a particular geographic region, and that the values for the major regions are based on a limited number of samples. Nonetheless, these AWM data should be useful as reference points and to supplement ecological data developed in other studies, especially in terms of how sensitive a taxon’s AWM values might be to geographic location, what environmental variable it might be most responsive to, and how widely it is distributed. Achnanthes minutissima is often used as an indicator of physical disturbance or toxic substances (Barbour et al., 1999). In the ANS data set, A. minutissima was abundant (>30–80%) in several samples, many of which were from disturbed sites. The nature of the disturbance varied considerably, however: below the tailrace of large dams, below industrial outfalls with abnormally high metal concentrations, in habitat highly disturbed by log drives, barge traffic, and other activities. It was also in samples from less disturbed sites, usually in smaller, higher gradient

rivers. Thus, populations of A. minutissima may be good indicators of disturbance in lowland, slowerflowing rivers in the eastern United States, but not necessarily good in terms of diagnosing the type of disturbance. Implications and recommendations for design of diatom-based water quality monitoring programs This study provides information relevant to key questions concerning diatom-based monitoring programs, particularly in terms of developing new ecological data appropriate for sites to be monitored, and applicability of currently existing information: How should a calibration set be designed to provide appropriate autecological data on diatom taxa for making environmental assessments? How large a geographic area is appropriate? Can data from large and small scale data sets both be used? What water chemistry and other environmental characteristics should be measured? Over what range? How applicable are ecological data from other studies? We propose some general guidelines and recommendations based on our results. Calibration sets should be designed to meet the objectives of the ultimate goal for which they are being developed. This means that the calibration samples should contain as many as possible of the taxa occurring in the sites to be assessed and that the quality of the environmental data of most importance in assessments be maximized. These two objectives should determine the geographic size of a calibration data set and the parameters to measure. The geographic size of a calibration study area could be relatively large if all the individual sites within it meet the above criteria; if they do not, it might need to be restricted to a relatively small area. Because natural geochemical characteristics (e.g., pH, alkalinity, hardness, Cl) can strongly influence distributions of taxa and their AWM values, these parameters should be measured and their range minimized, unless they are the primary variable of interest (e.g., pH for acid mine drainage studies). Variation in major physical factors (e.g., river size, substrate, light) should be minimized as well. If these factors are measured, their influence can be taken into account when assessing the influence of human activities to make any disturbance signal clearer. The ranges of

55 variables of most interest (e.g., phosphorus concentration) should be as great as possible, however, so that distributions of taxa along these gradients can be characterized as accurately as possible. Our results suggest that regions larger than the four intermediate-scale regions we defined for this study would be inappropriate for developing calibration sets, but do not suggest how much improvement might result from going to even smaller regions. Use of Level III ecoregions or USGS Water Resource regions for defining boundaries of calibration-set regions would be consistent with our results, as long as the criteria discussed above are met. Based on the DCCA results (Figs. 3 and 4), major-ion chemistry is more important over larger scales, and factors related to human influences on water quality (SRP, BOD/DO) become more important at intermediate scales. This suggests that the spatial scale for calibration sets used to develop indicators of basic geochemical characteristics (e.g., pH, hardness, chloride) could be larger than indicators for factors significantly influenced by human activities (e.g., phosphorus). It also suggests that ecological data for major-ion chemistry from large calibration sets might be more reliable than data on human-influenced characteristics. Clearly, calibration sets should take into account the local environmental context, particularly regarding interaction of natural characteristics and human influences. Another consideration is that these results and implications may apply more to larger rivers, of the size range included in this study, and less to calibration sets of smaller-size streams. When using or developing ecological data for particular parameters, it is important to be aware of the relative roles of natural vs. human influences on water bodies in the calibration set. For example, diatom taxa may respond differently to high P concentration in an area with naturally high values as compared to areas where the P is high due to discharges from wastewater treatment plants or industrial facilities. Errors in interpretation might result, therefore, if P AWM values from calibration sets with a significant number of samples from naturally high P sites were used in assessments of the effects of P

loading for sites in regions with naturally low P values. Acknowledgements This research was funded primarily by the US Environmental Protection Agency through an Exploratory Research Grant (R821676), and also by the Patrick Center’s Endowment for Innovative Research. We thank Marina Potapova for providing many useful comments on the data analysis and the manuscript. Peggy Burkman entered diatom counts into computer files, Robin Davis helped with technical editing, and Jamie Carr prepared the map.

References Barbour, M. T., J. Gerritsen, B. D. Snyder & J. B. Stribling, 1999. Rapid bioassessment protocols for use in streams and wadeable rivers: periphyton, benthic macroinvertebrates, and fish. 2nd edn., 841-B-99–002, U.S. EPA, Office of Water, Washington, DC. Biggs, B. J. F., 1996. Pattern of benthic algae of streams. In Stevenson, J. R., M. L. Bothwell & R. L. Lowe (eds), Algal Ecology: Freshwater Benthic Ecosystems. Academic Press, San Diego: 31–56. Borcard, D., P. Legendre & P. Drapeau, 1992. Partialling out the spatial component of ecological variation. Ecology 73: 1045–1055. Charles, D. F., 1996. Use of algae for monitoring rivers in the United States: some examples. In Whitton, B. A. & E. Rott (eds), Use of Algae for Monitoring Rivers II. Institut fu¨r Botanik, Universita¨t Innsbruk, Innsbruck, Austria: 109–118. Cholnoky, B. J., 1968. Die O¨kologie der Diatomeen in Binnengewa¨ssern. J. Cramer, Lehre, Germany. Commission for Environmental Cooperation, 1997. Ecological Regions of North America: Toward a Common Perspective. Commission for Environmental Cooperation, Montreal, Quebec, Canada, 71 pp. Map (scale 1:12,500,000). Descy, J. P., 1979. A new approach to water quality estimation using diatoms. Nova Hedwigia 64: 305–323. Gurtz, M. E., 1994. Design of biological components of the National Water-Quality Assessment (NAWQA) Program. In Loeb, S. L. & A. Spacie (eds), Biological Monitoring of Aquatic Systems. CRC Press, Boca Raton, FL: 323–354. Hill, M .O., 1979. TWINSPAN – A FORTRAN Program for Arranging Multivariate Data in an Ordered Two-way Table by Classification of the Individuals and Attributes. Cornell University, Ithaca, 90 pp. Hohn, M. H., 1961. Determining the pattern of the diatom flora. Journal Water Pollution Control Federation 33: 48–53.

56 Lange-Bertalot, H., 1979. Pollution tolerance of diatoms as a criterion for water quality estimation. Nova Hedwigia 64: 285–304. Leland, H. V., 1995. Distribution of phytobenthos in the Yakima River basin, Washington, in relation to geology, land use and other environmental factors. Canadian Journal of Fisheries and Aquatic Sciences 52: 1108–1129. Leland, H. V., L. R. Brown & D. K. Mueller, 2001. Distribution of algae in the San Joaquin River, California, in relation to nutrient supply, salinity, and other environmental factors. Freshwater Biology 46: 1139–1167. Leland, H. V. & S. Porter, 2000. Distribution of benthic algae in the Upper Illinois River basin in relation to geology and land use. Freshwater Biology 44: 279–301. Lepsˇ , J. & P. Sˇmilauer, 2003. Multivariate Analysis of Ecological Data Using CANOCO. Cambridge University Press, Cambridge, 269 pp. Line, J. M., C. J. F. Ter Braak & H. J. B. Birks, 1994. WACALIB version 3.3 – a computer program to reconstruct environmental variables from fossil assemblages by weighted averaging and to derive sample-specific errors of prediction. Journal of Paleolimnology 10: 147–152. Lowe, R. L., 1974. Environmental requirements and pollution tolerance of freshwater diatoms. EPA/670/4–74/005, U.S. EPA, Cincinnati, OH. 334 pp. Lowe, R. L. & Y. Pan, 1996. Benthic algal communities as biological monitors. In Stevenson, R. J., M. L. Bothwell & R. L. Lowe (eds), Algal Ecology: Freshwater Benthic Ecosystems. Academic Press, San Diego: 705–739. Munn, M. D., R. W. Black & S. J. Gruber, 2002. Response of benthic algae to environmental gradients in an agriculturally dominated landscape. Journal of the North American Benthological Society 21: 221–237. Omernik, J. M., 1987. Ecoregions of the Conterminous United States. Map (scale 1:7,500,000). Annals of the Association of American Geographers 77: 118–125. Omernik, J. M., 1995. Ecoregions: a spatial framework for environmental management. In Davis, W. S. & T. P. Simon (eds), Biological Assessment and Criteria: Tools for Water Resource Planning and Decision Making. Lewis Publishers, Boca Raton FL: 49–62. Pan, Y. D., R. J. Stevenson, B. H. Hill, P. R. Kaufmann & A. T. Herlihy, 1999. Spatial patterns and ecological determinants of benthic algal assemblages in Mid-Atlantic streams, USA. Journal of Phycology 35: 460–468. Pan, Y. D., R. J. Stevenson, B. H. Hill & A. T. Herlihy, 2000. Ecoregions and benthic diatom assemblages in Mid-Atlantic Highlands streams, USA. Journal of the North American Benthological Society 19: 518–540. Pan, Y., R. J. Stevenson, B. H. Hill, A. T. Herlihy & G. B. Collins, 1996. Using diatoms as indicators of ecological conditions in lotic systems: a regional assessment. Journal of the North American Benthological Society 15: 481–495. Patrick, R., 1950. Biological measure of stream conditions. Sewage and Industrial Wastes 22: 926–938. Patrick, R., 1951. A proposed biological measure of stream conditions. Internationale Vereinigung Fu¨r Theoretische und Angewandte Limnologie,Verhandlungen 9: 299–307.

Patrick, R., 1973. Use of algae, especially diatoms, in the assessment of water quality. In Biological Methods for the Assessment of Water Quality, ASTM STP 528: 76–95. Patrick, R., 1994. Rivers of the United States. Vol. 1. Estuaries. John Wiley & Sons, Inc, New York, 825 pp. Patrick, R., 1995. Rivers of the United States. Vol. 2. Chemical and Physical Characteristics. John Wiley & Sons, Inc, New York, 237 pp. Patrick, R., 1996. Rivers of the United States. Vol. 3. The Eastern and Southeastern States. John Wiley & Sons, Inc, New York, 829 pp. Patrick, R., 1998a. Rivers of the United States. Vol. 4. Part A: The Mississippi River and Tributaries North of St. Louis. John Wiley & Sons, Inc, New York, 408 pp. Patrick, R., 1998b. Rivers of the United States. Vol. 4. Part B: The Mississippi Tributaries South of St. Louis. John Wiley & Sons, Inc, New York, 488 pp. Patrick, R., 2000. Rivers of the United States. Vol. 5. Part A: The Colorado River. John Wiley & Sons, Inc, New York, 264 pp. Patrick, R., 2003. Rivers of the United States. Vol. 5. Part B: The Gulf of Mexico. John Wiley & Sons, Inc, New York, 272 pp. Potapova, M. & D. F. Charles, 2002. Benthic diatoms in USA rivers: distributions along spatial and environmental gradients. Journal of Biogeography 29: 167–187. Potapova, M. & D. F. Charles, 2003. Distribution of benthic diatoms in U.S. rivers in relation to conductivity and ionic composition. Freshwater Biology 48: 1311–1328. Data available at http://diatom.acnatsci.org/autecology/. Prygiel, J., B. A. Whitton & J. Bukowska (eds), 1999. Use of Algae for Monitoring Rivers III. Jean Prygiel, Agence de l’Eau Artois-Picardie, Douai, France, 271 pp. Seaber, P. R., F. P. Kapinos & G. L. Knapp, 1987. Hydrologic Unit Maps: U.S. Geological Survey Water-Supply Paper 2294. 63 pp. (http://water.usgs.gov/GIS/huc.html). Stevenson, R. J., 1997. Scale dependent determinants and consequences of benthic algal heterogeneity. Journal of the North American Benthological Society 16: 248–262. Stevenson, R. J. & Y. Pan, 1999. Assessing environmental conditions in rivers and streams with diatoms. In E. F. Stoermer & J. P. Smol (eds), The Diatoms: Applications for the Environmental and Earth Sciences. Cambridge University Press, Cambridge: 11–40. ter Braak, C. J. F. & P. F. M. Verdonschot, 1995. Canonical correspondence analysis and related multivariate methods in aquatic ecology. Aquatic Sciences 57: 255–289. ter Braak, C. J. F., 1986. Canonical correspondence analysis: a new eigenvector technique for multivariate direct gradient analysis. Ecology 67: 1167–1179. ter Braak, C. J. F. & P. Sˇmilauer, 1998. CANOCO Reference Manual and User’s Guide to CANOCO for Windows: Software for Canonical Community Ordination (version 4). Microcomputer Power, Ithaca. van Dam, H., A. Mertens & J. Sinkeldam, 1994. A coded checklist and ecological indicator values of freshwater diatoms from the Netherlands. Netherlands Journal of Aquatic Ecology 28: 117–133.

57 Whitton, B. A. & G. Friedrich, (eds), 1991. Use of Algae for Monitoring Rivers. Du¨sseldorf, Germany, Dr. Eugen Rott, Institut fu¨r Botanik AG Hydrobotanik, Universita¨t Innsbruck, Sternwaartestrabe 15 A-6020 Innsbruck Austria, 193 pp. Whitton, B. A. & E. Rott, (eds), 1996. Use of Algae for Monitoring Rivers II. Institut fu¨r Botanik, Universita¨t Innsbruk, Innsbruck, Austria.

Whitton, B. A. & M. G. Kelly, 1995. Use of algae and other plants for monitoring rivers. Australian Journal of Ecology 20: 45–56. Williams, L. G., 1964. Possible relationships between planktondiatom species numbers and water-quality estimates. Ecology 45: 809–823.

Hydrobiologia (2006) 561:59–69  Springer 2006 R.J. Stevenson, Y. Pan, J.P. Kociolek & J.C. Kingston (eds), Advances in Algal Biology: A Commemoration of the Work of Rex Lowe DOI 10.1007/s10750-005-1604-4

Short-term effects of elevated velocity and sediment abrasion on benthic algal communities Steven N. Francoeur1,* & Barry J. F. Biggs2 1

Department of Biology, Eastern Michigan University, 316 Mark Jefferson, Ypsilanti, MI 48197, USA National Institute of Water and Atmospheric Research, P.O. Box 8602, Christchurch, New Zealand (*Author for correspondence: E-mail: [email protected]) 2

Key words: periphyton, disturbance, flood, suspended sediment, biomechanics, biomass removal, scour

Abstract Of the mechanisms that remove benthic algae during flood disturbances, relatively little is known about the effects of sediment scour. We investigated suspended sediment scour using naturally colonized benthic algal communities exposed to realistic velocities and suspended sediment concentrations in a laboratory flowtank. Increased velocity alone removed benthic algal biomass, and high suspended sediment concentrations further increased algal removal. Efficacy of biomass removal by velocity and suspended sediments was community-specific; communities with a tightly adherent cohesive mat physiognomy were resistant to removal, despite taxonomic similarity to easily disturbed communities. In addition, some taxa were more susceptible to removal by disturbance than others. The duration of scour and physical refugia on the substratum also influenced algal biomass removal. Our results indicate that suspended sediment scour may be an important mechanism for algal removal during flood events, and some variability in biomass removal among flood events may be the result of differences in suspended sediment load.

Introduction Flood disturbance can greatly reduce lotic benthic algal biomass and alter community composition (e.g., Grimm & Fisher, 1989; Uehlinger, 1991; Biggs, 1996; Peterson, 1996; Francoeur et al., 1998). In certain streams, flood-induced regulation of benthic algal biomass can be the dominant organising factor, affecting the distribution and abundance of algae and consequently higher trophic levels (e.g., Resh, et al., 1988; Poff & Ward, 1989; Fisher & Grimm, 1991; Biggs, 1995). Removal of benthic algal biomass by flood disturbance can occur by three mechanisms: (1) direct removal of algae by the elevated shear stress caused by increased water velocity, (2) abrasion of algae from substrata by mobilized sediments, and (3) molar action of tumbling gravel/cobble substrata upon which algae grow. These 3 mech-

anisms are interrelated because both sediment suspension and mobilization of large substrata are dependent on increased water velocity. Because streams differ in fine sediment supply and bed sediment stability, the intensity of suspended sediment abrasion and molar action will vary among streams, even for flood events of equal discharge and velocity (see Duncan & Biggs, 1998; Biggs et al., 1999; Biggs et al., 2001). Recent research and current theory suggest that velocity increases alone may not be sufficient to explain the degree of benthic algal biomass removal during floods (Horner et al., 1990; Uehlinger, 1991; Biggs, 1996). In laboratory experiments, increased water velocity (up to 1.5 m s)1) by itself could not remove a tightly adherent, basal layer of algae (Biggs & Thomsen, 1995). The role of molar action (e.g., gravel/cobble movement) in determining lotic algal biomass and community composition has

60 also received theoretical consideration (Biggs et al., 2001), and experimental investigation has shown that algae can be removed by the molar action of tumbling cobbles (Power & Stewart, 1987). Suspended sediment scour during floods can drastically alter benthic algal communities (Uehlinger, 1991; Biggs, 1996; Jowett & Biggs, 1997). Grimm & Fisher (1989) hypothesized that the lack of difference in the amount of chlorophyll a removed by floods between cobble-boulder riffles and adjacent fine gravel/sand runs was the result of suspended sediment abrasion. Additionally, low algal biomass on faces of boulders exposed to flow has been attributed to suspended sediment scour (Blinn & Cole, 1991). In contrast, areas of relatively high biomass may persist on large substrata after flooding (e.g., Uehlinger, 1991; Francoeur et al., 1998), perhaps in part because they are elevated above the zone of sediment scour. Few efforts have been made to test the assumption that suspended sediment abrasion can remove benthic algal biomass from substrata or to quantify the magnitude of suspended sediment scouring. Relatively small increases in total inorganic suspended sediments (from 6 to 25 mg l)1) caused short-lived (<1 day) increases in benthic algal loss rates and concomitant reductions of biomass in laboratory streams, but did not appear to alter algal community composition (Horner et al., 1990). Heinlein (2000) reported that suspended sediment scour (up to 10,000 mg l)1) was less influential than water velocity (1.8 m s)1) in removal of benthic algae during simulated flood events. A combination of increased velocity and an unspecified amount of suspended sediments caused physical injury to benthic diatoms (Blenkinsopp & Lock, 1994). The objective of this study was to investigate the importance of suspended sediment abrasion as a mechanism for removal of lotic algae at

velocities and suspended sediment loads typical for rivers in flood. Additional objectives were to document how the duration of suspended sediment scour affected the removal of algae, and investigate possible variability is resistance to suspended sediment scour by algal communities of different physiognomies.

Methods Design of three experiments Molded plastic sheets (0.4 m  0.56 m) consisting of 119 3.8-cm diameter hemispheres with shallowly textured surfaces designed to mimic rock surfaces in a cobble-bed river (Biggs et al., 1998a; Biggs & Thomsen, 1995) were bonded to Plexiglas baseplates and incubated in an unshaded, shallow (30– 41 cm), moderate velocity (0.42 m s)1 ± 0.06 SD), high nutrient (21.7 lg l)1 DRP, 5790 lg l)1 NO3, 8 lg l)1 NH4) run in the Cust River (Canterbury, New Zealand; see Biggs et al., 1998a; Biggs & Thompson, 1995, for more site information). Plastic substrata were anchored to concrete blocks and laid in 4 columns across the channel. The upstream end of each substratum was marked, ensuring that a uniform direction of flow would be maintained during natural colonization and laboratory experiments. Substrata were deployed on 10 March, 2002. No spates occurred during the incubation period. To reduce biomass and alter community physiognomy, some substrata were periodically brushed in-situ with a paintbrush. Tables 1 and 2 summarize the community generation process and the state of algal communities immediately prior to experimentation. Experiment #1 was designed to assess effects of suspended sediment and community type.

Table 1. Summary of benthic algal communities used in experiments. Biomass is given as mean±1 SD Experiment #

Retrieved

Initial biomass (mg chl a m)2)

Deployed

Brushed

Loose

10 March

–

18–19 March

61±12

Tight

10 March

20, 31 March

3–4 April

70±14

2

10 March

20 March

26–27 March

3

10 March

20 March

25 March

1

110±19 73±8

61 Table 2. Initial taxonomic composition of algal communities Experiment #

Taxon Rank Category

Taxa (mean rank)

Dominant (7–8)

Melosira varians (8.0)

Sub-dominant (6–7) Abundant (5–6)

Navicula avenacea (6.9) –

1 Loose

Tight

2

3

Common (4–5)

Cosmarium spp. (4.6), N. cf. gregaria (4.6), Synedra ulna (4.5)

Dominant (7–8)

M. varians (7.8)

Sub-dominant (6–7)

N. avenacea (6.1)

Abundant (5–6)

cf. Lyngbya sp. (5.6)

Common (4–5)

Ulothrix spp. (4.8), N. cf. gregaria (4.3)

Dominant (7–8)

Ulothrix spp. (7.8), M. varians (7.2)

Sub-dominant (6–7) Abundant (5–6)

– N. avenacea (5.8)

Common (4–5)

cf. Lyngbya sp. (5.0), N. cf. gregaria (4.8), S. ulna (4.7)

Dominant (7–8)

Ulothrix spp. (8.0)

Sub-dominant (6–7)

N. avenacea (6.5)

Abundant (5–6)

–

Common (4–5)

N. cf. gregaria (4.5), S. ulna (4.5)

Following incubation in the Cust River, substrata were transferred to the laboratory (30 min) in covered plastic boxes with a small amount of water at the bottom to keep that atmosphere humid. Substrata were not submerged during transport because of potential removal of benthic algae by water motion. Upon arrival at the laboratory, substrata were submerged in a large tub of untreated, high-quality groundwater from the local aquifer. All substrata were used within 8 h of collection. Immediately prior to experimentation, algae were sampled from 4 randomly-chosen hemispheres and surrounding flat areas from each substratum by scraping with a scalpel and toothbrush. The material from each substratum was composited, homogenized with a blender (Biggs, 1987; Biggs & Kilroy, 2000), and split for chlorophyll a and taxonomic analysis. Manipulations of velocity and suspended sediments were conducted using a modified version of the large flow tank initially described by Biggs & Thomsen (1995). Briefly, the flow tank was a fully enclosed Plexiglas structure (1.5 m  1.4 m  0.4 m) consisting of an upper and a lower test area, in which water is recirculated in a closed vertical loop (loop volume=162 l). A 3 hp (2.2 kW) motor and twin 0.17 m diameter propellers circulated the water. Water velocities in the

flow tank were measured at the height of the substratum hemispheres ( 4 cm from base of substrata), which approximate free-stream velocities. More detailed information, including finescale velocity profiles, can be found in Biggs & Thomsen (1995). Individual substrata were placed into the upper test area. A premeasured amount (0, 926, 2778, or 6481 mg dry wt l)1) of dried, commerciallyavailable New Zealand river sand (cumulative retention in sieve mesh: 3000 lm 0%, 699 lm 8.5%, 500 lm 15.8%, 250 lm 55.7%, 180 lm 77.2%, 63 lm 99.4%) was placed in the lower portion of the loop, and the tank was filled with untreated, highquality groundwater. This orientation of substratum and sand ensured that sediments contacting the substratum were truly suspended, and not acting as bedload. Each substratum was subjected to a stepped velocity treatment. Water velocity was increased through a series of 5 increments: 0.5, 0.8, 1.1, 1.4, and 1.8 m s)1. Each velocity increment was maintained for 5 min. After the final velocity increment, substrata were removed from the flow tank and attached algae were sampled as previously described. At the higher velocities (‡ 1.4 m s)1), all sediments were in suspension, but some deposition of sand was observed behind the downstream faces of hemispheres. Sediment

62 and organic matter were cleaned out of the flow tank between runs. Each combination of suspended sediment load and community type (brushed, tightly attached communities and unbrushed, loosely attached communities) was run in duplicate. Experiment #2 was designed to assess effects of suspended sediment and scour duration. Substrata were retrieved, transported, and initially sampled as previously described. Individual substrata were then placed in the flow tank with a premeasured amount of sediment (0, 926, or 6481 mg dry wt l)1), and subjected to a velocity of 1.8 m s)1 for 60 min. Triplicate 60-ml samples for suspended chlorophyll analysis were removed from the water column using a syringe attached to a sampling tube that protruded into the water column after 5, 15, 30, 45, and 60 min. The water column was mixed by the propellers, ensuring representative samples (Biggs & Thomsen, 1995). At the end of the velocity treatment, substrata were removed from the flow tank and attached algae were sampled as previously described. The flow tank was cleaned of sediment and organic matter between runs. Each suspended sediment treatment was run in duplicate. Experiment #3 was designed to assess spatial distribution of scour. Substrata were retrieved and transported as previously described. Predisturbance sampling consisted of scraping and brushing the front (the upstream side of the hemispheres) and rear (the downstream side) faces of 4 randomly chosen hemispheres from each substratum by with a scalpel and toothbrush. Individual substrata were then placed in the flow tank with a premeasured amount of sediment (926 or 6481) mg dry wt l)1, and subjected to a velocity of 1.4 m s)1 for 60 min. At the end of the velocity treatment, substrata were removed from the flow tank and attached algae were again sampled from the front and rear faces of 4 randomly-chosen hemispheres. Sample analysis Chlorophyll samples were immediately filtered (Whatman GFC) and stored frozen in darkness until analysis. Chlorophyll a was measured spectrophotometrically following hot EtOH extraction (Biggs, 1987; Biggs & Kilroy, 2000). Phaeopigments were corrected for by acidification. Samples

for microscopy were preserved with glutaraldehyde (5% final concentration) and stored refrigerated in darkness until examination at a magnification of 400 (or 780, if necessary). Taxa were identified to the lowest practical taxonomic level, usually to species (25 taxa), but often only to genus (20 taxa). Algal community composition was assessed by defining the taxon with the greatest biovolume in each sample (a visual integration of size and frequency), than rating all other taxa in that sample on an 8-point linear scale (8=dominant, 1=very rare) in relation to the dominant taxon (e.g., Biggs et al., 2000; Biggs & Kilroy, 2000; Biggs & Smith, 2002). Statistical analysis Percent removal of benthic algal biomass was calculated from pre-disturbance and postdisturbance measurements of attached chlorophyll a on each substratum for Experiments #1 and 2. Percent biomass removal in Experiment 3 was calculated separately for front and back faces of the hemispheres. Percent biomass removal at 5, 15, 30, 45, and 60 min of scour duration (Experiment #2) was calculated from the pre-disturbance measurements of attached chlorophyll a and the total amount of suspended chlorophyll a at each time interval. In Experiment #2, percent biomass removal values calculated from attached chlorophyll a were somewhat less than the corresponding values calculated from suspended chlorophyll a (60 min duration), but these variables were strongly correlated (r=0.75, n=6). The effect of suspended sediment concentration on percent biomass removal calculated from attached chlorophyll a (Experiments #1 and 2) was assessed using linear regression. Repeated measures analysis of variance was used to assess the effects of scour duration and suspended sediment concentration on percent biomass removal calculated from suspended chlorophyll a (Experiment #2). Algal community composition in Experiment 1 was summarized by conducting principal component analysis (PCA) of the 1–8 relative biovolume rankings of all taxa. Preliminary analyses indicated that linear multivariate models (e.g., PCA) would be more appropriate for these data than unimodal models (e.g., detrended correspondence analysis) because of the short gradient lengths ( £ 1.5 SD

63 units) recovered by unimodal models (see ter Braak, 1995; ter Braak & Smilauer, 1998). Principal components were computed from sample correlations (i.e., centered and standardized PCA), which resulted in an equal weighting of all taxa. Sample scores and the 12 taxon vectors that explained the greatest amount of variance in PCA axes 1 and 2 were combined as a biplot. All ordinations were done using CANOCO 4.0. MANOVA was used to test the hypotheses that algal communities of differing types (brushed, tightly attached vs. unbrushed, loosely attached) or of differing experimental treatment (before vs. after disturbance) occupied consistently different regions of the 2-D ordination space (i.e., displayed consistent differences in algal community composition).

Results Effects of suspended sediment and community type Suspended sediments increased the removal of algal biomass in Experiment #1, relative to removal by sediment-free water at the same velocity (Fig. 1). The ability of suspended sediments and shear stress to remove algal biomass varied with respect to community type. Biomass losses of the unbrushed, loosely attached communities were substantial in sediment-free water (40% of predisturbance biomass), and removal increased

Chlorophyll removed (%)

90 80 70 60 50

strongly with increasing suspended sediment concentration (r2=0.65, p=0.003). In contrast, removal of tightly adherent communities was low (25%), and was not related to suspended sediment concentration (r2=0.09, p=0.4). Forty-five algal taxa were identified in Experiment #1 communities. Although the loosely and tightly attached communities were dominated by the same taxa (Table 2), the initial community composition differed between these communities. Loosely and tightly attached communities occupied different regions of 2-D ordination space (MANOVA, p<0.001). In particular, Melosira varians, Navicula capitoradiata, and Cosmarium spp. were relatively more dominant in loosely attached communities, whereas Ulothrix spp., Stigeoclonium sp., Gloeocystis spp., cf. Lyngbya sp. and Gomphoneis minuta var. cassieae were more dominant in tightly attached communities (Fig. 2). Disturbance shifted both the loosely (MANOVA, p<0.001) and tightly attached (MANOVA, p<0.001) communities to different regions of 2-D ordination space, by reducing the relative biovolume of Ankistrodesmus spp., Coelastrum spp., Cosmarium spp., Staurastrum spp., and Eunotia spp. and increasing the relative biovolume of Ulothrix spp., Stigeoclonium sp., Gloeocystis spp., cf. Lyngbya sp. and Gomphoneis minuta var. cassieae (Fig. 2). The level of suspended sediments during the disturbance (0, 926, 2778, or 6487 mg L)1) had no discernable effect on algal community composition (data not shown). Experiment #2 also showed that suspended sediment scour could increase removal of benthic algal biomass, regardless of whether biomass losses were quantified using attached (Fig. 3, r2=0.62, p=0.04) or suspended (Fig. 4, repeated measures ANOVA psediment=0.013) chlorophyll a.

40

Effect of scour duration

30 20 10 0

0 1000

3000

5000

7000

Suspended sediments (mg/l) Figure 1. Proportion (%) of biomass removed by experimental disturbance as a function of suspended sediment loads during Experiment #1. Circles are loosely-attached communities, squares are tightly-attached communities. Best-fit lines (leastsquares linear regression) are given for each community type.

Biomass removal rates were not constant over the 60 min exposure. Most biomass losses (83–100% of maximum observed removal) occurred within the first 5 min of disturbance (Fig. 4). Over the period from 5 to 60 min, biomass removal slightly increased with disturbance duration (repeated measures ANOVA ptime = 0.03), apparently reaching an asymptote after 30 min (Fig. 4). The temporal pattern of biomass removal did not vary

64

Chlorophyll Chlorophyllremoved removed(%) (%)

80 60

50 70 40 60 30 50 20 40 10 30 00

10

20 30 926

40

50 60 6481

70

Disturbancesediments duration (min) Suspended (mg/l)

Figure 2. PCA biplot of sample scores and taxon vectors for pre- and post-disturbance communities from Experiment #1. Open circles are pre-disturbance loosely attached communities, dark circles are post-disturbance loosely attached communities, open squares are pre-disturbance tightly-attached communities, dark squares are post-disturbance tightly-attached communities. The first two PCA axes captured 19.9 and 9.0 % of the variance in taxa data, respectively. Taxon abbreviations are as follows: Ankistro = Ankistrodesmus spp., cf. Lyng = cf. Lyngbya sp., Coelastr = Coelastrum spp., Cosmariu = Cosmarium spp., Eunotia = Eunotia spp., Gloeocys = Gloeocystis spp., Gomphone = Gomphoneis minuta var. cassieae, Melosira = Melosira varians, Navcapit = Navicula capitoradiata, Staurast = Staurastrum spp., Stigeocl = Stigeoclonium sp., and Ulothrix = Ulothrix spp.

with respect to sediment treatment (repeated measures ANOVA psediment*time = 0.7). Spatial distribution of scour The front faces of hemispheres lost a greater percentage of their pre-disturbance biomass than the rear faces, in both tested sediment concentrations (Fig. 5). Deposition of algal cells on the rear faces of hemispheres during disturbance may have contributed to the post-disturbance biomass measurements, but this is likely to have been only a

small bias, as much the algal biomass on the rear faces of hemispheres consisted of filaments attached to these areas. Although quantified on only two substrata, this pattern was visually obvious on other substrata, particularly those exposed to suspended sediment scour.

Discussion Suspended sediment scour increased removal of benthic algal biomass above that resulting from

65 60

60

Chlorophyll removed (%)

Chlorophyll removed (%)

70

50 40 30 20 10 0

0 1000

3000

5000

7000

Suspended sediments (mg/l) Figure 3. Proportion (%) of biomass removed by experimental disturbance as a function of suspended sediment loads during Experiment #2. Best-fit line calculated by least-squares linear regression.

water velocity alone. The increased velocity used in our experiments removed 20–40% of initial biomass and an additional 0–40% of the initial biomass was removed by sediment scour. Clearly, suspended sediment scour can be an important biomass removal mechanism within a realistic range of water velocities and suspended sediment concentrations. Because floods of similar magnitude within the same river may have widely varying suspended sediment loads (e.g., Hicks & Griffiths, 1992; Hicks et al., 2000), variability in suspended sediment load (presumably resulting in

Chlorophyll removed (%)

80 70 60 50 40 30

0

10

20

30

40

50

60

70

Disturbance duration (min) Figure 4. Proportion (%) of biomass removed by experimental disturbance as a function of suspended sediment load and time during Experiment #2. Note that biomass removal was 0% at time 0. Smoothing lines generated by LOWESS (tension = 1.0). ¤ = 0 mg l)1,  = 926 mg l)1, + = 6481 mg l)1.

50 40 30 20 10 0

926

6481

Suspended sediments (mg/l) Figure 5. Proportion (%) of biomass removed by experimental disturbance as a function of position on the substratum during Experiment #3. Upstream faces of hemispheres = dark bars, downstream faces of hemispheres = open bars.

variable scour intensity) may be another reason why the biological effects of flood disturbance are not perfectly correlated to flood discharge (e.g., Biggs & Close, 1989; Peterson, 1996). Communities with a relatively tightly adherent physiognomy were highly resistant to both velocity-based removal and suspended sediment scour. This suggests that the factors which confer resistance to velocity disturbance (e.g., Peterson, 1996; Biggs et al., 1998b) could also reduce susceptibility to suspended sediment scour. In a similar study, Heinlein (2000) found that high biomass (75 mg chlorophyll a m)2) communities were more resistant to suspended sediment scour than low biomass (22 mg chlorophyll a m)2) communities of similar taxonomic composition. Biomassinduced variation in resistance is unlikely to account for the observed differences in tightly adherent and loosely attached communities, as these communities had similar initial biomass (Table 1). The large community-specific differences in resistance observed in Experiment #1 were not unexpected, given the differing pre-adaptation to disturbance in the brushed and unbrushed communities. However, the existence of a large difference in resistance between communities with similar initial biomass and dominant taxa (Tables 1 & 2) was surprising. Large communityspecific differences in resistance have been reported in the literature, but only between benthic algal communities that differed greatly in composition

66 and physiognomy (i.e., communities dominated by filamentous green algae and filamentous diatoms vs. communities dominated by adnate diatoms) (Biggs & Thomsen, 1995). There are several potential explanations for this discrepancy. First, although both loosely and tightly adherent communities were dominated by Melosira varians and Navicula avenacea (Table 1), the communities did differ with respect to several other less-abundant taxa (Fig. 2; Table 1). Benthic algal taxa vary markedly in their ability to resist removal by flood disturbance (e.g., Peterson et al., 1990; Peterson & Stevenson, 1992). Differential resistance of these taxa may have caused some of the disparity in response of the loosely and tightly adherent communities. It is also possible that within the loosely and tightly adherent communities, algae (of the same species) varied with respect to physiological state. Attachment strength of the benthic algal communities is linked to the physiological health of the cells in the lower layers of the community (Peterson, 1996). It is also possible that the physical structure of the benthic algal mats themselves (e.g., extent of extracellular polysaccharide binding between cells and substratum, mat density/thickness) contributed to differential resistance. Peterson et al. (1994) found that thick, cohesive cyanobacterial mats were much more resistant to a natural flood event (with suspended sediment scour) than were diatomdominated communities. Our tightly attached communities also had a relatively cohesive mat structure, and included a common cyanobacterial filament (cf. Lyngbya sp.) which was lacking in the loosely attached communities. A generalized relationship seems evident between community composition and resistance to flood disturbance: upright filamentous algae are less resistant than adnate diatoms and cohesive mats including filamentous cyanobacteria. However, variability in spate resistance among periphyton communities can be related to other factors, not just the identity of the dominant taxon. Although we observed a statistically significant effect of disturbance duration on biomass removal, most biomass was removed during the first 5 min of disturbance. Biomass losses appeared to increase slightly from 5 to 30 min, with little or no further net removal of biomass after 30 min. These temporal dynamics were similar, regardless of

suspended sediment concentration. Biggs & Thomsen (1995) also found that short (10 min) disturbance exposures were sufficient to cause maximal biomass removal, and reported no strong temporal effects during sediment-free disturbances of benthic algal communities. In contrast, increasing the physical intensity of a disturbance event, either by increasing velocity (Biggs & Thomsen, 1995) or increasing sediment load (this study) generally resulted in greater biomass removal, suggesting that a short period of high physical intensity is more effective at removing algal biomass than long events of low intensity. Our observations of the spatial distribution of biomass removal suggest that the physical structure of the bed is important in determining resistance to suspended sediment disturbance. After experimental disturbances, biomass removal was noticeably more complete on the upper upstream faces of hemispheres, while thick algal growths remained attached to more sheltered areas, such as the corresponding downstream faces of hemispheres. This pattern was especially strong in treatments with elevated suspended sediment concentrations; in these cases, the upper front faces of hemispheres were often completely denuded of all organic matter, and the plastic substrata itself was abraded, yet algae remained attached to sheltered areas. Similar spatial differences in benthic algal biomass between upstream and downstream faces of boulders in the Colorado river have been attributed to spatial differences in suspended sediment scour (Blinn & Cole, 1991), and other natural refugia have been identified at micro (Krejci & Lowe, 1986; Miller et al., 1987; Bergey & Resh, 1994) and meso/macro scales (Uehlinger, 1991; Francoeur et al., 1998; Bergey, 1999). Because the magnitude of biomass removal is partially dependent upon the shape and texture of the substratum upon which benthic algae grow, it is important to replicate the shapes of natural substrata; estimates of biomass losses caused by flow and suspended sediment scour obtained from experiments in smooth, tile-lined artificial channels will not be representative of natural disturbance events. Quantitative examination of the effects of substratum shape/texture on benthic algae resistance may serve to further our understanding of lotic disturbance.

67 The benthic algal communities used in this study were generally dominated by filamentous algae (Melosira and Ulothrix). Cross comparison of communities with a wider variety of taxonomic compositions may reveal that the relative importance of velocity and suspended sediment scour varies as a function of algal community composition/physiognomy. A community composed of thin layer of diatoms (e.g., Cocconeis) would be quite resistant to removal by elevated velocities (see Biggs & Thomsen, 1995; Peterson, 1996; Stevenson, 1996), but still vulnerable to removal or damage by suspended sediment abrasion (e.g., Blenkinsopp & Lock, 1994). Thus, the importance of suspended sediment scour may be greater in such communities than in the filament-dominated communities examined in the present study. Although our experiments were conducted in a laboratory flow tank, they were designed to replicate natural conditions during a sedimentmobilising flood. The shape and texture of substrata were similar to a cobble river bottom (Biggs & Thomsen, 1995; Biggs et al., 1998a). Water velocities (e.g., Biggs, 1995) and suspended sediment loads (Hicks & Griffiths, 1992; Hicks et al., 2000) were within the wide ranges observed during floods in New Zealand rivers. Our suspended sediment loads were similar to those observed during a managed reservoir release (4000 mg l)1, Robinson et al., 2004) and after low impact timber harvesting (800 mg l)1, Stott et al., 2001), but were considerably less than suspended sediment loads associated with use of earth moving equipment for timber harvest or heavy construction (up to 30,000 mg L)1, Hicks & Griffiths, 1992). Our experimental sediments were New Zealand river sands (84% between 63 and 500 lm). Particles of this size are transported by rivers, particularly lower in the river water column (Hicks & Griffiths, 1992). Slightly larger sands (500–1000 lm) have been used for experimental disturbance of lotic invertebrate communities (Bond & Downes, 2003). In some river systems, clays are the predominant sediment size class (e.g., 80–85% <63 lm, Laubel et al., 1999), and some forms of human disturbance can result in suspended sediments of very small particle sizes (e.g., <1 lm, Winterbourn & Ryan, 1994). These clay particles likely have less abrasive potential than do sand particles, and may become incorporated into periphyton communities

(Graham, 1990; Winterbourn & Ryan, 1994). Conducting similar experiments with a wider range of suspended sediment concentrations and sediment sizes would help refine estimates for algal biomass removal under different physical conditions. Acknowledgements The authors wish to thank Julianne Heinlein for generous discussions of her previous work in this area and Joe Holomuzki for field assistance. This study was funded by US National Science Foundation grant INT-0107360 (SNF) and New Zealand Foundation for Research, Science and Technology program COIX-0308 ‘‘Water Allocation: Effects on Instream Values’’ (BJFB).

References Bergey, E. A., 1999. Crevices as refugia for stream diatoms: Effect of crevice size on abraded substrates. Limnology and Oceanography 44: 1522–1529. Bergey, E. A. & V. H. Resh, 1994. Effects of burrowing by a stream caddisfly on case-associated algae. Journal of the North American Benthological Society 13: 379–390. Biggs, B. J. F., 1987. Effects of sample storage and mechanical blending on the quantitative analysis of river periphyton. Freshwater Biology 18: 197–203. Biggs, B. J. F., 1995. The contribution of disturbance, catchment geology and landuse to the habitat template of periphyton in stream ecosystems. Freshwater Biology 33: 419–438. Biggs, B. J. F., 1996. Patterns in benthic algae of streams. In Stevenson, R. J., M. L. Bothwell, & R. L. Lowe (eds), Algal Ecology: Freshwater Benthic Ecosystems. Academic Press, San Diego (CA): 31–56. Biggs, B. J. F. & M. E. Close, 1989. Periphyton biomass dynamics in gravel bed rivers: The relative effects of flows and nutrients. Freshwater Biology 22: 209–231. Biggs, B. J. F., M. J. Duncan, A. M. Suren & J. R. Holomuzki, 2001. The importance of bed sediment stability to benthic ecosystems of streams. In Mosley, M. P. (ed.), Gravel-Bed Rivers, 5. New Zealand Hydrological Society, Wellington (NZ): 423–450. Biggs, B. J. F., S. N. Francoeur, A. D. Huryn, R. Young, C. J. Arbuckle & C. R. Townsend, 2000. Trophic cascades in streams: effects of nutrient enrichment on autotrophic and consumer benthic communities under two different fish predation regimes. Canadian Journal of Fisheries and Aquatic Sciences 57: 1380–1394. Biggs, B. J. F., D. G. Goring & V. I. Nikora, 1998a. Subsidy and stress responses of stream periphyton to gradients in

68 water velocity as a function of community growth form. Journal of Phycology 34: 598–607. Biggs, B. J. F. & C. Kilroy, 2000. Stream periphyton monitoring manual. New Zealand Ministry for the Environment/ NIWA, Christchurch (NZ). Biggs, B. J. F. & R. A. Smith, 2002. Taxonomic richness of stream algae: Effects of flood disturbance and nutrients. Limnology and Oceanography 47: 1175–1186. Biggs, B. J. F., R. A. Smith & M. J. Duncan, 1999. Velocity and sediment disturbance of periphyton in headwater streams: Biomass and metabolism. Journal of the North American Benthological Society 18: 222–241. Biggs, B. J. F., R. J. Stevenson & R. L. Lowe, 1998b. A habitat matrix conceptual model for stream periphyton. Archiv fu¨r Hydrobiologie 143: 21–56. Biggs, B. J. F. & H. A. Thomsen, 1995. Disturbance in stream periphyton by perturbations in shear stress: Time to structural failure and differences in community resistance. Journal of Phycology 31: 233–241. Blenkinsopp, S. A. & M. A. Lock, 1994. The impact of stormflow on river biofilm architecture. Journal of Phycology 30: 807–818. Blinn D. W. & G. A. Cole, 1991. Algal and invertebrate biota in the Colorado river: Comparisons of pre- and post-dam conditions. In Colorado River Ecology and Dam Management. National Academy Press, Washington (DC): 102–123. Bond, N. R. & B. J. Downs, 2003. The independent and interactive effects of fine sediment and flow on benthic invertebrate communities characteristic of small upland streams. Freshwater Biology 48: 455–465. Duncan, M. J. & B. J. F. Biggs, 1998. Substrate stability vs flood frequency and its ecological implications for headwater streams. In Wheater, H. S. & C. Kirby (eds), Hydrology in a Changing Environment, 1. John Wiley and Sons Ltd, Chichester (UK): 347–355. Fisher, S. G. & N. B. Grimm, 1991. Streams and disturbance: Are cross-system comparisons useful? In Cole, J. G. Lovett, & S. Findlay (eds), Comparative Analyses of Ecosystems: Patterns, Mechanisms, and Theories. Springer-Verlag, New York (NY): 196–221. Francoeur, S. N., B. J. F. Biggs & R. L. Lowe, 1998. Microform bed clusters as refugia for periphyton in a flood-prone headwater stream. New Zealand Journal of Marine and Freshwater Research 32: 363–374. Graham, A. A., 1990. Siltation of stone-surface periphyton in rivers by clay-sized particles from low concentrations in suspension. Hydrobiologia 199: 107–115. Grimm, N. B. & S. G. Fisher, 1989. Stability of periphyton and macroinvertebrates to disturbance by flash floods in a desert stream. Journal of the North American Benthological Society 8: 293–307. Heinlein, J., 2000. Flood Disturbance of Stream Periphyton: The Independent and Interactive Effects of Water Velocity, Suspended Sediment, and Initial Community Biomass. M.S. Thesis, Bowling Green State University, Bowling Green (OH). Hicks, D. M. & G. A. Griffiths, 1992. Sediment load. In Mosley, M. P. (ed.), Waters of New Zealand. New Zealand Hydrological Society, Wellington (NZ): 229–248.

Hicks, D. M., B. Gomez & N. A. Trustrum, 2000. Erosion thresholds and suspended sediment yields, Waipaoa River basin, New Zealand. Water Resources Research 36: 1129– 1142. Horner, R. R., E. B. Welch, M. R. Seeley & J. M. Jacoby, 1990. Responses of periphyton to changes in current velocity, suspended sediment and phosphorus concentration. Freshwater Biology 24: 215–232. Jowett, I. & B. J. F. Biggs, 1997. Flood and velocity effects on periphyton and silt accumulation in two New Zealand rivers. New Zealand Journal of Marine and Freshwater Research 31: 287–300. Krejci, M. E. & R. L. Lowe, 1986. Importance of sand grain mineralogy and topography in determining microspatial distribution of episammic diatoms. Journal of the North American Benthological Society 5: 211–220. Laubel, A., M. Svendsen, B. Kronvang & S. E. Larsen, 1999. Bank erosion in a Danish lowland stream system. Hydrobiologia 410: 279–285. Miller, A. R., R. L. Lowe & J. T. Rotenberry, 1987. Succession of diatom communities on sand grains. Journal of Ecology 75: 693–709. Peterson, C. G., 1996. Response of benthic algal communities to natural physical disturbance. In Stevenson, R. J., M. L. Bothwell, & R. L. Lowe (eds), Algal Ecology: Freshwater Benthic Ecosystems. Academic Press, San Diego (CA): 375– 402. Peterson, C. G., K. D. Hoagland & R. J. Stevenson, 1990. Timing of wave disturbance and the resistance and recovery of a freshwater epilithic microalgal community. Journal of the North American Benthological Society 9: 54–67. Peterson, C. G. & R. J. Stevenson, 1992. Resistance and resilience of lotic algal communities: Importance of disturbance timing and current. Ecology 73: 1445–1461. Peterson, C. G., A. C. Weibel, N. B. Grimm & S. G. Fisher, 1994. Mechanisms of benthic algal recovery following spates: Comparison of simulated and natural events. Oecologia 98: 280–290. Poff, N. L. & J. V. Ward, 1989. Implications of streamflow variability and predictability for lotic community structure: A regional analysis of streamflow patterns. Canadian Journal of Fisheries and Aquatic Sciences 46: 1805–1818. Power, M. E. & A. J. Stewart, 1987. Disturbance and recovery of an algal assemblage following flooding in an Oklahoma stream. American Midland Naturalist 117: 333–345. Resh, V. H., A. V. Brown, A. P. Covich, M. E. Gurtz, W. L. Hiram, G. W. Minshall, S. R. Reice, A. L. Sheldon, J. B. Wallace & R. C. Wissmar, 1988. The role of disturbance in stream ecology. Journal of the North American Benthological Society 7: 433–455. Robinson, C. T., U. Uehlinger & M. T. Monaghan, 2004. Stream ecosystem response to multiple experimental floods from a reservoir. River Research and Applications 20: 359–377. Stott, T., G. Leeks, S. Marks & A. Sawyer, 2001. Environmentally sensitive plot-scale timber harvesting: Impacts on suspended sediment, bedload and bank erosion dynamics. Journal of Environmental Management 63: 2–25. ter Braak, C. J. F., 1995. Ordination. In Jongman, R. H. G. C. J. F. ter Braak, & O. F. R. Tongeren Van (eds), Data

69 Analysis in Community and Landscape Ecology, 2. Cambridge University Press, Cambridge (UK): 91–173. ter Braak, C. J. F. & P. Smilauer, 1998. CANOCO Reference Manual and User’s Guide to CANOCO for Windows: Software for Canonical Community Ordination (Version 4). Microcomputer Power, Ithaca (NY). Stevenson, R. J., 1996. The stimulation and drag of current. In Stevenson, R. J., M. L. Bothwell, & R. L. Lowe (eds), Algal

Ecology: Freshwater Benthic Ecosystems. Academic Press, San Diego (CA): 321–340. Uehlinger, U., 1991. Spatial and temporal variability of the periphyton biomass in a prealpine river (Necker, Switzerland). Archiv fu¨r Hydrobiologie 123: 219–237. Winterbourn, M. J. & P. A. Ryan, 1994. Mountain streams in Westland, New Zealand: Benthic ecology and management issues. Freshwater Biology 32: 359–373.

Hydrobiologia (2006) 561:71–82  Springer 2006 R.J. Stevenson, Y. Pan, J.P. Kociolek & J.C. Kingston (eds), Advances in Algal Biology: A Commemoration of the Work of Rex Lowe DOI 10.1007/s10750-005-1605-3

The effects of pH on a periphyton community in an acidic wetland, USA Jennifer L. Greenwood1,2,3,* & Rex L. Lowe1,2 1

Bowling Green State University, Department of Biological Sciences, Bowling Green, OH, 43403, USA University of Michigan Biological Station, 9008 Biological Road, Pellston, MI, 49769, USA 3 U.S. EPA National Exposure Research Laboratory, 26 W. Martin Luther King Dr., Cincinnati, OH, 45268, USA (*Author for correspondence: E-mail: [email protected]) 2

Key words: diatom, desmid, peatland, fen, acidification, alkalinization

Abstract Despite the importance of peatlands, the algal ecology of peatlands and the periphyton communities which are abundant in these habitats are relatively understudied. We performed an in situ manipulation of pH in an intermediate fen in northern lower Michigan in order to examine how hydrogen ion concentrations structure an epiphytic algal community. Levels of pH were manipulated in enclosures from the control level (pH=5) to an acid treatment (pH=4) by adding H2SO4 and a neutral treatment (pH=7) by adding NaOH. Algal communities growing on sections of Chamaedaphne calyculata (L.) Moench stems were examined after 22 days of colonization. Chlorophyll a concentration was significantly greater only in the acid treatment (5.5 mg m)2) relative to the control (3.5 mg m)2). Taxa richness was lower in the acid treatment. The algal assemblages were dominated by filamentous green algae and a filamentous taxon, Mougeotia spp., was significantly greater in the acid treatment relative to the control. Increases in Zygnemataceae and Oedogonium spp. most likely account for the higher chlorophyll a in the acid treatment. Most treatment differences were detected in the neutral treatment, including increased abundances of Closterium polystichum Nygaard, Cosmarium sp., Peridinium inconspicuum Lemmermann, and Synedra acus Ku¨tz. Unexpectedly, there was no strong response of the desmid community. These data can be informative in the development of algal monitoring programs in peatlands when assessment of acidification is desired.

Introduction Peat-storing wetlands are important freshwater habitats. They are widely distributed, provide important habitats for often endemic flora and fauna and are in danger of destruction due to human uses, such as drainage for agriculture or excavation of peat (Parkyn et al., 1997; Charman, 2002). Benthic algal ecology in peatlands is drastically understudied compared to other aquatic systems such as lakes, streams, rivers, and non-peat-forming wetlands (Stevenson et al., 1996). Peatlands are physically and chemically unusual habitats, often with low pH levels and low

nutrient and ion concentrations (Crum, 1988; Vitt, 2000). This unique chemistry results in a distinctive algal flora. In-depth descriptions of the unique and highly diverse algal assemblages from peatlands have been the topic of several floristic investigations describing patterns of species distribution and, often, correlations with physical variables (e.g., Flensburg & Malmer, 1970; Kingston, 1982a; Yung et al., 1986; Mataloni, 1999; Poulı´ ckova´ et al., 2004). There have been a few paleolimnological investigations of algal communities in peatlands to reconstruct the developmental history of a particular aquatic ecosystem (Kingston, 1982b; Ru¨hland et al., 2000).

72 However, in situ experimental manipulations to study the mechanisms governing the distribution and composition of algal communities in peatlands are rare or non-existant. As the importance of using algae as biomonitors in wetlands increases (Stevenson et al., 2002), understanding the mechanisms driving algal community structure in peatlands also becomes more important. An abiotic factor important in the classification of peatlands is pH (Crum, 1988; Vitt, 2000). However, the role of pH in structuring periphyton communities has rarely been examined explicitly in peatlands (but see van Dam et al., 1981; Bellemakers & van Dam, 1992). In recent years, examinations of the effect of pH on aquatic systems or algal communities have mostly addressed the effects of anthropogenic acidification of surface waters (see Planas, 1996 for a review). In contrast with peatland habitats, community changes of algal flora due to acidification have been extensively documented in lakes (Yan, 1979; Turner et al., 1987; Kingston et al., 1990; Turner et al., 1991) and streams (van Dam & Mertens, 1995; Verb & Vis, 2000). Effects of acidification on algal communities have shown some general trends. Acidification often results in decreased species richness (e.g., Mu¨ller, 1980; Schindler et al., 1985; Turner et al., 1991). Also, acidification experiments in lakes often result in an increase in filamentous green algae in the family Zygnemataceae, especially the genera Mougeotia, Spirogyra, and Zygogonium (e.g., Turner et al., 1995a; Turner et al., 1995b). This increase has been attributed to an increased competitive advantage associated with decreased dissolved inorganic carbon at lower pH (Jackson et al., 1990; Klug & Fischer, 2000) or to a suppression of herbivores (Turner et al., 1987; Barmuta et al., 1990; Fairchild & Sherman, 1992). Generally, total algal biomass has been reported to increase in response to acidification (Mu¨ller, 1980), usually due to the blooms of filamentous green algae. In the opposite direction on the pH spectrum, alkalinization (or neutralization) in lakes has also been examined. Neutralization is generally conducted as a process to mitigate impacts by recent anthropogenic acidification, generally by liming (Hultberg & Andersson, 1982; Fairchild & Sherman,

1990; Olem, 1991; Renberg & Hultberg, 1992; Ho¨rnstro¨m, 2002). Neutralization can often reverse the changes in the algal community which resulted from acidification, again, particularly in the filamentous green algae. For example, several neutralization studies have found that Mougeotia is drastically reduced when pH is neutralized via liming from about pH 5 (Hultberg & Andersson, 1982; Jackson et al., 1990; Fairchild & Sherman, 1992). In contrast to the applied research on the effects of anthropogenic acidification in lentic systems, empirical research on how algal communities respond to effects of pH levels in peatlands has not been thoroughly examined. This experiment was designed to examine the response of a peatland algal community to a range of pH regimes, increased and decreased from ambient levels. The changes in periphyton algal assemblages resulting from variation in pH were examined in enclosures in Waldron Fen, an acidic peatland in northern lower Michigan. We hypothesized that increasing acidity would increase algal biomass due to increased filamentous green algal abundance, and would result in a decrease in species richness and a shift toward more acidophilic or acidobiontic species, especially desmids or acidophilic diatoms. We also hypothesized that neutralization would result in an increase in taxa richness and a decrease in algal biomass due to decreases in filamentous green algae.

Materials and methods Study site Waldron Fen is located in northern lower Michigan (Fig. 1) approximately 17 km northeast of Petoskey (4523¢ N, 8446¢ W). This wetland is an ‘‘intermediate fen’’ (Crum, 1988) where minerals and nutrients are scarce, and the open water has a pH of approximately 5. The common vascular plant flora is indicative of fen-type peatstoring wetlands. Chamaedaphne calyculata (L.) Moench and the sedge, Carex lasiocarpa Ehrh. are dominant species within the Sphagnum lawn which surrounds the pond of open water. Other vascular plants such as Vaccinium spp., Sarracenia purpurea L., Drosera spp., and Dulichium arundinaceum (L.)

73

Figure 1. Map of location of study site.

Britton are also present. Experiments were performed in an open water area of the fen. Here, phytoplankton densities were very low, and the dominant substrata for periphyton were Nymphaea odorata Aiton and Nuphar variegate Durand stems and underleaves, and submerged stems of C. calyculata dominated the margin of the open water.

pH manipulation Hydrogen ion concentration was manipulated in open-ended cylindrical enclosures which were embedded in the bottom of the fen (Fig. 2). Enclosures were constructed from welded wire fabric, 30 cm in diameter and 1.5 m tall. The top 1.1 m of the inside of the enclosures was lined with 4-mil (approximately 0.1 mm thick) window vinyl which was glued at the seams and attached to the wire fabric with clear silicone sealant. Enclosures were embedded approximately 20 cm into the sediment. Approximately 20 cm of the bottom of the enclosure remained open to the environment between the bottom of the vinyl lining and the sediment surface after enclosures were deployed in the fen. This allowed some interaction with the natural environment, while maintaining experimental pH levels. Flocculent sediment filled the

first 70 cm of enclosure, with about 50 cm of transparent water above the flock. Approximately 10 cm of the enclosure extended above the water surface. Three pH treatments: pH=4, pH=7 and the control, pH=5, with four replicates each, were established. Enclosures were placed in areas of the fen that had similar depths, currents, and distance from the margin of the fen. About 5 ml of 2.5% solutions of H2SO4 (for the pH=4 treatment) and NaOH (for the pH=7 treatment) were required to initially change the pH within the enclosures. The pH of each enclosure was checked every 2–3 days with a Corning Model 6 portable pH meter and adjusted as needed with enough 2.5% H2SO4 or 2.5% NaOH to maintain experimental pH levels. Water in the enclosure was stirred 10 times with a meter stick after dosing or at each visit if additional acid or base was not added. Algal substrata were deployed (see below) after pH within the enclosures stabilized. Macro-herbivores were not observed in the enclosures during the experiment. Chamaedaphne calyculata stems were employed as a standard substratum for sampling periphyton. This ericaceous shrub dominated the margin of the open water, and stems that were submerged often had thick growths of green algae. The other potential natural algal substrata to use from the open water area of the fen, Nymphaea or Nuphar plants,

74

Figure 2. Enclosures used in pH manipulation (not to scale).

would have decomposed more quickly than Chamaedaphne stems, potentially confounding experimental effects. The Chamaedaphne stems were also smooth and sturdy, providing a replicable yet natural substratum for algal colonization. Aerial Chamaedaphne stems (i.e., stems were not previously submerged and did not have periphyton growth) 4–6 mm in diameter were removed from live plants and cut into approximately 10 cm length segments. Sheets of 55  30 cm plastic canvas positioned vertically inside the enclosure on the northern side (to maximize direct sun exposure and minimize light interference from the vinyl lining) were used to stabilize the horizontal position of the stems by inserting one end of the stem into an opening in the canvas, allowing the stem to extend horizontally toward the center of the enclosure at more or less a right angle to the sheet of canvas. The plastic canvas was tied to the enclosure with strings and could be repositioned within the enclosure to maintain a consistent depth of the substrata of 5 cm from the surface. Algae were allowed to colonize for 22 days (July 28–August 19, 1996). A longer colonization period was infeasible for logistical reasons. Of ten stems deployed, three randomly selected stems were

harvested for sampling from each enclosure and pooled to produce one replicate. All samples were processed on 19 August 1996. Chamaedaphne stems were brushed clean with a toothbrush and rinsed thoroughly with deionized water. The resulting algal slurry (usually 50–100 ml) was homogenized with a tissue miser. For chlorophyll a analysis, between 20–50 ml of algal slurry was filtered through a Millipore 0.45 lm filter, filters were sonicated in 90% acetone buffered with magnesium carbonate for 10 min, then extracted in a dark freezer for 24 h. The pigment concentration was measured on a Turner 10-000R fluorometer and corrected for phaeophytin (APHA, 1998). For algal community analysis, 20 ml were preserved in 2.5% formalin and samples were analyzed by counting at least 500 cells or colonies per sample in a Palmer-Maloney nannoplankton counting chamber. Algal cells were identified to the lowest taxonomic level possible at 400 using Prescott (1962), Patrick & Reimer (1966, 1975), and Prescott et al. (1975–1983). Calcium (Ca), Soluble reactive phosphorus (SRP), nitrate-nitrogen (NO3–N) ammonia nitrogen (NH3–N) and silica (SiO2) were measured in all enclosures and in 4 areas in the open water on 8

75 August 1996. Concentrations were determined with a Technicon II Autoanalyzer (APHA, 1998). Differences among treatments in nutrients, chlorophyll a, taxa richness and absolute abundances of taxa were determined with ANOVA, and multiple comparisons were performed with a Tukey test with JMP (Release 5.0.1a, SAS Institute, Inc., 2002). All non-nutrient data were log-transformed prior to analysis. Below detection values for nutrients were considered zeroes during analysis. Only taxa and groups of taxa that were > 5% relative abundance were included in the analysis. The Bonferroni procedure was used in algal assemblage analysis to preserve the experimentwise Type I error rate of p=0.05 (Ott & Longnecker, 2001).

required to initially raise the pH to or above 7, with maintenance requiring an additional 3 ml every 2–3 days for the first 10 days of enclosure deployment. During the algal colonization period, approximately 0.5 ml of NaOH were added on the 3rd and 19th day to maintain pH=7. Mineral and nutrient concentrations were relatively low in all samples (Table 1). Concentrations of Ca, SRP, NH3)N and SiO2 were significantly different between ambient condition, control and treatments (Table 1) (ANOVA Ca: p = 0.02; SRP: p < 0.0001; NH3)N: p = 0.0003, SiO2: p = 0.03). NO3–N was not detected in any of the enclosures or in the ambient water column. Mean chlorophyll a levels were approximately double in the acid treatment compared to the control (ANOVA, p=0.05, Fig. 4). Chlorophyll a concentration in the neutral treatment was not significantly different from the acid or control treatment. The average taxa richness was significantly lower in the acid treatment compared to the neutral treatment (ANOVA, p=0.03, Fig. 5). The total number of taxa identified (47 for all treatments) in composited treatment samples was lower in the acid treatment (27 taxa identified) compared with the neutral (35 taxa) and control (32 taxa) treatments.

Results Enclosures were deployed for 21 days to allow pH levels to stabilize before adding the algal substrata (Fig. 3). For the acid treatment, an average of 3 ml of H2SO4 were required to initially decrease the pH to 4, with pH maintenance requiring an additional 0.5 ml every 2–3 days. For the neutral treatment, an average of 6 ml of NaOH were

11 10 9

Acid

Neutral

Control

Ambient

pH

8 7 6 5 4 3 -25

-20

-15

-10

-5

0

5

10

15

20

25

Days Since Algal Substrata Deployment Figure 3. Average pH levels in enclosures during the stabilization period before algal substrata were deployed and during algal substrata deployment in the acid, control, and neutral treatment enclosures and in the surrounding water (ambient). For each data point, n=4.

76 Table 1. Average concentrations and range of concentrations (in parentheses) of calcium (Ca), soluble reactive phosphorus (SRP), nitrate-nitrogen (NO3–N) ammonia nitrogen (NH4–N) and silica (SiO2) in the open water of Waldron Fen (Ambient) and in the experimental enclosures on 8 Aug 1996 Ca (mg l)1)

SRP (lg l)1)

NO3–N (lg l)1)

NH3–N (lg l)1)

SiO2 (lg l)1)

Ambient

2.25 A (2.0–3.0)

2.6 A (2.0–3.2)

bd (bd)

25 A (20–30)

bd A (bd)

Acid

2.26 A (0.05–3)

bd B (bd)

bd(bd)

bd B (bd)

60 B (bd-130)

Control

1.51 AB (0.05–2)

2.45 A (2–2.6)

bd (bd)

bd B (bd-23)

Bd A(bd)

Neutral

0.05 B (0.05–0.05)

3.8 C (3.2–4.4)

bd (bd)

bd B (bd)

bd A (bd)

bd=below detection and n=4 for all values. Capital letters indicate significant differences (ANOVA p < 0.05, Tukey p < 0.05) within each mineral or nutrient among the treatments and ambient concentrations.

8 7

30

b

a

ab

6 5

ab

4

b

3

Taxa richness

Chlorophylla mg m-2

25 20

a 15 10

2

5

1 0 Acid

Control

Neutral

0 Acid

Treatment Figure 4. Mean (±1 standard error) chlorophyll a concentration (mg m)2 of substratum) from substrata in pH-manipulated enclosures in Waldron Fen (Acid: pH=4.0; Control: pH = 5.2; Neutral: pH=7.0). Significant difference indicated by different letters (ANOVA, p = 0.05, Tukey p < 0.05). For each treatment, n = 4.

There were no significant differences in total algal abundance between treatments, although total abundance tended to be higher in the acid treatment (Fig. 6, ANOVA p>0.05). Responses of taxa to the treatments were analyzed only in individual taxa or groups of taxa (Cyanobacteria, dinoflagellates, diatoms, chrysophytes, Chlorococcales, Oedogoniales, Zygnematales, unicellular desmids, and filamentous desmids) that had greater than 5% relative abundance in any one replicate (Table 2). There were no significant differences in absolute abundances of the total number of cells for any of the groups of algae (Fig. 7). The only species that was more abundant in the acid treatment relative to the control was

Control

Neutral

Treatment Figure 5. Mean taxa richness (±1 standard error) from substrata exposed for three weeks in pH-manipulated enclosures in Waldron Fen (Acid: pH=4.0; Control: pH=5.2; Neutral: pH=7.0). Significant differences indicated by different letters above bars (ANOVA, p < 0.05, Tukey p < 0.05). For each treatment, n=4.

Mougeotia sp.3 (ANOVA p=0.005, Fig. 8a). There were significantly lower cell densities in the acid treatment of the species Closterium polystichum Nygaard (ANOVA p=0.002, Fig. 8b), Cosmarium sp.1 (ANOVA p<0.001, Fig. 8c), Peridinium inconspicuum Lemmermann (ANOVA p=0.0009, Fig. 8d), and Synedra acus Ku¨tzing (ANOVA p=0.0005, Fig. 8e).

Discussion This study is one of the first to examine the effects of acidification and neutralization simultaneously in periphyton from a peatland. Changes in pH

77 resulted in significant changes in the biomass and species composition of an attached algal community. The increase in biomass as chlorophyll a may have been due to increases in some filamentous green algae in the acid treatment, particularly Mougeotia spp. (Fig. 8a) and Oedogonium spp. (Fig. 7). Mougeotia spp. is consistently found to increase in response to acidification in many lake studies (Turner et al., 1987; Klug & Fischer, 2000), so the increase in Mougeotia in our study is not surprising. The results from our measures of abundance were likely more conservative than biovolume measurements would have been since a large proportion of cell densities were composed of high biovolume taxa, i.e. filaments. Also, algal assemblage data from peatlands are relatively unavailable; we are unaware of published biomass data for periphyton from peatlands. The lower taxa richness in the acid treatment was also anticipated (Fig. 5). A pH of 4.0 is a physically stressful habitat for many taxa due to

10 a

9 8

106 cells cm-2

7 6 5 a

4

a

3 2 1 0 Acid

Control

Neutral

Treatment Figure 6. Total abundance (±1 standard error) in cells cm)2 of substratum from substrata exposed for 22 days in pH-manipulated enclosures in Waldron Fen (Acid: pH=4.0; Control: pH=5.2; Neutral: pH=7.0). There were no significant differences within groups among the treatments (ANOVA p > 0.05). For each treatment, n=4.

Table 2. Average cell density of taxa and taxon groups with >5% relative abundance Cells cm)2

Acid

Control

Neutral

Filamentous Desmids

1345167

1107313

182040

Bambusina brebissonii Ku¨tz. Desmidium baileyi (Ralfs) Norstedt

374967

82694

18595

329074

962389

163446

Desmidium grevellii (Ku¨tz) De Bary Hyalotheca dissilens (Smith) Bre´b. ex Ralfs Unicellular Desmids

465242

45964

0

175884 49058

16266 179513

0 558570

*Closterium polystichum Nygaard

0

38715

159056

*Cosmarium sp.1

0

14850

165532

Staurastrum sp.6

22695

73562

96560

Zygnemataceae

1540524

172541

243830

Mougeotia sp.2

1114885

163332

159556

*Mougeotia sp.3

165852

0

13323

Spirogyra sp. Oedogoniales

226738 4372764

9209 1545725

69682 1449259

Oedogonium sp.2

921640

1388052

302752

Oedogonium sp.3

3309282

112864

1146507

Diatoms

35856

79589

207023

*Synedra acus Ku¨tz. Chrysophytes

0

6795

188940

131716

44563

7614

Dinobryon sp.

70207

40913

6345

Dinoflagellates *Peridinium inconspicuum Lemm.

12452 0

1808 1808

144129 137097

*p < 0.05, ANOVA in enclosures from the acid, control and neutral treatments.

78 7000

Absolute Abundance cells cm-2

6000 Acid Control Neutral

5000

4000

3000

2000

1000

0 Cy

Dn

Chr

Dt

Chl

Oed

Zyg

Uni

Fil

Algal Periphyton Taxonomic Groups

Figure 7. Mean absolute abundance (±1 standard error) of Cyanobacteria (Cy), dinoflagellates (Dn), chrysophytes (Chr), Chlorococcales (Chl), Oedogoniales (Oed), Zygnematales (Zyg), unicellular desmids (Uni), and filamentous desmids (Fil) in cells cm)2 of substratum from substrata exposed for 22 days in pH-manipulated enclosures in Waldron Fen (Acid: pH=4.0; Control: pH=5.2; Neutral: pH=7.0). There were no significant differences within groups among the treatments (ANOVA p>0.05). For each treatment, n=4.

such factors as increased metal toxicity or reduced bicarbonate availability, an important source of carbon for algae. Also, phosphorus concentrations were lower in the acid treatment, potentially affecting taxa richness. Further, lower taxa richness may have resulted from the dominance of filamentous green algae (Zygnemataceae, Oedogoniales, and desmids) in the acid treatment. This may create a situation where encountering new taxa during identification will be less likely since the counts will be saturated by the dominant filaments. There were no increases or appearances of known acidophilic or acidobiontic species (i.e., diatoms). Effects on individual algal species seemed strongest in the neutral treatment. Peridinium inconspicuum showed an increase in cell density with increasing pH (Fig. 8d) which is contrary to findings in previous acidification studies (Dixit & Smol, 1989; Ho¨rnstro¨m, 1999). Also, Yan (1979) considered this taxon indicative of acidic lakes if it constituted a large part of the phytoplankton (evidence from Canada and Sweden). Further, Hultberg & Andersson (1982) found a decrease in

P. inconspicuum in the plankton from limed lakes in Sweden. Perhaps the changes in P. inconspicuum are not a result of pH preferences. This taxon may be indifferent and/or tolerant of low pH levels, and may be able to exploit niches left open by pHsensitive taxa that were eliminated during acidification or neutralization. However, P. inconspicuum was part of the periphyton (i.e., not plankton) in this study. It is not clear if this taxon is tychoplanktonic or if perhaps it was physiologically stressed in the neutral treatment during the experiment and settled onto the substrata. Also, the lack of current in the enclosures may have reduced the suspension of cells, increasing settling onto the substrata. However, plankton samples were depauperate of algal cells, making comparisons between planktonic and periphytic community composition difficult. Diatoms were relatively rare and no well documented acidophilic or acidobiontic diatoms species appeared in the acid treatment. This is not surprising considering that SiO2 levels were below detection, except in the acid treatment, and potentially limiting to diatom growth (Table 1). The higher concentrations of silica

79 3.5

300

(a)

Mougeotia sp. 3

(b)

Closterium polystichum

250

a

2.5

106cells cm-2

106cells cm-2

3

2 1.5 1 0.5

b

200 150 100

b

50

ab

a

b 0

0

Acid

Control

Neutral

Acid

Treatment 300

300

(c)

Cosmarium sp. 1

250

b

200 150

(d)

Neutral

Peridinium inconspicuum

250

106cells cm-2

106cells cm-2

Control

Treatment

b

200 150 100

100

b

50

50

a

a

a 0

0 Acid

Control

Neutral

350

(e)

Acid

Control

Neutral

Treatment

Treatment

Synedra acus

300

b

106cells cm-2

250 200 150 100 50

a

b

0 Acid

Control

Neutral

Treatment Figure 8. Mean absolute abundance (±1 standard error) in cells cm)2 of substratum of the taxa Mougeotia sp. 3 (a), Closterium polystichum (b), Cosmarium sp. 1 (c), Peridinium inconspicuum (d), and Synedra acus (e) with significant differences between absolute abundances of taxa. Samples were from substrata exposed for three weeks in pH-manipulated enclosures in Waldron Fen (Acid: pH=4.0; Control: pH=5.2; Neutral: pH=7.0). Significant difference indicated by different letters above bars (ANOVA, p<0.05, Tukey p<0.05). For each treatment, n=4.

80 observed in the acid treatment, however, did not result in increases in diatoms, although chryophyte levels tended to be higher. Another consideration is that diatoms may have grown better under a lower light environment relative to the enclosures (Hill, 1996). However, in a related study of shade effects on algal communities in the fen, decreased light intensities did not result in increases in diatoms (Greenwood, 1998). In our study, there was an increase in Synedra acus (Fig. 8e) in the neutral treatment which corroborates previous findings. Renberg and Hultberg (1992) found an increase in S. acus in lakes in Sweden that had been limed with CaCO3 to a pH of 7.5. This taxon is also considered planktonic and, like P. inconspicuum, reasons for its presence on experimental substrata are unceratain. Desmids, as a group, show the highest diversity in acidic, low-nutrient habitats (Brook, 1981; Coesel, 1981; Coesel, 1982). In this study, filamentous desmids did show a trend of lower cell density with increasing pH. However, densities of 2 unicellular desmids, Closterium polystichum (Fig. 8b), and Cosmarium sp. 1 (Fig. 8c) were significantly higher in the neutral treatment. Even though there are species which are found in higher nutrient environments (Gerrath, 2003), it is interesting that there was not more of a decrease in unicellular desmid cell density with neutralization. There was a non-significant decrease in filamentous desmids with neutralization, including the complete disappearance of Desmidium grevellii and Hyalotheca dissilens. However the overall response of desmids to changes in pH were relatively more muted than expected. Potentially, the length of the experiment was insufficient for the algal community to fully respond to changes in pH. Response from the desmids and other algae may have been more robust with a longer colonization period. However, our results do suggest that periphyton community structure can change with changes in pH.

Conclusion We successfully changed attached algal assemblages with in situ manipulations of pH. Many of our results were consistent with findings from acidification and acidification mitigation studies,

with higher algal biomass measured as chlorophyll a and filamentous green algal cell density and lower taxa richness with decreasing pH. Cell densities of taxa usually considered acidophilic were lower in response to neutralization. These data could potentially be informative in the development of algal monitoring programs in peatlands when assessment of acidification is desired. With our short-term, small scale experiment, we did see significant experimental effects, however, a longerterm, larger scale study would better predict the effects of pH change on periphytic algal communities in a peatland.

Acknowledgements We gratefully acknowledge Jack and Nancy Waldron for allowing us access to their beautiful fen. Gina LaLiberte and Bob Pillsbury helped in the field and Mike Grant performed the nutrient analysis. This work is part of a Master’s thesis at Bowling Green State University by JLG. We would especially like to thank Jan Stevenson, Yangdong Pan, John Kingston and Pat Kociolek for organizing RexFest and this special issue. Two anonymous reviewers greatly improved an earlier version of this manuscript.

References American Public Health Association, 1998. Standard Methods for the Examination of Water and Wastewater. Water Environment Federation. Arlington, VA: 1183. Barmuta, L. A., S. D. Cooper, S. K. Hamilton, K. W. Kratz & J. M. Melack, 1990. Responses of zooplankton and zoobenthos to experimental acidification in a high-elevation lake (Sierra Nevada, California, USA). Freshwater Biology 23: 571–586. Bellemakers, M. J. S. & H. van Dam, 1992. Improvement of breeding success of the moor frog (Rana arvalis) by liming of acid moorland pools and the consequences of liming for water chemistry and diatoms. Environmental Pollution 78: 165–171. Brook, A. J., 1981. The Biology of Desmids. University of California Press, Berkeley pp 276. Charman, D., 2002. Peatlands and Environmental Change. John Wiley, Sons, Inc, New York, NY pp 301. Coesel, P. F. M., 1981. Classification of desmid assemblies in a Dutch broads area. Archiv fu¨r Hydrobiologie 91: 56–81. Coesel, P. F. M., 1982. Structural characteristics and adaptations of desmid communities. Journal of Ecology 70: 163–177.

81 Crum, H., 1988. A Focus on Peatlands and Peat Mosses. The University of Michigan Press, Ann Arbor pp 306 . Dixit, S. S. & J. P. Smol, 1989. Algal Assemblages in AcidStressed Lakes with Particular Emphasis on Diatoms and Chrysophytes. Acid Stress and Aquatic Microbial Interactions. S. S. Rao. CRC Press, Inc, Boca Raton, Florida pp 91–114. Fairchild, G. W. & J. W. Sherman, 1990. Effects of liming on nutrient limitation of epilithic algae in an acid lake. Water, Air, and Soil Pollution 52: 133–147. Fairchild, G. W. & J. W. Sherman, 1992. Linkage between epilithic algal growth and water column nutrients in softwater lakes. Canadian Journal of Fisheries and Aquatic Sciences 49: 1641–1649. Flensburg, T. & N. Malmer, 1970. Studies on mire vegetation in the Archaean Area of South-Western Go¨taland (south Sweden). Botaniska Notiser 123: 269–299. Gerrath, J. P., 2003. Conjugating green algae and desmids. In Wehr, J. D. & R. G. Sheath (eds) Freshwater Algae of North America, Academic Press New York: 353–81. Greenwood, J. L., 1998. The Effects of pH and Light on Periphyton Communities in a Michigan Peatland. Bowling Green State University Bowling Green, OH. Hill, W. R., 1996. Effects of Light. In Stevenson, R. J., M. L. Bothwell, & R. L. Lowe (eds), Algal Ecology, Freshwater Benthic Ecosystems. Academic Press, San Diego, CA: 121– 148. Ho¨rnstro¨m, E., 1999. Long-term phytoplankton changes in acid and limed lakes in SW Sweden. Hydrobiologia 394: 93–102. Ho¨rnstro¨m, E., 2002. Phytoplankton in 63 limed lakes in comparison with the distribution in 500 untreated lakes with varying pH. Hydrobiologia 470: 115–126. Hultberg, H. & I. B. Andersson, 1982. Liming of acidified lakes: induced long-term changes. Water, Air, and Soil Pollution 18: 311–331. Jackson, M. B., E. M. Vandermeer, N. Lester, J. A. Booth L. Molot & I. M. Gray, 1990. Effects of neutralization and early reacidification on filamentous algae and macrophytes in Bowland Lake. Canadian Journal of Fisheries and Aquatic Sciences 47: 432–439. Kingston, J. C., 1982a. Association and distribution of common diatoms in surface samples from northern Minnesota peatlands. Beiheft zur Nova Hedwigia 73: 333–346. Kingston, J. C., 1982b. Paleolimnology of a Lake and Adjacent Fen in Southeastern Labrador: Evidence from Diatom Assemblages. 7th International Diatom Symposium, Philadelphia, Otto Koeltz. Kingston, J. C., R. B. Cook, J. Kreis, G. Russel, K. E. Camburn, S. A. Norton, P. R. Sweets, M. W. Binford, M. J. Mitchell, S. C. Schinder, L. Shane & G. King, 1990. Paleoecological investigation of recent lake acidification in the northern Great Lakes states. Journal of Paleolimnology 4: 153–201. Klug, J. L. & J. M. Fischer, 2000. Factors influencing the growth of Mougeotia in experimentally acidified mesocosms. Canadian Journal of Fisheries and Aquatic Sciences 57: 538–547. Mataloni, G., 1999. Ecological studies on algal communities from Tierra del Fuego peat bogs. Hydrobiologia 391: 157–171.

Mu¨ller, P., 1980. Effects of artifical acidification on the growth of periphyton. Canadian Journal of Fisheries and Aquatic Sciences 37: 355–363. Olem, H., 1991. Liming Acidic Surface Waters. Lewis Publishers, Chelsea, Michigan, pp 331. Ott, R. L. & M. Longnecker, 2001. An Introduction to Statistical Methods and Data Analysis. Duxbury, Pacific Grove, CA, pp 1152. Parkyn, L. & R. E. H. A. P. Stoneman Ingram (eds), 1997. Conserving Peatlands. CAB International, Wallingford, UK, pp 500. Patrick, R. & C. W., Reimer, 1966. The Diatoms of the United States, Volume 1: Fragilariaceae, Eunotiaceae, Achnanthaceae, Naviculaceae. The Academy of Natural Sciences of Philadelphia. Philadelphia, PA: 688. Patrick, R. & C. W. Reimer, 1975. The Diatoms of the United States, Volume 2, Part 1: Entomoneidaceae, Cymbellaceae, Gomphonemaceae, Epithemiaceae. The Academy of Natural Sciences of Philadelphia. Philadelphia, PA: 213. Planas, D., 1996. Acidification effects. In Stevenson R. J., M. L. Bothwell & R. L. Lowe (eds), Algal Ecology: Freshwater Benthic Ecosystems. Academic Press, San Diego: pp. 497– 530. Poulı´ ckova´, A., P. Ha´jkova´, P. Krenkova´ & M. Ha´jek, 2004. Distribution of diatoms and bryophytes on linear transects through spring fens. Nova Hedwigia 78: 411–424. Prescott, G. W., 1962. Algae of the Western Great Lakes Area, With an Illustrated Key to the Genera of Desmids and Freshwater Diatoms. W.C. Brown Co, Dubuque, IA, pp 977. Prescott, G. W., H. Croasdale, W. C. Vinyard & C. E. d M. Bicudo, 1975–1983. A Synopsis of North American desmids. University of Nebraska Press, Lincoln and London. Renberg, I. & H. Hultberg, 1992. A paleolimnological assessment of acidification and liming effects on diatom assemblages in a Swedish lake. Canadian Journal of Fisheries and Aquatic Sciences 49: 65–72. Ru¨hland, K., J. P. Smol, J. P. P. Jasinski & B. G. Warner, 2000. Response of diatoms and other siliceous indicators to the developmental history of a peatland in the Tiksi Forest, Siberia, Russia. Arctic, Antarctic, and Alpine Research 32: 167–178. Schindler, D. W., K. H. Mills, D. F. Malley, D. L. Findlay, J. A. Shearer, I. J. Davies, M. A. Turner, G. A. Linsey & D. R. Cruikshank, 1985. Long-term ecosystem stress: the effects of years of experimental acidification on a small lake. Science 228: 1395–1401. Stevenson, R. J., M. L. Bothwell & R. L. Lowe (eds), 1996. Algal Ecology: Freshwater Benthic Ecosystems. Aquatic Ecology. Academic Press, San Diego, pp 753. Stevenson, R. J., P. V. McCormick & R. Frydenborg, 2002. Methods for Evaluating Wetland Condition #11: Using Algae to Assess Environmental Conditions in Wetlands. EPA822-R-02–021. USEPA, Washington, DC. Turner, M. A., E. T. Howell, M. Summerby, R. H. Hesslein, D. L. Findlay & M. B. Jackson, 1991. Changes in epilithon and epiphyton associated with experimental acidification of a lake to pH 5. Limnology and Oceanography 36: 1390– 1405.

82 Turner, M. A., M. B. Jackson, D. L. Findlay, R. W. Graham, E. R. DeBruyn & E. M. Vandermeer, 1987. Early Responses of Periphyton to Experimental Lake Acidification. Canadian Journal of Fisheries and Aquatic Sciences 44: 135–149. Turner, M. A., G. G. C. Robinson, B. E. Townsend, B. J. Hann & J. A. Amaral, 1995a. Ecological effects of blooms of filamentous green algae in the littoral zone of an acid lake. Canadian Journal of Fisheries and Aquatic Sciences 52: 2264–2275. Turner, M. A., D. W. Schindler, D. L. Findlay, M. B. Jackson & G. G. C. Robinson, 1995b. Disruption of littoral algal associations by Experimental Lake acidification. Canadian Journal of Fisheries and Aquatic Sciences 52: 2238–2250. van Dam, H. & A. Mertens, 1995. Long-term changes of diatoms and chemistry in headwater streams polluted by atmospheric deposition of sulphur and nitrogen compounds. Freshwater Biology 34: 579–600.

van Dam, H., G. Suurmond & C. J. F. ter Braak, 1981. Impact of acidification on diatoms and chemistry of Dutch moorland pools. Hydrobiologia 83: 425–459. Verb, R. G. & M. L. Vis, 2000. Comparison of benthic diatom assemblages from streams draining abandoned and reclaimed coal mines and nonimpacted sites. Journal of the North American Benthological Society 19: 274–288. Vitt, D. H., 2000. Peatlands: ecosystems dominated by bryophytes. In Shaw A. J. & B. Goffinet (eds), Bryophyte Biology. Cambridge University Press, Cambridge, New York: 476. Yan, N. D., 1979. Phytoplankton community of an acidified, heavy metal-contaminated lake near Sudbury, Ontario: 1973–1977. Water, Air, and Soil Pollution 11: 43–55. Yung, Y., P. Stokes & E. Gorham, 1986. Algae of selected continental and maritime bogs in North America. Canadian Journal of Botany 64: 1825–1833.

Hydrobiologia (2006) 561:83–94  Springer 2006 R.J. Stevenson, Y. Pan, J.P. Kociolek & J.C. Kingston (eds), Advances in Algal Biology: A Commemoration of the Work of Rex Lowe DOI 10.1007/s10750-005-1606-2

Food limitation affects algivory and grazer performance for New Zealand stream macroinvertebrates Joseph R. Holomuzki1,* & Barry J.F. Biggs2 1

Department of Evolution, Ecology, and Organismal Biology, The Ohio State University, 1680 University Drive, Mansfield, OH, 44906, USA 2 National Institute of Water and Atmospheric Research, P.O. Box 8602, Christchurch, New Zealand (*Author for correspondence: Tel: 419-755-4340; Fax: 419-755-4367; E-mail: [email protected])

Key words: herbivory, algal biomass and assemblage, Potamopyrgus antipodarum, Deleatidium spp., Pycnocentrodes aeris, growth, mayfly subimago emergence, caddisfly pre-pupation, grazer activity

Abstract We studied herbivory and grazer performance (i.e., fitness correlates) for the hydrobiid snail Potamopyrgus antipodarum, the leptophlebiid mayfly Deleatidium spp., and the conoesucid caddisfly Pycnocentrodes aeris, common, co-occurring algivores in many New Zealand streams. Grazing effects and costs of coexisting differed among these taxa reared at ambient densities in different combinations in microcosms with algal food conditions (on clay tiles) characteristic of heavily grazed streams. The prostrate diatoms Staurosirella leptostauron, Cymbella novazealandia, and Achnanthidium minutissimum were the dominant algal species on pre- and post-grazed tiles. The relative abundance of erect physiognomic forms, dominated by Synedra ulna and Fragilaria vaucheriae, were 2–3 higher in ungrazed controls and in snail alone treatments than in other grazer treatments. The green filamentous algae Mougeotia sp. and Stigeoclonium lubricum, and the cyanophyte Merismopedia glauca were present only in ungrazed controls. Grazers significantly reduced algal community biomass in treatments by 26–52% relative to controls, except snails alone. Snails (15–30%) burrowed into surrounding sand substrates, dampening their grazing impact on tiles. Caddisflies were more effective than mayflies or snails at removing algae because of higher foraging rates, a larger body size, and an abrasive sand-grained case. Algal biomass reductions did not affect grazer growth. However, pre-pupation rates of caddisflies and emergence rates of subimago mayflies were significantly higher in caddisfly-alone and mayfly-alone treatments, respectively, than in combined-species treatments. These results imply that a limited periphytic food supply (<0.3 mg AFDM cm)2) even over a relatively brief period ( £ 16 d) may have population-scale consequences for co-existing P. aeris and Deleatidium spp.

Introduction Herbivores can impact benthic algal standing crops and community composition in streams. Herbivores can severely reduce algal community biomass and shift taxonomic structure toward low-profile, tightly adhering taxa that are unavailable to many grazers (e.g., Stevenson, 1990; Uehlinger, 1991; Biggs et al., 1998a, b). As a

result, food shortages may occur over small (e.g., a reach) to large (e.g., whole streams) spatial scales (reviewed by Feminella & Hawkins, 1995), and possibly impose limitations on growth, survival, and behaviors of individuals. Herbivory by caddisflies, mayflies, and snails can significantly reduce algal biomass in streams (caddisflies: e.g., Kohler & Wiley, 1997; Anderson et al., 1999; mayflies: e.g., Karouna & Fuller, 1992;

84 Taylor et al., 2002; snails: Steinman et al., 1987; Lowe & Hunter, 1988; Hill et al., 1992a). However, the degree to which these grazer types can reduce their algal food supply depends on many interacting factors, including grazer mobility, feeding rate, density, body size, and mouthpart morphology. For example, some cased, larval caddisflies (Dicosmoecus) even at low densities can exert a greater impact on algal biomass than mayfly larvae and snails because of their relatively high mobility and consumption rates (Lamberti et al., 1987a; DeNicola et al., 1990). Mobile mayflies like Baetis will seek different foraging patches without caddisflies (Glossosoma) when algal resources are reduced (McAuliffe, 1984a, b). However, Baetis can depress algal resources as much as Glossosoma if interpatch movements are restricted (Kohler, 1992). The consequence of low food supplies is slower growth of Glossosoma and poorer survivorship of Baetis. However, snails are often the competitive dominants in streams, where they reach relatively high densities. Radular mouthparts allow snails to feed on low-profile, tightly attached algae growing in understory levels of periphyton mats that are unavailable to many collectorgatherers with bladelike or sweeping mouthparts (i.e., mayflies). Erect, tall-growing algae in the overstory are also dislodged, particularly by large snails (e.g., Elimia, Juga) as they maneuver through the mat, further decreasing feeding opportunities for other grazers. As a result, algal removal by snails can slow the growth and development of competitors (Hill, 1992) and decrease invertebrate diversity and densities in the benthos (Hawkins & Furnish, 1987; Rosemond et al., 1993). Here, we studied herbivory and grazer performance of the hydrobiid snail Potamopyrgus antipodarum (Gray), the leptophlebiid mayfly Deleatidium spp., and the stone-cased, conoesucid caddisfly Pycnocentrodes aeris (Wise). These New Zealand endemics graze periphytic algae (Winterbourn, 1970; Towns, 1983; Biggs and Lowe, 1994; Welch et al., 2000) and commonly co-occur in streams. However, these grazers differ in their mobility, mouthpart morphology, and abundance patterns. Deleatidium is a highly mobile, collector-browser found in a wide range of habitats and lotic systems, whereas Potamopyrgus and Pycnocentrodes are slow-moving scrapers that are generally most abundant in stable to mildly

flooded habitats (Scrimgeour and Winterbourn, 1989; Quinn & Hickey, 1990; Jowett et al., 1991; Death, 1995). However, Deleatidium often coexists with snails and caddisflies in low-order streams with low algal biomass, but does not predominate presumably because it is an inferior competitor for periphytic algae (Biggs et al., 1998b; Broekhuizen et al., 2002). Determining how these species tolerate or respond to low-resource conditions may help us better understand the population dynamics and local abundance patterns of these herbivores. We were particularly interested in learning how Potamopyrgus affected, and responded to, heterospecific grazers. Despite its small size (4–6 mm), P. antipodarum can significantly reduce algal biomass, modify algal assemblage, and displace other macroinvertebrates (Winterbourn & Fegley, 1989; Biggs & Lowe, 1994) in streams where densities exceed 1500 m)2. Potamopyrgus are now also found as a pest species in southeastern Australia (Schreiber et al., 1998, 2003), Europe (Statzner, 1981), the Laurentian Great Lakes (Zaranko et al., 1997), and the western United States (Richards et al., 2001; Kerans et al., 2005), where they are beginning to alter biotic structure and function in some regulated rivers. However, their populations seldom reach nuisance levels in New Zealand. Predation by native and non-native fishes (Sagar & Glova, 1995; Levri, 1998) and flooding (Quinn & Hickey, 1990; Jowett et al., 1991; Holomuzki & Biggs, 1999) appear to regulate population densities. However, food supply may influence population dynamics when shortages impose limitations on growth and survival of individuals, and this influence may be exacerbated when co-existing with grazing caddisflies and mayflies. Here, we examined how different combinations of these grazers affected algal biomass, assemblage, and physiognomy, and whether herbivory effects on algae were related to grazer activity. We also examined grazer growth, mayfly emergence, and caddisfly pre-pupation in single and mixedspecies treatments in microcosms with low algal food supplies characteristic of grazer-controlled systems. We predicted Potamopyrgus would be the least affected by low resource levels primarily because: (1) its rasping radula can collect low profile, adnate periphytic cells, and (2) its food-level requirements are relatively low (Broekhuizen et al., 2002).

85 Methods Experimental design We used 77 four-sided, clear-plastic jars (height 12 cm, mouth diameter 6.0 cm, 400 ml capacity) as rearing chambers (microcosms). Each microcosm contained a pre-colonized, algal-covered ceramic tile (4.74.7 cm), sand, and 150 ml of stream water. Tiles were colonized in a large (12.0 m long, 0.75 m wide), outdoor flume made of transparent acrylic plates supported by a steel frame (see Nikora et al., 1998 for details on flume design and hydraulics). The flume is situated on the banks of the East Branch of the Kaiapoi River at the Silverstream Research Facility 30 km north of Christchurch, New Zealand. The East Branch at Silverstream is hydrologically stable with high nitrogen concentrations (NO3–N: 3800 mg m)3), owing largely to sheep and dairy farms in the Canterbury Plains watershed, and low to moderate phosphorus levels (soluble reactive PO4–P [(SRP]): 2.5 mg m)3). Despite nitrogen loading, grazers appear to regulate algal biomass at Silverstream (Welch et al., 2000). Water in the flume (15–17 C; current: 20 cm s)1) was drawn from the adjacent river by pump and released into the river at the opposite end of the flume. Colonization of 400 tiles started on 23 January 2002, and the first set of 77 randomly selected tiles were removed from the flume 3 weeks later on 12 February. Tiles were picked-free of macroinvertebrates (i.e., chironomid midges [40% Chironomus spp.]) and immediately placed in microcosms after removal from the flume. Each tile, with colonized surface up, was placed on a thin layer (1 cm) of coarse sand collected from the stream. The sand provided a realistic substrate and case-building material for caddisflies. The tile and water (from adjacent river) in each microcosm were replaced on 16, 20, and 24 February. We chose a 4-d replacement period knowing these grazer types can significantly reduce algal biomass in relatively small enclosures in 2– 6 days (DeNicola et al., 1990; Feminella & Resh, 1991; Taylor et al., 2002). Our goal was not to starve animals, but to expose them to a fairly constant low-level food supply like those encountered in grazer-controlled New Zealand streams. Microcosms were aligned in rows of 11 on a Hayter orbital table (45 rotations per min) to

create a circular current (12 cm s)1). The table was next to a south-facing, translucent window to expose organisms to a natural photoperiod (16 h daylight). Treatments consisted of 8 different herbivore combinations: ungrazed controls (U), a snail alone (S), a mayfly alone (M), a caddisfly alone (C), a snail and mayfly (SM), a snail and caddisfly (SC), a mayfly and caddisfly (MC), and a snail, mayfly, and caddisfly (SMC). One, two, and three animals per microcosm correspond to 300, 600, and 900 individuals m)2, respectively. These densities lie within the limits of natural populations of these animals (Jowett et al., 1991; Holomuzki & Biggs, 1999), and density in the S-treatment is similar to that of Potamopyrgus at Silverstream (210 individuals m)2) (Holomuzki, unpublished data). Each treatment with grazers was replicated 10 times, whereas controls were replicated 7 times. All treatments were randomly assigned to microcosms. Animals were collected from the East Branch on 12 February 2002. Each animal was blotted on a paper towel for 5 s and damp mass was recorded to the nearest 0.1 mg on a Mettler AE 200 balance. Sizes of individuals ( x±1SE; snails with shell: 10.84±0.38 mg; mayflies 8.32±0.34 mg; caddisflies with case: 16.31±0.89) reflected average sizes found at Silverstream. We did not select mayfly larvae with large, dark wing-pads that would soon emerge after placement in microcosms. Weighed animals were gently poured from a small container with stream water into microcosms. Dead animals, emerged subimago mayflies, and pre-pupating caddisflies found at the end of each 4-day period were replaced. Caddisflies were considered prepupal if the anterior end of the case was sealed. Replacement animals were weighed prior to stocking as before.

Herbivory effects on algae Tiles removed from microcosms on 16 and 20 February were used to compare algal assemblage and physiognomy among treatments, whereas tiles removed on 24 and 28 February were used to estimate algal biomass in treatments. On each removal date, periphyton from the upper surface of tiles was removed with a razor blade and

86 soft-bristled brush and rinsed with distilled water into whirl-pak bags. Only microcosms in which all animals were alive and had not emerged or prepupated were used for algal assemblage and biomass determinations. Samples used for algal assemblage determinations were taken from 3 to 4 randomly chosen microcosms per treatment on each date, combined into a composite sample, and preserved in a 3% glutaraldehyde solution. Periphyton from three additional tiles removed from the flume on 12 and 16 February, but not placed in microcosms, was also preserved to compare pre- and post-grazed algal assemblages. Distilled water was added to preserved samples before cell enumerations to dilute the glutaraldehyde. Subsamples (0.1 ml water mounts) were taken from uniform suspensions and placed in a Palmer–Maloney counting chamber. At least 300 cells were counted per sample at 400 magnification. Cell counts were converted to relative abundances. Taxa were usually identified to species. Kendall’s coefficient of concordance (W) was used to compare relative abundances of algae among treatments. After identification, algae were categorized by physiognomy. Prostrate algae are those that grow with their long axis on the substrate, erect algae are non-branching with their long axis directed toward the water column, and stalked algae are multicellular with limited branches attached to a mucilaginous stalk (Lowe & LaLiberte, 1996). G-tests for goodness of fit with William’s correction (Sokal & Rohlf, 1995) were used to compare percentages of physiognomic forms across treatments. Samples used to estimate algal biomass were frozen and processed for ash-free dry mass (AFDM) within 24 h of collection. A General Linear Model (SYSTAT version 7.0) was used to compare AFDM among grazer treatments on both dates. Row was entered into the model as a blocking factor to examine possible light (window) effects on AFDM. Posthoc pairwise comparisons with p-values adjusted using the Bonferroni method (SYSTAT version 7.0) followed significant grazer treatment effects.

Grazer performance On 28 February, damp masses of all animals in microcosms were re-weighed. Growth for each

species was determined by subtracting initial mass at placement from final mass at the end of the experiment, or at pre-pupation, and dividing by the number of days the animal was in the microcosm. Thus, growth was expressed as mg damp mass day)1. ANOVA tested whether growth rates (log x+1 transformed) of each species differed among treatments. G-tests were used to determine whether percent mayfly emergence and mortality, and caddisfly pre-pupation, varied among treatments. We recorded grazing activity of animals in each microcosm to explain possible variation in algal biomass or assemblage among treatments. Animals were scored as actively grazing if seen on the upper surface of a tile during snapshot observations (3–5 s per microcosm). Observations were done diurnally (1400–1630 h) and nocturnally (2230–2330 h) every 2 days during each 4-day period. Nocturnal observations were made with a headlamp with a red filter to minimize photoreactive responses by animals (particularly mayflies). We ran repeated-measures (rm) ANOVAs to compare percent grazer activity (arcsine-transformed) of each taxon across treatments during day and night on observation dates. Diurnal and nocturnal activity were modeled as the within subjects (repeated) effect (SYSTAT version 7.0). We also counted the number of snails that burrowed into the sand at the end of each 4-day period to help explain possible grazing differences among treatments and to quantify snail responses to heterospecifics. Numbers of snails that burrowed in snail-containing treatments at the end of each 4-day period were analyzed using a Median test (Conover, 1980). The Median test evaluates differences among treatment medians, and is applicable to categorical and nonnormal data (Sokal & Rohlf, 1995).

Results Algal composition, physiognomy, and biomass Periphyton communities on pre-grazed tiles placed in microcosms on 12 and 16 February were dominated by prostrate diatoms (65%), particularly Staurosirella leptostauron, Cymbella novazealandia, and Achnanthidium minutissimum. On 12

87 physiognomic forms, dominated by Synedra ulna and Fragilaria vaucheriae, differed among treatments on 16 February (G=18.592, df=7, p<0.01), being 2–3 higher in ungrazed controls (U) and S-treatments than in other treatments (Fig. 1). Relative abundance of erect forms did not differ among treatments on 20 February (G=9.868, df=7, p>0.10). Similarly, stalked forms, dominated by Gomphonema parvulum, were equally abundant among treatments (16 February: G=10.100, df=7, p>0.10; 20 February: G=5.296, df=7, p>0.50). Mougeotia sp. and Merismopedia glauca (2%), and Stigeoclonium lubricum (2.6%), were found only in ungrazed controls on 16 and 20 February, respectively. Algal community biomass significantly differed among grazing treatments on both dates (24 February: F7,38=11.310, p<0.001; 28 February: F7,39=7.612, p<0.001). Neither snails nor mayflies alone significantly reduced algal biomass relative to controls on 24 February (Fig. 2a). Algal biomass reductions were again not detected in the

February, the filamentous green alga Mougeotia sp. and the cyanophyte Merismopedia glauca comprised <1% and 2%, respectively, of the pre-grazed community. On 16 February, the filamentous chlorophytes Ulothrix aequalis and Stigeoclonium lubricum comprised 11% and 6% of the pre-grazed algal community, respectively. Diatoms comprised ‡97% of all algae in grazed treatments on 16 and 20 February. Dominance rankings of the 10 most common diatoms were highly concordant among treatments on both dates (16 February: W=0.785, p<0.001; 20 February: W=0.858, p<0.001). The prostrate species Staurosirella leptostauron, Cymbella novazealandia, and Achnanthidium minutissima comprised  52–65% of the diatom communities across treatments. Prostrate forms dominated the diatom communities on both dates (‡67%, Fig. 1), and percent composition of prostrate forms did not differ among treatments (16 February: G=6.618, df=7, p>0.25; 20 February: G=3.352, df=7, p>0.75). However, the relative abundance of erect

Prostrate

Erect

Stalked

Form composition (%)

100

80

60

40

20

0 U

S

M

C

SM

SC

MC

SMC

Treatments Figure 1. Mean % composition of prostrate, erect, and stalked diatoms in grazing treatments (U=ungrazed controls; S=a snail alone; M=a mayfly alone; C=a caddisfly alone; SM=a snail and mayfly; SC=a snail and caddisfly; MC=a mayfly and caddisfly; SMC=a snail, mayfly, and caddisfly) (N=2 sample dates [16 and 20 February] per treatment).

88

(a)

0.4

24 February a ac

0.3

abc 0.2

AFDM (mg/cm2)

bc

b

b

b

b

0.1

0.0

U

S

M

C

SM

SC

CM

SMC

(b)

0.3

a

ac

28 February bc b

0.2

b

b

b b

0.1

0.0

U

S

M

C

SM

SC

CM SMC

Treatments Figure 2. Mean (+1SE) algal biomass as determined by AFDM in grazer treatments on (a) 24 and (b) 28 February. Lowercase letters over bars indicate significantly different means using Bonferroni pairwise comparisons. N=7 and 10 for ungrazed controls (U) and snails alone (S), respectively, on each date. However, sample sizes in other treatments on both dates varied from 4 to 9, depending on mayfly mortality and emergence and caddisfly pre-pupation. See ‘Methods’ or Figure 1 legend for meanings of other grazer treatment abbreviations.

S-treatment on 28 February (Fig. 2b). Caddisflies were more effective than snails and mayflies at removing algae on both dates. Moreover, algal biomass did not differ among caddisfly-containing treatments, indicating added snails and/or mayflies had no detectable impact on algae (Fig. 2a, b). Row (i.e., positioning) had no affect on AFDM on both dates (both p ‡ 0.23).

Grazer performance Average snail growth in treatments ranged from 0.092 to 0.210 mg d)1, but did not significantly differ among treatments (F3,36=0.63, p=0.600). Only 1 snail died during the experiment. Average caddisfly growth was 3–4 greater than that of snails, ranging from 0.410 to 0.637 mg d)1. Like

89 snails, caddisfly growth was similar among treatments (F3,36=1.06, p=0.376). However, prepupation rates significantly differed among caddisfly-containing treatments (G=45.01, df=3, p<0.001), being highest in the C-treatment (Fig. 3a). Mass at pre-pupation, however, did not differ significantly between caddisflies in the Ctreatment ( x ± 1SE: 6.12 ± 0.37 mg without case, n=6) and the other caddisfly-containing treatments (7.69 ± 0.75 mg, n=6) (F1,10=2.86, p=0.122). No caddisflies died during the experiment. Like snails and caddisflies, no differences in mayfly growth were detected among treatments

(F3,18=0.839, p=0.49). Average growth was 0.158 mg d)1 (1SE=0.023, n=22). No mayflies died between 12 and 16 February. Thereafter,  30% of mayflies in mayfly-containing treatments died each 4-day period. Death rates over these periods did not differ among treatments (G=1.128, df=3, p>0.75) and were likely due to handling. However, subimago emergence rates were significantly higher in the M-treatment than in other mayfly-containing treatments (G=8.13, df=3, p<0.05) (Fig. 3b). Grazing activity differed significantly among taxa (F2,189=222.28, p<0.001). Caddisflies were

Pre-pupation (%)

(a) 80

60

40

20

0

C

CM

SC

M

SM

CM

SMC

Emerged (%)

(b) 50 40 30 20 10 0 SMC

Treatments Figure 3. Percent (a) pre-pupation of caddisflies and (b) emergence of subimago mayflies in grazer treatments.

90 found on tiles  6 more often than the other two grazers (Fig. 4). Grazing activity by each taxon was similar among treatments (rm ANOVA: all p ‡ 0.35), indicating heterospecifics did not influence feeding activity. Aggressive interactions were not observed in any mixed-species treatment. Grazing activity by caddisflies and snails was similar between day and night (both p ‡ 0.148), but mayflies tended to spend more time grazing at night than during the day (F1,21=3.876, p=0.062; Fig. 4). On average, 9 ( 23%) snails were found beneath the sand in all treatments each 4-day period. Burrowing activity tended to be higher in the SMC-treatment (30%) than in other snail-containing treatments (15–25%), but not significantly (Median test: v2=7.177, df=3, p>0.05).

Discussion Algal responses to herbivory Even though grazer effects tend to be underestimated in small-scale experiments (Kohler & Wiley, 1997; Taylor et al., 2002), our results strongly

suggest P. aeris and Deleatidium spp. at ambient densities can depress algal biomass, even in the low food conditions of our study (<0.3 mg AFDM cm)2). Caddisflies and mayflies significantly reduced community algal biomass by 26–52% over 4 d periods. Other studies have shown that these grazer types can significantly reduce algal biomass over similar durations when dispersal is restricted (Feminella & Resh, 1991; Anderson et al., 1999; Taylor et al., 2002). We also found that caddisflies were more effective at removing algal biomass than mayflies, probably because of their higher grazing activity, scraping mouthparts, and abrasion from case drag when foraging. Likewise, Deleatidium spp. were more effective than Potamopyrgus at reducing algal biomass. Welch et al. (2000) reported similar results, but found grazing rates to vary widely for both taxa. The negligible effect Potamopyrgus had on algal community biomass was unexpected, considering its ability to act as a keystone species in some habitats and the results of some other snail grazing experiments. Hill et al. (1992a), Rosemond (1993), and Rosemond et al. (2000) reported Elimia clavaeformis to reduce algal biomass by >45% in  3–7 week periods. The 4 d turnover period of

100

Grazing activity (%)

80

Day a

Night 60

40 b b

20

0

Caddis

Snail

Mayfly

Taxa Figure 4. Mean (+1SE) percent grazing activity of caddisflies, snails, and mayflies based on daytime and nighttime snapshot observations. Lowercase letters over bars indicate significantly different means of overall activity using Bonferroni pairwise comparisons. N=8 observations for each taxon and time.

91 tiles in our experiment may have been too short to detect significant grazing effects by Potamopyrgus. However, Vaughn et al. (1993) also found that ambient densities of Physella virgata, another small snail, did not significantly decrease algal biomass even after 35 d. Small body size, combined with low energetic requirements (Broekhuizen et al., 2002), likely partly explain the nonsignificant effects of Potamopyrgus on algal biomass. However, burrowing activity and use of alternative food sources may also partly account for this result. On average, 23% of snails burrowed into surrounding sand substrates each 4 d period. Potamopyrgus apparently commonly reside in subsurface sediments in real streams (Statzner, 1981), where they consume epipsammic algae and heterotrophic biofilms (bacteria and fungi) (Rounick & Winterbourn, 1983). Potamopyrgus may have used these alternative food sources in microcosms, dampening their grazing impacts on periphyton-covered tiles. Reductions in community algal biomass only slightly affected algal community composition and physiognomy, probably because prostrate diatoms were the dominant algae on pre- and post-grazed tiles. Studies reporting assemblage changes from grazing often show these changes result from declines in overstory forms and loosely attached microalgae (e.g., Colletti et al., 1987; Steinman et al., 1987; Lowe & Hunter, 1988; Feminella & Resh, 1991; Tuchman & Stevenson, 1991). Both forms were uncommon on our replacement tiles, perhaps because of grazing by chironomid midges during colonization. However, the relative frequency of erect, physiognomic forms of diatoms was significantly lower in our mayfly- and caddisfly-containing treatments compared to ungrazed controls. Moreover, all three grazer types significantly reduced the abundance of filamentous green algae, corroborating the results of similar studies (Jacoby, 1987; Lamberti et al., 1987a; Walton et al., 1995; Anderson et al., 1999; Opsahl et al., 2003). Though rare, the filamentous green algae Mougeotia sp. and Stigeoclonium lubricum, and the cyanophyte Merismopedia glauca, were found only in our ungrazed treatments. Welch et al. (2000) similarly found Potamopyrgus antipodarum, Deleatidium spp., and Pycnocentrodes aeris to effectively remove filamentous chlorophytes in experimental streams.

Grazer performance costs Our experimental food levels reflect the low algal biomass conditions in heavily grazed New Zealand streams (e.g., Biggs et al., 2000; Welch et al., 2000; Broekhuizen et al., 2002), where densitydependent interactions among grazers are predicted to occur. Subimago mayfly emergence and caddisfly pre-pupation were significantly higher in single than in mixed-species treatments, results that suggest food limitation in our microcosms. Feminella and Resh (1990) similarly found pupation rate and pupal size of the caddisfly Helicopsyche borealis to decrease with algal biomass. Slowed developmental rates may prolong the vulnerability of larval stages to aquatic predators (e.g., Bowlby & Roff, 1986) and to physical disturbance (flooding and drought) (e.g., Williams, 1996; Holomuzki, 1997) and delay adult reproduction. Sizes of pre-pupating caddisflies were similar in our caddisfly-containing treatments, likely owing to the similar growth rates among treatments. However, studies finding food shortage effects on growth also show that food depletion is a density-dependent response occurring over 3–7 weeks (caddisflies: Lamberti et al., 1987b; Hill & Knight, 1988; Kohler, 1992; mayflies: Hill & Knight, 1987; snails: Rosemond, 1993; Hill et al., 1992b). Tiles (i.e., periphyton) in our treatments were periodically replaced when food levels reached 0.1–0.2 mg AFDM cm)2 in caddisfly- and mayfly-containing treatments. Moreover, even our SMC-treatment contained densities within the range of natural densities of these organisms. Scrimgeour et al. (1991) suggest that mayflies are capable of consuming diatoms even at very low biomass levels (0.11 mg cm)2). However, our results suggest these grazers will incur performance costs if these food conditions persist even over relatively brief periods ( £ 16 d). The finding that emergence rates were highest in mayfly-alone treatments adds support to the hypotheses of Biggs et al. (1998b) and Broekhuizen et al. (2002) that food limitation may regulate Deleatidium spp. more than Potamopyrgus. Deleatidium spp. is ubiquitous in New Zealand streams, but predominates in large rivers and in low-order, high gradient streams where flooding is frequent and intense (Quinn & Hickey, 1990; Harding et al., 1997). Its dominance in these

92 systems stems from continuous breeding and hatching over an annual cycle, and high recruitment, dispersal, and death rates (Winterbourn, 1974; Armstrong, 1996), characteristics of ‘‘ruderal’’ taxa with limited competitive abilities (sensu Grime, 1974). In contrast, Potamopyrgus and Pycnocentrodes spp. dominate in low gradient streams with predominantly grassland (open canopy) riparian areas (Quinn & Hickey, 1990; Harding et al., 1997). Allochthonous input may be greatly reduced in these non-forested systems, preventing Deleatidium spp. from using fine detrital material as an alternative food source (Winterbourn, 1974), and increasing the likelihood of food limitation as a population regulator. Finally, food limitation may be exacerbated in nutrientpoor streams that constrain periphyton productivity and in streams with high densities of grazing chironomids that compete for food and predaceous fish (e.g., trout) that can restrict macrograzer foraging activity (McIntosh & Townsend, 1994). Results from our tile colonization period and those of Biggs et al. (2000) suggest chironomids can significantly depress periphyton biomass in New Zealand streams. It remains unclear how important food limitation is in controlling P. antipodarum. Nutrient levels and food availability play some role in controlling populations, considering the species can thrive in streams draining catchments with intense human land use (e.g., grazing, urbanization, forestry) (Schrieber et al., 2003). Food limitation may be particularly important in bedrock-dominated streams with limited hyporheic areas that prevent snails from using alternative food sources in subsurface sediments. Understanding how food supply affects local population dynamics of these snails is critical given their ability to dramatically alter biotic structure and function in some habitats. Considering Potamopyrgus, Deleatidium, and Pycnocentrodes are common inhabitants of many New Zealand streams, careful manipulations that separate the effects of food limitation, flow disturbance, and physical habitat features (i.e., riparian cover, bedform, sediment depth) and their interactions will contribute to our understanding of population regulation and distributions of these taxa, and of the importance of herbivory in streams outside north-temperate latitudes.

Acknowledgements We thank Steve Francoeur for help colonizing tiles and starting the experiment, Rex Lowe and Jennifer Ress for algal enumerations, Jan Stevenson and Steve Kohler for constructive editorial comments, and Alan Stokes for his technical design work on the flume. This study was funded by an International Studies Grant from The Ohio State University to JRH and grants C01X0813 and C01X0308 (Water allocation: effects on instream values) from the New Zealand Foundation for Research, Science and Technology to BJFB.

References Anderson, E. L., G. M. Schimek, E. B. Welch, J. M. Jacoby & R. R. Horner, 1999. Periphyton removal related to phosphorus and grazer biomass level. Freshwater Biology 41: 633–651. Armstrong, P.M., 1996. The effects of stream bed stability and flow variability on the production of aquatic invertebrates. MSc. Thesis, University of Canterbury, Christchurch, New Zealand. Biggs, B. J. F., S. N. Francoeur, A. D. Huryn, R. Young, C. J. Arbuckle & C. R. Townsend, 2000. Trophic cascades in streams: effects of nutrient enrichment on autotrophic and consumer benthic communities under two different fish predation regimes. Canadian Journal of Fisheries and Aquatic Science 57: 1380–1394. Biggs, B. J. F., C. Kilroy & R. L. Lowe, 1998a. Periphyton development in three valley segments of a New Zealand grassland river: test of a habitat matrix conceptual model within a catchment. Archiv fu¨r Hydrobiologie 143: 147–177. Biggs, B. J. F. & R. L. Lowe, 1994. Responses of two trophic levels to patch enrichment along a New Zealand stream continuum. New Zealand Journal of Marine and Freshwater Research 28: 119–134. Biggs, B. J. F., R. J. Stevenson & R. L. Lowe, 1998b. A habitat matrix conceptual model for stream periphyton. Archiv fu¨r Hydrobiologie 143: 21–56. Bowlby, J. N. & J. C. Roff, 1986. Trophic structure in southern Ontario streams. Ecology 67: 1670–1679. Broekhuizen, N., S. Parkyn, D. Miller & R. Rose, 2002. The relationship between food density and short term assimilation rates in Potamopyrgus antipodarum and Deleatidium sp. Hydrobiologia 477: 181–188. Colletti, P. J., D. W. Blinn, A. Pickart & V. T. Wagner, 1987. Influence of different densities of the mayfly grazer Heptagenia criddlei on lotic diatom communities. Journal of the North American Benthological Society 6: 270–280. Conover, W. J., 1980. Practical Nonparametric Statistics 2nd edn. John Wiley, Sons, New York.

93 Death, R. G., 1995. Spatial patterns in benthic invertebrate community structure: products of habitat stability or are they habitat specific?. Freshwater Biology 33: 455–467. DeNicola, D. M., C. D. McIntire, G. A. Lamberti, S. V. Gregory & L. R. Ashkenas, 1990. Temporal patterns of grazer–periphyton interactions in laboratory streams. Freshwater Biology 23: 475–489. Feminella, J. W. & C. P. Hawkins, 1995. Interactions between stream herbivores and periphyton: a quantitative analysis of past experiments. Journal of the North American Benthological Society 14: 465–509. Feminella, J. W. & V. H. Resh, 1990. Hydrologic influences, disturbance, and intraspecific competition in a stream caddisfly population. Ecology 71: 2083–2094. Feminella, J. W. & V. H. Resh, 1991. Herbivorous caddisflies, macroalgae, and epilithic microalgae: dynamic interactions in a stream grazing system. Oecologia 87: 247–256. Grime, J. P., 1974. Vegetation classification by reference to stratigies. Nature 250: 26–31. Harding, J. S., M. J. Winterbourn & W. F. McDiffett, 1997. Stream faunas and ecoregions in South Island, New Zealand: do they correspond? Archiv fu¨r Hydrobiologie 140: 289–307. Hawkins, C. P. & J. K. Furnish, 1987. Are snails important competitors in stream ecosystems?. Oikos 49: 209–220. Hill, W. R., 1992. Food limitation and interspecific competition in snail-dominated streams. Canadian Journal of Fisheries and Aquatic Sciences 49: 1257–1267. Hill, W. R., H. L. Boston & A. D. Steinman, 1992a. Grazers and nutrients simultaneously limit lotic primary productivity. Canadian Journal of Fisheries and Aquatic Sciences 49: 504–512. Hill, W. R. & A. W. Knight, 1987. Experimental analysis of the grazing interaction between a mayfly and stream algae. Ecology 68: 1955–1965. Hill, W. R. & A. W. Knight, 1988. Concurrent grazing effects of two stream insects on periphyton. Limnology and Oceanography 33: 15–26. Hill, W. R., S. C. Weber & A. J. Stewart, 1992b. Food limitation of two lotic grazers: quantity, quality, and size-specificity. Journal of the North American Benthological Society 11: 420–432. Holomuzki, J. R., 1997. Habitat-specific life histories and foraging by stream-dwelling American toads. Herpetologica 53: 445–453. Holomuzki, J. R. & B. J. F. Biggs, 1999. Distributional responses to flow disturbance by a stream-dwelling snail. Oikos 87: 36–47. Jacoby, J. M., 1987. Alterations in periphyton characteristics due to grazing in a Cascade foothill stream. Freshwater Biology 18: 495–508. Jowett, I. G., J. Richardson, B. J. F. Biggs, C. W. Hickey & J. M. Quinn, 1991. Microhabitat preferences of benthic invertebrates and the development of generalized Deleatidium spp. habitat suitability curves, applied to four New Zealand streams. New Zealand Journal of Marine and Freshwater Research 25: 187–199. Karouna, N. K. & R. L. Fuller, 1992. Influence of four grazers on periphyton communities associated with clay tiles and leaves. Hydrobiologia 245: 53–64.

Kerans, B. L., M. F. Dybdahl, M. M. Gangloff & J. E. Jannot, 2005. Potamopyrgus antipodarum: distribution, density, and effects on native macroinvertebrate assemblages in the Greater Yellowstone Ecosystem. Journal of the North American Benthological Society 24: 123–138. Kohler, S. L., 1992. Competition and the structure of a benthic stream community. Ecological Monographs 62: 165–188. Kohler, S. L. & M. J. Wiley, 1997. Pathogen outbreaks reveal large-scale effects of competition in stream communities. Ecology 78: 2164–2176. Lamberti, G. A., L. R. Ashkenas, S. V. Gregory & A. D. Steinman, 1987a. Effects of three herbivores on periphyton communities in laboratory streams. Journal of the North American Benthological Society 6: 92–104. Lamberti, G. A., J. W. Feminella & V. H. Resh, 1987b. Herbivory and intraspecific competition in a stream caddisfly population. Oecologia 73: 75–81. Levri, E. P., 1998. The influence of non-host predators on parasite-induced behavioral changes in a freshwater snail. Oikos 81: 531–537. Lowe, R. L. & R. D. Hunter, 1988. Effects of grazing by Physa integra on periphyton community structure. Journal of the North American Benthological Society 5: 221–229. Lowe, R. L. & G. D. LaLiberte, 1996. Benthic stream algae: distribution and structure. In Hauer, F. R. & G. A. Lamberti (eds), Methods in Stream Ecology. Academic Press, San Diego. McAuliffe, J. R., 1984a. Resource depression by a stream herbivore: effects on distributions and abundances of other grazers. Oikos 42: 327–333. McAuliffe, J. R., 1984b. Competition for space, disturbance, and the structure of a benthic stream community. Ecology 65: 894–908. McIntosh, A. R. & C. R. Townsend, 1994. Interpopulation variation in mayfly anti-predator tactics: differential effects of contrasting predatory fish. Ecology 75: 2078–2090. Nikora, V. I., D. G. Goring & B. J. F. Biggs, 1998. Silverstream eco-hydraulics flume: hydraulic design and tests. New Zealand Journal of Marine and Freshwater Research 32: 607– 620. Opsahl, R. W., T. Wellnitz & N. L. Poff, 2003. Current velocity and invertebrate grazing regulate stream algae: results of an in situ electrical exclusion. Hydrobiologia 499: 135–145. Quinn, J. M. & C. W. Hickey, 1990. Magnitude of effects of substrate particle size, recent flooding, and catchment development on benthic invertebrates in 88 New Zealand rivers. New Zealand Journal of Marine and Freshwater Research 24: 387–410. Richards, D. C., L. D. Cazier, B. L. Kerans & G. T. Lester, 2001. Competition between the invasive New Zealand mud snail, Potamopyrgus antipodarum and the threatened Bliss Rapids snail, Taylorconcha serpenticola. Bulletin of the North American Benthological Society 18: 271. Rosemond, A. D., 1993. Interactions among irradiance, nutrients, and herbivores constrain a stream algal community. Oecologia 94: 585–594. Rosemond, A. D., P. J. Mulholland & S. H. Brawley, 2000. Seasonally shifting limitation of stream periphyton: response of algal populations and assemblage biomass and

94 productivity to variation in light and nutrients. Canadian Journal of Fisheries and Aquatic Sciences 57: 66–75. Rosemond, A. D., P. J. Mulholland & J. W. Elwood, 1993. Top-down and bottom-up control of stream periphyton: effects of nutrients and herbivores. Ecology 74: 1264–1280. Rounick, J. S. & M. J. Winterbourn, 1983. Leaf processing in two contrasting beech forest streams: effects of physical and biotic factors on leaf breakdown. Archiv fu¨r Hydrobiologie 96: 448–474. Sagar, P. M. & G. J. Glova, 1995. Prey availability and diet of juvenile brown trout (Salmo trutta) in relation to riparian willows (Salix spp.) in three New Zealand streams. New Zealand Journal of Marine and Freshwater Research 29: 527–537. Schrieber, E. S. G., A. Glaister, G. P. Quinn & P. S. Lake, 1998. Life history and population dynamics of the exotic snail Potamopyrgus antipodarum (Prosobranchia: Hydrobiidae) in Lake Purreumbete, Victoria, Australia. Australian Journal of Marine and Freshwater Research 49: 73–78. Schreiber, E. S. G., G. P. Quinn & P. S. Lake, 2003. Distribution of an alien aquatic snail in relation to flow variability, human activities and water quality. Freshwater Biology 48: 951–961. Scrimgeour, G. J., J. M. Culp, M. L. Bothwell, F. J. Frona & M. H. McKee, 1991. Mechanisms of algal patch depletion: importance of consumptive and non-consumptive losses in mayfly-diatom systems. Oecologia 85: 343–348. Scrimgeour, G. J. & M. J. Winterbourn, 1989. Effects of floods on epilithon and benthic macroinvertebrate populations in an unstable New Zealand river. Hydrobiologia 171: 33–44. Sokal, R. R. & F. J. Rohlf, 1995. Biometry 3rd edn. W. H. Freeman and Company, New York. Statzner, B., 1981. The relationship between ‘‘hydraulic stress’’ and microdistributions of benthic macroinvertebrates in a lowland running water system, the Schierensebrooks (North Germany). Archiv fu¨r Hydrobiologie 91: 192–218. Steinman, A. D., C. D. McIntire, S. V. Gregory, G. A. Lamberti & L. Ashkenas, 1987. Effect of herbivore type and density on taxonomic structure and physiognomy of algal assemblages in laboratory streams. Journal of the North American Benthological Society 6: 175–188. Stevenson, R. J., 1990. Benthic algal community dynamics in a stream during and after a spate. Journal of the North American Benthological Society 9: 277–288.

Taylor, B. W., A. R. McIntosh & B. L. Peckarsky, 2002. Reachscale manipulations show invertebrate grazers depress algal resources in streams. Limnology and Oceanography 47: 895– 899. Towns, D. R., 1983. Life history patterns of six sympatric species of Leptophlebiidae (Ephemeroptera) in a New Zealand stream and the role of interspecific competition in their evolution. Hydrobiologia 99: 37–50. Tuchman, N. C. & R. J. Stevenson, 1991. Effects of selective grazing by snails on benthic algal communities. Journal of the North American Benthological Society 10: 430–443. Uehlinger, U., 1991. Spatial and temporal variability of the periphyton biomass in a prealpine river (Necker, Switzerland). Archiv fu¨r Hydrobiologie 123: 219–237. Vaughn, C. C., F. P. Gelwick & W. J. Matthews, 1993. Effects of algivorous minnows on production of grazing stream invertebrates. Oikos 66: 119–128. Walton, S. P., E. B. Welch & R. R. Horner, 1995. Stream periphyton response to grazing and changes in phosphorus concentration. Hydrobiologia 302: 31–46. Welch, E. B., E. L. Anderson, J. M. Jacoby, B. J. F. Biggs & J. M. Quinn, 2000. Invertebrate grazing of filamentous green algae in outdoor channels. Verhandlungen der Internationalen Vereinigung fu¨r theoretische und angewandte Limnologie 27: 2408–2414. Williams, D. D., 1996. Environmental constraints in temporary fresh waters and their consequences for the insect fauna. Journal of the North American Benthological Society 15: 634–650. Winterbourn, M. J., 1970. The New Zealand species of Potamopyrgus (Gastropoda: Hydrobiidae). Malacologia 10: 283–321. Winterbourn, M. J., 1974. The life histories, trophic relations and production of Stenoperla prasina (Plecoptera) and Deleatidium sp. (Ephemeroptera) in a New Zealand river. Freshwater Biology 4: 507–524. Winterbourn, M. J. & A. Fegley, 1989. Effects of nutrient enrichment and grazing on periphyton assemblages in some spring-fed, South Island streams. New Zealand Natural Sciences 16: 57–65. Zaranko, D. T., D. G. Farara & F. G. Thompsen, 1997. Another exotic mollusk in the Laurentian Great Lakes: the New Zealand native Potamopyrgus antipodarum (Gray) 1943) (Gastropoda, Hydrobiidae). Canadian Journal of Fisheries and Aquatic Sciences 54: 809–814.

Hydrobiologia (2006) 561:95–110  Springer 2006 R.J. Stevenson, Y. Pan, J.P. Kociolek & J.C. Kingston (eds), Advances in Algal Biology: A Commemoration of the Work of Rex Lowe DOI 10.1007/s10750-005-1607-1

Benthic diatom communities in subalpine pools in New Zealand: relationships to environmental variables Cathy Kilroy1,*, Barry J. F. Biggs1, Wim Vyverman2 & Paul A. Broady3 1

NIWA, 8602, Christchurch, New Zealand Laboratory of Protistology and Aquatic Ecology, Department of Biology, University of Gent, Krijgslaan 281 – S8, B-9000, Gent, Belgium 3 School of Biological Sciences, University of Canterbury, Christchurch, New Zealand (*Author for correspondence: E-mail: [email protected]) 2

Key words: diatom community composition, subalpine pools, pH gradient, BIO-ENV analysis, endemism

Abstract In spite of their potential use as indicators of both present and past environmental conditions, little is known about the diatom communities in the many small water bodies at high altitudes in New Zealand. We sampled benthic diatoms at 20 sites in a typical subalpine mire pool/tarn complex near Arthur’s Pass in South Island, New Zealand in the austral spring 2001. The aims were to characterise the diatom communities, including identification of a possible endemic component, and to investigate relationships with environmental variables. The community at genus level was consistent with the peat-bog diatom flora reported from elsewhere except for the common occurrence of the Tasmania/New Zealand endemic genus Eunophora. At the species level, 27 of the 52 most common taxa appear to correspond to known species from the Northern Hemisphere and are therefore presumed to be cosmopolitan in their distribution. Just two taxa are known from the Southern Hemisphere only, however identification of the remaining common species proved problematic. Analysis using the BIO-ENV procedure of the PRIMER computer program confirmed an expected strong association between diatom community composition and pH, with water conductivity and gilvin also important. Weighted averaging regression and cross-validation using C2 software enabled selection of four diatom species as potentially sensitive indicators of certain pH levels. Neither species of Eunophora showed a strong preference for pH or for any of the other environmental variables measured, indicating that other factors are determining their distributions. The strength of the species–environment relationships found in this small survey suggests good potential for monitoring current conditions and for palaeoecological applications. Extension of the dataset with information from other alpine/subalpine areas is desirable, as is the compilation of a regional diatom identification guide for these habitats.

Introduction Diatom communities respond quickly to environmental changes because of their short life-cycles, rapid dispersal and colonisation and the large number of species with differing tolerances to physical and chemical variables (Lotter et al., 1999), notably pH (Battarbee et al., 1999).

Consequently diatoms are well-established tools both for monitoring present changes in water condition (Stevenson & Pan, 1999; Winter & Duthie, 2000), and for tracking past environmental changes inferred from communities preserved in sediments (Moser et al., 1996; Smol & Cumming, 2000). High-latitude and high-altitude aquatic ecosystems have been identified as being

96 especially sensitive to environmental changes (Pienitz & Smol, 1993; Vincent & Pienitz, 1996; Sommaruga-Wograth et al., 1997; Smol & Cumming, 2000), and over the past 10–15 years there have been numerous studies to establish diatom community composition and environmental preferences in lakes and ponds in the Northern Hemisphere, especially in arctic regions (e.g. Douglas & Smol, 1995; Fallu et al., 2000; Laing & Smol, 2000; Lim et al., 2001; Michelutti et al., 2003). In the Southern Hemisphere, however, equivalent information is still relatively sparse, though comprehensive data are available for Tasmanian and Antarctic lakes (Vyverman et al., 1995; Verleyen et al., 2003 and references therein). In New Zealand, small lakes, tarns and mire pools are characteristic features of alpine and subalpine landscapes, especially in the Southern Alps, South Island. Recent estimates from the NZMS 1:50 000 map database (Land Information New Zealand) indicate that in the South Island there are over 7500 small lakes and ponds (<10 ha area) located at over 600 m a.s.l. (H. Hurren, pers. comm.). Despite their prominence in the landscape and potential biodiversity and conservation values, small subalpine/alpine waterbodies have been the focus of surprisingly few studies. Taxonomic work over the past few years in the Australasian region suggests that these habitats support significant numbers of diatom taxa that appear to be endemic to the Southern Hemisphere (Vyverman et al., 1997, 1998; Sabbe et al., 2001; Kilroy et al., 2003), and for which only general autecological data are available. Thus, to understand the biodiversity of high-altitude diatom communities and to exploit their value as indicators of environmental changes and of natural habitats in this region (Kociolek & Stoermer, 2001), more information is required on community composition and the environmental conditions that influence individual taxa. Data from unmodified areas are especially important in order to provide a reference point in future studies of both human impacts and climate change (Battarbee et al., 1997). Analysis of data from a survey in January 2001 of 71 small lakes, tarns and mire pools in predominantly unimpacted South Island catchments at altitudes up to 1520 m a.s.l. characterised their general limnology and benthic diatom community composition at the genus level (K. Vanhoutte,

pers. comm.). The present investigation focuses on a system of subalpine pools, which includes two of the sites sampled in the earlier survey. The aims of this more detailed study were: 1) to characterise diatom community composition in a typical unimpacted supalpine mire pool complex in New Zealand, including recognition of potential Southern Hemisphere endemic taxa; 2) to investigate the importance of a range of environmental variables in explaining the distributions of diatom communities and species over a relatively small area without the confounding large-scale influences of climate and geology; and 3) to identify diatom species that could potentially be used as indicators of specific water chemistry conditions in these environments. We also aimed to obtain preliminary information on the environmental preferences of species in the known endemic genus Eunophora (Vyverman et al., 1998)

Site description The investigation was conducted in a wetland area approximately 1000 m  200 m occupying a shelf on the northern flank of Bealey Spur (43 02¢ S, 171 35¢ E) near Arthur’s Pass, South Island, New Zealand, at about 1030 m a.s.l. (Fig. 1). The locality is typical of the many small tarn/mire pool complexes that occur throughout the Southern Alps, generally in basins originally shaped by glaciation (Lowe & Green, 1987). Bealey Spur lies to the east of the main divide of the Southern Alps and runs roughly east to west. Basement geology is greywacke (hard siliceous sedimentary rock laid down in the Jurassic/Triassic), which underlies much of the northern section of the Southern Alps. Acid ‘brown soils’ are typical of the eastern ranges of this part of the Southern Alps (Morgan, 2001), however, soils in the vicinity of the pools sampled were largely organic. Annual rainfall in the area is approximately 2000 mm. Until the late 1970s Bealey Spur, including the wetland area, was pastoral leasehold land and there was low-density sheep grazing in the adjacent subalpine tussock land up to about 1800 m a.s.l. Nevertheless, there is little evidence of attempts to ‘improve’ the land for grazing by planting exotic grass and herbaceous species, as was common practice in other subalpine grassland areas in New

97

Figure 1. Sketch map of the study site at Bealey Spur wetland area, near Arthur’s Pass, South Island, New Zealand, showing general landscape features and relative locations of the 20 sampling sites. Note that not all pools in the area are shown.

Zealand. The remains of Nothofagus stumps around the wetland suggest forest burning in the past. A large fire in the 1880s affected the spur immediately to the east, but there has been no significant burning in the area since then (McLeod, 1974). In 1978, the land was transferred to the (now) Department of Conservation, which provides protection against any further development. A board-walked track crosses the eastern edge of the wetland area (Fig. 1). Other than this, the area is largely unimpacted by human activities and therefore a suitable location for a baseline study of benthic algal communities. In the wetland itself, a deep (>1.5 m), relatively large tarn at the western end is flanked by a series of smaller, shallow pools. Elevation falls about 20 m towards the boardwalk, where there are numerous pools within a predominantly waterlogged area. All the pools have well-defined vertical peat margins and water depth is typically <50 cm. Three small streams drain the wetland, one to the east and

two to the north (Fig. 1). Vegetation is a mixture of tussock (Chionochloa spp.) and subalpine shrubs (Hebe spp.) in well-drained areas on the flanks of the wetland, giving way to tussock and sedge species mixed with typical alpine wetland species including Dracophyllum sp., Gleichenia dicarpa, Lycopodium sp. and Sphagnum sp. The area is situated more or less on the tree line, with large stands of mountain beech (Nothofagus solandri var. cliffortioides) on the lower slopes to the north of the wetland, and scattered patches on the higher ground to the south. The substrate in most pools comprises soft algal and other organic detritus ranging in depth from a few millimetres to 1.5 m or more. The thinnest organic layers are in small pools directly connected to streams. Deeper organic substrates resemble those described for mire pools elsewhere (Foster & Fritz, 1987). In some pools the top 10 mm is consolidated into a firm cyanobacterial mat, laminated into an upper red-brown layer, with a bright green layer beneath, similar in

98 appearance to microbial mats described from Antarctica (Vincent et al., 1993; Sabbe et al., 2004).

Methods Algae Qualitative samples of benthic algae were collected from 20 pools on the same day in the early austral spring 2001. Sites were selected to cover the whole of the wetland area and to include a wide range of pool size. All samples were from water depths between 0.2 and 0.4 m. Two small pools (sites 13 and 14) were directly connected to the western stream draining the wetland to the north. No other site had any obvious surface inflow or outflow. Samples were taken approximately 0.5 m from the pool margins. At each site 4 or 5 small cores (10 mm diameter) from the top 15 mm of substrate were combined into a single sample and preserved in glutaraldehyde (final concentration of 2.5%) within 6 h of collection. Subsamples were examined under an inverted microscope at magnifications up to 400. A scale of 1 (rare) to 8 (dominant) was used to assess the relative abundance of major groups of algae and other material using the method described in Biggs & Kilroy (2000). The groups were: cyanobacteria, fine trichomes <5 lm diameter (cf. Leptolyngbya spp.); coarse trichomes >7–10 lm diameter (e.g. Tolypothrix, Hapalosiphon); colonial unicellular cyanobacteria; desmids; diatoms; green filamentous algae; organic detritus; inorganic material (silt/ sand). Scores assigned to diatoms are an estimate of the relative abundance of living cells (i.e. containing chloroplasts). Further subsamples were oxidised in concentrated sulphuric acid followed by hydrogen peroxide to remove all organic material, then rinsed in distilled water. Drops of the resulting suspension of diatoms were dried onto coverslips and mounted onto glass slides using Naphrax (Northern Biological Supplies, UK). Slides were examined at 1000 on a Leica DMLB microscope with differential interference contrast optics, and counts made across transects. At least 600 valves were counted per sample. Diatoms were identified to species level where possible, using a range of texts (see Table 1). Where no

satisfactory identification could be made, species were assigned numbers (e.g. Kobayasiella sp. 1) and photographed for further investigation (Zeiss Axiocam mounted on a Leica DMLB, and Leica S440 scanning electron microscope). Morphospecies were placed in one of the following categories: 1) indistinguishable from species described and recorded from the Northern Hemisphere (i.e. likely to be cosmopolitan); 2) known species recorded to date only in the Southern Hemisphere; 3) closely resembling a known species; 4) genus only, unknown species.

Environmental variables Water pH, conductivity and temperature were measured in the field using a TPS WP-81 meter (TPS Pty Ltd, Australia). Water samples were collected into acid-cleaned polyethylene bottles and sub-samples filtered in the field for subsequent analyses of nutrients, dissolved organic carbon (DOC), and major anions and cations. In the laboratory, a subsample was filtered prior to measuring absorbance at 440 nm using the method of Davies-Colley et al. (1993). This provides a convenient measure of the UV-absorbing yellow colour in water caused by dissolved humic substances, referred to here as gilvin (g440) but also known as Gelbstoff or yellow substance (Yacobi et al., 2003). Nitrate–nitrogen, ammonium, dissolved reactive phosphorus (DRP), total dissolved nitrogen (TDN) and total dissolved phosphorus (TDP) were determined on a Technicon 2 Autoanalyser, after UV irradiation in respectively alkaline or acidic, conditions (Downes 2001). Dissolved organic nitrogen (DON) and phosphorus (DOP) were determined as TDN minus (NH4+ nitrate N), and TDP minus DRP, respectively. Analyses for major ions (K, Na, Ca, Mg, Si, Cl, SO4) and DOC were undertaken using standard methods (APHA, 1998) at the Landcare Research Environmental Chemistry Laboratory, Palmerston North. Numerical analyses The diatom dataset was square-root transformed in order to downweight the effects of dominant taxa (>20% in many samples) (Laing & Smol,

Krammer (1997) Lange-Bertalot & Moser (1994) Lange-Bertalot & Moser (1994) Vyverman et al. (1997) Krammer & Lange-Bertalot (1991–1997) Moser et al. (1998) Foged (1979) Krammer & Lange-Bertalot (1991–1997)

Encyonopsis cf. blanchensis Krammer

Brachysira cf. brebissonii Ross (form 1)

Brachysira cf. brebissonii Ross (form 2) Eunophora oberonica Vyverman, Sabbe & Mann

Frustulia rhomboides (Ehrenberg) De Toni

Frustulia cf. nana Moser, Lange-Bertalot & Metzelin

Frustulia magaliesmontana Cholnoky

Eunotia bilunaris v. mucophila Lange-

Lange-Bertalot (1996, 1999)

Tabellaria flocculosa (Roth) Kutzing Brachysira neoexilis Lange-Bertalot

Kobayasiella cf. madumensis (Jorgensen)

Krammer & Lange-Bertalot (1991–1997) Krammer (1992) Cleve (1881) Vyverman et al. (1998) Lange-Bertalot & Metzelin (1996) Lange-Bertalot (1996, 1999) Krammer & Lange-Bertalot (1991–1997)

Williams & Round Eunotia cf. rhomboidea Hustedt

Pinnularia cf. oriunda Krammer

Amphora bergrennii Cleve (= Eunophora sp. 1)

?Chamaepinnularia sp.

Kobayasiella subtilissima (Cleve) Lange-Bertalot

Frustulia rhomboides var. saxonica

Lange-Bertalot & Moser (1994) Krammer & Lange-Bertalot (1991–1997) Krammer & Lange-Bertalot (1991–1997)

Brachysira cf. lehmanniae Lange-Bertalot & Moser Stenopterobia delicatissima (Lewis) Van Heurck

Stenopterobia denestriata (Hustedt) Krammer

(Rabenhorst) De Toni

Foged (1979) Williams & Round (1987)

Fragilariforma virescens (Ehrenberg)

Krammer (1992)

Pinnularia cf. anglica Krammer

Frustulia rhomboides var. elongatissima Manguin

Krammer (1992)

Pinnularia subgibba Krammer

Lange-Bertalot

Patrick & Reimer (1966) Krammer & Lange-Bertalot (1991–1997) Lange-Bertalot & Moser (1994)

Neidium iridis (Ehrenberg) Cleve

Bertalot & Norpel

1

Lange-Bertalot & Moser (1994)

Brachysira wygaschii Lange-Bertalot

1

3 1

0.7

0.8 0.8

0.9

0.9

1 1

1.0

1.3

1.2

1.2

1.3

1.3

1.3

1.3

1.5

2.5 2.1

2.9

2.9

4.0

4

2

3

3

1

1

3

1

3

1 1

1

1

1

4.5 4.4

1

5.7 5.1

6.1

6.1

3

3 2

3

3

10.6 9.9

10.9

11

11 7

11

19

12

14

16

15

2

8

17

8

9

13 11

16

13

18

15

20

13 16

16

13

15

19

13

at Bealey

samples)

4

Number of occurrences

% (Over all 20

1

Krammer (1997)

Encyonema neogracile Krammer

ID category

Kobayasiella sp. 1

References

Species

Table 1. List of the 52 most common benthic diatom species in the Bealey mire/tarn complex

5.9

5.7 5.3

6.0

5.6

5.5

5.4

6.2

5.4

6.8

6.5

5.5

5.4

5.3

6.5 6.2

5.3

5.4

5.5

6.0

5.7

6.0 5.5

5.3

5.1

5.3

5.8

6.3

pH optima

3.3

2.5 3.8

2.9

3.1

3.0

3.7

2.8

3.5

1.9

2.7

3.3

3.5

4.1

2.4 2.6

3.7

3.3

2.8

2.7

2.8

2.7 2.7

2.8

4.6

3.6

2.9

Continued on p.100

8.7

7.3 7.4

7.8

7.8

8.5

7.5

9.9

7.9

22.2

12.4

7.9

7.5

7.7

15.5 9.9

7.5

7.2

7.7

8.8

8.7

8.0 6.9

7.6

7.6

7.1

8.3

2.7

optima

(lS cm)1) 9.5

Gilvin (g440)

Conductivity optima

99

Krammer & Lange-Bertalot (1991–1997)

Rossithidium cf. pusillum (Grunow)

0.1 0.1 0.1 0.1 0.1

1 4 4 1 4

Eunotia implicata Norpel, Lange-Bertalot & Alles

Eunotia sp.

Navicula sp. Pinnularia microstauron Ehrenberg (Cleve) Krammer (1992)

2

5 4

5

6

1

2

1

3

2 3

3

4

1

7

3

8

3 4

11

9

6

9

12

occurrences at Bealey

Number of

6.0

6.25 6.1

5.6

5.5

6.85

6.85

6.85

6.7

6.5 5.2

5.5

5.3

6.85

5.8

5.3

6.4

6.8 6.8

6.1

5.9

6.5

6.3

6.0

optima

pH

8.8

10.8 10.0

7.6

6.6

22.2

21.6

22.2

17.3

10.5 6.7

7.3

6.3

22.2

10.2

6.5

12.6

19.7 19.7

10.7

7.9

13.2

10.7

8.5

optima (lS cm)1)

Conductivity

3.1

2.7 3.2

3.4

2.3

1.9

1.9

1.9

2.4

2.7 3.0

2.8

2.7

1.9

2.7

3.0

2.3

2.1 2.1

2.6

3.1

2.7

2.8

2.8

(g440) optima

Gilvin

Identification categories: 1. Indistinguishable from known species with a worldwide distribution; 2. known from Southern Hemisphere only; 3. close to a known species (minor differences); 4. unknown species in a known genus. Optima for pH, conductivity and gilvin calculated are calculated using weighted averaging. Optima are in bold for potential indicator species (see text).

Encyonopsis sp. 2

0.1

4 Krammer & Lange-Bertalot (1991–1997)

0.1

1

Krammer & Lange-Bertalot (1991–1997)

Eunotia diodon Ehrenberg

0.1

Encyonopsis sp. 1

1

0.1

4 Krammer & Lange-Bertalot (1991–1997)

0.1 0.1

0.2

0.2

0.2

0.2

0.2

0.3

0.4 0.4

0.4

0.5

0.5

0.5

3 4

Cymbella naviculiformis (Auerswald) Cleve

Krammer (1997) Moser et al. (1998)

Encyonopsis cf. delicatissima (Hustedt) Krammer Adlafia sp.

3

1

1

3

1

1

1 4

3

1

1

1

0.6

all 20 samples)

category

3

% (Over

ID

Diadesmis sp.

Krammer & Lange-Bertalot (1991–1997)

Achnanthes marginulata Grunow in Cleve & Grunow Krammer & Lange-Bertalot (1991–1997)

Krammer & Lange-Bertalot (1991–1997)

Eunotia cf. glacialis Meister

Encyonopsis cf. aequalis (W. Smith) Krammer

Krammer & Lange-Bertalot (1991–1997)

Nupela paludigena (Scherer) Lange-Bertalot

Eunotia praerupta Ehrenberg

Krammer & Lange-Bertalot, (1991–1997) Lange-Bertalot & Moser (1994)

Achnanthidium minutissimum (Kutzing) Czarnecki

Lange-Bertalot & Moser (1994) Krammer & Lange-Bertalot, (1991–1997), Vyverman (1996)

Krammer & Lange-Bertalot (1991–1997)

Eunotia incisa Gregory

Navicula gottlandica Grunow in Van Heurck ?Rossithidium sp. 1

Lange-Bertalot & Moser (1994)

Brachysira styriaca (Grunow) Ross

Brachysira cf. brebissonii Ross (form 3)

Krammer & Lange-Bertalot (1991–1997)

Stenopterobia curvula (W. Smith) Krammer

Round & Bukhtiyarova

References

Species

Table 1. (Continued)

100

101 2000). All of the environmental variables except water depth, pH, DOC, gilvin and Na had skewed distributions and were log-transformed for subsequent analyses. A Pearson correlation matrix with Bonferroni-adjusted probabilities (SYSTAT, v. 10) identified several significantly correlated sets (P< 0.05). Redundant variables were selected by running preliminary BIO-ENV analyses (see below) for each and retaining the variables that yielded the highest rank correlation with the diatom dataset. The following variables were eliminated: DOC (correlated with gilvin), DRP, NH4, TDN, DON (all correlated with TDP), Ca (correlated with conductivity), Mg (correlated with Ca). Relationships between diatom community composition and environmental variables were explored using the BIO-ENV procedure in the University of Plymouth’s program PRIMER (Clarke & Warwick, 2000). BIO-ENV was considered to be suitable for exploring the present dataset because of the small number of samples from a restricted area. The procedure requires no prior assumptions about the nature of the relationships (i.e., linear or unimodal) of biota to environmental variables (Clarke & Ainsworth, 1993), and therefore was expected to produce a realistic result within a low range of environmental conditions. BIO-ENV compares a similarity matrix of the biotic data with similarity matrices of all combinations of associated environmental data and computes a Spearman rank correlation coefficient qs for each combination. The coefficient ranges from )1 (ranks in complete opposition) to 1 (ranks in complete agreement). Matrices were constructed using Bray Curtis similarities for the diatom data, and normalised Euclidean distance for the environmental dataset. The environmental variables identified as having the closest correlations with the biotic data can be visualised by comparing non-metric multidimensional scaling (MDS) plots of the two datasets. MDS plots are based on matrices of ranked similarities and configure the sites in a specified number of dimensions (usually two) such that similar sites are close together and dissimilar sites are far apart. Values of qs > 0.8 generally produce a close visual match, providing that the stress for both plots is low (Clarke & Ainsworth, 1993). Stress is a measure of how well the data are represented in the

number of dimensions specified: values <0.1 indicate a good representation; values >0.1 and <0.2 are acceptable (Clarke & Warwick, 2000). Note that currently there is no means in the BIO-ENV procedure of establishing whether a relationship is statistically significant, though this may be introduced in a future version of the software (Clarke & Warwick, 2000). We used BIO-ENV to look for relationships between three pairs of datasets: the diatom community data vs. environmental variables; the live algae relative abundance assessments vs. environmental variables and the diatom community data vs. live algae data (as environmental variables). Because the above analysis confirmed strong environmental relationships with the diatom community, we used weighted averaging regression analysis (WA, Birks et al., 1990) to investigate the strength of models for predicting pH, conductivity and gilvin from the diatom community data, using C2 software (Juggins, 2003). The software calculates species optima and tolerances (respectively, the average and standard deviation of the environmental parameter over all sites where a taxon occurs, weighted by the relative abundance of the taxon at each site.) The predictive capability of the resulting models was assessed using the jackknife (‘leave-one-out’) cross-validation procedure. These techniques are standard for the construction of transfer functions for use in palaeolimnological studies (e.g. Hall & Smol, 1992). In the present case, the procedure is relevant because it also enables a preliminary identification of taxa that may be suitable as indicators of particular conditions because of their narrow tolerance ranges. For this dataset criteria for potential indicator species are based on two of the criteria used by Fallu et al. (2000). Our criteria were 1) occurrence in at least 8 of the 20 sites, and 2) tolerance to the variable of interest <0.75  the mean tolerance for all the species. Untransformed species and environmental data were used for the WA analyses.

Results Environmental characteristics Pool size varied from approximately 6400 m2 to about 8 m2. The pool water was comparable in its

102 chemistry with similar South Island water bodies sampled in the 2001 survey (Vanhoutte et al., 2006). pH ranged from moderately acidic to almost neutral (pH 5.1–6.85), and in general declined across the wetland from west to east. Conductivity was very low (5.7–22.2 lS cm)1). Gilvin ranged from g440 = 1.54 (optically clear water) to 5.25 (lightly brown-stained). A high proportion of the dissolved nitrogen in most pools was in organic form, with total dissolved nitrogen ranging from 178 to 607 lg l)1, and the inorganic component (NH4 + NO3) ranging from 10.8 to 61.2 lg l)1. Of the cations measured, Na predominated in most sites (though at low levels, range 0.31–1.01 mg l)1), consistent with the expected largely ombrotrophic hydrological regime. Ca was very low at most sites, but was noticeably higher at sites 13 and 14 (which were connected to a stream) and in pools 15 and 19, suggesting that the latter two sites also received some stream inputs (range over all sites 0.14–3.21 mg l)1). Complete data for all the sites are available from the author on request. Community composition In the inverted microscope assessments, diatoms were assigned a score of 5 or more at 18 of the 20 sites (Table 2), with the lowest score (3) at site 1. Diatoms dominated at sites 8, 9 and 12. Detritus (mainly decaying sphagnum and other vegetation) was dominant in all other samples except sites 2 and 19 where the main constituent was fine cyanobacterial trichomes (Leptolyngbya sp.). Green algae other than desmids were uncommon, though a species of Oedogonium was recorded at 17 sites. Eleven genera made up 95% of the diatom community overall. These were (in order of abundance): Brachysira, Frustulia, Kobayasiella, Encyonema, Encyonopsis, Eunophora, Eunotia, Pinnularia, Neidium, Tabellaria and Stenopterobia. Eighty-one diatom morphospecies in 31 genera were distinguished. Total species richness at individual sites ranged from 18 (site 12) to 40 (site 13). Fifty-two common species (with an abundance over all 20 sites of >0.1%) are listed in Table 1 along with calculated optima for pH, conductivity and gilvin, and the identification category (see Methods). Of these, 27 were indistinguishable from known cosmopolitan taxa (category 1) and

two (Eunophora spp.) were known only from the Southern Hemisphere. The remaining taxa were problematic in that they did not correspond exactly to any species in the available literature. Some of these indeterminate species were very common and had distinctive distinguishing features. A complete taxonomic treatment of the species found will be presented elsewhere. The seven most common diatom species: Encyonema neogracile, Kobayasiella sp. 1, Brachysira wygaschii, Encyonopsis cf. blanchensis, two forms of Brachysira cf. brebissonii, and Eunophora oberonica, accounted for large proportions of the community at most sites (typically >50%), but there were marked shifts in composition across the wetland. For example, E. cf. blanchensis occurred in high relative abundances only at sites with low pH, while Encyonema neogracile tended to occur in the pools with higher pH. At site 13, a pool in a stream, these seven taxa comprised only 6.5% of the community, which was dominated by Fragilariforma virescens and Tabellaria flocculosa. Two species of Eunophora were recorded over the whole wetland: Eunophora sp. 1 (Vyverman et al., 1998, see below) (at 14 sites) and E. oberonica (at 16 sites). The latter species was abundant, comprising >13% of the community at four sites. Note that although site 13 differed from the other sites in its much higher conductivity and lower relative abundances of the common diatom taxa, we consider that the site lies at one end of a natural continuum in the Bealey wetland, from isolated water bodies to flowing water. Therefore the site was retained in the community – environment analyses. To check the influence of this site, BIO-ENV analyses were also run omitting site 13. Diatom community – environment relationships The BIO-ENV analysis on all data produced a best rank correlation between the diatom and environmental matrices of qs = 0.763 for a combination of pH and conductivity. Addition of gilvin caused the match to deteriorate very slightly (qs = 0.757). pH only yielded qs = 0.739, conductivity only, qs = 0.548 and gilvin only, qs = 0.293. A comparison of MDS plots derived from the diatom dataset and from pH + conductivity shows good visual correspondence (Fig. 2). Addition of gilvin (Fig. 2c) increases the distance between sites 11

103 Table 2. Relative abundance of main substrate constituents in raw samples from 20 pools at Bealey Spur Site

Relative abundance of main constituents

Diatoms

Desmids

Cyanobacteria

Green

Organic

algae

detritus

Inorganic

Fine

Coarse

Colonial

trichomes

trichomes

unicells

1 2

1 8

1 1

2 5

3 5

4 1

3 0

8 6

3 0

3

2

1

7

6

1

2

8

1

4

3

2

4

6

2

1

8

3

5

2

1

0

7

1

1

8

4

6

1

0

4

7

3

2

8

1

7

3

3

4

7

2

1

7

8

8

7

2

6

8

5

1

6

2

9 10

2 4

2 0

3 3

8 4

5 2

2 1

6 8

1 3

11

4

1

3

6

3

2

8

2

12

7

1

4

8

4

2

3

2

13

0

0

0

6

2

1

8

1

14

0

0

0

6

5

1

8

3

15

2

2

3

6

5

3

8

0

16

1

0

1

5

3

0

8

2

17 18

2 5

0 0

3 3

5 5

1 3

2 0

8 8

3 3

19

8

2

3

7

6

5

4

2

20

4

0

2

7

2

2

8

1

The relative abundance scale used ranges from 8 (dominant) to 1 (rare).

and 12, which corresponds to the biotic MDS, as does increased separation of sites 4, 5 and 6 from sites 3, 7 and 16. However, to offset this, site 1 is positioned much farther from the biologically similar 4, 5 and 6 (Fig. 2a). The best result with site 13 omitted was qs = 0.738 for pH only, followed by qs = 0.685 for a combination of pH, conductivity and gilvin. Histograms of the abundance of common species vs. sites arranged in order of pH show that several taxa are strongly influencing the separation of communities (Fig. 3). In particular, Encyonema neogracile, Encyonopsis cf. blanchensis, Brachysira neoexilis, and Neidium iridis showed marked clustering. Both BIO-ENV analyses involving the relative abundance assessment data on the whole algal community (with and without site 13) gave low maximum values for rank correlation (qs < 0.355). Weighted averaging regression and calibration produced a strong model for predicting pH. The

best result used simple WA regression (no tolerance down-weighting) with classical deshrinking (Birks et al., 1990). Jackknife-derived predicted pH values matched the measured values well (Fig. 4, R2 = 0.91, RMSEP = 0.18). The residuals plotted against predicted values (Racca & Prairie, 2004) indicated no bias in the model (Fig. 4). A similar model for predicting conductivity produced predicted values that correlated well with the measured values (R2 = 0.88, RMSEP = 1.55, data not shown) but there was considerable bias in the predictions as a result of the high conductivity at site 13. A model for gilvin performed poorly (data not shown). Four diatom taxa met the criteria specified for potential use as indicator species for pH: Encyonopsis cf. blanchensis (pH 5.1), Brachysira wygaschii (5.3), Eunotia bilunaris var. mucophila (5.4) and Frustulia rhomboides var. elongatissima (6.5). Neither species of Eunophora showed up as a potentially good indicator of pH. Separate BIO-

104 Stress: 0.1

13 1

11

64 5

16 17

9

15 20

(a)

72

18 12 Stress: 0.01

13 4 56 37 16 1 2 8

17 18 11 14 15 2019

10 9 12

(b)

Stress: 0.02

13 14 15 20 19

(c)

Discussion

3

8 19 10

14

For example, the best correlation for E. oberonica (untransformed data) was qs = 0.540 (a combination of conductivity, NO3, Na and Cl).

4 5 6 10 9 11 8 17 18 2 16

3 7 1

12

Figure 2. Non-metric multidimensional scaling (MDS) plots for (a) square-root transformed diatom community data, (b) pH and conductivity, (c) pH, conductivity and gilvin. The circled groups on the best-fitting abiotic plot (b) indicate sites that have similar pH and conductivity. The same sites are grouped on the biotic plot (a) to highlight the relatively good visual match between the two plots. In (c) the right-hand group of 7 sites are separated along a gradient of gilvin. This appears to improve the match with (a), except for the position of site 1 (see text). A stress value of 0.1 for the biotic MDS indicates a good representation of the 20 sites in two dimensions. Stress values for the abiotic plots are low because a maximum of 3 variables was used to construct the plots.

ENV analyses run for the two species showed no strong correlation with any of the environmental variables measured (either singly or in combination).

The diatom communities found in the Bealey Spur pools are generally consistent at the genus level with the characteristic ‘peat-bog flora’ found in other regions and dominated by species of Eunotia, Pinnularia, Frustulia and Brachysira. This flora appears to have a global distribution (Scherer, 1988 and references therein). For example, Gaiser & Johansen (2000) studied the diatom communities in South Carolina lowland wetlands and reported similar assemblages, which also included Neidium, Kobayasiella, Stenopterobia, Encyonema and Encyonopsis. Studies in Tierra del Fuego (Mataloni, 1999) and Japan (Watanabe et al., 2000) have yielded the same genera and species. The taxa found evidently reflect an association with the dystrophic, low-nutrient, acidic conditions typical of mire-pool environments, and this association appears to be independent of latitude and altitude (temperature) (Scherer, 1988). However, as Vyverman et al. (1995) found in Tasmania, there did seem to be a distinctive regional element in this New Zealand dataset. The occurrence of two species of Eunophora at Bealey represents a major difference between this New Zealand peat-bog flora and those reported from elsewhere. Eunophora was first found as a common component of the diatom flora in oligotrophic and dystrophic lakes in the Tasmanian highlands (Vyverman et al., 1995). The genus currently includes three species (E. oberonica, E. tasmanica and E. indistincta) plus a fourth species (Eunophora sp. 1) previously described as Amphora berggrenii Cleve, which, prior to the Tasmanian survey had been known only from fossil material in New Zealand (Vyverman et al., 1998). Although Eunophora was widespread in Tasmania and common in some samples, Vyverman et al. (1998) reported that they had ‘never found thriving populations of E. oberonica...’, though live cells of Eunophora sp. 1 were ‘a little more abundant’. In the 2001 survey in South Island, New Zealand, Eunophora occurred at >50% of the sites sampled, mainly in alpine pools and

105 50

20 Encyonema neogracile

Frustulia cf. nana

0

0

50

20 Kobayasiella sp. 1

Eunotia bilunaris

0

0

50

20 Brachysira wygaschii

Neidium iridis

0

0 20

50

Relative abundance (%)

Encyonopsis cf. blanchensis

Brachysira neoexilis

0

0

50

20 Brachysira cf. brebissonii f. 1

Kobayasiella madumensis

0

0

50

20 Brachysira cf. brebissonii f. 2

0

Pinnularia subgibba

0 20

50 Eunophora oberonica

Pinnularia cf. anglica

0

0

50

20 Frustulia rhomboides var. elongatissima

Frustulia rhomboides

0

0

50

20 Eunotia rhomboidea

Frustulia magaliesmontana

0

0

50

20 Pinnularia cf. oriunda

Tabellaria flocculosa

0

0

50

20 Eunophora sp. 1

Fragilariforma virescens

0

0 5 4 6 7 3 1 16 18 17 2 8 11 10 12 9 19 20 15 14 13

5 4 6 7 3 1 16 18 17 2 8 11 10 12 9 19 20 15 14 13

Figure 3. Relative abundances at each site of the 22 most common diatom taxa plotted with sites arranged in order of increasing pH. Note the different vertical scales in the two columns.

tarns with organic sediments and water with pH <6 (K. Vanhoutte, pers. comm.). Relative abundances of up to 11% were recorded for E. oberonica (Edwards Valley, Arthurs Pass) and up to 2.5% for Eunophora sp. 1 (Staircase Saddle, Fiordland) and both these samples contained many live specimens. In the present study at Bealey Spur, examination of preserved material showed high relative abundances of live cells of both E. oberonica (up to 15%) and Eunophora sp. 1 (up to >5.5%) at several sites.

In Figure 5 we plot the relative abundances of both species in sites grouped by the pH classes used to examine the Tasmanian populations (Vyverman et al., 1998). This shows that at Bealey, the pH preferences of the two species cover a broad range, but differ. However, the optimal pH ranges found in Tasmania and Bealey are different again. Vyverman et al. (1998) found Eunophora oberonica in highest abundance in their most acidic sites (pH < 4.8, not represented in the present survey), whereas at Bealey the highest abundances

106

Figure 4. Observed pH values at the 20 pool sites plotted against predicted values calculated from a WA model (see text). The righthand graph shows that there is no bias in the residuals.

were at pH 5.2–6.4. Conversely Eunophora sp. 1 had highest abundance in lower pH environments at Bealey (pH < 5.2) compared to pH > 6 in Tasmania. Thus, although Eunophora is generally restricted to dystrophic, acidic to slightly acidic, low nutrient waters, in cool areas (mainly high altitude) its distribution within these environments appears to be driven by factors other than water chemistry. In ongoing investigations at Bealey Spur, we have observed that thriving populations of live E. oberonica tend to occur in pools where the substrate is consolidated into a firm cyanobacterial mat, though this is not always the case. The cyanobacterial mats resemble those found in Antarctic freshwater systems (Sabbe et al., 2004) and comprise a mixture of fine trichomes (Leptolyngbya sp.), closely packed colonial unicells, and larger taxa including Dichothrix, Stigonema and Scytonema. They are organised into a yellow-brown pigmented layer overlying a bright green zone, where most of the live diatom cells are located. At Bealey, such mats were best developed at sites 8, 9 and 12, all of which had relatively high concentrations of E. oberonica. Other diatoms common at these sites (generally >15% relative abundance) included Encyonema neogracile, Kobayasiella sp. 1, and Brachysira cf. brebissonii form 2. Given that cultures of E. oberonica have shown extremely slow rates of cell division (one division every 2–3 weeks, V.

Figure 5. Mean relative abundance (%, with standard error bars) of Eunophora oberonica and Eunophora sp. 1 occurring at Bealey in the pH classes used by Vyverman et al. (1998). pH classes are: 3, 4.81–5.2; 4, 5.21–5.6; 5, 5.61–6.0; 6, 6.01–6.4; 7, 6.41–6.8; 8, 6.81–7.2. Classes 1 and 2 (pH 4.0–4.8) did not occur at Bealey. Classes 5 and 8 were represented by only one site each at Bealey.

107 Chepurnov, pers comm.) it is possible that very stable physical conditions may be required for the development of large populations. The cyanobacterial mats may provide suitably stable conditions. Work is underway to investigate this possibility. To date, Eunophora species have been reported only from Tasmania and New Zealand (Vyverman et al., 1998, K. Sabbe, pers. comm.). Thus, the occurrence of Eunophora species in New Zealand as a major component of the otherwise globally distributed ‘peat-bog flora’ (at the genus level) raises some interesting questions. Is Eunophora confined to New Zealand and Tasmania because these are the only locations where suitable conditions exist for populations to thrive? Has the geographic isolation of suitable habitats in New Zealand and Tasmania, combined with slow growth rates, ensured that the genus has retained a restricted distribution? Does Eunophora form part of a regionally distinctive high-altitude diatom flora that has been less subject to invasions by cosmopolitan taxa than communities at low-altitude locations populated by humans? Given that many cosmopolitan species also appear to be thriving at the sites sampled in the present study, what are the specific conditions that allow Eunophora species to do well in the same environment? A complicating factor is that the pH data discussed above suggest not only that E. oberonica and E. sp. 1 have different environmental preferences, but that these vary with location. Studies on live populations of both species, and the cosmopolitan taxa associated with them, are probably the only way to begin answering the last of these questions. With regard to the diatom communities as a whole, although the sensitivity of diatom community composition to pH has long been known and is well documented (see Battarbee et al., 1999 for a review), this is the first time such a strong relationship between diatom community composition and pH has been demonstrated in New Zealand. The South Island-wide survey of 71 sites in 2001 also showed pH (and related variables) to be the main environmental factor influencing diatom community composition, though that analysis was based on genus level data (K. Vanhoutte, pers. comm.). Reducing our dataset to genus level and re-running the BIO-ENV procedure resulted in a correlation coefficient of qs = 0.682 for a

combination of pH, conductivity and gilvin. This confirms the result found with the species data, but demonstrates it less convincingly. Diatom assemblages have been used to produce robust models for predicting alkalinity, conductivity and DOC in other studies (Fritz et al., 1999; Fallu et al., 2000; Davies et al., 2002). In this study, there appeared to be a response to conductivity along a very small range (5.7 – 22 lS cm)1). Conductivity was strongly correlated to Ca, which is in turn correlated with alkalinity at several sites at Bealey (C. Kilroy, unpublished data). Therefore this result is also consistent with findings elsewhere. However, the conductivity – diatom community relationship was strongly influenced by site 13, where conductivity was almost twice as high as the next highest value. The BIO-ENV procedure run omitting site 13 confirmed that pH was by far the most important variable influencing diatom communities at the remaining 19 sites. The relatively poor model for gilvin is to be expected given the confounding effect of a very strong pH gradient over a small number of sites. Nevertheless, there is evidence for a link between gilvin and community composition at Bealey. Figure 3 shows that deterioration of the diatom community–environment match by addition of gilvin is largely due to inconsistency in the positioning of site 1. (Compare the relative positions of sites 1 and 4, 5, 6 in Figures 3a, b, c, and note the increased separation in c). This pool was observed to have dried out later in the summer. A tendency to desiccate probably explained the low relative abundance of live diatoms at site 1 (Table 2), but may also have influenced the water chemistry– diatom community relationship. Indeed, re-running BIO-ENV omitting site 1 produced an optimum correlation coefficient of qs = 0.799 for a combination of pH, conductivity and gilvin. These discrepancies highlight the fact that a larger dataset would be more likely to yield robust models, a point also made following other studies involving relatively small datasets (Peinitz & Smol, 1993, 22 lakes; Battarbee et al., 1997, 24 lakes and streams). Thus it would be desirable to expand the present dataset with information from different locations. There is increasing interest in the use of diatom communities for inferring past environmental

108 conditions in New Zealand although to date most emphasis has been placed on salinity gradients (Goff et al., 2000; Cochran, 2002). However, the first transfer function for inferring trophic status (as chlorophyll a), conductivity and pH in medium to large New Zealand lakes is now completed (Reid, 2005). The present study represents an early step towards extending the potential range of palaeoecological studies into alpine and treeline zones. Interpretation of inferred pH, conductivity and gilvin from cores taken from subalpine mire pools is likely to be difficult, given the complex way in which mires develop over time and interact with climate (Foster & Wright, 1990). Multiple proxies to provide independent chronologies will almost certainly be necessary. For example, a chironomidbased transfer function for temperature is currently under development (M. Reid, pers. comm.). In the meantime, the results from this relatively small survey also indicate that regional identification guides are needed in order to capitalise on the environmental indicator capacity of high-altitude diatom communities in New Zealand.

Acknowledgements This work was funded by the New Zealand Foundation for Research, Science and Technology (contract no. C01X0308). All collections on Department of Conservation estate were made under a collecting permit issued by the Minister for Conservation to NIWA. We thank Michael Reid and two anonymous referees for constructive reviews, and Greg Kelly for drafting the map of the study site.

References APHA., 1998. Standard Methods for the Analysis of Water and Wastewater, 20th edn. American Public Health Association, Washington DC. Battarbee, R. W., R. J. Flower, S. Juggins, S. T. Patrick & A. C. Stevenson, 1997. The relationship between diatoms and surface water quality in the Høylandet area of NordTrøndelag, Norway. Hydrobiologia 348: 69–80. Battarbee, R. W., D. F.Charles, S. S. Dixit & I. Renberg, 1999. Diatoms as indicators of surface water acidity. In Stoermer E. F. & J. Smol (eds), The Diatoms: Applications for the

Environmental and Earth Sciences. Cambridge University Press 85–127. Biggs, B. J. F. & C. Kilroy, 2000. Stream Periphyton Monitoring Manual. Published by NIWA for the New Zealand Ministry for the Environment. 226 p. (available online at www.niwa.co.nz/ncwr/tools/periphyton/). Birks, H. J. B., J. M. Line, S. Juggins, A. C. Stevenson & C. J. F. ter Braak, 1990. Diatoms and pH reconstruction. Philosophical Transactions of the Royal Society, London Series B 327: 263–278. Clarke, K. R. & M. Ainsworth, 1993. A method of linking multivariate community structure to environmental variables. Marine Ecology Progress Series 92: 205–219. Clarke, K. R. & R. M. Warwick, 2000. Change in Marine Communities: An Approach to Statistical Analysis and Interpretation, Plymouth Marine Laboratory, PRIMER-E Ltd. Cleve, P.T., 1881. On some new and little known diatoms. Kongl. Svenska Vetenskaps-Akademiens Handlingar. Bandet 18(5): 3–6. Cochran, U. A., 2002. Detection of large Holocene earthquakes in the sedimentary record of Wellington, New Zealand, using diatom analysis. Unpublished PhD Thesis, Victoria University, Wellington. Davies, S. J., S. E. Metcalfe, M. E. Callabero & S. Juggins, 2002. Developing diatom-based transfer functions for Central Mexican lakes. Hydrobiologia 467: 199–213. Davies-Colley, R. J., W. N. Vant & D. G. Smith, 1993. Colour and Clarity of Natural Waters. Science and Management of Optical Water Quality. Ellis Horwood, Chichester. Douglas, M. S. V. & J. P. Smol, 1995. Periphytic diatom assemblages from high Arctic ponds. Journal of Phycology 31: 60–69. Downes, M., 2001. NIWA Christchurch Auto-analyser Methods. NIWA Internal Report No. 103. NIWA, Christchurch, New Zealand. Fallu, M. -A., N. Allaire & R. Pienitz, 2000. Freshwater diatoms from northern Que´bec and Labrador (Canada). Species–environment relationships in lakes of boreal forest, forest-tundra and tundra regions. Bibliotheca Diatomologica 45: 1–200. Foged, N., 1979. Diaotms in New Zealand, the North Island. Bibliotheca Phycologica 47: 1–225. Foster, D. R. & H. E. Wright, 1990. Role of ecosystem development and climate change in bog formation in central Sweden. Ecology 71: 450–463. Foster, D. R. & S. C. Fritz, 1987. Mire development, pool formation and landscape processes on patterned fens in Dalarna, Central Sweden. Journal of Ecology 75: 409–437. Fritz, S. C., B. F. Cumming, F. Gasse & K. R. Laird, 1999. Diatoms and indicators of hydrologic and climatic change in saline lakes. In Stoermer E. F. & J. Smol (eds), The Diatoms: Applications for the Environmental and Earth Sciences. Cambridge University Press: 41–72. Gaiser, E. E. & J. Johansen, 2000. Freshwater diatoms from Carolina Bays and other isolated wetlands on the Atlantic Coastal plain of South Carolina, U.S.A, with descriptions of seven taxa new to science. Diatom Research 15: 75–130.

109 Goff, J. R., H. L. Rouse, S. L. Jones, B. W. Hayward, U. Cochran, W. McLea, W. W. Dickinson & M. S. Morley, 2000. Evidence for earthquake and tsunami about 3100 – 3400 years ago, and other catastrophic saltwater inundations recorded in a coastal lagoon. New Zealand. Marine Geology 170: 231–249. Hall, R. & J. P. Smol, 1992. A weighted-averaging regression and calibration model for inferring total phosphorus concentration from diatoms in British Colombia (Canada) lakes. Freshwater Biology 27: 417–434. Juggins, S., 2003. C2 User Guide: Software for Ecological and Palaeocological Data Analysis and Visualization. University of Newcastle, Newcastle-upon-Tyne, UK, 69 pp. Kilroy, C., K. Sabbe, E. A. Bergey, W. Vyverman & R. Lowe, 2003. New species of Fragilariforma (Bacillariophyta) from New Zealand and Australia. New Zealand Journal of Botany 41: 535–554. Kociolek, J. P. & E. F. Stoermer, 2001. Taxonomy and ecology: a marriage of necessity. Diatom Research 16: 433–442. Krammer, K., 1992. Pinnularia, eine Monogrphie der europaischen Taxa. Biobliotheca Diatomologica 26: 1–353. Krammer, K., 1997. Die cymbelloiden Diatomeen. Teil 2. Bibliotheca Diatomologica 37: 1–469. Krammer, K. & H. Lange-Bertalot, 1991–1997. Susswasserflora von Mitteleuropa. 2 Bacillariophyceae. Teil 1–4 Fischer, Stuttgart. Laing, T. E. & J. P. Smol, 2000. Factors influencing diatom distributions in circumpolar treeline lakes of northern Russia. Journal of Phycology 36: 1035–1048. Lange-Bertalot, H., 1996. Kobayasia bicuneus gen. et sp. nov. Iconographica Diatomologica 4: 277–287. Lange-Bertalot, H., 1999. Kobayasiella nom. nov. ein neuer Gattungsname fur Kobayasia Lange-Bertalot 1996. Iconographica Diatomologica 6: 272–275. Lange-Bertalot, H. & G. Moser, 1994. Brachysira. Mongraphie der Gattung. Bibliotheca Diatomologica 29: 1–212. Lange-Bertalot, H. & D. Metzeltin, 1996. Indicators of Oligotrophy. 800 taxa representative of three ecologically distinct lake types. Iconographica Diatomologica 2: 1–390. Lim, D. S. S., M. S. V. Douglas & J. P. Smol, 2001. Diatoms and their relationship to environmental variables from lakes and ponds on Bathurst Island, Nunavut, Canadian High Arctic. Hydrobiologia 450: 215–230. Lotter, A. F., R. Pienitz & R. Schmidt, 1999. Diatoms as indicators of environmental change near arctic and alpine treeline. In Stoermer, E. F. & J. Smol (eds), The Diatoms: Applications for the Environmental and Earth Sciences. Cambridge University Press: 205–226. Lowe, D. J & J. D. Green, 1987. Origins and development of the lakes. In Viner A. B. (ed.), Inland Waters of New Zealand. DSIR, Wellington, pp. 1–64. Mataloni, G., 1999. Ecological studies on algal communities from Tierra del Fuego peat bogs. Hydrobiologia 391: 157–171. McLeod, D., 1974. Kingdom in the Hills: The Story of a Struggle. Whitcombe and Tombs, Christchurch. Michelutti, N., A. Holtham, M. S. V. Douglas & J. P. Smol, 2003. Periphytic diatom assemblages from ultra-oligotrophic and UV transparent lakes and ponds on Victoria Island and

comparisons with other diatom surveys in the Canadian Arctic. Journal of Phycology 39: 465–480. Morgan, R., 2001. Soil formation processes. In Sturman, A. & R. Spronken-Smith (eds), The Physical Environment: A New Zealand Perspective. Oxford University Press: 350–375. Moser, G., H. Lange-Bertalot & D. Metzeltin, 1998. Insel der Endemiten. Geobotanisches Phanomen Neukaledonien. Bibliotheca Diatomologica 38: 1–464. Moser, K. A., G. M. MacDonald & J. P. Smol, 1996. Applications of freshwater diatoms to geographical research. Progress in Physical Geography 20: 21–52. Patrick, R. & C. W. Reimer, 1966, 1975. The diatoms of the United States. Monographs of the Academy of Natural Sciences of Philadelphia, Volume 1. Fragilariaceae, Eunotiaceae, Achnanthaceae, Naviculaceae, p. 1–688; Volume 2: Entomoneidae, Cymbellaceae, Gomphonemaceae, Epithemiaceae, p. 1–213. Pienitz, R. & J. P. Smol, 1993. Diatom assemblages and their relationship to environmental variables in lakes from the boreal forest – tundra ecotone near Yellowknife, Northwest Territories, Canada. Hydrobiologia 269/270: 391–404. Racca, J. and Y. Prairie, 2004. Apparent and real bias in numerical transfer functions in palaeolimnology. Journal of Paleolimnology 31: 117–124. Reid, M., 2005. Diatom-based models for reconstructing past water quality and productivity in New Zealand lakes. Journal of Paleolimnology 33: 13–38. Sabbe, K., D. A. Hodgson, E. Verleyen, A. Taton, A. Wilmotte & K. Vanhoutte, 2004. Salinity, depth and the structure and composition of microbial mats in continental Antarctic lakes. Freshwater Biology 49: 296–319. Sabbe, K., K. Vanhoutte, R. L. Lowe, E. A. Bergey, B. J. F. Biggs, S. N. Francoeur, D. Hodgson & W. Vyverman, 2001. Six new Actinella (Bacillariophyta) species from Papua New Guinea, Australia and New Zealand: further evidence for widespread diatom endemism in the Australasian region. European Journal of Phycology 36: 321–340. Scherer, R., 1988. Freshwater diatom assemblages and ecology/ palaeoecology of the Okefenokee swamp/marsh complex, Southern Georgia, U.S.A. Diatom Research 3: 129–157. Smol, J. P. & B. F. Cumming, 2000. Tracking long-term changes in climate using algal indicators in lake sediments. Journal of Phycology 36: 986–1011. Sommaruga-Wograth, S., K. A. Koinig, R. Schmidt, R. Sommaruga, R. Tessadri & R. Psenner, 1997. Temperature effects on the acidity of remote alpine lakes. Nature 387: 64–76. Stevenson, R. J. & Y. Pan, 1999. Assessing environmental conditions in rivers and streams with diatoms. In Stoermer, E. F. & J. Smol (eds), The Diatoms: Applications for the Environmental and Earth Sciences. Cambridge University Press: 11–40. Vanhoutte, K., E. Verleyen, C. Kilroy, K. Sabbe, R. Dasseville & W. Vyverman, 2006. Catchment characteristics and chemical limnology of small lakes, tarns and mire pools in New Zealand (South Island) and Tasmania. Journal of Marine and Freshwater Research 57: 83–93. Verleyen, E., D. A. Hodgson, W. Vyverman, D. Roberts A. McMinn, K. Vanhoutte & K. Sabbe, 2003. Modelling

110 diatom response to climate induced fluctuations in the moisture balance in continental Antarctic lakes. Journal of Paleolimnology 30: 195–215. Vincent, W. & R. Pienitz, 1996. Sensitivity of high-latitude freshwater ecosystems to global change: temperature and ultraviolet radiation. Geoscience Canada 23: 231–236. Vincent, W., M. Downes, R. W. Castenholz & C HowardWilliams, 1993. Community structure and pigment organisation of cyanobacteria-dominated microbial mats in Antarctica. European Journal of Phycology 28: 213–221. Vyverman, W., K. Sabbe & R. Vyverman, 1997. Five new species of Biremis (Bacillariophyta) from Tasmania. Phycologia 36: 91–102. Vyverman, W., R. Vyverman, D. Hodgson & P. Tyler, 1995. Diatoms from Tasmanian mountain lakes: a reference dataset (TASDIAT) for environmental reconstruction and a systematic autecological study. Bibliotheca Diatomologica 33: 1–193. Vyverman, W., R. Vyverman, V. S. Rajendran & P. Tyler, 1996. Distribution of benthic diatom assemblages in Tasmanian highland lakes and their possible use as indicators of

environmental changes. Canadian Journal of Fisheries and Aquatic Sciences 53: 493–508. Vyverman, W., K. Sabbe, D. G. Mann, C. Kilroy, R. Vyverman, K. Vanhoutte & D. Hodgson, 1998. Eunophora gen nov. (Bacillariophyta) from Tasmania and New Zealand: description and comparison with Eunotia and amphoroid diatoms. European Journal of Phycology 33: 95–111. Watanabe, M. M., S. Mayama, M. Hiroki & H. Nozaki, 2000. Biomass, species composition and diversity of epipelic algae in mire pools. Hydrobiologia 421: 91–102. Williams, D. M. & F. E. Round, 1987. Revision of the genus Fragilaria. Diatom Research 2: 267–288. Winter, J. & H. Duthie, 2000. Epilithic diatoms as indicators of stream total N and total P concentration. Journal of the North American Benthological Society 19(1): 32–49. Yacobi, Y. Z., J. J. Alberts, M. Takacs & M. McElvaine, 2003. Absorption spectroscopy of colored dissolved organic carbon in Georgia (USA) rivers: the impact of meolecular size distribution. Journal of Limnology 62: 41–46.

Hydrobiologia (2006) 561:111–117  Springer 2006 R.J. Stevenson, Y. Pan, J.P. Kociolek & J.C. Kingston (eds), Advances in Algal Biology: A Commemoration of the Work of Rex Lowe DOI 10.1007/s10750-005-1608-0

The relationships among disturbance, substratum size and periphyton community structure Mark R. Luttenton1,* & Cathy Baisden1,2 1

Biology Department and Annis Water Resources Institute, Grand Valley State University, Allendale, MI 49401, USA Current address: South Hill Academy, 50 Spencer Street, Battle Creek, MI 49014, USA (*Author for correspondence: Tel.: 414-331-2503; E-mail: [email protected]) 2

Key words: periphyton, disturbance, substratum size, tile substrata

Abstract Numerous studies have determined the effects of physical disturbance on periphyton, however, the substrata used have varied in size among studies. In this study we examined the influence of substratum size on the change in periphyton exposed to three levels of disturbance. Periphyton communities were established in a large greenhouse tank on square unglazed tiles that were either 2.5, 5, or 7.5 cm on a side. Following community development sets of tiles were randomly assigned to controlled disturbances removing 25%, 50%, or 75% of the community or an undisturbed control. Following disturbance, changes in periphyton density were associated with both disturbance treatment and tile size as was taxa richness. Experimental results and direct observation revealed that algal growth was most concentrated along the edge of the tile and progressively declined toward the center. Thus, substratum size influences colonization and pre-disturbance community structure, which then affects the extent of periphyton community change due to different levels of disturbance.

Introduction Physical disturbance has been widely acknowledged to have significant effects on the structure and function of periphyton communities (Peterson, 1996). Indeed, Biggs (1996) suggested that periphyton biomass loss can occur primarily through disturbance rather than other processes. Scour during high flow regimes (Peterson & Stevenson, 1992; Biggs, 1995), substratum movement (Douglas, 1958), sediment abrasion (Horner et al., 1990) and grazers (Power, 1992; Steinman, 1996) are common mechanisms that impose disturbance on periphyton communities either singly or in combination. Ultimately, the affects of a disturbance event will be determined by characteristics of the disturbance (e.g. magnitude) and the structure and condition of the periphyton community

at the time of the disturbance (Biggs & Close, 1989). Our understanding of how disturbance affects periphyton communities has drawn heavily from experimental manipulations conducted in both the laboratory and field (e.g., Steinman et al., 1987; Tuchman & Stevenson, 1991). Many of these studies have exposed periphyton colonized on clay tiles to various disturbance regimes in an attempt to identify if a certain level of a disturbance (e.g., number of grazers) results in a statistically significant change in a community or algal populations. Although our understanding of the role of disturbance in structuring periphyton has been advanced by studies using tiles as substrata, some basic questions appear to remain regarding the design and application of these studies. One factor that deserves additional consideration is whether

112 tile size influences the results of a disturbance experiment. That is, will a periphyton community attached to a 2.52.5 cm tile be affected to the same extent as a periphyton community attached to a 55 cm tile when exposed to similar levels of disturbance? This experiment was conducted to determine if substratum size can influence the results of experiments using tile substrata to assess the affects of disturbance on periphyton communities exposed to controlled levels of disturbance. During community analysis, a related question emerged regarding the influence of substratum size on periphyton community structure prior to disturbance.

Materials and methods This study was conducted in a large (111 m) fiberglass tank located in a climate controlled greenhouse. The tank was filled with water from the nearshore zone of Lake Michigan after 15 small (2.52.5 cm), medium (55 cm) and large (7.57.5 cm) (45 total) unglazed white ceramic tiles had been randomly placed in the tank. Tiles were incubated for 5 weeks after the tank was inoculated with periphyton scraped from several rocks collected from a backwater area along the Grand River, Michigan. Inoculation was achieved by combining the periphyton in a large beaker and vigorously agitating the tank water as the inoculum was added. Periphyton from the Grand River was selected because of the great species richness, whereas Lake Michigan water was used because it had low nutrient concentrations. Tiles of each size were randomly assigned in triplicate to one of four treatments including; no disturbance (control), 25% disturbance, 50% disturbance, and 75% disturbance (Fig. 1a). Disturbance treatments were implemented by physically scraping the tile surface with a razor blade. The area disturbed formed a rectangle with one edge of the area coinciding with one edge of the tile. For example, the 50% disturbance treatment resulted in the algae on one side of the tile being completely removed whereas the algae on the adjoining half remained intact. Immediately after the disturbance was imposed, algae on the undisturbed section was removed using a razor blade and soft bristle brush,

homogenized and preserved in 4% formalin. Algal communities were analyzed quantitatively using a Palmer–Malony counting chamber at 400 magnification with organisms grouped taxonomically by genus. At least 500 cells were identified to genus level and counted from each sample in random fields of view to determine cell densities (cells/cm2, based on the whole tile area) and taxa richness. Differences among treatments were determined using 2-way ANOVA (p=0.05) and a Tukey test (p=0.05). A second experiment was conducted to evaluate possible mechanisms to better explain differences among control substrata observed during the initial experiment. The tank used during the first experiment was drained and refilled with Lake Michigan water after randomly placing five small, medium and large (15 total) unglazed white ceramic tiles in the tank. Prior to submersion, a single glass cover slip was affixed to each tile with a small drop of silicon sealer. One edge of the cover slip was aligned with one edge of the tile and passed through the middle of the tile. After a 5 week colonization period, each cover slip was removed and mounted to slides with a syrup medium (Stevenson, 1984a, b). Algae were counted along three transects: one at the edge, one halfway between the edge and middle, and one at the middle of the cover slip; these locations corresponded to similar locations on the tile. Transects ran parallel to the edge of the tiles. A minimum of 300 cells were tallied for each region from randomly selected fields along each transect (Fig. 1b).

Results Total algal abundances decreased with increasing disturbance but were inversely related to tile size (Fig. 2). Differences in algal densities associated with disturbance treatments, substratum size and the interaction term were all significant (Table 1). Multiple comparisons could not separate disturbance treatments but densities associated with tile size were different among all treatments (Table 1). To better understand the relationship between substratum size and differences in algal densities, we analyzed differences among control treatments of small, medium and large substrata. Cell densities on large tiles and medium were only 22.4% and 48.9%

113 Disturbance Treatments 25%

50%

75% Edge-EffectTransects

(a)

(b)

Figure 1. Illustration of disturbance treatments (a) and coverglass placement on tiles and location of 3 transects (dotted rectangles) examined (b).

of densities on small tiles respectively. One-way ANOVA and Tukey’s test indicated that total algal abundances on small, medium and large control treatments all differed significantly (p<0.001). Regression analysis was used to further evaluate the interaction between tile size and disturbance. Algal densities were highly correlated with

disturbance on small (R2=0.95, p<0.001), medium (R2=0.89, p<0.001) and large (R2=0.81, p<0.001) tiles. In addition, comparing the slopes of each regression line revealed that they were all significantly different (F=92.9) indicating that algal loss due to the disturbance was not proportional among tile sizes. For example, cell densities

2 Small tiles Medium tiles

1.5

Cells/cm2 x 106

Large tiles

1

0.5

0 Control

25%

50%

75%

Treatment Figure 2. Total algal abundance on tiles of three different sizes exposed to four levels of disturbance.

114 Table 1. Two-way ANOVA comparing algal densities on tiles of different sizes exposed to four levels of disturbance Source of

df

F

p

variation Disturbance

3

99.86

0.001

Size

2

242.24

0.001

Interaction

6

2.52

0.046

Disturbance

Cont

Size

Small

25%

50%

Medium

75% Large

Treatments with a common underline are not significantly different (Tukey test, p=0.05).

on small, medium and large tiles after a 25% disturbance were 68, 56 and 49% of the control treatment densities. A total of 36 genera were identified, 9 of which were called abundant because they accounted for at least 1% of the total number of cells in two or more replicates. Scenedesmus sp., Coelospherium sp. and coccoid blue-green algae dominated all treatments, each accounting for 10–30% of each sample. The total number of genera ranged from 12 to 28 (Fig. 3), however all treatments were dominated by the same four taxa. Small tiles

supported fewer taxa than large tiles (Fig. 3). The number of taxa associated with levels of disturbance differed significantly from the control and 25% disturbance treatments differing from each other and the other treatments (Table 2). Significant differences were also observed among size treatments. However, multiple comparison tests could only separate treatments on large tiles (Table 2). The interaction term was not significant (Table 2). Algal density was not evenly distributed across the surface of the experimental tiles. Algal densities were highest within a 3 mm strip along the edge of the tiles and progressively declined from the edge to the center of the tile (Fig. 4). Two-way ANOVA indicated that cell densities differed significantly among each location (Table 3). Tile size had no influence on algal densities and there was no interaction between location and size.

Discussion The fact that the disturbance regime imposed during this study resulted in a significant difference among treatments is certainly no surprise. How-

30 Small tiles Medium tiles

25

Large tiles

Taxa Richness

20

15

10

5

0

Control

25%

50%

75%

Treatment Figure 3. Species richness on tiles of three different sizes exposed to four levels of disturbance.

115 Table 2. Two-way ANOVA comparing taxa richness on tiles exposed to four levels of disturbance Source of

df

F

p

variation

Table 3. One-way ANOVA comparing algal densities at three locations on tiles of three different sizes Source of

df

F

p

variation

Disturbance

3

39.93

0.001

Location

2

21.33

0.001

Size

2

9.71

0.001

Size

2

0.75

0.485

Interaction

6

1.50

0.208

Interaction

4

0.37

0.822

Distance

Edge

Disturbance

Cont

Size

Small

25%

50% Medium

Midway

Center

75% Large

Treatments with a common underline are not significantly different (Tukey test, p=0.05).

Treatments with a common underline are not significantly different (Tukey test, p=0.05).

ever, it is somewhat surprising that these data suggest that tile size did influence the results of the experiment. Both the ANOVA interaction term and analysis of regression slopes indicated that substrata size did affect the outcome and may partially determine the magnitude of disturbance required to realize a significant result. Thus, when disturbance experiments employ tile substrata, consideration should be given to the size of the substrata selected. In addition, tile size may need to be considered when scaling results of experiments up to algae on rock in natural ecosystems.

Because naturally occurring substrata are often quite variable in size, experimental tiles may approximate the surface area of only a fraction of natural substrata. Consequently, the results of experiments in which tile size is not manipulated may not completely reflect potential changes in algal communities on natural substrata, simply due to size-mediated conditions and responses. Decreasing cell density and decreasing disturbance effects with tile size is likely due to higher colonization of tiles near edges than in the middle. Our colonization experiment showed algal densities were greatest within a strip approximately

40 Small tiles Medium tiles

Cells/cm2 x104

30

Large tiles

20

10

0 Edge

Midway

Middle

Treatment Figure 4. Total algal abundance along the edge, near the center and midway between the edge and center of clay tiles of three different sizes.

116 3 mm wide along the edge of tiles regardless of tile size, and that density decreased with increasing distance from the edge. The difference in algal abundances with tile size can be explained by this edge effect, with 3 mm wide edges accounting for 47, 23 and 15% of the total surface area of small, medium and large tiles, respectively. Differing effects of disturbance with tile size could also be related to the ratio of edge area:total surface area as well. For example, a 25% disturbance would have relatively less effect on total cell densities on small tiles than on total algal abundance on a large substratum. In this case, the 25% disturbance treatment removed 32, 44 and 51% of the algae on small, medium and large tiles respectively. Colonization patterns (Korte & Blinn, 1983) and the microdistribution of algae on artificial substrata (Tuchman & Stevenson, 1980; Steinman & McIntire, 1986; McCormick, 1991) have received considerable attention. Other studies have examined microscale distribution on natural substrata (Hamilton & Duthie, 1984; Krejci & Lowe, 1986; Miller et al., 1987) or have evaluated structural–functional relationships (Lock et al., 1984). Periphyton community development near edges may be accelerated by disruption of currents near edges (Stevenson, 1983). Many have observed greater accrual on substrata at the upstream edge (Tuchman & Stevenson, 1980; Korte & Blinn, 1983; McCormick, 1991). Even in the lentic mesocosms of our experiments, convection currents may be sufficient to generate modest edge effects. Greater taxa richness on larger substrata is consistent with some early studies conducted by Patrick (1967, 1976). Patrick (1967) found more species on substrata with larger surface area and compared this to processes of island biogeography, immigration and competition (MacArthur & Wilson, 1967). The results of our study may not be completely explained by the same theory because our system was lentic whereas Patrick’s (1967) study was conducted in a flowing system, where immigration is probably more important. Large substrata may also have greatest habitat diversity as a balance between edge and center of tiles, which may support greater biodiversity than small substrata. The decrease in species richness among disturbance treatments may have been due to the loss of rare taxa or loss of habitat. Patrick (1976) found that 70–80% of the taxa reported in her

study occurred as four individuals or fewer. Increasing levels of disturbance would systematically remove these rare taxa, especially if they have patchy distributions. Relating results of experiments to natural systems is critical for linking cause-effect relations determined in experiments to patterns observed in nature. Future studies should continue to address issues related to scaling results of experiments up to field scale patterns. Acknowledgements We wish to express our appreciation to R. J. Stevenson and an anonymous reviewer for helpful comments on a previous draft. A. Steinman and D. DeNicola assisted with some of the statistical analysis. References Biggs, B. J. F., 1995. The contribution of disturbance, catchment geology and land use to the habitat template of periphyton in stream ecosystems. Freshwater Biology 33: 419–438. Biggs, B. J. F., 1996. Patterns in benthic algae of streams. In Stevenson R. J., M. L. Bothwell & R. L. Lowe (eds), Algal Ecology: Freshwater Benthic Ecosystems. Academic Press, Inc, San Diego, CA: 31–56. Biggs, B. J. F. & M. E. Close, 1989. Periphyton biomass dynamics in gravel bed rivers: The relative effects of flows and nutrients. Freshwater Biology 22: 209–231. Douglas, B., 1958. The ecology of the attached diatoms and other algae in a small stony stream. Journal of Ecology 46: 295–322. Hamilton, P. B. & H. C. Duthie, 1984. Periphyton colonization of rock surfaces in a boreal forest stream studied by scanning electron microscopy and track autoradiography. Journal of Phycology 20: 525–532. Horner, R. R., E. B. Welch, M. R. Seeley & J. M. Jacoby, 1990. Responses of periphyton to changes in current velocity, suspended sediment and phosphorus concentrations. Freshwater Biology 24: 215–232. Korte, V. L. & D. W. Blinn, 1983. Diatom colonization on artificial substrata in pool and riffle zones studied by light and scanning electron microscopy. Journal of Phycology 19: 332–341. Krejci, M. E. & R. L. Lowe, 1986. Importance of sand grain mineralogy and topography in determining micro-spatial distribution of epipsammic diatoms. Journal of the North American Benthological Society 5: 211–220. Lock, M. A., R. R. Wallace, J. W. Costerton, R. M. Ventullo & S. E. Charlton, 1984. River epilithon: Toward a structural– functional model. Oikos 42: 10–22.

117 MacArthur, R. H. & E. O. Wilson, 1967. The Theory of Island Biogeography. Princeton University Press, Princeton, N.J 203 pp. McCormick, P., 1991. Spatial considerations in the study of benthic algal colonization in streams. Transactions of the American Microscopical Society 110: 279–288. Miller, A. R., R. L. Lowe & J. T. Rotenberry, 1987. Succession of diatom communities on sand grains. Journal of Ecology 75: 693–709. Patrick, R., 1967. The effect of invasion rate, species pool, and size of area on the structure of the diatom community. Proceedings of the National Academy of Science 58: 1335– 1342. Patrick, R., 1976. The formation and maintenance of benthic diatom communities. Proceedings of the American Philosophical Society 120: 475–484. Peterson, C. G., 1996. Response of algae to natural physical disturbance. In Stevenson R. J., M. L. Bothwell & R. L. Lowe (eds), Algal Ecology: Freshwater Benthic Ecosystems. Academic Press, Inc, San Diego, CA: 373–402. Peterson, C. G. & R. J. Stevenson, 1992. Resistance and resilience of lotic algal communities: Importance of disturbance timing and current. Ecology 73: 1445–1461. Power, M. E., 1992. Hydrologic and trophic controls of seasonal algal blooms in northern California rivers. Archiv fu¨r Hydrobiologie 125: 385–410. Steinman, A. D., 1996. Effects of grazers on freshwater benthic algae. In Stevenson R. J., M. L. Bothwell & R. L. Lowe

(eds), Algal Ecology: Freshwater Benthic Ecosystems. Academic Press, Inc, San Diego, CA: 341–373. Steinman, A. D. & C. D. McIntire, 1986. Effects of current velocity and light energy on the structure of periphyton assemblages in laboratory streams. Journal of Phycology 22: 352–361. Steinman, A. D, C. D. McIntire, S. V. Gregory, G. A. Lamberti & L. Ashkenas, 1987. Effect of herbivore type and density on taxonomic structure and physiognomy of algal assemblages in laboratory streams. Journal of the North American Benthological Society 6: 189–197. Stevenson, R. J., 1983. Effects of current and conditions simulating autogenically changing microhabitats on benthic algal immigration. Ecology 64: 1514–1524. Stevenson, R. J., 1984a. How currents on different sides of substrates in streams affect mechanisms of benthic algal accumulation. International Revue der gesamten Hydrobiologia 69: 241–262. Stevenson, R. J., 1984b. Procedures for mounting algae in syrup medium. Transactions of the American Microscopical Society 97: 320–321. Tuchman, M. L. & R. J. Stevenson, 1980. Comparison of clay tiles, sterilized rock, and natural substrate diatom communities in a small stream in southeastern Michigan, USA. Hydrobiologia 75: 73–79. Tuchman, N. C. & R. J. Stevenson, 1991. Effects of selective grazing by snails on benthic algal communities. Journal of the North American Benthological Society 10: 430–443.

Hydrobiologia (2006) 561:119–130  Springer 2006 R.J. Stevenson, Y. Pan, J.P. Kociolek & J.C. Kingston (eds), Advances in Algal Biology: A Commemoration of the Work of Rex Lowe DOI 10.1007/s10750-005-1609-z

Relationships between environmental variables and benthic diatom assemblages in California Central Valley streams (USA) Yangdong Pan1,*, Brian H. Hill2, Peter Husby3, Robert K. Hall4 & Philip R. Kaufmann5 1

Environmental Science and Resources, Portland State University, Portland, Oregon, USA USEPA, ORD, NHEERL/MED, Duluth, MN, USA 3 USEPA, Region 9, Laboratory (PMD-2), Richmond, CA, USA 4 USEPA, Region 9, Water Division (WTR-2), San Francisco, CA, USA 5 USEPA, ORD, NHEERL/WED, Corvallis, OR, USA (*Author for correspondence: E-mail: [email protected]) 2

Key words: canonical correspondence analysis, cluster analysis, UPGMA, TWINSPAN, classification tree analysis

Abstract This study examines distributional patterns of benthic diatom assemblages in relation to environmental characteristics in streams and rivers in the California Central Valley ecoregion. Benthic diatoms, water quality, and physical habitat conditions were characterized from 53 randomly selected sites. The stream sites were characterized by low mid-channel canopy cover and high channel substrate embeddedness. The waters at these sites were enriched with minerals and turbidity varied from 1.3 to 185.0 NTU with an average of 13.5 NTU. A total of 249 diatom taxa were identified. Average taxa richness was 41 with a range of 7–76. The assemblages were dominated by Staurosira construens (11%), Epithemia sorex (8%), Cocconeis placentula (7%), and Nitzschia amphibia (6%). Multivariate analyses (cluster analysis, classification tree analysis, and canonical correspondence analysis) all showed that benthic diatom assemblages were mainly affected by channel morphology, in-stream habitat, and riparian conditions. The 1st CCA axis negatively correlated with mean wetted channel width (r=)0.66) and thalweg depth (r=)0.65) (Table 4). The 2nd axis correlated with % coarse substrates (r=0.60). Our results suggest that benthic diatoms can be used for assessing physical habitat alterations in streams.

Introduction Anthropogenic activities such as damming and diverting waters for irrigation of agricultural land, urban water use, and flood control have substantially altered lotic environments worldwide, especially physical habitat conditions (Mount, 1995). The intensities of these activities and their effects on the lotic ecosystems are most noticeable in semi-arid and arid regions with increasing agricultural and urban development such as in California Central Valley ecoregion. Benthic diatoms have been an integral part of both national and

state stream assessment programs (Kroeger et al., 1999; Stevenson & Bahls, 1999). Using diatoms as indicators of water quality has a long history. Researchers have shown that changes in diatom assemblages are often associated with eutrophication, acidification, heavy metal contamination, and pesticides (see review by Stevenson & Pan, 1999). Assessment of physical habitat conditions with diatoms, however, has received less attention. Several conceptual models have suggested that physical habitat conditions may closely associate with benthic diatom assemblage structure in streams (Biggs, 1996; Stevenson, 1997; Biggs

120 et al., 1998). In their habitat matrix model, Biggs et al. (1998) extended the CSR model developed by Grime (1977) for terrestrial plants to benthic algae. In a given stream site with open canopy, diatom assemblages may be largely predicted by habitat stability and resource supply. Empirical studies have shown the close association between diatom life strategies/composition and habitats. For instance, diatom assemblages in habitats with excessive and periodic fine sediment deposition are often dominated by the species with relatively high motility (e.g., Gyrosigma, Surirella, Nitzschia) while epipsammon, diatoms attached on unstable sandy substrates, often consists of small adnate species such as Achnanthidium minutissimum (Miller et al., 1987). Recently, several diatom metrics (e.g., % Achnanthes/Achnanthidium, % biraphids) have been suggested as potential indicators of stream physical habitat conditions (Stevenson & Bahls, 1999). Kutka & Richards (1996) reported that % bank erosion and overall riparian conditions correlated well with these diatom metrics in a Minnesota agricultural basin. The main objective of this study was to identify distributional patterns of benthic diatom assemblages in relation to physical habitat conditions and other environmental variables in California Central Valley ecoregion. A better understanding of changes in benthic diatom assemblages in relation to environmental characteristics will eventually enhance development of numerical criteria for physical, chemical, and biological attributes in the region and state.

extensive modifications, sampled ‘‘streams’’ in this study include ditches, drains, natural streams and rivers, and isolated water-bodies which do not connect to any downstream channels. Natural streams on the eastern and central portion of the valley have narrow riparian habitat with agriculture and levee construction next to the streams, and constructed conveyances (ditches) as having little to no habitat. In some water districts, constructed conveyance habitat is lined with concrete or sprayed with herbicides to control vegetation. Streams on the western side of the valley are predominantly ephemeral or intermittent. Land use on the western side of the valley is predominantly rangeland with a few valley oak. Sampling sites were selected using a tessellation stratified design, described by Stevens & Olsen (1999) and Stevens (1997), to represent the two main populations of interest: natural streams and man-made waterways (Hall et al., 2000). The stream population in the region was estimated based on digitized versions of the USGS 1:100,000 scale topographic maps. A study reach (40 mean wetted channel width), ranging from 150 to 500 m, was sampled around each of the selected sample sites. Eleven cross-section transects were set up in each study reach by dividing the reach into 10 equal length intervals (includes transects at the start and end of each reach). Periphyton, water quality, and physical habitat conditions were characterized from 53 unique sites during the baseflow from early August to late September in 1994 and 1995. Sampling and sample analysis

Materials and methods Study area and design The Central Valley ecoregion is surrounded by the Coast Ranges on the west and the Sierra Nevada foothills on the east. The ecoregion covers two major drainage basins, the Sacramento River in the north, and San Joaquin River in the south. The climate ranges from semiarid in the north to arid in the south. Streams in the ecoregion have been significantly modified by human activities, including damming and diverting waters for irrigation of agricultural land, urban water use, and flood control (Saiki, 1984; Mount, 1995). Due to these

Periphyton samples were collected from each of the 9 transects (excludes transects at the start and end of each reach) and combined into either a depositional (pool/glide) or erosional habitat (riffle/run) composite sample. Transects with no visible water movement were defined as depositional habitat, those with visible water movement were considered erosional habitat. At each transect, periphyton were collected from a 12 cm2 area of stream bed using a 1.5 cm long piece of 3.9 cm diameter PVC pipe as a template. For fine substrate, periphyton were suctioned into a 60 ml syringe; in coarser substrate, periphyton were scraped off using a toothbrush and rinsed with

121 distill water. The end result was one composite periphyton sample for erosional habitats and one for depositional habitats for each stream site. Depositional samples were collected at only few sites and thus this study was based on erosional samples. Samples were processed for chl a, ash-free dry mass, and species composition. The subsample of the preserved periphyton suspension (final concentration of 4% formalin) was acid-cleaned and mounted in NAPHRAX to enumerate diatom species (Patrick & Reimer, 1966). A minimum of 500 diatom valves was counted at 1000 magnification unless the sample was silty with sparse cells. Sites with less than 250 valve counts were excluded from the analysis. Patrick & Reimer (1966, 1975) and Krammer & Lange-Bertalot (1986; 1988; 1991a, b) were used as primary references for diatom taxonomy. Diatom taxa mobility classification follows Stevenson & Bahls (1999). A 4-l sample of water was collected at a flowing portion near the middle of each stream. Within 48–72 h of collection, water samples were split into aliquots and preserved for different assays (see Table 1). The aliquots for dissolved metals were filtered (0.45 lm pore size) and preserved with concentrated HNO3. The aliquots for nutrient analysis were preserved with concentrated H2SO4. Base cations and metals were determined by ICP spectrophotometry. Anions were measured by ion chromatography. Nutrients were determined with a flow-injection analyzer. Detailed information on the analytical procedures used for each of the analyses can be found in USEPA (1987). Vegetative cover over the stream was measured at each of the 11 cross section transects using a convex spherical densiometer. At each transect, epifaunal substrate (e.g., grain size, embeddedness) and presence and proximity of 11 categories of human activities (i.e., row crops, pasture, dams and revetments, buildings, pavement, roadways, pipes, landfill or trash, parks/lawns, logging operations, and mining activities) were estimated (Kaufmann & Robison, 1998). Proximity-weighted riparian disturbance indices were calculated by tallying the number of riparian stations at which a particular type of disturbance was observed, weighting by its proximity to the stream, and then averaging over all stations on the reach (Kauf-

Table 1. Summary of selected environmental and biotic variables Variables

Median Minimum Maximum

Aluminum (lg l)1) Boron (lg l)1)

116 76

48 6

Calcium (mg l)1)

20.65

1.77

Iron (lg l)1)

56

6

Magnesium (mg l)1)

6.33

Manganese (lg l)1)

4

0.359 BD

297 4300 90.1 810 60.4 649

Potassium (mg l)1)

1.565

0.612

Selenium (lg l)1)

0.8

0.7

5.0

Sodium (mg l)1) Chloride (mg l)1)

14.65 8.00

1.57 0.66

242.00 365.00

25

Sulfate (mg l)1)

5.70

0.59

468.00

Nitrate (mg l)1)

0.08

0.01

22.00

Ammonia (mg l)1)

0.10

0.10

TP (lg l)1) Turbidity (NTU) Chl-a (lg cm)2)

100 13.5 0.22

29 1.3 BD

1.50 3600 185.0 4.37

AFDM (mg cm)2) Mean bank angle (degree)

0.84 42

Mean bankside

82

0

100

9

0

100

0

100

0.05 15

23.60 95

canopy cover (%) Mean mid-channel canopy cover (%) Mean embeddedness (%)

100

% filamentous algae cover

0.05

0.00

0.88

Total riparian disturbance Non-ag. riparian

3.42 2.70

0.00 0.00

6.38 5.67

0.71

0.00

1.50

0.75

)1.86

1.73

Log10 relative bed stability

)2.41

)3.94

4.13

Mean thalweg depth (cm) Mean wetted

82.61 5.85

5.00 0.29

334.88 59.57

disturbance Agriculture riparian disturbance Log10 erodible substrate diameter

stream width (m) BD: below detection.

mann & Robison, 1998). Stream habitat characterization included thalweg, mean wetted width, mean reach cross-section, width/depth, slope, residual pool area, and surficial channel substrates. Field methods and metric calculation, respectively, are described in more detail by Kaufmann & Robison (1998) and Kaufmann et al. (1999).

122 Data analysis We used two approaches to examine the relationships between diatom assemblages and environmental variables in the region. In a region where landscapes have been substantially modified by human activities, it is often difficult or expensive to adequately characterize environmental variation. On the other hand, benthic diatom assemblages may integrate both temporal and spatial variation of environmental conditions. With well-developed sampling and analytical protocols (Stevenson & Balhs, 1999), we can characterize benthic diatom assemblages much better than the environmental variables especially human-related disturbance. The first approach was an indirect analysis of relationships between diatom species composition and environmental factors. We characterized diatom distributional patterns using clustering methods. Two-way indicator species analysis (TWINSPAN), a divisive cluster method, uses a ‘top-down’ approach. The division starts with the entire data set based on ‘global’ differences in the data set and successively divides it into smaller groups (Hill et al., 1975). A flexible, unweighted pair-group method (UPGMA) clustering method, an agglomerative hierarchical method, uses a ‘bottom-up’ approach, which starts with individual sites and repeatedly combines them into larger groups. Rare taxa, defined as <1% relative abundance with <3 sites occurrences, were excluded from the analyses. The UPGMA was performed using the Bray–Curtis distance with a b-value of )0.25. To better characterize each diatom-based cluster group, we performed an indicator species analysis. This analysis identifies a set of indicator species affiliated with each stream site group (Dufrene & Legendre, 1997). Statistical significance of each species indicator value was tested using a Monte Carlo permutation test (999 permutations, p<0.05). TWINSPAN, UPGMA, and indicator species analysis were performed using PC-ORD v.4 statistics software (McCune & Mefford, 1999). A classification tree analysis was used to identify a set of environmental predictors, which can best discriminate diatom-based groups. The method is an attractive alternative to linear and additive regression methods. This binary recursive

partitioning method selects a set of hierarchically organized environmental variables that reveals relative importance of each selected variable and their interactions with relation to response variables (Clark & Pregibon, 1993). The classification tree analysis was performed using S-plus statistics software (MathSoft, 2000). For all analyses, continuous environmental variables, except pH, were log-transformed. Proportional environmental variables were double transformed with square root and then arcsine. The second approach is a direct analysis on the relationship between diatom assemblages and environmental variables. Canonical correspondence analysis (CCA) focuses on the portion of the diatom assemblages that co-varies with measured environmental variables (ter Braak, 1986). Water quality variables and physical habitat variables were selected using a forward-selection option. The importance of these variables in relation to diatom assemblages was tested using a Monte Carlo permutation test (999 permutations, p<0.05). CCA was performed using CANOCO v.4 statistics software (ter Braak & Smilauer, 1998).

Results Environmental conditions Sampled stream sites varied considerably in stream types, physical habitat conditions, and water quality (Table 1). Approximately 58% of the sites were in the San Joaquin River Basin. A large proportion of the sites (66%) were ditches and drains. Only 28% of sites were regular, natural streams. Of these natural sites, stream sizes ranged from 1st to 8th order. Mean wetted stream width varied from 0.29 m to 59.57 m, and mean thalweg depth ranged from 0.05 to 3.35 m. The study stream sites were characterized by low mid-channel canopy cover and high channel substrate embeddedness. Where present, vegetative ground cover consisted of grasses, forbs and bramble, mid-story consisted predominantly of shrub willow, and the upper-story/canopy consisted of cottonwoods and some eucalyptus trees. In the man-made conveyances the larger mean substrate grain size is attributed to the presence of

123 hardpan, a hard caliche layer below valley floor sediments. Channel substrate and composition for ditches and drains is predominantly fine sand and silty-clay with only 15% of sites having a substrate of coarse sand and gravel. In natural streams, 86% had a substrate composed of coarse sand and larger material. Median % fine substrates (<0.08 mm) and % embeddedness were 95 and 100%, respectively. Stream waters were overall mineral enriched. The median Ca2+ concentrations were )1 20.65 mg l with a maximum value of 90.10 mg l)1. The median Na+ concentrations were 14.65 mg l)1 with a maximum value of 242.00 mg l)1. Turbidity varied from 1.3 to 185.0 NTU with a median value of 13.5 NTU. Total phosphorus concentrations varied from 29 to 3600 lg l)1 with a median value of 100 lg l)1. Diatom-based stream site classification A total of 249 diatom taxa were identified. Average taxa richness was 41 with a range of 7–76. On average, the assemblages were dominated by taxa in 2 genera, Nitzschia (23%) and Navicula (18%). Staurosira construens Ehr. (11%), Epithemia sorex Ku¨tz. (8%), Cocconeis placentula Ehr. (7%), and Nitzschia amphibia Grun. (6%) were the most common species. Most of taxa occurred in <3 sites with <1% relative abundance. After deleting these ‘rare’ taxa, 65 taxa were used for data analyses. The assemblages were characterized by a relatively high abundance of motile diatom taxa. Average relative abundances of diatom taxa with high mobility (e.g., Nitzschia) or medium mobility (e.g., Navicula) were 25 and 35%, respectively. (Fig. 1). The UPGMA divided 53 sites into four groups based on visual examination of the dendrogram (Fig. 2). Indicator species analysis showed that each group was characterized by different sets of indicator taxa with maximum relative abundance and frequency (Table 2). Group A was characterized by Amphora veneta Ku¨tz. and Nitzschia amphibia. A total of 11 indicator species including Navicula cryptotenella Lange-Bert., Cymbella silesiaca Bleisch, and Achnanthidium minutissimum (Ku¨tz.) Czarnecki had significant indicator values for Group B. Group C was characterized by a centric diatom (Cyclotella meneghiniana Ku¨tz.)

and a chain-forming biraphid (Diadesmis confervacea Ku¨tz.) and other two taxa. Group D was characterized by two araphids (Staurosira construens, Synedra parasitica (W. Sm.) Hust.) and one monoraphid (Cocconeis placentula). For the comparison, the TWINSPAN also produced four groups. Group A was characterized by Aulacoseira granulata (Ehr.) Simonsen, Epithemia turgida (Ehr.) Ku¨tz., and Staurosira construens. Nitszchia constricta (Ku¨tz.) Ralfs was the only significant indicator taxon for Group B. Group C was characterized by Amphora pediculus (Ku¨tz.) Grun., Diatoma vulgare Bory, Rhoicosphenia abbreviate (Agardh) Lange-Bert., Nitszchia inconspicua Grun. and three Navicula taxa. Group D was characterized by Planothidium lanceolatum (Bre´b. ex Ku¨tz.) Lange-Bert., Cymbella silesiaca, and other four taxa. The correspondence between the stream groups classified using the UPGMA and TWINSPAN was poor (Table 3). For example, 19 sites in UPGMA Group B were split equally between the TWINSPAN group C and D. The UPGMA groups corresponded poorly to the Sacramento and San Joaquin drainage basins and ‘streams’

Figure 1. The State of California showing sampling locations.

124 Distance (Objective Function) 4.7E-02

3.2E+00

CA014S CA038S CA186S CA108S CA055S CA104S CA152S CA056S CA250S CA138S CA200S CA026S CA066S CA084S CA094S CA096S CA106S CA040S CA044S CA045S CA273S CA046S CA071S CA022S CA212S CA142S CA223S CA227S CA032S CA050S CA042S CA357S CA065S CA374S CA302S CA317S CA347S CA341S CA033S CA213S CA156S CA286S CA068S CA372S CA280S CA309S CA164S CA079S CA091S CA322S CA377S CA369S CA244S

6.4E+00

9.5E+00

1.3E+01

Group3 A

B

C

D7

Figure 2. Dendrogram of diatom-based stream site classification using the UPGMA method. The dashed line is the cutting line for defining 4 diatom-based stream site groups.

types (i.e., natural stream, man-made conveyances). The TWINSPAN groups showed a much better correspondence to the drainage basins (Table 3). All three sites in Group A and all sites in Group D except one were located in the San Joaquin River Basin. Most of the sites in Group C were in the Sacramento River Basin.

The classification tree analysis showed that hierarchically organized physical habitat variables might be important in relation to the variability among diatom-based groups (Fig. 3). Channel morphology (mean bankfull width), in-stream habitat (relative stream bed stability) and riparian conditions (riparian disturbance index) were

125 Table 2. Summary of indicator species analysis showing indicator taxa, relative abundance, relative frequency, and indicator value for each diatom-based stream group classified using the UPGMA method Taxa

Relative abundance

Relative frequency

Indicator value

Group

Group

Group

A Amphora veneta Ku¨tz. Nitzschia amphibia Grun.

B

C

D

A

B

C

D

A

B

C

D

77

7

8

9

73

37

50

36

56

2

4

3

66

8

13

14

100

74

83

73

66

6

11

10

Planothidium lanceolatum (Bre´b. ex Ku¨tz.) Lange-Bert. Achnanthidium minutissimum (Ku¨tz.) Czarnecki

5

56

20

19

36

89

83

73

2

50

17

14

12

70

3

15

45

89

42

55

5

62

1

8

Caloneis bacillum (Grun.) Cl.

23

55

13

9

45

79

42

27

10

43

5

3

4

67

7

22

27

95

42

64

1

64

3

14

Cymbella silesiaca Bleisch C. sinuata Greg.

12

68

0

21

18

63

0

27

2

43

0

6

Diatoma vulgare Bory Fragilaria capucina Desm.

3 20

89 56

5 19

4 6

9 18

63 74

8 42

9 36

0 4

56 41

0 8

0 2

Navicula cryptotenella Lange-Bert.

15

3

73

2

10

45

95

8

36

7

69

0

N. viridula (Ku¨tz.) Ku¨tz. emend. V. H. Nitzschia dissipata (Ku¨tz.) Grun.

9

64

13

14

27

63

42

27

2

40

5

4

13

73

5

9

45

79

33

27

6

58

2

2

Rhoicosphenia abbreviate (Agardh) Lange-Bert.

15

57

3

24

55

79

33

64

8

45

1

16

8

10

76

6

55

84

92

64

4

9

70

4

6

5

90

0

18

47

58

0

1

2

52

0

29 13

7 2

64 83

0 2

45 18

21 16

67 50

0 9

13 2

2 0

43 41

0 0

Cyclotella meneghiniana Ku¨tz. Diadesmis confervacea Ku¨tz. Nitzschia calida Grun. Rhopalodia gibberula (Ehr.) O. Mu¨ll. Cocconeis placentula Ehr.

6

30

3

61

55

100

58

91

3

30

2

56

Staurosira construens Ehr.

3

7

8

83

36

84

58

91

1

6

5

75

Synedra parasitica (W. Sm.) Hust.

6

6

1

88

9

26

8

55

1

1

0

48

The bold numbers are significant indicator values (p<0.05, Monte Carlo permutation test).

Table 3. Correspondence among diatom-based stream groups, drainage basins, and ‘‘stream’’ types TWINSPAN group A

Drainage basin

B

C

D

Sacramento

‘‘Stream’’ type San Joaquin

Ditch/drain

Natural

Isolated

UGPMA group A (11)

0

8

2

1

7

4

10

0

1

B (19) C (12)

0 0

1 8

9 2

9 2

9 5

10 7

15 6

4 5

0 1

D (11)

3

1

0

7

2

9

4

6

1

A (3)

0

3

2

1

0

B (18)

11

7

11

6

1

C (13)

11

2

9

3

1

D (19)

1

18

13

5

1

TWINSPAN group

The numbers in the parenthesis are the total number of sampled sites for each group. The numbers in the table show the distribution of the sites among different categories of each classification.

126 identified as important variables to discriminate among the diatom-based groups classified using the UPGMA (Fig. 3). For the TWINSPAN groups, riparian conditions (mean bankside canopy cover) channel morphology (mean bankfull width, mean thalweg depth) and in-stream habitat conditions (erodible substrate diameters) were identified. Relative importance of water quality and physical habitat conditions on diatom assemblages CCA with a forward selection option identified channel morphology (mean wetted stream width, mean thalweg depth, mean bank angle), in-stream habitat (% coarse substrates), riparian conditions (non-agricultural riparian disturbance index), and

(a) UPGMA classification Mean bankfull width<5.4 m |

LRBSTST<-1.92 A Riparian disturbance index <3.98 B

C

D

(b) TWINSPAN classification Mean bankside canopy cover <80% |

Mean bankfull width<17.0 m

LTEST<1.01 Mean bankfull width<10.5 m

D

A

Mean thalweg depth <34.67 cm

D

B B

C

Figure 3. Classification tree analysis showing hierarchically organized physical habitat variables which were important in relation to the variability among diatom-based groups. LRBSTST: log10 relative bed stability, LTEST: log10 erodible substrate diameters.

cations (manganese) as important in relation to diatom assemblages (Monte Carlo permutation test, p<0.05) (Fig. 4). These variables explained a total of 21% variance in diatom species data set. The 1st CCA axis negatively correlated with mean wetted channel width (r=)0.66) and thalweg depth (r=)0.65) (Table 4). The 2nd axis correlated with % coarse substrates (r=0.60). Bivariate correlation analysis showed that % Cymbella significantly correlated with mean bank angle (r=0.51, p=0.0001, n=53). Percent of Navicula increased with mean wetted width (r=0.40, p=0.03, n=53).

Discussion Our analyses showed that stream diatom assemblages in the Central Valley ecoregion were mainly associated with physical habitat conditions. The classification tree analysis indicated that the variability among the diatom-based stream groups could be best accounted for by the variables associated with riparian conditions, in-stream habitat, and channel morphology. Canonical correspondence analysis indicated that variables associated with these same physical habitat categories could explain the largest amount of the variability in the diatom data set. It is not surprising that physical habitat conditions accounted for most of the variability in the diatom assemblages. Streams in this ecoregion have been extensively modified by human activities including intensive agriculture and urbanization (Mount, 1995). Most of these activities involve direct or indirect alteration of natural flow regimes and channel morphology in streams and rivers. Our results were consistent with several studies conducted in the region. As part of the same Regional Environmental Monitoring and Assessment Program (REMAP), Griffith et al. (2003) found that channel morphology and substrates were the major environmental variables related to aquatic macroinvertebrates. In a separate study, Brown & May (2000) reported that stream-size gradient associated with combined agricultural and urban land use in the basin was the major environmental gradient related to macroinvertebrates found on large woody debris. In contrast, Leland et al. (2001) reported that algae in the San Joaquin

127

+1.0

(a) CCA ordination site plot showing the TWINSPAN groups

PCTBIGR

XBKA XDEPTH

Manganes

XWIDTH

-1.0

W1HNOAG

-1.0

+1.0

+1.0

(b) CCA ordination site plot showing the UPGMA groups

PCTBIG

XBKA XDEPTH

Manganes

XWIDTH

-1.0

W1HNOAG

-1.0

+1.0

Figure 4. Ordination site plot of canonical correspondence analysis showing environmental variables identified by a forward selection method. PCTBIGR: % of coarse substrates (>16 mm), W1HNOAG: mean Non-agricultural riparian disturbance index, XWIDTH: mean wetted stream width, XDEPTH: mean thalweg depth, XBKA: mean bank angle (degrees).

River and its major tributaries were strongly associated with salinity and nutrients. The difference in patterns observed between Leland et al.

and those found in our study may derive from different sampling designs. Their sampling sites, mainly located along the main stem of the San

128 Table 4. Correlation coefficients between selected environmental variables and the 1st two CCA axes Variables

CCA ordination I

II

% streams as coarse substrates Manganese

)0.18 0.28

0.60 0.04

Mean bank angle (degree)

)0.14

0.34

Mean wetted stream width

)0.66

)0.11

Mean thalweg depth

)0.65

0.10

Non-agricultural riparian disturbance

0.18

)0.37

% variance of species data explained

5.3

4.9

Joaquin River, reflect large river habitat conditions while most of our sampling sites were 1st-order ditches and drains with only a few large river sites. In addition, their study included several sites in the Sierra Nevada foothills, which produces stronger gradients of salinity and nutrients in their data set. Because stream types and physical habitat conditions varied so much among our sites, it was expected that diatom-based classification would yield several discrete stream groups that might correspond to the major environmental characteristics. Two classification methods (UPGMA and TWINSPAN) commonly used in ecology and bioassessment yielded substantially different memberships for each group (Table 3). Neither diatom-based stream classification corresponded well with stream types, indicating that withinstream-type variability of environmental variables in relation to diatom assemblages may be high. The TWINSPAN groups corresponded better with the drainage basin than the UPGMA groups. Two basins are different in natural environmental settings. For example, the San Joaquin is more xeric than the Sacramento River with higher soil salinity and erodibility. The TWINSPAN method based on the ‘top-down’ approach seems to capture basin-related variability among stream sites better than the UPGMA method. Indicator species analysis identified sets of indicator taxa for each diatom-based stream groups. However, interpretation of these indicator taxa in relation to the environmental variables identified by the classification tree method is difficult. For example, a total of 11 indicator taxa

were identified for the UPGMA-Group B (Table 2). Most of these taxa have been classified as either non-motile (e.g., Staurosira) or stalked taxa (e.g., Achnanthidium, Cymbella, Rhoicosphenia). The classification tree model indicated that these sites may have larger channel sizes and higher in-stream bed stability (Fig. 3). However, non-motile taxa (Staurosira and Cocconeis) were also indicative for Group D that was characterized by relatively lower in-stream bed stability. The difficulty in relating indicator taxa to a particular set of environmental variables may partly be due to the complex interactions among environmental variables related to diatoms, and possible mismatches between diatoms and measured environmental variables across both spatial and temporal scales. For example, in this study samples were collected from 9 different locations in each 150– 450 m long study reach and combined as one sample. This mixture of the samples from the multiple sampling locations, each possibly with different substrates and microhabitats, may decrease our ability to infer ecological processes which may be responsible for location-specific diatom assemblages, and consequently lower diatom sample response sensitivity. The composite samples, aiming at assessing overall reach-scale conditions, may reflect only strong and coarsescale environmental conditions. Poor association between diatom assemblages and water chemistry may reflect overall impaired water quality in this region. Despite the high variability of nutrients among sites, overall nutrient concentrations may be high enough for diatom growth. Bothwell (1989) suggested that areal periphyton biomass, dominated by diatoms, may )1 peak around 28 lg PO3– based on three long4 P l term artificial stream enrichment studies. Recently, use of soluble reactive phosphorus concentrations to indicate nutrient status in stream field studies has been questioned (Dodds, 2003). Dodds recommended that total phosphorus (TP) may be a better indicator of nutrient limitation in surface water studies. Unlike lentic systems, the relationship between TP and lotic trophic conditions has not been adequately quantified. Dodds et al. (1998) suggested that streams may be classified as oligotrophic if TP<29 lg l)1. In this study, the median concentration of TP was 100 lg l)1 with a range from 29 lg l)1 to 3600 lg l)1. The rela-

129 tionship between nutrients and lotic diatom assemblages in this ecoregion may be further confounded by applications of herbicides in streams or near-riparian areas. Herbicides have been used to control excessive algal growth in some streams, especially constructed conveyance habitats.

Acknowledgements Field work and data collection were undertaken by California Division of Game and Fish and USEPA Region 9, with funding from USEPA’s Office of Research and Development as part of the REMAP program. Susanna DeCelles identified and enumerated all benthic diatom samples. The data analyses and preparation of this manuscript were supported by USEPA STAR stream classification grant to the first author (R-82949801). We thank Dr. R. J. Stevenson and an anonymous reviewer for their comments on earlier versions of this paper.

References Biggs, B. J. F., 1996. Patterns in benthic algae of streams. In Stevenson, R. J., M. L. Bothwell, & R. L. Lowe (eds), Algal Ecology. Academic Press, San Diego, California: 31–51. Biggs, B. J. F., R. J. Stevenson & R. L. Lowe, 1998. A habitat matrix conceptual model for stream periphyton. Archive fu¨r Hydrobiologie 143: 21–56. Bothwell, M. L., 1989. Phosphorus-limited growth dynamics of lotic periphytic diatom communities: areal biomass and cellular growth rate responses. Canadian Journal of Fisheries and Aquatic Sciences 46: 1293–1301. Brown, L. R. & J. T. May, 2000. Macroinvertebrate assemblages on woody debris and their relations with environmental variables in the Lower Sacramento and San Joaquin River drainages, California. Environmental Monitoring and Assessment 64: 311–329. Clark, L. A. & D. Pregibon, 1993. Tree-based models. In Chambers, J. M. & T. J. Hastie (eds), Statistical Models in S. Wadsworth & Brooks, Pacific Grove, CA: 377–420. Dodds, W. K., J. R. Jones & E. B. Welch, 1998. Suggested classification of stream trophic state: distribution of temperate stream types by chlorophyll, total nitrogen, and phosphorus. Water Resources Research 32: 1455–1462. Dodds, W. K., 2003. Misuse of inorganic N and soluble reactive P concentrations to indicate nutrient status of surface waters. Journal of North American Benthological Society 22: 171–181.

Dufrene, M. & P. Legendre, 1997. Species assemblages and indicator species: the need for a flexible asymmetrical approach. Ecological Monographs 67: 345–366. Griffith, M. B., P. Husby, R. K. Hall, P. R. Kaufmann & B. H. Hill, 2003. Analysis of macroinvertebrate assemblages in relation to environmental gradients among lotic habitats of California’s Central Valley. Environmental Monitoring and Assessment 82: 281–309. Grime, J. P., 1977. Evidence for the existence of three primary strategies in plants and its relevance to ecological theory. American Naturalists 111: 1169–1194. Hall, R. K., A. Olsen, D. Stevens, B. Rosenbaum, G. Wolinsky, P. Husby & D. Heggem, 2000. EMAP design and river reach file 3 (RF3) as a sample frame in the Central Valley, California. Environmental Monitoring and Assessment 64: 69– 80. Hill, M. O., R. G. H. Bunce & M. W. Shaw, 1975. Indicator species analysis, a divisive polythetic method of classification and its application to a survey of native pinewoods in Scotland. Journal of Ecology 63: 597–613. Kaufmann, P. R. & E. G. Robison, 1998. Physical habitat characterizaton. In Lazorchak, J. M., D. J. Klemm, & D. V. Peck (eds), Environmental Monitoring and Assessment Program – Surface Waters: Field Operations and Methods for Measuring the Ecological Condition of Wadeable Streams. EPA/620/R-94/004F. Office of Research and Develop., U.S. Envir. Protection Agency, Washington, D.C,: 77–118. Kaufmann, P. R., P. Levine, E. G. Robison, C. Seeliger & D. V. Peck, 1999. Quantifying Physical Habitat in Wadeable Streams. EPA 620/R-99/003. Environmental Monitoring and Assessment Program. U.S. Environmental Protection Agency, Corvallis, OR. Krammer, K. & H. Lange-Bertalot, 1986. Bacillariophyceae. 1.Teil: Naviculaceae. VEB Gustav Fisher Verlag, Jena. Krammer, K. & H. Lange-Bertalot, 1988. Bacillariophyceae. 2. Teil: Epithemiaceae, Bacillariaceae, Surirellaceae. VEB Gustav Fisher Verlag, Jena. Krammer, K. & H. Lange-Bertalot, 1991a. Bacillariophyceae. 3 Teil: Centrales, Fragilariaceae, Eunotiaceae, Achnanthaceae. VEB Gustav Fisher Verlag, Jena. Krammer, K. & H. Lange-Bertalot, 1991b. Bacillariophyceae. 4 Teil: Achnanthaceae, Kritische Erganzungen zu Navicula (Lineolatae) und Gomphonema. VEB Gustav Fisher Verlag, Jena. Kroeger, S., E. Fensin, K. Lynch & M. Vander Borgh, 1999. United States Water Quality Programs that Use Algae as a Biological Assessment Tool. North Carolina Division of Water Quality, Raleigh, NC. Kutka, F. J. & C. Richards, 1996. Relating diatom assemblage structure to stream habitat quality. Journal of North American Benthological Society 15: 469–480. Leland, H. V., L. R. Brown & D. K. Mueller, 2001. Distribution of algae in the San Joaquin River, California, in relation to nutrient supply, salinity and other environmental factors. Freshwater Biology 46: 1139–1167. MathSoft., 2000. S-Plus: Guide to Statistics. Data Analysis Products Division. MathSoft, Inc. Seattle, Washington 875 pp.

130 McCune, B. & M. J. Mefford, 1999. PC-ORD. Multivariate Analysis of Ecological Data, Version 4. MjM Software Design. Gleneden Beach, Oregon, USA. Miller, A. R., R. L. Lowe & J. T. Rotenberry, 1987. Microsuccession of diatoms on sand grains. Journal of Ecology 75: 693–709. Mount, J. F., 1995. California Rivers and Streams: The Conflict between Fluvial Process and Land Use. University of California Press, Berkeley, CA. Patrick, R. & C. W. Reimer, 1966. The Diatoms of the United States. Vol. 1. Monographs of the Academy of Natural Sciences of Philadelphia, No.13, Philadelphia. Pennsylvania, USA. Patrick, R. & C. W. Reimer, 1975. The Diatoms of the United States. Vol. 1. Part 1. Monographs of the Academy of Natural Sciences of Philadelphia, No.13, Philadelphia. Pennsylvania, USA. Saiki, M. K., 1984. Environmental conditions and fish faunas in low elevation rivers on the irrigated San Joaquin Valley Floor, California. California Department of Fish & Game 70: 145–157. Stevens, D. L. Jr. & A. R. Olsen, 1999. Spatially restricted surveys over time for aquatic resources. Journal of Agricultural, Biological, and Environmental Statistics 4: 415–428. Stevens, D. L. Jr., 1997. Variable density grid-based sampling designs for continuous spatial populations. Environmetrics 8: 167–95.

Stevenson, R. J., 1997. Scale-dependent determinants and consequences of benthic algal heterogeneity. Journal of North American Benthological Society 16: 248–262. Stevenson, R. J. & Y. Pan, 1999. Monitoring environmental changes using diatoms in stream and river communities. In Stoermer E. F. & J. P. Smol (eds), The Diatoms: Applications to Environmental and Earth Sciences. Cambridge University Press: 11–40. Stevenson, R. J. & L. L. Bahls, 1999. Periphyton protocols. Rapid bioassessment protocols for use in Wadeable Streams and Rivers: Periphyton. In Barbour, M. T., J. Gerritsen, & B. D. Snyder (eds), Benthic Macroinvertebrates, and Fish. 2nd ed. EPA 841-B-99–002 United States Environmental Protection Agency, Washington, DC: 6-1–622. ter Braak, C. J. F., 1986. Canonical correspondence analysis: a new eigenvector technique for multivariate direct gradient analysis. Ecology 67: 1667–1679. ter Braak, C. J. F. & P. Smilauer, 1998. CANOCO Reference Manual and User’s Guide to Canoco for Windows: Software for Canonical Community Ordination (version 4). Microcomputer Power, Ithaca, NY. USEPA (US Environmental Protection Agency)., 1987. Handbook of Methods for Acid Deposition Studies: Laboratory Analysis for Surface Water Chemistry. EPA 600/4-87/ 026. Office of Research and Development, US Environmental Protection Agency, Washington, DC.

Hydrobiologia (2006) 561:131–147  Springer 2006 R.J. Stevenson, Y. Pan, J.P. Kociolek & J.C. Kingston (eds), Advances in Algal Biology: A Commemoration of the Work of Rex Lowe DOI 10.1007/s10750-005-1610-6

Response of periphytic algae to gradients in nitrogen and phosphorus in streamside mesocosms Steven T. Rier1,2,* & R. Jan Stevenson1 1

Department of Zoology, Michigan State University, Michigan, 48824, USA Department of Biological & Allied Health Sciences, Bloomsburg University, Hartline Science Center, 400 East 2nd Street, Bloomsburg, PA 17815, USA (*Author for correspondence: E-mail: [email protected]) 2

Key words: periphyton, diatoms, nutrient limitation, Monod model, growth rates, peak biomass

Abstract In this study we manipulated both nitrogen and phosphorus concentrations in stream mesocosms to develop quantitative relationships between periphytic algal growth rates and peak biomass with inorganic N and P concentrations. Stream water from Harts Run, a 2nd order stream in a pristine catchment, was constantly added to 36 stream-side stream mesocosms in low volumes and then recirculated to reduce nutrient concentrations. Clay tiles were colonized with periphyton in the mesocosms. Nutrients were added to create P and N concentrations ranging from less than Harts Run concentrations to 128 lg SRP l)1 and 1024 lg NO3-N l)1. Algae and water were sampled every 3 days during colonization until periphyton communities reached peak biomass and then sloughed. Nutrient depletion was substantial in the mesocosms. Algae accumulated in all streams, even streams in which no nutrients were added. Nutrient limitation of algal growth and peak biomass accrual was observed in both low P and low N conditions. The Monod model best explained relationships between P and N concentrations and algal growth and peak biomass. Algal growth was 90% of maximum rates or higher in nutrient concentrations 16 lg SRP l)1 and 86 lg DIN l)1. These saturating concentrations for growth rates were 3–5 times lower than concentrations needed to produce maximum biomass. Modified Monod models using both DIN and SRP were developed to explain algal growth rates and peak biomass, which respectively explained 44 and 70% of the variance in algal response.

Introduction Nuisance algal blooms resulting from N and P contamination have become a major problem in many streams (USEPA, 1998; Dodds & Welch, 2000). Protection and remediation of these resources is therefore contingent on a better understanding of the quantitative relationships between algae and inorganic nutrients. However, most studies of nutrient effects on periphytic algae in streams have only assessed the presence or absence of nutrient limitation through enriching streams or stream habitats with nutrient

diffusing substrates (e.g., Pringle & Bowers, 1984; Hepinstall & Fuller, 1994; Pan & Lowe, 1994; Francoeur et al., 1999), once-through artificial stream channels (e.g., Mundie et al., 1991; Hill et al., 1992), or whole-stream enrichments (e.g., Petersen et al., 1993; Slavik et al., 2004). Nutrient limitation has also been investigated by applying regression approaches to field survey data (Lohman et al., 1992; Dodds et al., 1997, 2002; Biggs, 2000), and these empirical methods confirm the experimental evidence that N and P are both important determinants of stream community structure and function.

132 Two studies have examined the response of periphytic algae across experimentally manipulated P gradients (Horner et al., 1983; Bothwell, 1989) to determine the effects of specific nutrient concentrations on periphytic algal growth. Bothwell (1989), examining both cellular growth rates and maximum areal biomass, demonstrated that the amount of P needed to saturate cellular growth rates during early stages of colonization was two orders of magnitude lower than concentrations needed to produce maximum areal biomass. The difference in nutrient concentrations is probably due the increase in mat thickness through time, which can decrease nutrient diffusion and nutrient availability within the mat (Stevenson, 1990; Stevenson et al., 1991; Stevenson & Glover, 1993). No one has developed quantitative relations between growth or peak biomass of periphytic algae and specific inorganic N concentrations. One potential difficulty in relating N and P concentrations to algal growth rates in thin biofilms is that all but the most oligotrophic streams have nutrient concentrations that exceed levels needed to saturate cellular growth rates. For example, Biggs (1990) found that algal growth rates were not limited by P concentrations in his low nutrient control reach, which resulted in no difference in growth rates above and below a P discharge. Bothwell (1988) using artificial stream mesocosms demonstrated that growth rate saturation occurred between 0.3 and 0.6 lg P l)1. Therefore, to study the full range of algal growth responses to nutrients in most streams, one must remove nutrients from the stream water. Artificial stream mesocosms with partial recirculation can substantially reduce NO 3 and SRP concentrations in water (Mulholland et al., 1991), thereby providing a way to examine algal response to nutrient concentrations lower than ambient levels. The purpose of this experiment was to examine the response of benthic algal growth rates and maximum areal biomass across experimentally manipulated N and P gradients. We utilized partially recirculating streams with stream water from a forested catchment to reduce nutrient levels yet provide natural environments for periphyton colonization (Lamberti & Steinman, 1993). Our first objective was to test the hypothesis that algal responses to N would be the same as P, i.e. algal growth rates and peak biomass would reach a

maximum at some saturating nutrient concentration and that saturating nutrient concentrations would be higher for peak biomass than for growth rates. In addition, we hypothesized that saturating nutrient concentrations would be higher for N than P because of higher cell quotas for N than P (e.g., Redfield, 1958; Droop, 1974). The second objective was to determine whether ratios, concentrations, or loading rates of total or soluble N and P best predicted benthic algal growth and peak biomass.

Materials and methods This experiment was carried out between May 17 and June 3, 1999 at the University of Louisville stream research facility located in the Bernheim Research Forest 50 km south of Louisville Kentucky. Manipulations were carried out in 36 partially re-circulating stream mesocosms that consisted of closed 3.51.5 m loops of 10 cm diameter PVC with the top piece cut lengthwise (Fig. 1). These loops were filled with stream water from Harts Run, a pristine 2nd order stream contained entirely within the Bernheim research forest. Current was generated in each loop by forcing air into the bottom of one end of the loop through a hose with a Sweet Water model S45 Blower. Air bubble lift produced current velocities of 25 cm s)1. We minimized turbulence by placing flow stabilizers, made from 15, 1.5-cm diameter plastic tubes, parallel to the flow at the head of each channel. Temperature was maintained in each mesocosm by half immersing each loop in water baths made from concrete blocks lined with plastic film. Six channels were placed in each water bath (total of six baths). Stream water was continuously exchanged in each bath at a rate high enough to maintain water in the recirculating channels at stream temperature. Stream water was exchanged in each channel at a rate of 7 ml s)1 by adding water from 200 l head tanks with plastic aquarium tubing. Water was allowed to overflow through a hole drilled in the end of each channel. This allowed for complete turnover of water every 4 h. We used partial re-circulation to allow microorganisms inhabiting the low nutrient channels to draw down N and P to levels lower than that found in the ambient

133

Figure 1. Diagram of the partially re-circulating stream system used in the experiment.

stream water. This method was shown to reduce nutrient concentrations in source water (Mulholland et al., 1991). Stock solutions with different concentrations of NaNO3 and KH2PO4 were delivered from 1 l containers to each channel using peristaltic pumps (Manostat model STD) set at a drip rate of 0.3 ml min)1. We added N and P throughout the experiment at a rate that would elevate ambient nutrient concentrations by the amounts shown in Table 1, not taking into account uptake within the channels. Nutrient treatments were assigned randomly throughout the 36 channels. We used a total of three replicate channels for each treatment and used three channels as controls, which received no nutrient additions. Twenty unglazed ceramic tiles (29.2 cm2) were placed at the downstream end of each channel where turbulence was lowest. Two layers of greenhouse shade cloth, supported by 1 cm2 wire mesh screen were placed over the tiles to simulate light regimes in nearby streams. The shades

allowed 12% of incident photosynthetically active radiation (LiCorTM Model Li 189 light meter) to reach the tiles. A large species pool of potential algal colonists was collected from streams varying in nutrient conditions by adding 500 ml of an algal inoculum to each stream mesocosm one day prior to the start of experiment. This inoculum was produced by scraping rocks from three streams differing in the level of human impact in the watershed: Harts Run, the low nutrient stream where the experiment was run; Wilson Creek with moderate agriculture in the watershed and moderate concentrations of N and P (R. J. Stevenson unpublished data); and the Middle Fork of Beargrass Creek, located within the city of Louisville, Kentucky. Beargrass Creek has an urban watershed and periodically has high nutrient concentrations (R. J. Stevenson, unpublished data). Scrapings from the three streams were combined in a single container and homogenized before aliquots were added to each experimental stream. In using this approach, we

Table 1. Nutrient addition treatments utilized in the study Nitrogen concentration(s) (lg l)1)

Phosphorus concentration(s) (lg l)1)

Control channels

0

0

Nitrogen treatments

0, 16, 64, 256, 512, 1024

50

Phosphorus treatments

500

0, 2, 8, 32, 128

All values are projected concentrations above ambient stream water for each treatment not taking into account uptake within each channel.

134 assumed that the ‘‘assemblage’’ that resulted from combining algae from these three different streams actually represented the pool of potential colonists to all systems in this region and that differences in community structure between these streams are primarily a function of local conditions, such as N and P concentrations, and not dispersal. We randomly collected six tiles from different channels on May 18 (Day 0), which were immediately frozen for later analysis of initial chlorophyll a. Two tiles were then collected from each channel on days 3, 6, 10, 13, and 16. Periphyton on tiles was frozen for later analyses of chlorophyll a and mat chemistry. Material was scraped from thawed samples into beakers with a toothbrush and razor blade. Samples were diluted to known volumes with deionized water, homogenized with a Biospec model m133\1281-0 biohomogenizer, and then sub-sampled with a mechanical pipette while continuously mixing on a magnetic stir plate. Chlorophyll a was determined sprectrophotometerically on sub-samples with a Spectronic Genesys 2 spectrophotometer after a 24 hr extraction in 90% buffered acetone (APHA, 1998). Sub-samples of the periphyton mats were also assayed for total P (mat P) and total N (mat N). Homogenized suspensions of mat material were analyzed using the methods outlined for total N and total P (APHA, 1998). Results were expressed as mg of N or P per g AFDM. Water samples were collected from the natural stream and from each channel on days 1, 3, 6, 8, 12, and 14 with acid washed 125 ml polypropylene bottles. Water samples were immediately packed in ice and returned to the lab for analysis. Nitrate+nitrite N, ammonium N (NHþ 4 ), and silica (Si) concentrations were estimated on a Skalar auto-analyzer according to standard methods (APHA, 1998). Since nitrate+nitrite N consists mainly of nitrate in this stream (R. Schultz, unpublished data), we considered this a measure of nitrate N (NO 3 ). Dissolved inorganic nitrogen concentrations were obtained by adding þ nitrate N (NO 3 ) and ammonium N (NH4 ) concentrations. Soluble reactive P (SRP), total P (TP), and total N (TN) was also determined according to standard methods using a Hitachi U-200 1 spectrophotometer (APHA, 1989).

Data analysis We estimated algal growth rates (l) from each channel between day 3 and day 6 using the following equation: l ¼ ðloge B2  loge B1 Þ=ðt2  t1 Þ where B is algal biomass (chlorophyll a) and t is time. Peak biomass (PB) was also calculated for each channel using chlorophyll a. PB is defined as the maximum areal biomass attained by a periphyton community during development (Bothwell, 1989). We estimated PB by averaging the two highest biomass estimates for chlorophyll a from each channel. For the growth rate analysis, we used an average of nutrient concentrations from days 1, 3, and 6 and for peak biomass we used an average of DIN and SRP over days 6, 8, and 10. Three date averages enabled more precise characterization of nutrient conditions within each channel by reducing day-to-day variation. Growth rates and PB were related to N and P concentrations with linear, log-linear, and Monod (1950) models using the linear and nonlinear regression procedure in Systat 10 (SPSS Science, Chicago, Illinois). If linear or log-linear models explained more variation than the Monod models, then results would indicate that nutrients were not high enough to saturate growth rates and peak biomass. SRP, TP, mat P (PB only) and PO4-P loading concentration (PLOAD) and DIN, NO3-N, NH4-N, TN, mat N (PB only) and NO3-N loading concentration (NLOAD) were used in the models to predict l and PB. The linear model took the form l or PB ¼ a þ bS where a was a constant, b was the rate of increase in l or PB per unit S, which was either the measured or loaded nutrient concentration in the channels. The log-linear model took the form l or PB ¼ a þ b loge ðSÞ where log(S) was the natural logarithm of S. The Monod (1950) equation took the forms l ¼ lmax ðS=ðS þ Kl ÞÞ PB ¼ PBmax ðS=ðS þ KPB ÞÞ

135 where Kl and KPB were the half saturation constants for the particular nutrient manipulation and lmax or PBmax were the maximum growth rate or maximum PB respectively. With respect to the Monod (1950) model, we used a ‘‘working’’ definition of saturating nutrient concentrations as those when l and PB attained 90% of their respective maxima. Relations between l and PB and S were only determined when the other nutrient was assumed to be above saturating conditions; i.e. the relation between l and PB and N was determined with results from treatments with PLOAD>49 lg PO4-P l)1 and algal response to P was determined with NLOAD>499 lg NO3-N l)1. Coefficients of determination (r2) (Sokal & Rohlf, 1995) for the models were used to compare the goodness of fit of the different models as proportion of error in l or PB explained by the model. Predictive models of l and PB that simultaneously considered both N and P were assessed using DIN and SRP concentrations, TP and TN concentrations, and PLOAD and NLOAD. Linear, log-linear, and modified Monod models were used. Multiple regression and nonlinear regression were used to calculate the model parameters and determine goodness of fit with coefficients of determination (Systat 10). The linear equation took the form l or PB ¼ a þ bN SN þ bP SP where the subscripts N and P denote nutrient concentrations and the rate of increase in l or PB per unit of each specific nutrient. The log-linear equation took the form l or PB ¼ a þ loge bN SN þ loge bP SP The modified Monod equation took the forms: l ¼ lmax ðSN =ðSN þ KlN ÞÞ  ðSP =ðSP þKlP ÞÞ PB ¼ PBmax ðSN =ðSN þ KPBN ÞÞ  ðSP =ðSP þ KPBP ÞÞ

where the subscripts N and P denote nutrient concentrations and half saturation constants specific to those nutrients. Molar ratios of DIN:SRP, TN:TP, and NLOAD:PLOAD were also used to develop predictive models of l and PB in different nutrient conditions using simple linear regression.

Results Temperature and water chemistry Water temperatures in the re-circulating channels averaged approximately 20 C. Stream water )1  and NO 3 ranged from 17.5 to 38.0 lg NO3 -N l )1 SRP ranged from 1.5 to 7.3 lg l . Both nutrients generally decreased throughout the experiment due to a lack of rain with the exception of day 14, which immediately followed a moderate rain (Table 2). During the period we defined as the early growth stage (days 1, 3, and 6), stream water )1 NO (±1 SE) and 3 -N averaged 30.1±5.6 lg l )1 SRP averaged 4.8±1.4 lg l (±1 SE). In the period we defined as the peak biomass stage (days 6, 8, and 10), stream NO 3 -N concentrations averaged 19.1±0.5 lg l)1 (±1 SE) and SRP averaged 2.6±0.3 lg l)1 (±1 SE). Ammonium and silica concentrations fluctuated throughout the experiment ranging from 6.8 to 51.0 lg )1 NHþ and 10.53–11.99 mg Si l)1. 4 -N l Nutrient depletion was evident in recirculating channels for all nutrients and was related to stage of colonization and nutrient concentration (Figs. 2 þ and 3). In the control channels NO 3 and NH4 uptake were apparent from day 3 to the end of the experiment (Table 2). Starting on day 3, NO 3 -N and NHþ -N concentrations were significantly 4 lower (p<0.05) in control channels than in ambient stream water (Harts Run). Nitrate concentrations in recirculating streams decreased during this time from ambient levels to a range varying )1 between 3.7 and 8.5 lg NO 3 -N l . SRP uptake was not apparent in the control channels, which may be due to ambient levels that were at or near detection limits. No difference in silica concentrations was observed between control channels and ambient stream water (Table 2). Algal response Diatoms dominated algal communities in all treatments throughout the experiment. Although relative abundances differed between treatments and despite a mixture to potential colonists from three different streams with varying levels of nutrient enrichment, only 14 diatom species dominated algal species composition in both low and high N and P treatments (Table 3) (K. Manoylova

136 Table 2. Stream water and control channel (mean ± 1 SE) nutrient concentrations measured throughout the experiment Day of experiment

)1 NO 3 -N (lg l )

)1 NHþ 4 -N (lg l )

SRP (lg l)1)

Si (mg l)1)

Stream

Control

Stream

Control

Stream

Control

Stream

Control

water

channels

water

channels

water

channels

water

channels

1

38.0

34.7±1.5

6.8

6.9±1.3

4.4

4.3±0.2

11.99

12.1±0

3 6

33.0 19.4

16.4±1.2 3.7±1.3

37.9 51.0

26.9±1.7 10.4±0.9

7.3 2.6

4.3±0.5 2.5±0.2

10.86 10.72

10.8±0.07 11.9±0.02

8

19.9

8.5±0.2

24.8

21.3±0.2

3.1

3.3±0.4

10.97

10.4±0.29

10

18.1

7.2±1.1

19.9

12.9±2.0

2.0

1.4±0.1

10.71

10.6±0.04

12

17.5

5.3±0.3

42.4

21.8±0.5

1.5

2.5±0.1

10.53

10.0±0.04

14

28.9

7.5±0.3

24.8

15.4±0.4

3.7

4.8±0.8

10.88

10.5±0.12

& R. J Stevenson, unpublished manuscript). Biomass (chlorophyll a) changed little between days 0 and 3 (Fig. 4). The largest increase in algal biomass was between days 3 and 6. This increase was used to calculate growth rates. Peak biomass was reached between days 6 and 10 in all treatments. Average biomass for these two dates was used to estimate peak biomass in each channel. After day 10 there was a sharp decline in biomass in all treatments followed by a leveling off between days 13 and day 16. Starting on day 6, we found that algal biomass increased generally with increasing N and P concentration. Control treatments consistently had the lowest biomass of all treatments on all dates beginning on day 6 (Fig. 4).

Table 3. Common diatom species (abundance > 5%) found in both high and low N and P treatments (K. Manoylov and R. J Stevenson, unpublished data) Achnanthidium deflexa (Hustedt) Kobayasi Achnanthidium minutissimum (Ku¨tz.) Czarnecki Cymbella excisa Ku¨tz. Cymbella cistula (Ehrenberg) Kirchner Encyonema minutum (Hilse in Rabh.) D.G. Mann Fragilaria crotonensis Kitton Synedra nana Meist. Fragilaria vaucheriae (Ku¨tzing) Petersen Meridion circulare (Grev.) Ag. Nizschia acicularis (Ku¨tz) W. Sm. Nizschia dissipata (Ku¨tz) Grun. Nizschia linearis (Ag.ex W. Sm) W. Sm. Nizschia palea (Ku¨tz) W. Sm. Ulnaria ulna (Nitzsch) Compe`re

Algal growth rates were high in all treatments and ranged from 0.8 d)1 to 2.1 d)1. Growth rates displayed saturating kinetics with respect to DIN concentration (Fig. 5), i.e. a rapid increase in growth rates with nutrient concentration at low concentrations (between 10 and 100 lg DIN l)1) and then no change in growth rate at high nutrient concentrations. In fitting the N data to the Monod equation, Kl was 9.59 ± 2.60 lg DIN l)1 (±1 SE) and lmax was 1.77 ± 0.05 d)1 (±1 SE). The model explained 59% of the variance in growth rates along the N gradient. Growth rates saturated (90% lmax) at 86 lg DIN l)1 according to the Monod equation. Although considerable variability in algal growth rates occurred along the P gradient, estimates did show saturating kinetics (Fig. 5). Kl was 1.56±0.37 lg SRP l)1 (±1 SE) and a lmax was 1.83±0.06 d)1 (±1 SE) for the Monod (1950) equation. This model could explain approximately 58% of the variance. Growth rates saturated (90% lmax) at 16 lg SRP l)1. Peak biomass ranged from 0.35 lg cm)2 to 9.86 lg chlorophyll a cm)2 along the DIN gradient. When peak biomass was related to the DIN gradient, with the Monod (1950) equation, Kl was 34.23±9.57 lg DIN l)1 (±1 SE) and PBmax was 7.72±0.49 lg chlorophyll a cm)2 (±1 SE). Saturation (90% lmax) occurred at 308 lg DIN l)1. This model explained 77% of the variance in PB. Although the Monod model explained 77% of the variance in PB as a function of DIN, some evidence indicated that PB did not saturate across the N gradient (Fig. 6). Peak biomass increased rapidly with increasing DIN between 12

137

NO3--N (μg / L)

1600

(a)

1400 1200 1000 800

Expected

600 400 200 0 140

SRP (μg / L)

120

(b)

ted pec

Ex

100 80 60 40 20 0 100

NH4+-N (μg / L)

(c) 80 60 40 20 0

(d)

Si (mg / L)

11

10

9

8

0

20

40

60

80

100

120

140

Phosphorus Treatment Figure 2. Nutrient concentrations measured in each treatment across the phosphorus gradient in both the early growth stage (darkened circle) and peak biomass stages (open square) of community development. Panel a = NO 3 -N, panel b = SRP, panel c = NHþ 4 -N, panel d = Si.

138 1200

NO3--N (μg / L)

1000

(a)

ted pec

Ex

800 600 400 200 0 120

(b)

SRP (μg / L)

100 80

Expected

60 40 20 0

NH4+-N (μg / L)

30

(c)

25 20 15 10 5 0 12

(d)

Si (mg / L)

11 10 9 8 0

200

400

600

800

1000

1200

Nitrogen Treatment Figure 3. Nutrient concentrations measured in each treatment across the nitrogen gradient in both the early growth stage (darkened þ circle) and peak biomass stages (open square) of community development. Panel a = NO 3 -N, panel b = SRP, panel c = NH4 -N, panel d = Si.

139 and 100 lg N l)1, but then continued to increase linearly along the remainder of the DIN gradient. However, a log-linear model of PB as a function of DIN only explained 2% more variation than the Monod model (Table 4), even though the loglinear model does not saturate at high N and should fit the suspected pattern. PB across the P gradient increased sharply between 2 and 20 lg SRP l)1 and then did not change with higher SRP concentration. Kl was

4.20 ± 1.78 lg P l)1 (±1 SE) and PBmax was 8.65 ± 0.90 lg chlorophyll a cm)2 (±1 SE) when fit to the Monod equation (Fig. 6). Saturation of peak biomass occurred at 38 lg SRP l)1. This model explained 53% of the variance. The Monod model usually explained as much or more variation in l and PB as a function of nutrient concentrations as the log-linear model, and it always explained nutrient effects better than the linear model (Tables 4 and 5). For models

Figure 4. Algal biomass (Chlorophyll a) throughout the experiment in different nitrogen and phosphorus treatments.

140

2.00

μ (d-1)

1.75

1.50

1.25

Nitrogen 1.00 0

200

400

600

800

1000

DIN (μg / L) 2.00

μ (d-1)

1.75

1.50

1.25

Phosphorus 1.00 0

20

40 SRP (μg / L)

60

80

Figure 5. Algal growth rates (l) based on chlorophyll a as a function of dissolved inorganic nitrogen (DIN) and soluble reactive phosphorus (SRP). Curve represents the best fit of the data to the Monod (1950) equation.

relating algal attributes to one nutrient when the other was assumed to be saturating, coefficients of determination (r2) varied from values less than 0.035 for l as a function of NH4 concentration to 0.810 for the Monod model explaining PB as a function of TN. Growth rates were explained by P measures (SRP, TP, PLOAD) about the same with the log-linear model as the Monod model

(Table 4). In contrast, coefficients of determination were about 0.10 greater for Monod model than for the log-linear model when l was related to N measures (DIN, NO3, TN, NLOAD). For relations between PB and nutrients, P models were slightly better with the Monod equation than log-linear equations, but the best fit for N models varied slightly between Monod and log-linear equations.

141

Chlorophyll a (μg / cm2)

10 8 6 4 2

Nitrogen

0 0

200

400

600

800

1000

DIN (μg / L) Chlorophyll a (μg / cm2)

10 8 6 4 2 0

Phosphorus 0

20

40

60

80

SRP (μg / L) Figure 6. Peak algal biomass (chlorophyll a) as a function of dissolved inorganic nitrogen (DIN) and soluble reactive phosphorus (SRP). Curve represents the best fit of the data to the Monod (1950) equation.

Soluble nutrient concentrations (SRP and DIN) explained as much or more variation in l with the Monod model as TP or TN, but the relationship between soluble nutrients and PB was somewhat less than for TP and TN (Table 4). The log-linear model explained more variation than the Monod model for l and PB as a function of loading rates (PLOAD and NLOAD). Using loglinear models, PLOAD and NLOAD explained more variation in l than measured nutrient concentrations, but only PLOAD explained more variation in PB. NLOAD explained less variation

in PB than measured N concentrations (DIN, NO3, and TN). Mat N and P concentrations explained little variation in peak biomass. Mat N varied from 153 to 569 mg N g)1 AFDM and mat P varied from 12 to 107 mg P g)1 AFDM. Peak biomass was better related to mat P than mat N with (r2=0.365 and 0.106 for log-linear models using mat P and mat N, respectively, Table 4). Monod models with nutrient concentrations were better able to explain l and PB than models between molar ratios of nutrients (DIN:SRP,

142 Table 4. Coefficients of determination (r2) for linear, log-linear, and Monod models of algal growth rate (l), and peak biomass (PB) as a function of concentrations of soluble reactive phosphorus (SRP), total phosphorus (TP), PO4-P or NO3-N loading concentrations in experimental channels (PLOAD & PLOAD), dissolved inorganic nitrogen concentration (DIN), NO3-N, NH4-N, total nitrogen (TN), mat phosphorus (Mat P) and mat nitrogen (Mat N) Algal attribute

Nutrient

r2 Linear

Log-linear

Monod

l

SRP

0.465

0.590

0.585

l

TP

0.485

0.583

0.572

l

PLOAD

0.527

0.630

0.420

l l

DIN NO 3

0.206 0.207

0.432 0.484

0.504 0.571

l

NHþ 4

l

TN

0.192

0.335

l

NLOAD

0.285

0.607

0.401

PB

SRP

0.344

0.511

0.528

PB

TP

0.403

0.580

0.602

PB

PLOAD

0.303

0.634

0.646

PB PB

DIN NO 3

0.579 0.565

0.805 0.784

0.774 0.687

PB

NHþ 4

0.628

0.689

0.671

PB

TN

0.595

0.798

0.810

PB

NLOAD

0.681

0.685

0.573

PB

Mat P

0.321

0.365

PB

Mat N

N.S.

N.S.

N.S.

N.S.

N.S. 0.427

0.361 N.S.

N.S. denotes those relationships where one or more of the parameters estimated were not significant in the regression (p<0.05).

TN:TP, and NLOAD:PLOAD, Table 5). r2 for all relations between l and PB and molar ratios were less than 0.10 and not significant (p>0.05), whereas the modified Monod model explained 43% of the variation in l and 70% of the variation in PB when all treatments in the study were combined in a single analysis. lmax was 1.792±0.065 d)1 (±1 SE), Kl-N was 7.067±2.38 lg N l)1 (±1 SE), and Kl-P was 1.027±0.355 lg P l)1 (±1 SE) in the modified Monod model for l. PBmax was 9.179±0.865 lg chlorophyll a cm)2 (±1 SE), Kmax-N was 36.950±12.356 lg N l)1 (±1 SE), and Kmax-P was 4.210±1.438 lg P l)1 (±1 SE) in the modified Monod model for PB.

Discussion In this study, we evaluated relationships between growth rates and peak biomass of diatom-dominated periphyton with N and P concentrations. Drawing down nutrients by partial re-circulation

of stream water in experimental streams made it possible to create environments with sufficiently low nutrient concentrations to observe nutrient limitation of stream periphyton for both N and P. Our results demonstrated that the dynamics of N limitation are similar to those observed in other studies established for P limitation (Horner et al., 1983; Bothwell, 1985, 1988, 1989). Growth rates were N limited below 86 lg DIN l)1 and P limited below 16 lg SRP l)1. Above these nutrient concentrations, little change in growth rates was observed. Peak biomass also responded to increases in nutrient concentrations with rapid increases at low nutrient concentrations and little PB increase above a saturating concentration (308 lg DIN l)1 and 38 lg SRP l)1). The molar ratios of saturating nutrients concentrations for growth rates (12:1) and peak biomass (18:1) were similar to the Redfield (1958) ratio (16:1). P saturation of growth rates occurred at concentrations that were considerably higher than those observed by Bothwell (1988). Several differ-

143 Table 5. Coefficients of determination (r2) for linear, log-linear, Monod, or modified Monod models (if two nutrient variables are involved) of algal growth rate (l), and peak biomass (PB) as a function of concentrations of soluble reactive phosphorus (SRP), total  phosphorus (TP), PO3 4 -P or NO3 -N loading concentrations in experimental channels (NLOAD & PLOAD), dissolved inorganic þ nitrogen concentration (DIN), NO 3 -N, NH4 -N, total nitrogen (TN) Algal attribute

Nutrient

r2 Linear

Log-linear

Monod 0.436

l

DIN & SRP

0.207

0.362

l

TN & TP

0.213

0.306

0.377

l

NLOAD & PLOAD

0.397

0.464

0.488

l l

TN:TP DIN:SRP

N.S. N.S.

l

NLOAD:PLOAD

N.S.

PB

DIN & SRP

0.404

0.663

0.701

PB

TN &TP

0.477

0.659

0.744

PB

NLOAD & PLOAD

0.603

0.637

0.783

PB

TN:TP

PB

DIN: SRP

N.S.

PB

NLOAD: PLOAD

N.S.

N.S.

N.S. denotes those relationships where one or more of the parameters estimated were not significant in the regression (p<0.05).

ences in the chemical and physical conditions of our channels and in Bothwell’s experiment could account for our higher saturating P levels. Current velocities in our systems were 50% slower than in Bothwell’s experiment (1988). Since higher current velocities can enhance nutrient uptake (e.g., Whitford & Schumacher, 1964; Horner & Welch, 1981; Horner et al., 1983; Horner et al., 1990), high current velocities in Bothwell’s experiments may have allowed for a greater flux of nutrients between the algal cells and the overlying water, thereby reducing the nutrient concentrations required to saturate algal growth. Bothwell reported DIN concentrations that ranged from 60 to 80 lg N l)1 during the winter and spring, but only 20 lg N l)1 during the summer. NHþ 4 -N was undetectable throughout his study. Since we show saturation for stream algae above 86 lg N l)1, N limitation could have constrained growth responses at high P concentrations in Bothwell’s study. Although, we attempted to estimate cellular growth in the absence of density effects by using the change in biomass between days 3 and 6, a 100-fold increase in biomass between these two dates may have resulted in density effects in the later stages of this period. Stevenson (1990) has shown that algal density can reduce growth rates. In addition, higher nutrient concentrations are

required to saturate accrual at high biomass than low biomass (Bothwell, 1988; this study). In general, algal growth in this experiment was extraordinarily high with respect to other values reported in the literature for stream periphyton (Bothwell, 1985, 1988, 1989; Biggs, 1990; Stevenson et al., 1991; Humphrey & Stevenson, 1992) and similar to only a few phytoplankton observations (e.g., Hecky & Fee, 1981). A number of factors could have contributed to these high growth rates. Average temperature during this experiment was 20 C, which has been shown to be within an optimal range for algal growth (Rhee & Gotham, 1981). Current velocity over the tiles was 25 cm s)1, which may have been high enough to overcome some diffusion limitation (Whitford, 1960) without creating excess drag (see review by Stevenson, 1996). Saturation with respect to the other nutrient in question in all treatments except the controls, likely contributed to the high growth rates by preventing a shift to limitation by the other nutrient. Initially seeding the experimental channels with a diverse species pool representing a wide range of nutrient conditions may have increased chances that species were present in each treatment that were best adapted to grow and compete under the given nutrient conditions. Using chlorophyll a in the algal growth rate

144 calculations may have introduced an upward bias in our estimates because concentrations of chlorophyll a per cell can change rapidly in response to changes in nutrient availability (Rosen & Lowe, 1984) and shade adaptation induced by increases in mat thickness (Hill & Boston, 1991; Dodds et al., 1999). Peak biomass of diatom-dominated periphyton saturated at higher nutrient concentrations than growth rates and at P concentrations similar to those in other studies. Bothwell (1989) found PB saturating between 25 and 50 lg P l)1 and Horner et al. (1983) found saturation at 25 lg P l)1. Dodds et al. (2002), using data from field surveys, showed saturation breakpoints in maximum chlorophyll a (analogous to peek biomass) at 27–100 lg l)1 TP and 308 lg DIN l)1. This is the only report of a saturating nutrient concentration that we found for DIN. Although peak biomasses along both the N and P gradients did produce a statistically significant (p<0.05) fit to the Monod model, our data did show some evidence that saturation of peak biomass accrual by N may not be as complete as for P. The log-linear model for PB as a function of DIN had a slightly higher r2 than the Monod model. Plots of data did indicate a continuing increase in PB at higher DIN concentrations. Since algae bind excess P as polyphosphate (Stevenson & Stoermer, 1982) and cells do leak nutrients (Olsen, 1989; Borchardt et al., 1994), periphyton might leak N from cells at higher rates than that of P, which could maintain the demand for higher and higher N supply to saturate peak biomass. Although increases in ammonia concentrations in our recirculating stream channels with nitrate addition and colonization time may be partially due to heterotrophic decomposition of organic N, it is also possible that nitrate reduction in cells and subsequent leakage of ammonia may also be important. The conversation of nitrate to ammonia subsequent leakage into the water column likely explains why DIN, and TN were the best predictors of peak biomass using the Monod model. Ultimately, high N turnover may lead to a higher N:P ratio required to saturate peak biomass. One limitation of the single nutrient and modified Monod model compared to the linear and log-linear models is that it forces the relationship between nutrient concentration and algal growth

or peak biomass through the origin. Therefore, the Monod model assumes that growth rates or biomass will equal 0 at an N or P concentration of 0. It does not account for instances where the algal community declines (i.e., negative growth rates) as N or P approaches zero, which would produce a negative growth rate. Our modified Monod model predicted growth rates and peak biomass relatively accurately when compared to linear and log-linear models and when compared to single nutrient models. The modified Monod model, that predicts algal performance based on N and P concentrations, may not be appropriate for single species growth rates because algal populations should be limited by one nutrient only (e.g., Droop, 1974), although some experimental evidence indicates that populations may be limited by more than one nutrient at a time (Stevenson & Pan, 1995). Assemblages composed of many populations may be simultaneously limited by more than one nutrient because different species may be limited by different nutrients. This model should serve as a valuable starting point for comparing algal growth responses to variability in N and P concentrations in other habitats. Complications often hinder our attempts to quantify relationships between nutrients and algae in streams as well as other habitats. The depletion  þ of PO3 4 , NO3 , and NH4 , as well as Si, by algal uptake often generates negative relationships (e.g., Rier & Stevenson, 2001) or no relationship between these nutrients and algal biomass in surveys of algae and nutrients in streams. By default, we may be left with the circular relationship between algal biomass (either planktonic or benthic) and water column TP concentrations, where most of that TP is particulate P in the form of suspended algae originating from the benthos. Molar ratios of water column N:P were not correlated with algal growth rates in our study or Stelzer & Lamberti (2001), and should not be expected to be correlated unless both nutrients were depleted below saturating concentrations. Biggs (2000) showed relations in field surveys between periphyton biomass and soluble inorganic nutrient concentrations by characterizing nutrient availability as the annual average concentration. This characterization includes an assessment of soluble inorganic nutrient concentrations that occur during periods of periphyton recolonization,

145 which probably better reflect the nutrient availability that enables development of peak biomasses in streams than nutrient concentrations measured during periods of peak biomass. Even in our recirculating streams, the commonly higher correlations between loading concentrations (PLOAD and NLOAD) and growth rates and peak biomass, vs. measured nutrient concentrations with algal responses, indicated that nutrient availability was difficult to characterize with soluble inorganic nutrient concentrations. The lack of relation between N and P concentrations in periphyton mats may be due to variability in measures of mat nutrients or complex relations between cell nutrient concentrations and growth rates. Mat nutrient concentrations have been suggested as potential indicators of nutrient status in streams and have been related to water column nutrient concentrations (Humphrey & Stevenson, 1992; Biggs, 1995). N:P ratios in mats may indicate relative limitation by N or P availability (Shanz & Juon, 1983; Borchardt, 1996). Complex relations between mat nutrient concentrations and algal growth or peak biomass are probably related to accrual of excess nutrients (i.e., high N:AFDM or P:AFDM ratios), when cells are limited by other nutrients or other resources (Stevenson & Stoermer, 1982). At peak biomass, either N or P concentrations may be limiting further accrual of periphyton. Therefore, mat N or P may provide some information about trophic status of streams, but concentrations of watercolumn nutrients provided better predictors of peak biomass under the controlled settings of this study. Many challenges remain in solving the relations between algae and nutrients. Differing responses of species to different temperature, light, and current environments surely affect our prediction of periphyton growth as a function of nutrients. Not only do different diatoms respond to nutrients differently, but also filamentous algae add complexity by responding to nutrients independently and producing additional substratum for epiphytic algal colonization. Even in our study, where nutrients in water from a stream draining a relatively pristine forested catchment were decreased in recirculating streams, we were not able to stop diatom growth with low nutrients. Can benthic diatoms accumulate under any natural nutrient

conditions accrual?

where

grazers

do

not

constrain

Acknowledgements Major Waltman, Brian Wade and Kalina Manoylova helped to build the experimental stream facility at Bernheim Research Forest and to sample during the experiment. Brian Wade and Rich Schultz helped to analyze algal and water samples. Walter Dodds provided helpful suggestions for improving the manuscript. We would like to thank The University of Michigan Biological Station where pilot studies were conducted. This research was supported by the Water and Watershed Program with funds from the United States Environmental Protection Agency’s STAR Program, Agreement Number R824783. References American Public Health Association, 1998. Standard methods for the examination of water and wastewater, 19th edn. American Public Health Association, Washington, DC. Biggs, B. J. F., 1990. Use of relative specific growth rates of periphytic diatoms to assess enrichment of a stream. New Zealand Journal of Marine and Freshwater Research 24: 9–18. Biggs, B. J. F., 1995. The contribution of disturbance, catchment geology, and land use to the habitat template of periphyton in stream ecosystems. Freshwater Biology 33: 419–438. Biggs, B. J. F., 2000. Eutrophication of streams and rivers: dissolved nutrient–chlorophyll relationships for benthic algae. Journal of the North American Benthological Society 19: 17–31. Borchardt, M. A., J. P. Hoffmann & P. W. Cook, 1994. Phosphorus uptake kinetics of Spirogyra fluviatilis (Charophyceae) in flowing water. Journal of Phycology 30: 403–417. Borchardt, M. A., 1996. Nutrients. In Stevenson, R. J., M. L. Bothwell & R. L. Lowe (eds), Algal Ecology: Freshwater Benthic Ecosystems. Academic Press, San Diego, California: 184–228. Bothwell, M. L., 1985. Phosphorus limitation of lotic periphyton growth rates: an intersite comparison using continuousflow troughs (Thompson River, British Columbia). Limnology and Oceanography 30: 527–542. Bothwell, M. L., 1988. Growth rate responses of lotic periphytic diatoms to experimental phosphorus enrichment: the influence of temperature and light. Canadian Journal of Fisheries and Aquatic Sciences 45: 261–270. Bothwell, M. L., 1989. Phosphorus-limited growth dynamics of lotic periphytic diatom communities: Areal biomass and

146 cellular growth responses. Canadian Journal of Fisheries and Aquatic Sciences 46: 1293–1301. Dodds, W. K., B. J. F. Biggs & R. L. Lowe, 1999. Photosynthesis-irradiance patterns in benthic algae: variations as a function of assemblage thickness and community structure. Journal of Phycology 35: 42–53. Dodds, W. K. & E. Welch, 2000. Establishing nutrient criteria in streams. Journal of the North American Benthological Society 19: 186–196. Dodds, W. K., V. H. Smith & K. Lohman, 2002. Nitrogen and phosphorus relationships to benthic algal biomass in temperate streams. Canadian Journal of Fisheries and Aquatic Sciences 59: 865–874. Dodds, W. K., V. H. Smith & B. Zander, 1997. Developing nutrient targets to control benthic chlorophyll levels in streams: a case study of the Clark Fork River. Water Research 31: 1738–1750. Droop, M. R., 1974. The nutrient status of algal cells in continuous culture. Journal of Phycology 9: 264–272. Francoeur, S. N., B. J. F. Biggs, R. A. Smith & R. L. Lowe, 1999. Nutrient limitation of algal biomass accrual in streams: seasonal patterns and a comparison of methods. Journal of the North American Benthological Society 18: 242–260. Hecky, R. E. & E. J. Fee, 1981. Primary production and rates of algal growth in Lake Tanganyika. Limnology and Oceanography 26: 532–547. Hepinstall, J. A. & R. L. Fuller, 1994. Periphyton reactions to different light and nutrient levels and the response of bacteria to these manipulations. Archiv fu¨r Hydrobiologie 131: 161–173. Hill, W. R. & H. L. Boston, 1991. Community-development alters photosynthesis irradiance relations in stream periphyton. Limnology and Oceanography 36: 1375–1389. Hill, W. R., H. L. Boston & A. D. Steinman, 1992. Grazers and nutrients simultaneously limit lotic primary productivity. Canadian Journal of Fisheries and Aquatic Sciences 49: 504–512. Horner, R. R. & E. B. Welch, 1981. Stream periphyton development in relation to current velocity and nutrients. Canadian Journal of Fisheries and Aquatic Sciences 38: 449–457. Horner, R. R., E. B. Welch & R B. Veenstra, 1983. Development of nuisance periphytic algae in laboratory streams in relation to enrichment and velocity. In Wetzel, R. G. (ed.), Periphyton of freshwater ecosystems. Junk Publishers, The Hague: pp. 121–134. Horner, R. R., E. B. Welch, M. R. Seeley & J. M. Jacoby, 1990. Responses of periphyton to changes in current velocity, suspended sediment and phosphorus concentration. Freshwater Biology 24: 215–232. Humphrey, K. P. & R. J. Stevenson, 1992. Response of benthic algae to pulses in current and nutrients during simulations of subscouring spates. Journal of the North American Benthological Society 11: 37–48. Lamberti, G. A. & A. D. Steinman (eds), 1993. Research in artificial streams: Application, uses, and abuses. Journal of the North American Benthological Society 6: 92–104. Lohman, K., J. R. Jones & B. D. Perkins, 1992. Effects of nutrient enrichment and flood frequency on periphyton biomass in northern Ozark streams. Canadian Journal of Fisheries and Aquatic Sciences 49: 1198–1205.

Monod, J., 1950. La technique de culture continue, theorie et applications. Annales de l’Institut Pasteur, Paris 79: 390– 410. Mulholland, P. J., A. D. Steinman, A. V. Palumbo, J. W. Elwood & D. B. Kirschtel, 1991. Role of nutrient cycling and herbivory in regulating periphyton communities in laboratory streams. Ecology 73: 966–982. Mundie, J. H., K .S. Simpson & C. J. Perrin, 1991. Responses of stream periphyton and benthic insects to increases in dissolved inorganic phosphorus in a mesocosm. Canadian Journal of Fisheries and Aquatic Sciences 48: 2061–2072. Olsen, Y., 1989. Evaluation of competitive ability of Staurastrum luetkemuellerii (Chlorophyceae) and Microcystis aeruginosa (Cyanophyceae) under P limitation. Journal of Phycology 25: 486–499. Pan, Y. & R. L. Lowe, 1994. Independent and interactive effects of nutrients and grazers on benthic algal community structure. Hydrobiologia 291: 201–209. Peterson, B. J., L. Deegan, J. Helfrich, J. Hobbie, M. Hullar, B. Moller, T. E. Ford, A. Hershey, A. Hiltner, G. Kipphut, M. A. Lock, D. M. Fiebig, V. McKinley, M. C. Miller, R. Vestal, R. Ventullo & G. Volk, 1993. Biological responses of a tundra river to fertilization. Ecology 74: 653–672. Pringle, C. M. & J. A. Bowers, 1984. An in situ substratum fertilization technique: diatom colonization on nutrientenriched, sand substrata. Canadian Journal of Fisheries and Aquatic Sciences 41: 1247–1251. Redfield, A. C., 1958. The biological control of chemical factors in the environment. American Scientist 46: 205–222. Rhee, G.-Y. & I. J. Gotham, 1981. The effect of environmental factors on phytoplankton growth: temperature and the interactions of temperature with nutrient limitation. Limnology and Oceanography 26: 635–648. Rier, S. T. & R. J. Stevenson., 2001. Relation of environmental factors to density of epilithic lotic bacteria in 2 ecoregions. Journal of the North American Benthological Society 20: 588–600. Rosen, B. H. & R. L. Lowe, 1984. Physiological and ultrastructural responses of Cyclotella meneghiniana (Bacillariophyta) to light intensity and nutrient limitation. Journal of Phycology 20: 173–182. Schanz, F. & H. Juon, 1983. Two different methods of evaluating nutrient limitations of periphyton bioassays using water from the Rhine River and eight of its tributaries. Hydrobiologia 102: 187–195. Slavik, K., B. J. Peterson, L. A. Deegan, W. B. Bowden, A. E. Hershey & J. E. Hobbie, 2004. Long-term responses of the Kuparuk River ecosystem to phosphorus fertilization. Ecology 85: 939–954. Sokal, R. R. & F. J. Rohlf, 1995. Biometry: The principles and practice of statistics in biological research 3rd ed. W. H. Freeman and Company, New York. Stelzer, R. S. & G. A. Lamberti, 2001. Effects of N:P ratio and total nutrient concentration on stream periphyton community structure, biomass, and elemental composition. Limnology and Oceanography 46: 356–367. Stevenson, R. J., 1990. Benthic algal community dynamics in a stream during and after a spate. Journal of the North American Benthological Society 9: 277–288.

147 Stevenson, R. J., 1996. The stimulation and drag of current. In Stevenson R. J., M. L. Bothwell & R. L. Lowe (eds), Algal Ecology: Freshwater Benthic Ecosystems. Academic Press, San Diego, California: 321–340. Stevenson, R.J. & E. F. Stoermer, 1982. Luxury consumption of phosphorus by five Cladophora epiphytes in Lake Huron. Transactions of the American Microscopical Society 101: 151–161. Stevenson, R. J., C. G. Peterson, D. B. Kirschtel, C. C. King & N. C. Tuchman, 1991. Density-dependent growth, ecological strategies, and effects of nutrients and shading on benthic diatom succession in streams. Journal of Phycology 27: 59–69. Stevenson, R. J. & R. Glover, 1993. Effects of algal density and current on ion transport through periphyton communities. Limnology and Oceanography 38: 1276–1281.

Stevenson, R. J. & Y. Pan, 1995. Are evolutionary tradeoffs evident in responses of benthic diatoms to nutrients? In D. Marino (ed.), Proceedings of the 13th International Diatom Symposium 1994, pp. 71–81. USEPA, 1998. National Strategy for the Development of Regional Nutrient Criteria. EPA 822-R-98-002. Office of Water, U.S. Environmental Protection Agency, Washington, D.C. Whitford, L. A., 1960. The current effect and growth of freshwater algae. Transactions of the American Microscopical Society 79: 302–309. Whitford, L. A. & G. J. Schumacher, 1964. Effects of current on respiration and mineral uptake of Spirogyra and Oedogonium. Ecology 45: 168–170.

Hydrobiologia (2006) 561:149–165  Springer 2006 R.J. Stevenson, Y. Pan, J.P. Kociolek & J.C. Kingston (eds), Advances in Algal Biology: A Commemoration of the Work of Rex Lowe DOI 10.1007/s10750-005-1611-5

Comparing effects of nutrients on algal biomass in streams in two regions with different disturbance regimes and with applications for developing nutrient criteria R. Jan Stevenson1,*, Steven T. Rier2, Catherine M. Riseng3, Richard E. Schultz4 & Michael J. Wiley3 1

Department of Zoology, Michigan State University, East Lansing, MI, 48824, USA Department of Biological & Allied Health Sciences, Bloomsburg University, Bloomsburg, PA, 17815, USA 3 School of Natural Resources and the Environment, The University of Michigan, Ann Arbor, MI, 48109, USA 4 Department of Biological Sciences, University of Louisville, Louisville, KY, 40292, USA (*Author for correspondence: E-mail: [email protected]) 2

Key words: algae, benthic, nutrients, streams, regions, biomass

Abstract Responses of stream algal biomass to nutrient enrichment were studied in two regions where differences in hydrologic variability cause great differences in herbivory. Around northwestern Kentucky (KY) hydrologic variability constrains invertebrate biomass and their effects on algae, but hydrologic stability in Michigan (MI) streams permits accrual of high herbivore densities and herbivory of benthic algae. Multiple indicators of algal biomass and nutrient availability were measured in 104 streams with repeated sampling at each site over a 2)month period. Many measures of algal biomass and nutrient availability were positively correlated in both regions, however the amount of variation explained varied with measures of biomass and nutrient concentration and with region. Indicators of diatom biomass were higher in KY than MI, but were not related to nutrient concentrations in either region. Chl a and % area of substratum covered by Cladophora were positively correlated to nutrient concentrations in both regions. Cladophora responded significantly more to nutrients in MI than KY. Total phosphorus (TP) and total nitrogen (TN) explained similar amounts of variation in algal biomass, and not significantly more variation in biomass than dissolved nutrient concentrations. Low N:P ratios in the benthic algae indicated N as well as P may be limiting their accrual. Most observed responses in benthic algal biomass occurred in nutrient concentrations between 10 and 30 lg TP l)1 and between 400 and 1000 lg TN l)1.

Introduction Problems with high algal biomass due to nutrient enrichment in streams and other habitats has been the subject of considerable concern (Carpenter et al., 1998; Smith et al., 1999) and have lead to efforts by the US Environmental Protection Agency to develop nutrient criteria that prevent nuisance growths of algae (USEPA, 1999). Many

studies have documented effects of nutrients on algae (Bourassa & Cattaneo, 1998; Dodds et al., 1998; Biggs, 2000), but relationships in regional studies are seldom precise or accurate enough to predict benefits of specific nutrient regulations for streams throughout a region. In addition, many factors could affect algal–nutrient relationships in streams within a region and among regions with different climate, geology, water chemistry,

150 and hydrology (Biggs et al., 1990, 1998; Stevenson, 1997b). Understanding differences in algal– nutrient relationships between very different regions would help establish the range of possibilities that can be expected and the factors that regulate that range. Disturbance regime has been hypothesized to be a major factor regulating between-region differences in algal–nutrient relationships in streams. High current velocities associated with elevated discharge following rains scour algae directly or indirectly by tumbling substrata (Power & Stewart, 1987; Biggs & Thomsen, 1995). Disturbance regimes with low intensity or high predictability (Poff & Ward, 1989) support development of high grazer densities that constrain algal growths (Wootton et al., 1996; Riseng et al., 2004). Alternatively, high intensity and frequency disturbance regimes could constrain algal growth (Biggs, 1995; Clausen & Biggs, 1997). Shading and regional factors such as water chemistry and temperature may also affect the ability of large growths to occur by limiting photosynthesis or colonization of sites by species capable of producing nuisance growths. Lack of precision in relationships between algal biomass and nutrient concentrations may also be due to variability in estimates of algal biomass and nutrient concentrations in streams. Algal biomass varies greatly during colonization after storm disturbances (Fisher et al., 1982). Patchy growths of macroalgae throughout a reach and even a riffle make assessment of algal biomass difficult when sampling. The phyla and growth forms of algae (i.e., functional group) vary seasonally from diatoms to filamentous green and blue-green algae (Blum, 1957; Hynes, 1970). Each functional group has the potential to develop very different biomasses. Nutrient concentrations vary: diurnally with microbial metabolism; daily with weatherrelated hydrologic factors and with increasing biomass and nutrient uptake during periphyton community development after storms; and seasonally with human activities and metabolism of terrestrial vegetation in watersheds (Meyer et al., 1988; Kim et al., 1992). A sampling approach with extensive and repeated sampling during the period of algal community development should reduce error variance in algal–nutrient relationships by

accounting for spatial and temporal variability in nutrient concentrations. The objectives of this study were to relate benthic and suspended algal biomass to nutrient concentrations in streams of two regions with very different disturbance regimes and where very different relationships were expected. We expected that algal–nutrient relationships would vary with type of algae, the nutrient parameter, and the region studied such that more precise relationships could be quantified if these factors were accounted for. Changes in algal biomass along nutrient gradients were studied for thresholds and additional justification for establishing nutrient criteria to prevent nuisance growths of benthic algae.

Methods Field methods Benthic algal biomass was assessed in 104 streams in two regions of the north-central region in the United States (Fig. 1). These regions were chosen because of observed differences in grazing effects on benthic algae that have been related to disturbance regimes in the streams (Riseng et al., 2004). Benthic invertebrate biomass was significantly lower in streams of the Kentucky–Indiana region (KY) than Michigan (MI) (Riseng et al., 2004), t-test, p<0.001). Biomass of invertebrates was usually less than 2 g m)2 in the KY and greater than 3 g m)2 in MI. Invertebrates assemblages were dominated by mayflies in the KY region and by caddisflies in MI. The KY region included 46 streams of the unglaciated Knobs region of northwestern Kentucky and southeastern Indiana, which is typified by rapid runoff due to steep topography and the underlying limestone bedrock (Burroughs, 1926; McGrain, 1983). Fifty-eight streams were located in the glaciated region of Michigan’s Lower Peninsula (MI region), where precipitation slowly percolates to the groundwater and thereby, contributes to stable, consistent discharge rates (Wiley et al., 1997). Streams were chosen to represent a wide range of nutrient conditions, which were predicted based on land use/land cover types. Land use/land cover ranged from relatively pristine

151

N MI

(a) IN

(b) KY

0 100

km

500

Figure 1. Location of streams sampled in northwestern Kentucky (KY) and southeastern Indiana (IN) and in Michigan (MI, USA).

forested and wetland areas to urban streams. Most stream sites flowed through a mixture of agricultural and forested land. Stream size ranged from 1st to 4th order (Strahler, 1952) with most streams being 2nd and 3rd order. Sampling was conducted during 2-month sampling periods for each stream, May and June in KY and July and August in MI, during both 1996 and 1997. The sampling periods were chosen to control for temporal differences in water temperature, stream flow, and observed periods of Cladophora growth caused by latitudinal and hydrologic differences between the regions. Flow and Cladophora blooms were commonly gone in many KY streams by early summer. We restricted our sampling to riffles to minimize habitat variation among study sites and to concentrate on areas where nuisance algal proliferations tend to be greatest. We conducted 145 assessments of algal biomass, nutrient concentrations, and habitat features in 104 streams, with 74 assessments in 1996 and 71 in 1997. Forty-one of the stream sites sampled in 1996 were re-sampled in 1997 because of the limited number of sites in the regions targeted. Eight

pairs of sites (four from each region) were from the same stream in a watershed, but the lower sites were always at least 5 river kilometers downstream of the upper site. We assumed all 145 assessments were independent because flood flows in the same stream among years are commonly assumed to be independent for hydrologic and population time-series analysis in rivers (Cobb et al., 1992; Gordon, 1995; Kohler & Wiley, 1997). Riseng et al. (2004) found significant inter-annual differences in grazer and filter-feeder populations at our revisited sites (paired t-test, p<0.05) and that peak discharge and benthic chlorophyll levels were not correlated among sites between years. Each assessment involved 8–9 visits to KY streams and £ 4 visits to MI streams during a 2-month sampling period. A higher sampling frequency was used in KY versus MI because hydrologic variability was less and sites were much farther apart in MI than KY. During these weekly or biweekly visits, discharge was assessed with a Marsh–McBirney current meter; pH and conductivity were determined with a YSI meter (YSI Incorporated, Yellow Springs, Ohio, USA); water temperature was determined with a thermometer;

152 and canopy cover was assessed with a spherical canopy densiometer. Samples for nutrient and chloride assessments were collected in 2 125-ml acid-rinsed polyethylene bottles. Water in one sample was filtered in the field through 0.45-lm pore-size filters to measure dissolved nutrients. Nutrient samples were stored on ice until returning to the lab where they were frozen until analysis. Rapid periphyton surveys were also conducted during the weekly or biweekly visits to a site. This protocol was used for assessing algal biomass visually at larger spatial scales than practical when scraping algae on rocks and for distinguishing biomass composed of major phyla and growth forms of algae. A viewing bucket was constructed with the bottom cut out of a plastic tub and replaced with a clear plastic circle with a watertight seal. Use of a viewing bucket reduced glare from sunlight on the water surface and improved visibility of the bottom in these streams with relatively high light levels. A 50-point grid was marked on the bottom plastic circle to facilitate characterizing the percent of stream bottom in different conditions. Visual assessments were taken at nine locations throughout the targeted riffle at each site. Each assessment involved submerging the bottom of the viewing bucket in the water and characterizing: (1) % of the bottom covered by macroalgae (e.g., Cladophora); (2) the average cover of suitable substrata (>2 cm diameter); and (3) thickness of the each different type of microalgae (e.g., diatoms and blue-green algae). Thickness of microalgae was characterized with a ranking system in which: 0 indicated the rock felt rough and had very little or no periphyton; 0.5 indicated the rock felt slippery, but no periphyton could be seen; 1 indicated periphyton was visible in a very thin biofilm; 2 indicated periphyton with a thickness >thin biofilm but £1 mm; 3 indicated periphyton with a thickness >1 mm but £5 mm; 4 indicated periphyton with a thickness >5 mm but £2 cm; and 5 indicated periphyton >2 cm thick. The average diatom thickness rank and chl a (method described later) were determined separately on 50 rocks to assess the reliability of visual assessments of diatom biomass. Algae on rocks were sampled during the last two weeks of the 2-month sampling period and more than 5 days after a moderate storm event. Benthic algae were sampled using a spoon and

toothbrush from the tops of 15 rocks randomly selected in the riffle at each site. Subsamples for different assays were separated in the field and stored on ice for chl a and mat chemistry analysis and preserved with M3 for AFDM assays. Estimates of rock surface areas from which algae were scraped were made in the field by measuring the upper surface of rocks. Water samples were analyzed for Cl, NO3+NO2 (NOx), and NH4 using a Skalar auto-analyzer, for soluble reactive P (SRP) using a Hitachi U-2001 spectrophotometer, and for alkalinity according to standard methods (APHA, 1998). To determine TP and TN concentrations, particulate matter in water samples was oxidized with persulfate and analyzed for SRP and NOx (D’Elia et al., 1977; APHA, 1998). Mat chemistries were determined by subsampling portions of mats and analyzing the samples for TN and TP as described for water samples. Mat chemistries were calculated as the proportion of N and P in periphyton samples (including algae, meiofauna, bacteria, detritus, and silt) by dividing the mass of N and P in samples by the AFDM. AFDM was measured after drying samples in aluminum pans and ashing at 500 C (APHA, 1998). Chlorophyll a was extracted from algal samples with 90% buffered acetone and measured spectrophotometrically (APHA, 1998).

Data analysis Simple correlation, linear and non-linear regression, and multivariate regression approaches were used to determine relationships between algal biomass and nutrient concentrations in streams of the two regions. All statistical analyses were calculated using SYSTAT version 10 (Wilkinson, 1990). Variables with skewed distributions were natural-log transformed to produce more even and normal distributions. Descriptive statistics and t-tests were used to determine whether significant differences in environmental conditions occurred between regions. Covariance matrices for nutrient concentrations and algal biomass were calculated independently to evaluate differences among regions. Correlations between the three algal biomass attributes (benthic chl a, diatom rank, average Cladophora cover during the

153

10.00 Chla (mg/cm2)

sampling period) and average nutrient concentrations at sites were calculated independently for each region and for both regions combined to determine the importance of region for explaining variation in algal biomass–nutrient relationships. A fourth algal biomass attribute, maximum Cladophora cover at a site during a sampling period, was included in assessment of the probability of different average and maximum Cladophora covers occurring in streams with different nutrient concentrations. Linear and non-linear regression analyses were used to test the hypothesis that non-linear models explained significantly more variation in algal biomass–nutrient relationships than linear models. The Monod equation was modified for the nonlinear model,

0.10

B ¼ Bmax  ðS=ðS þ Ks ÞÞ

10.00 AFDM (mg/cm2)

where biomass (B) are predicted as a function of the maximum biomass possible (Bmax), nutrient concentrations (S), and a nutrient concentration at which half of Bmax would occur (Ks). The Monod model was selected a priori because we hypothesized that algal biomass would increase rapidly and then saturate with progressive increases in nutrient concentrations in the same way as growth rates. Stepwise multiple regression was used to determine whether TN and TP explained more variation in algal biomass than either nutrient alone or whether other environmental factors such as canopy cover accounted for unexplained variation in biomass–nutrient relationships. TN and TP were used in detailed analyses because these parameters are recommended for nutrient criteria development (USEPA, 1999) and they indicate nutrient availability as well or better than other parameters (Dodds, 2003) The Z-test (Zar, 1974) was used to compare the correlations among analyses and to determine whether differences in the amount of variance explained were statistically significant. Difference in algal responses to nutrients between regions was determined by direct comparison of regression coefficients with standard errors of coefficients and t-statistics (Zar, 1974). Analysis of covariance was not used to compare relationships between regions because error variances could theoretically be very different in the two regions.

1.00

1.00

0.10 0

1

2 3 Diatom Rank

4

Figure 2. Relationships between chl a and AFDM and diatom rank determined by assessment of individual rocks to assess the precision of visual assessments. Lines on figures produced by lowess smoothing (Wilkinson, 1990).

Results Accuracy of rapid periphyton survey Estimates of diatom biomass using thickness rank explained over 58 and 46% of the variation in diatom biomass estimated, respectively, with ln-transformed chl a and ln-transformed AFDM of periphyton on rocks (Fig. 2). The non-linear pattern in log-transformed chl a with increasing diatom rank increased variance explained to 71% by using polynomial regression. The non-linear pattern in AFDM with increasing diatom rank

10

300 200 100

100

10

100 10 1

10.0 1.0 KY

4000 3000 2000 1000 0 15

1.000 0.100 0.010

MI

5000

N/L)* NOx ((μgg N/L)

10.000

6000 5000 4000 3000 2000 1000 0

KY

MI

Region Figure 3. Comparison of non-nutrient abiotic factors between KY and MI. Asterisks indicate mean values were different among ecoregions (t-test, p<0.005).

was not statistically significant and explained little additional variation compared to the linear model. Comparison of stream conditions between KY and MI Averages of almost all non-nutrient parameters, except Cl, were significantly different between regions (t-tests, p<0.05), but ranges of parameters overlapped greatly (Fig. 3). Results indicated that our study streams were moderately buffered with alkalinity averaging 109 and 187 mg l)1 and pH 8.04 and 7.94 in the KY and MI, respectively. Conductivity was lower in KY than MI. Average water temperatures during the sampling periods were lower in KY than MI (14.2 and 18.0 C, respectively), streams were smaller in KY than MI (discharge was 0.32 and 1.52 m3 s)1, respectively), but ranges in both temperature and discharge overlapped in the two regions. Ranges in canopy cover (0–>80% cover) and chloride concentrations (0.5–123.5 mg l)1) also overlapped greatly between the two regions.

100.0

Silica (ml/L)*

100.0 0

TN (μg ( g N/L)

100 90 80 70 60 50 40 30 20 10 0

NH4 ((μg g N/L)

1200 1000 800 600 400 200 0

Canopy Cover (%)**

0

Discharge (m3/s)*

Chloride (mg/L)

Conductivity (μS/cm)*

0

P O4 ((μg g P/L)

20

1000

400

TP (μg P/L)

30

Alkalinity (mg/L)*

Water Temp (ºC)* *

154

10.0 1.0 KY

10 5 0

MI

KY

MI

Region Figure 4. Comparison of nutrient concentrations between KY and MI. Asterisks indicate mean values were different among ecoregions (t-test, p<0.005).

Average concentrations of TP, SRP, TN and NH4 and ratios of N:P in water and P:AFDM did not differ between regions and wide ranges in nutrient conditions were observed (Fig. 4). However, NOx, Si, mat N:P ratios, and mat N:AFDM ratios differed between regions, even though their ranges overlapped greatly. Many observations of SRP and TP concentrations were less than 10 lg P l)1 and some were greater than 100 lg l)1. SRP averaged 62% of TP in KY and only 36% of TP of MI. All but four TP concentrations in KY were less than 100 lg l)1. Figures illustrating algal response to TP were limited to less than 100 to optimize use of space and more clearly illustrate changes in algal biomass with TP. Silica concentrations ranged from 1.7 to 12.5 mg l)1 with averages of 6.5 and 8.0 in the two regions. Molar N:P ratios averaged 103 and 84 in the water and 11.3 and 9.2 in periphyton in KY and MI, respectively. NOx and TN concentrations were often less than 50 and 100 lg N l)1 respectively, and as high as 5497 lg TN l)1. NOx was 60 and 83% of TN in the KY and MI, respectively. NH4, ranging from less than 5–277 lg N l)1, was less than 5% of TN in both regions.

155 Relations between nutrients and environmental factors Nutrient concentrations were highly correlated with each other, but were not highly correlated with alkalinity, canopy cover, and discharge. Pearson correlations among nutrient concentrations were higher among KY than MI streams, with all r being greater than 0.6 in KY and less than 50% of r being greater than 0.6 in MI (Table 1). Conductivity and Cl were correlated to nutrients in the MI and KY, but pH was correlated significantly with nutrient concentrations only in KY streams. Even though ln(TP) and ln(TN) in streams were correlated with each other in KY and MI (r=0.69 and r=0.62, respectively), many high N sites had low P concentrations. However, low N sites with high P were absent from the possible combinations of nutrient conditions in both regions. SRP and TP were more highly correlated in KY than MI (Z-test, p<0.05). NOx was highly correlated with TN in both regions and more highly correlated with TN than NH4. Determining algal biomass relations with nutrients Variability in most measures of algal biomass was great in both regions (Fig. 5). Benthic chl a varied from <0.1 lg cm)2 to >40 lg cm)2. Average diatom rank at a site varied from less than 0.5 to about 2.0. Average and maximum Cladophora cover at a site varied from 0 to 80%. Benthic chl a and diatom rank were significantly greater in KY (averaging 13.0 lg chl a cm)2 and thickness rank=0.92) than MI (averaging 4.1 lg chl a cm)2 and rank=0.55). Correlations between all measures of N or P concentrations and measures of algal biomass were positive (Table 2) and more than half were statistically significant (p<0.05, without Bonferroni correction). Diatom rank was not significantly correlated to nutrient concentration in either region, except for NH4. Diatom abundances ranged widely from low to high ranks at all nutrient concentrations in KY, but ranged little in MI streams unless nutrient concentrations were really high (Fig. 4). Correlations between nutrients and algal attributes were higher for specific regions than when

Figure 5. Individual relationships between benthic chl a, diatom thickness rank, and average Cladophora cover and average total nitrogen (TN) and total phosphorus (TP) concentrations in streams in two regions, the Knobs (shaded circles) and the glaciated region (squares circles). The lines indicate the best fit of linear regression for streams in the Knobs (solid line) and in the glaciated region (dashed line). Statistical significance of the relationships can be found in Table 4.

data from both regions were combined (Table 2), especially for the most highly correlated algal attributes. For example, correlations between Cladophora cover and total nutrients averaged 0.354 when data for both regions were combined; they were significantly less (Z-test, p<0.05) than the average r=0.608 in MI where Cladophora cover was most highly correlated with TP and TN. Benthic chl a was better correlated to all nutrients in KY streams than when streams from MI were included in calculations. NO3, SRP, TN, and TP were similarly correlated with measures of algal biomass, except for low correlations between SRP and Cladophora cover in MI. Correlations tended to be higher between measures of algal biomass and NH4 than other nutrient concentrations. Algal biomass

156 Table 1. Correlations among water chemistry parameters in streams of the Knobs and glaciated region pH

Alk

CanCov

Cond

Cl

NOx

NH+ 4

TN

Alkalinity

0.568

Canopy cover

0.368

0.12

Conductivity Chloride

0.515 0.278

0.87 0.279

)0.007 0.120

NO3

0.235

0.150

0.203

0.299

0.758

NH4

0.353

0.185

0.336

0.339

0.675

TN

0.272

0.205

0.218

0.346

0.786

0.984

0.663

SRP

0.407

0.158

0.308

0.348

0.582

0.600

0.733

0.639

TP

0.418

0.248

0.255

0.421

0.649

0.645

0.749

0.694

Alkalinity Canopy cover Conductivity

Knobs Knobs Knobs

0.456

Knobs Knobs

0.602

Knobs Knobs Glaciated

0.054 0.331

Glaciated Glaciated

0.111

0.243

0.082

0.099

0.814

NO3

0.093

0.087

0.013

0.528

0.460

Glaciated Glaciated

NH4

0.168

0.088

0.083

0.649

0.662

0.281

TN

0.204

0.014

)0.081

0.601

0.554

0.852

0.435

)0.029

0.383

0.029

0.514

0.455

0.284

0.616

0.333

0.308

0.010

0.107

0.644

0.588

0.465

0.659

0.616

TP

Knobs 0.960

)0.039 0.081 0.269

Region

Knobs

Chloride

SRP

SRP

correlations with total nutrients were about the same as with dissolved nutrients. Silica concentrations were negatively correlated with benthic chl a and Cladophora cover in KY and when data for both regions were combined. Multivariate and non-linear models did not explain significantly more variation in algal biomass than univariate linear models using TN and TP. If r2 of multivariate and non-linear models were greater than for univariate linear models, then they were not more than 20% greater.

Glaciated Glaciated Glaciated 0.482

Glaciated

Only 4 of 36 correlations were statistically significant (p<0.05) between algal biomass measures and either water column N:P ratios or periphyton P/AFDM, N/AFDM, or N:P ratios (Table 3). Significantly positive correlations were observed between chl a and N/AFDM in MI and both regions together, between diatom rank and N/AFDM in both regions together, and between chl a and mat N:P ratio in MI streams. Water column N:P ratios were not related to algal biomass in either region.

Table 2. Correlations between measures of benthic algal biomass and nutrient indicators independently in streams of the Knobs and glaciated region and jointly with data from both regions Region

Biomass variable

ln(TN)

ln(TP)

ln(NOx)

ln(NH4)

ln(SRP)

ln(Si) )0.348

Knobs

ln (ben chl a)

0.433

0.416

0.399

0.521

0.369

Knobs

ln (dia rank)

0.177

0.124

0.141

0.336

0.136

0.116

Knobs

ln (Clad cov)

0.289

0.224

0.267

0.200

0.176

)0.379

Glaciated

ln (ben chl a)

0.298

0.320

0.164

0.528

0.436

0.176

Glaciated

ln (dia rank)

0.223

0.298

0.167

0.471

0.142

)0.059

Glaciated

ln (Clad cov)

0.548

0.668

0.469

0.523

0.129

0.048

Both

ln (ben chl a)

0.269

0.229

0.313

0.393

0.318

)0.263

Both Both

ln (dia rank) ln (Clad cov)

0.126 0.368

0.104 0.340

0.181 0.355

0.280 0.361

0.139 0.145

)0.065 )0.217

Correlation coefficients (r) in bold indicate statistical significance with p<0.05.

157 Table 3. Correlations between measures benthic algal biomass, nutrient ratios in the water, mat nutrient concentrations and ratios, and other stream chemistry and habitat parameters (CanCov, canopy cover; Cl, chloride; and cond, conductivity) Region

Biomass variable

Water N:P

Mat P/AFDM

N/AFDM

N:P

)0.045 0.039

)0.184 )0.036

)0.156 )0.091

0.100 )0.097

Knobs Knobs

ln (ben chl a) ln (dia rank)

Knobs

ln (Clad cov)

0.049

)0.047

)0.118

)0.127

Glaciated

ln (ben chl a)

)0.058

0.246

0.412

0.111

Glaciated

ln (dia rank)

)0.08

)0.070

)0.254

Glaciated

ln (Clad cov)

)0.141

)0.115

Both

ln (ben chl a)

0.028

0.185

CanCov

ln (Cl)

ln (cond)

0.374 0.351

0.540 0.057

0.419 0.090

0.196

0.422

0.538

)0.23

0.314

0.386

)0.193

)0.228

0.327

0.421

0.163

0.244

)0.072

0.583

0.589

0.463

0.273

)0.129

0.260

0.091

Both

ln (dia rank)

0.041

0.180

0.322

0.211

0.114

0.100

)0.036

Both

ln (Clad cov)

)0.020

0.008

0.123

0.154

0.073

0.501

0.493

Correlation coefficients (r) in bold indicate statistical significance with p<0.05.

Canopy cover was positively correlated (p<0.05) to chl a and diatom rank in KY, but was not related to algal biomass in other cases (Table 3). Even Cladophora cover was not related significantly to canopy cover. Correlations between chl a/AFDM ratios and canopy cover were r=0.444 in KY and r=)0.014 in MI. Chloride concentration and conductivity were positively correlated (p<0.05) with measures of algal biomass in both KY and MI. Correlations between these indicators of human disturbance and algal biomass were usually higher than correlations between nutrient concentrations and algal biomass (Table 3). Comparison of biomass–nutrient models between regions Algal biomass was usually lower in MI than KY, as indicated by the significantly lower constants in MI than KY in TP models (Fig. 5, Table 4). However, responses of benthic algal biomass to nutrients seemed to be due to changes in Cladophora versus epilithic diatom biomass, and these were higher in MI than KY. Chl a increased with nutrients in both regions at approximately the same rate, approximately 0.6 lg chl a cm)2 per ln (lg l)1) of both TP and TN. Diatom thickness rank did not increase significantly with nutrients in either region, however Cladophora cover did. TN and TP models indicated that Cladophora cover was lower in low nutrient conditions in the MI than KY, but increased faster with nutri-

ent enrichment in MI. The constants for models relating TN and TP with Cladophora cover were significantly lower in the MI than KY for both TN and TP models, but slopes for these models were higher in the MI than KY. The slope for the Cladophora/TP model was not significantly different than 0.0 in KY (p=0.058). Cladophora cover Due to differences in Cladophora/nutrient models between regions, differences in percent cover in different nutrient regimes were examined more closely (Figs. 6 and 7). The probability of observing average Cladophora cover <1% in a stream was negatively related (Spearman rank correlations, p>0.05) to nutrient concentrations in MI, but not in KY. However, Cladophora accrual was severely constrained in low nutrient streams in both regions. Average Cladophora cover was less than 1% in all streams of MI having less than 300 lg TN l)1 and 10 lg TP l)1. Thirty and fifty percent of KY streams in that low nutrient category (300 lg TN l)1 and 10 lg TP l)1) had, respectively, average and maximum Cladophora cover <1%. In addition, only 1 of 25 of these low nutrient KY streams had average Cladophora cover >10% and only 3 of 25 had maximum cover >10%. The probability of maximum Cladophora being >40% increased significantly (Spearman rank correlations, p>0.05) with both TN and TP conditions in both regions. The probability of average Cladophora cover being greater than 10%

158 Table 4. Linear regression statistics for relationships between six measures of algal biomass and nutrient conditions in streams of KY and MI Dependent variable

Independent variable

Region

Constant

Constant (SE)

Slope

Slope (SE)

Model p

ln (chl a) ln (chl a)

ln (TN) ln (TN)

KY MI

)2.098 )3.859

1.035 1.625

0.62 0.639

0.16 0.247

<0.001 0.012

ln (chl a)

ln (TN)

both

)2.389

1.07

0.535

0.164

0.001

ln (chl a)

ln (TP)

KY

0.408

0.422

0.521

0.239

<0.001

ln (chl a)

ln (TP)

MI

)1.908

0.813

0.72

0.256

0.006

ln (chl a)

ln (TP)

both

)0.167

0.471

0.42

0.153

0.007

ln (diatom rnk)

ln (TN)

KY

0.298

0.214

0.049

0.033

0.137

ln (diatom rnk)

ln (TN)

MI

0.155

0.188

0.041

0.028

0.15

ln (diatom rnk) ln (diatom rnk)

ln (TN) ln (TP)

both KY

0.317 0.529

0.172 0.089

0.035 0.03

0.026 0.029

0.18 0.299

ln (diatom rnk)

ln (TP)

MI

0.245

0.093

0.06

0.03

0.053

ln (diatom rnk)

ln (TP)

both

0.467

0.074

0.027

0.024

0.269

ln (avg Clad cov)

ln (TN)

KY

)1.287

0.953

0.369

0.146

0.014

ln (avg Clad cov)

ln (TN)

MI

)6.717

1.875

1.179

0.611

<0.001

ln (avg Clad cov)

ln (TN)

Both

)2.583

<0.001

ln (avg Clad cov)

ln (TP)

KY

ln (avg Clad cov) ln (avg Clad cov)

ln (TP) ln (TP)

MI Both

0.367 )3.667 )0.317

0.881

0.564

0.136

0.4

0.251

0.13

0.848 0.385

1.567 0.48

0.273 0.125

0.058 <0.001 <0.001

The relations between biomass and TP differ from those illustrated in Figure 4 due to limiting the TP range to <100 lg l)1. SE, standard error. Bold statistics indicate a statistically significant difference (p<0.05) between KY and MI using either pooled estimates of variance or individual estimates of variance. Bold statistics in italics indicate a statistically significant relationship between KY and MI using pooled estimates of variance only.

increased with TN and TP conditions in MI, but not KY. Between 20 and 50% increases in the likelihood of high Cladophora cover (>20% average and >40% maximum cover) were observed in MI when TP exceeded 30 lg TP l)1 and 1000 lg TN l)1. Similarly great increases were not observed in KY, but clear evidence of positive relationships between likelihood of high Cladophora cover and TP and TN concentrations during peak accrual periods were observed there (Figs. 6 and 7).

Discussion Algal–nutrient relations Variation in benthic algal biomass among streams was related to nutrient concentrations, but some relationships changed with region and between diatoms and the macroalga Cladophora. Total benthic algal biomass increased in both MI and KY with increasing concentrations of both soluble

and total nutrient concentrations. However, most of the predictable increase in biomass was probably due to an increase in Cladophora cover. Epilithic diatom biomass did not increase significantly with increasing nutrient concentrations in either region. Algal biomass was similarly related to TN, TP, and soluble nutrients (except NH+ 4 concentration). Dodds (2003) recently reviewed the rationale for total nutrients being better indicators of nutrient supply than soluble nutrients and clearly relates nutrient demand, uptake, and empirical evidence to this issue. However, Biggs (2000) was able to relate benthic algal biomass to soluble nutrients by using annual means of dissolved nutrient concentrations as an indicator of nutrient supply and accounting for time of accrual after disturbance. Our two-month averages in nutrient concentrations in KY streams may also have provided an indicator of nutrient supply. However, the lower variability in algal–nutrient relations in our study than in other large-scale regional studies (Bourassa & Cattaneo, 1998; Dodds et al., 1998, 2002; Biggs,

159

Figure 6. Histograms representing the percentage of streams in different nutrient categories and in different regions with increasing cover of Cladophora. Width of the bands in each bar represents the percentage of streams in a nutrient category with specific ranges of average Cladophora cover. Codes for those ranges are in the legend.

2000) is probably due to other factors as well, such as distinguishing responses of diatoms and Cladophora and controlling for effects of in-stream spatial variation and regions. Differing responses of diatoms and Cladophora may be due to differences in physiological requirements for nutrients or to differences in density-related constraints of nutrient supply through thick diatom and Cladophora accumulations. Accrual of even thick growths of diatoms is saturated in experiments around 30 lg P l)1 (Bothwell, 1989; Rier & Stevenson, accepted), and the negative effects of mat density on solute mixing, nutrient supply rate, and benthic algal growth rates are great (Stevenson et al., 1991; Stevenson & Glover, 1993). Little is known about effects of specific nutrient concentrations on Cladophora growth rates in streams (Dodds & Gudder, 1992). The ability of Cladophora to develop thick, deep growths in streams (Dodds, 1991) and thick

epiphyte covers may increase the demand of Cladophora for nutrients more than diatom-dominated periphyton. The lack of response of diatoms to nutrients in either region was unexpected, especially in KY where diatom accrual was sometimes great. Greater variability in diatom biomasses in KY than MI were observed. Many KY streams had less than 30 lg P l)1 and 300 lg N l)1, which have been shown to limit accrual of peak biomasses of diatoms in experimental settings (Bothwell, 1989; Rier & Stevenson, accepted). Even though error in nutrient assessments of streams was not sufficient to mask other algal–nutrient relationships, it may have masked diatom–nutrient relationships. Alternatively, high biomasses of algae can accumulate in low nutrient conditions if loss rates are low (e.g., Stevenson et al., 2002). Differences in benthic diatoms between regions in our study were probably due to grazing by

160

Figure 7. Histograms representing the cumulative frequency of streams in different nutrient categories and in different regions with increasing cover of Cladophora. Width of the bands in each bar represents the percentage of streams in a nutrient category with specific ranges of maximum Cladophora cover. Codes for those ranges are in the legend.

invertebrates. Benthic invertebrate biomass was much lower in streams of KY than MI. Low invertebrate biomass in the KY has been related to the more intense natural disturbance regime due to flooding and summer drought (Riseng et al., 2004). Invertebrate grazers are commonly associated with regulation of diatom-dominated periphyton (Lamberti & Resh, 1983; Feminella & Hawkins, 1995) and it is well documented in MI streams (Kohler & Wiley, 1997). Although averages of many environmental conditions were statistically different between regions, their ranges overlapped greatly. It’s unlikely that the statistically significant differences in water temperature, pH, alkalinity, conductivity, and canopy cover were sufficient to be biologically significant and affect diatom or Cladophora biomass. Ranges in nutrient concentrations were very similar in the two regions. Relationships between algae and nutrient concentrations varied greatly between regions and

were probably related to inaccuracies in measuring nutrient availability, herbivory, and scouring by floods. Correlations between measures of algal biomass and nutrients were higher when observations were constrained to one region or another versus when data for both regions were combined. Although diatom biomass did not increase with nutrients in either region, Cladophora cover did increase with nutrients and more rapidly in MI than KY. The soluble fraction of TP was greater in KY than MI, so the greater responses of Cladophora to low nutrients in KY than MI may be due to a higher fraction of bioavailable P. Grazers can consume Cladophora (Dudley & D’Antonio, 1991), thus lower % cover of Cladophora in low nutrient streams of MI than KY (Figs. 5 and 6) may have been due to higher grazing pressure. Higher likelihood of extensive growths in high nutrient streams of the MI than KY may be due to less frequent flood disturbances in MI.

161 Disturbance regime has been related to algal biomass in streams by Biggs (1995), who suggested that high disturbance regimes constrain algal biomass. This prediction seems to conflict with observations of lowest algal biomass in the low disturbance regimes of our glaciated region and below dams in other studies (Wootton et al., 1996). However, the mechanisms regulating the algal–nutrient relationships in different disturbance regimes and the range of disturbance regime probably account for differences in predictions. The Biggs model was developed in very high disturbance rivers draining high gradient mountain streams, which probably have more frequent and intense disturbances than the Knobs region of KY. Frequent, intense disturbances probably reduce the ability of periphyton to recolonize between storm events by scouring algae off substrata and preventing substantial recolonization. In the very low disturbance regimes of groundwater-fed, hydrologically stable streams in glaciated regions, we find high densities of grazers constraining diatom accrual. Thus, the greatest response of benthic algal biomass to nutrients is most likely at intermediate levels of disturbance, where sufficient time occurs between storm events for algal recolonization but not enough time for regrowth of sufficient invertebrate densities to constrain algal accrual. (Fig. 8) Nutrient criteria The primary purpose of establishing nutrient criteria is to prevent nuisance growths of algae. Nuisance growths may be related to benthic diatoms as well as proliferation of filamentous green algae in streams (Biggs, 2000). Response of Cladophora to nutrients should be given special consideration when establishing nutrient criteria to prevent in-stream nuisance algal growths. Filamentous green algae are prevalent in high biomass conditions (Chetelat et al., 1999), affect aesthetics, may cause local DO-depletion, and alter habitats in ways that affect invertebrates and fish (Dudley et al., 1986). Two important issues for establishment of nutrient criteria are identifying which nutrient limits algal growth and determining the concentration of that nutrient enabling nuisance growths. Phosphorus is usually identified as the nutrient

limiting algal growth in streams, however experiments show that N may also be a factor (Francoeur, 2001). Strength of correlations between biomass and nutrients and nutrient ratios has been used as evidence for likely N or P limitation. Using strength of correlations to indicate limiting nutrients in streams is problematic in streams because high benthic biomass may result in depletion of the most limiting nutrient in the water column. The negative relationship between Si and algal biomass in KY where high diatom biomasses develop demonstrates that nutrient demand is sufficient to deplete water column resources. The high correlation between NH+ 4 concentration and algal biomass could indicate N limitation; NH+ 4 concentrations were very low, relative to NOx and TN and NH+ 4 is assumed to be the preferred form of dissolved N. More likely, NH+ 4 is a good indicator of nutrient loading and supply by human activities, like Cl and conductivity. Water and mat N:P ratios, P/AFDM, and N/ AFDM were poorly related to algal biomass in our study. The positive correlation between N/AFDM and diatom biomass was probably due to correspondence in regional differences in both N/ AFDM and algal biomass rather than an indicator of N regulation of algal biomass. Both N/AFDM and diatom biomass were lower in MI than KY, but the low biomass in MI was most likely due to herbivory. Low N:P ratios in stream water were not related to algal biomass, which was probably again caused by the complexity in relationships between nutrient ratios, loading rates, absolute water column concentrations, and uptake rates by benthic algae (Stelzer & Lamberti, 2001; Snyder et al., 2002). Inclusion and variability in detritus, bacteria, and meiofauna may affect use of N:P ratios in periphyton samples to estimate algal N:P ratios in field studies. Low average N:P ratios in periphyton indicated the possibility of N-limitation of algal accrual. Average N:P ratios in periphyton were 11 and 9 in KY and MI, which is less than Redfield ratio and much less than the 103 and 84 N:P ratios in KY and MI waters. Humphrey & Stevenson (1992) showed molar N:P ratios decreasing from 13.4 to 3.1 during periphyton community development in experimental streams, which was hypothesized to be due to greater cellular retention and within-mat

Peak Biomass

Peak Biomass

162

100 90 80 70 60 50 40 30 20 10 0

100 90 80 70 60 50 40 30 20 10 0

(a) Total Biomass Grazer Resistant Biomass (e.g. Cladophora) Grazer Susceptible Biomass (e.g diatoms)

} Grazer Controlled

(b)

Low Grazing & Relatively Low Disturbance

High Disturbance

}

Grazer Controlled

Low Grazing & Relatively Low Disturbance

Total Biomass Grazer Resistant Biomass (e.g. Cladophora) Grazer Susceptible Biomass (e.g diatoms)

High Disturbance

0 1 2 3 4 5 6 7 8 9 10

Disturbance Regime Figure 8. Conceptual model of effects of physical disturbance regime on peak biomass of grazer susceptible algae, grazer resistant algae (such as Cladophora), and all benthic algae in relatively low (a) and high (b) nutrient conditions. Grazer density is hypothesized to be high in low disturbance regimes, where grazers regulate accumulation of grazer-susceptible algae (such as diatoms). In high nutrients and low disturbance, grazers cannot regulate density of grazer resistant forms, such as Cladophora. Grazer density decreases with disturbance (Riseng et al. 2004) and enables accrual of both grazer-susceptible and resistant forms in habitats with intermediate levels of disturbance. In high disturbance, accrual of both grazer-susceptible and grazer-resistant forms are constrained physical disturbance.

recycling of P than N. Thus, initial algal accrual after disturbance may be regulated by P availability, but sustained accrual to reach high biomass may depend on N availability. In addition, N and P limitation may vary seasonally with seasonal differences in uptake and retention of nutrients by

terrestrial vegetation in temperate zones (Meyer et al., 1988). Criteria for N as well as P may be useful for constraining accumulations of nuisance algal growths. We did not find thresholds in algal biomass–nutrient relations, which can be predicted

163 based on models and can be helpful in establishing criteria for protecting valued ecological attributes (Stevenson, 1997a; Muradian, 2001; Stevenson et al., in press). We found some evidence of diatoms and Cladophora escaping from grazing pressure in high nutrient concentrations in the glaciated MI region, but little evidence of saturating nutrient concentrations at high N and P levels. The great spatial and temporal variability in measures of algal biomass and nutrient availability may have masked non-linear relationships between biomass and nutrients. In addition, saturation of algal growth by N and P at high concentrations may be complicated by the increasing nutrient demand that likely results as biomass increases and constrains nutrient transport and availability (Stevenson & Glover, 1993; Rier & Stevenson, accepted). Without threshold responses in algal–nutrient relationships, we need to rely more on the reference approach for criteria development (Hughes & Larsen, 1986; Hughes, 1995). Streams in both MI and KY with the lowest levels of human disturbance in watersheds (reference streams, n=6) had TP £ 11 lg l)1 and TN £ 400 lg l)1. Nutrient concentrations in our study streams increased with % agriculture and decreased with % forest in both regions (Riseng et al., 2004; Wang & Stevenson, accepted). Given our observed probability of Cladophora cover in low nutrient, reference streams, Cladophora accrual would be very low in the most natural streams. Streams with low nutrient conditions in the EMAP study of the Mid-Atlantic region (unpublished data) also had very little occurrence of filamentous algae. In MI, diatom rank was constrained to 0.5 in reference streams, which corresponded to a chl a concentration of 1.0 lg cm)2; but a wider range of diatom ranks were observed in KY. Figure 5 indicates benthic chl a would be about 1 and 2 lg cm)2 in reference streams of MI and KY, respectively, which is very similar to the definition of oligotrophic streams by Dodds et al. (1998). However, the TP and TN concentrations in our reference streams are about half of the concentrations recommended by Dodds et al. (1998) as the oligotrophic–mesotrophic boundary. High algal biomasses, for example average Cladophora cover >20% or >40% maximum cover were rare (less than 10% of streams) in both

regions, if TP was less than 30 lg l)1 and TN was less than 1000 lg l)1. The 30 lg TP l)1 target was also recognized by Dodds et al. (1998) as a concentration that constrained chl a to less than 15 lg chl a cm)2 in a Cladophora dominated stream. Higher probabilities of more extensive Cladophora growths were observed with increasing nutrient levels. Thus, 30 lg TP l)1 and 1000 lg TN l)1 could be considered as targets to prevent a high probability of nuisance accrual of Cladophora. Thus, different nutrient criteria could be established to support stream-specific management goals. The concept of ‘‘tiered aquatic life uses’’ has been suggested to enable protection of high quality waters as well as goals for restoring impaired ecosystems (S.B. Davies & S.K. Jackson, submitted). To protect naturally low levels of productivity and algal biomass in streams that are hydrogeomorphically similar to our study streams, nutrient concentrations should probably be constrained to £10 lg TP l)1 and 400 lg TN l)1. To prevent nuisance growths of Cladophora, higher criteria in the range of 30 lg TP l)1 and 1000 lg TN l)1 may be satisfactory. More research is necessary to refine these criteria, but these concentrations should provide a starting point for adaptive management of nutrients in streams. In conclusion, relationships between algal biomass and nutrient concentrations in streams varied with the type of algae and the region. Epilithic diatom biomass did not respond as much to nutrient enrichment as Cladophora cover. Some evidence suggested N as well as P limitation of high biomass accrual. Different P and N criteria can be identified to support high quality conditions in selected streams and to restore minimum acceptable levels of biological condition. Further refinement of algal–nutrient relationships is warranted to relate algal problems to N as well as P and to other valued ecological attributes, such as taste and odor in drinking water, low DO, and impaired habitat for invertebrates and fish.

Acknowledgements This research was sponsored by the EPA-STAR Water and Watersheds Program, Grant number R824783. Joe Holomuzki provided assistance in the field and with ideas about study design. Laurel

164 Nida, Carrie DeJaco, and Jennifer Slate helped with field work. Brian Wade provided additional laboratory assistance. Yangdong Pan, Alan Steinman, and people working in Stevenson’s lab provided valuable reviews of the paper. References APHA, 1998. Standard Methods for the Evaluation of Water and Wastewater, 20th edn. American Public Health Association, Washington, DC. Biggs, B. J. F., 1995. The contribution of flood disturbance, catchment geology and land use to the habitat template of periphyton in stream ecosystems. Freshwater Biology 33: 419–438. Biggs, B. J. F., 2000. Eutrophication of streams and rivers: dissolved nutrient–chlorophyll relationships for benthic algae. Journal of the North American Benthological Society 19: 17–31. Biggs, B. J. F., M. J. Duncan, I. G. Jowett, J. M. Quinn, C. W. Hickey, R. J. Davies-Colley & M. E. Close, 1990. Ecological characterisation, classification, and modeling of New Zealand rivers: an introduction and synthesis. New Zealand Journal of Marine and Freshwater Research 24: 277–304. Biggs, B. J. F. & H. A. Thomsen, 1995. Disturbance of stream periphyton by perturbations in shear stress: time to structural failure and differences in community resistance. Journal of Phycology 31: 233–241. Biggs, B. J. F., R. J. Stevenson & R. L. Lowe, 1998. A habitat matrix conceptual model for stream periphyton. Archiv fu¨r Hydrobiologie 143: 21–56. Blum, J. L., 1957. An ecological study of the algae of the Saline River, Michigan. Hydrobiologia 9: 361–408. Bothwell, M. L., 1989. Phosphorus-limited growth dynamics of lotic periphytic diatom communities: areal biomass and cellular growth rate responses. Canadian Journal of Fisheries and Aquatic Sciences 46: 1293–1301. Bourassa, N. & A. Cattaneo, 1998. Control of periphyton biomass in Laurentian streams (Quebec). Journal of the North American Benthological Society 17: 420–429. Burroughs, W. G., 1926. The Geography of the Kentucky Knobs. The Kentucky Geological Survey, Frankfort, Kentucky. Carpenter, S. R., N. F. Caraco, D. L. Correll, R. W. Howarth, A. N. Sharpley & V. H. Smith, 1998. Nonpoint pollution of surface waters with phosphorus and nitrogen. Ecological Applications 8: 559–568. Chetelat, J., F. R. Pick, A. Morin & P. B. Hamilton, 1999. Periphyton biomass and community composition in rivers of different nutrient status. Canadian Journal of Fisheries and Aquatic Sciences 56: 560–569. Clausen, B. & B. J. F. Biggs, 1997. Relationships between benthic biota and hydrological indices in New Zealand streams. Freshwater Biology 38: 327–342. Cobb, D. G., T. D. Galloway & J. F. Flannagan, 1992. Effects of discharge and substrate stability on density and species

composition of stream insects. Canadian Journal of Fisheries and Aquatic Sciences 49: 1788–1795. Davies, S. B. & S. K. Jackson, submitted. The biological condition gradient: a conceptual model for interpreting detrimental change in aquatic ecosystems. Ecological Applications. D’Elia, C. F., J. A. Steudler & N. Corwin, 1977. Determination of total nitrogen in aqueous samples using persulfate digestion. Limnology and Oceanography 22: 760–764. Dodds, W. K., 1991. Micro-environmental characteristics of filamentous algal communities of flowing freshwaters. Freshwater Biology 25: 199–209. Dodds, W. K., 2003. Misuse of inorganic N and soluble reactive P concentrations to indicate nutrient status of surface waters. Journal of the North American Benthological Society 22: 171–181. Dodds, W. K. & D. A. Gudder, 1992. The ecology of Cladophora. Journal of Phycology 28: 415–427. Dodds, W. K., J. R. Jones & E. B. Welch, 1998. Suggested criteria for stream trophic state: distributions of temperate stream types by chlorophyll, total nitrogen and phosphorus. Water Research 32: 1455–1462. Dodds, W. K., V. H. Smith & K. Lohman, 2002. Nitrogen and phosphorus relationships to benthic algal biomass in temperate streams. Canadian Journal of Fisheries and Aquatic Sciences 59: 865–874. Dudley, T. L., S. D. Cooper & N. Hemphill, 1986. Effects of macroalgae on a stream invertebrate community. Journal of the North American Benthological Society 5: 93–106. Dudley, T. L. & C. M. D’Antonio, 1991. The effects of substrate texture, grazing, and disturbance on macroalgal establishment in streams. Ecology 72: 297–309. Feminella, J. W. & C. P. Hawkins, 1995. Interactions between stream herbivores and periphyton: a quantitative analysis of past experiments. Journal of the North American Benthological Society 14: 465–509. Fisher, S. G., L. J. Gray, N. B. Grimm & D. E. Bush, 1982. Temporal succession in a desert stream ecosystem following flash flooding. Ecological Monographs 52: 93–110. Francoeur, S. N., 2001. Meta-analysis of lotic nutrient amendment experiments: detecting and quantifying subtle responses. Journal of North American Benthological Society 20: 358–368. Gordon, N. D., T. A. McMahon & B.L. Finlayson, 1995. Stream Hydrology: An Introduction for Ecologists. John Wiley and Sons, Inc, New York. Hughes, R. M., 1995. Defining acceptable biological status by comparing with reference conditions. In Davis, W. S. & T. P. Simon (eds), Biological Assessment and Criteria: Tools for Water Resource Planning and Decision Making. Lewis Publishers, Boca Raton, Florida, USA: 31–47. Hughes, R. M. & D. P. Larsen, 1986. Regional reference sites: a method for assessing stream potentials. J. M. Omernik Environmental Management 10: 629–635. Humphrey, K. P. & R. J. Stevenson, 1992. Responses of benthic algae to pulses in current and nutrients during simulations of subscouring spates. Journal of the North American Benthological Society 11: 37–48.

165 Hynes, H. B. N., 1970. The Ecology of Running Waters. University of Toronto Press, Toronto. Kim, B. K. A., A. P. Jackman & F. J. Triska, 1992. Modeling biotic uptake by periphyton and transient hyporrheic storage of nitrate in a natural stream. Water Resources Research 28: 2743–2752. Kohler, S. L. & M. J. Wiley, 1997. Pathogen outbreaks reveal large-scale effects of competition in stream communities. Ecology 87: 2164–2176. Lamberti, G. & V. H. Resh, 1983. Stream periphyton and insect herbivores: an experimental study of grazing by a caddisfly population. Ecology 64(5): 1124–1135. McGrain, P., 1983. The Geologic Story of Kentucky. Kentucky Geological Survey. University of Kentucky, Lexington, Kentucky. Meyer, J. L., W. H. McDowell, T. L. Bott, J. W. Elwood, C. Ishizaki, J. M. Melack, B. L. Peckarsky, B. J. Peterson & P. A. Rublee, 1988. Elemental dynamics in streams. Journal of the North American Benthological Society 7: 410–432. Muradian, R., 2001. Ecological thresholds: a survey. Ecological Economics 38: 7–24. Poff, N. L. & J. V. Ward, 1989. Implications of streamflow variability and predictability for lotic community structure: a regional analysis of streamflow patterns. Canadian Journal of Fisheries and Aquatic Sciences 46: 1805–1818. Power, M. E. & A. J. Stewart, 1987. Disturbance and recovery of an algal assemblage following flooding in an Oklahoma stream. American Midland Naturalist 117: 333–345. Rier, S. T. & R. J. Stevenson, 2006. Response of periphytic algae to gradients in nitrogen and phosphorus in streamside mesocosms. Hydrobiologia 561: 131–147. Riseng, C. M., M. J. Wiley & R. J. Stevenson, 2004. Hydrologic disturbance and nutrient effects on the structure of benthic communities in Midwestern streams: a covariance structure analysis. Journal of the North American Benthological Society 23: 309–326. Smith, V. H., G. D. Tilman & J. C. Nekola, 1999. Eutrophication: impacts of excess nutrient inputs on freshwater, marine, and terrestrial ecosystems. Environmental pollution 100: 179–196. Snyder, E. B., C. T. Robinson, G. W. Minshall & S. R. Rushforth, 2002. Regional patterns in periphyton accrual and diatom assemblage structure in a heterogeneous nutrient landscape. Canadian Journal of Fisheries and Aquatic Science 59: 564–577. Stelzer, R. S. & G. A. Lamberti, 2001. Effects of N:P ratio and total nutrient concentration on stream periphyton community structure, biomass, and elemental composition. Limnology and Oceanography 46: 356–367.

Stevenson, R. J., 1997a. Resource thresholds and stream ecosystem sustainability. Journal of the North American Benthological Society 16: 410–424. Stevenson, R. J., 1997b. Scale-dependent determinants and consequences of benthic algal heterogeneity. Journal of North American Benthological Society 16: 248–262. Stevenson, R. J., C. G. Peterson, D. B. Kirschtel, C. C. King & N. C. Tuchman, 1991. Density-dependent growth, ecological strategies, and effects of nutrients and shading on benthic diatom succession in streams. Journal of Phycology 27: 59–69. Stevenson, R. J. & R. Glover, 1993. Effects of algal density and current on ion transport through periphyton communities. Limnology & Oceanography 38: 1276–1281. Stevenson, R. J., Y. Pan & P. Vaithiyanathan, 2002. Ecological assessment and indicator development in wetlands: the case of algae in the Everglades, USA. Verhandlungen Internationale Vereinigung fu¨r Theoretische und Andgewandte Limnologie 28: 1248–1252. Stevenson, R. J., J. Alba-Tercedor, B. Bailey, M. Barbour, C. Couch, S. Dyer, F. Fulk, J. Harrington, M. Harass, C. J. Hawkins, C. Hunsaker, R. Johnson & K. Thornton. in press. Interpreting results of ecological assessments. In Barbour, M., K. Thornton, R. Preston & S. Norton (eds), Ecological Assessment of Aquatic Resources: Linking Science to Decision-Making. Society of Environmental Toxicology and Contamination Publication. Strahler, A. N., 1952. Hyposometric (area-altitude) analysis of erosional topography. Bulletin of the Geological Society of America 63: 1117–1142. USEPA, 1999. Nutrient Criteria Technical Guidance Manual: Rivers and Streams. EPA-822-D-99–003. United States Environmental Protection Agency, Washington, D.C. Wang, Y. K., R. J. Stevenson & L. Metzmeier, 2005. Development and evaluation of a diatom-based index of biotic integrity for the Interior Plateau Ecoregion. Journal of the North American Benthological Society 24: 990–1008. Wiley, M. J., S. L. Kohler & P. W. Seelbach, 1997. Reconciling landscape and local views of aquatic communities: lessons from Michigan trout streams. Freshwater Biology 37: 133–148. Wilkinson, L., 1990. SYSTAT: The System for Statistics. SYSTAT, Inc, Evanston, IL. Wootton, J. T., M. S. Parker & M. E. Power, 1996. Effects of disturbance on river food webs. Science 273: 1558–1561. Zar, J. H., 1974. Biostatistical Analysis. Prentice-Hall, Inc, Edgewood Cliffs, NJ.

Hydrobiologia (2006) 561:167–177  Springer 2006 R.J. Stevenson, Y. Pan, J.P. Kociolek & J.C. Kingston (eds), Advances in Algal Biology: A Commemoration of the Work of Rex Lowe DOI 10.1007/s10750-005-1612-4

Differential heterotrophic utilization of organic compounds by diatoms and bacteria under light and dark conditions Nancy C. Tuchman1,*, Marc A. Schollett1, Steven T. Rier1,2 & Pamela Geddes1 1

Department of Biology, Loyola University Chicago, 6525 N. Sheridan Road, Chicago, Illinois, 60626, USA Department of Biological and Allied Health Sciences, Bloomsburg University, Bloomsburg, Pennsylvania, 17815, USA (*Author for correspondence: E-mail: [email protected]) 2

Key words: diatoms, bacteria, heterotrophy, organic substrates

Abstract The heterotrophic utilization of organic substrates by diatoms is likely an important survival strategy when light levels are too low for photosynthesis. The objectives of this study were: (1) to determine if heterotrophic utilization of a large array of organic compounds by eight common freshwater benthic diatom taxa was light-dependent, and (2) to determine if organic substrate utilization patterns differed between darkgrown diatoms and bacteria as a possible means of reducing competition by niche separation. Eight lightand dark-grown diatom taxa and five bacterial species were incubated in 96-well Biolog Microtiter plates with each well containing 1 of 95 different organic substrates. Oxidation rates of each organic substrate were measured through time. There was a substantial increase in the number of organic substrates oxidized by diatoms grown in the dark compared to their light-grown counterparts, indicating that the transport systems for these molecules may be light activated. Therefore, diatoms likely only utilize these metabolically expensive uptake mechanisms when they are necessary for survival, or when substrates are plentiful. A principal components analysis indicated discernible differences in the types of organic-C substrates utilized by dark-grown diatoms and bacteria. Although bacteria were able to oxidize a more diverse array of organic substrates including carboxylic acids and large polymers, diatoms appeared to more readily utilize the complex carbohydrates. By oxidizing different organic substrates than bacteria, heterotrophically metabolizing diatoms may be reducing direct competition and enhancing coexistence with bacteria.

Introduction Photosynthesis is the primary means by which algae sequester carbon for cell maintenance and growth. However, algal cells often inhabit environments where levels of solar radiation are too low to support autotrophic metabolism. For example, algae are often located below polar ice sheets (Palmisano et al., 1985), buried in sediments (Wasmund, 1987), or shaded at the base of thick periphyton communities (Jorgenson et al., 1983; Burkholder et al., 1990; Liehr et al., 1990; Dodds, 1992; Tuchman, 1996; Johnson et al., 1997). If algae in irradiance-limited habitats are not able to

sequester light by other means such as stalk formation or motility (Johnson et al., 1997), they may resort to either entering a state of dormancy or utilizing heterotrophic metabolism (e.g., Stadelmann, 1962; Tuchman, 1996; Zhang et al., 2003). Under light limited conditions, many algae are capable of heterotrophically metabolizing a diverse range of organic carbon sources including pyruvates, acetate, lactate, ethanol, saturated fatty acids, glycollate, glycerol, hexoses, and amino acids (e.g., Parker et al., 1961; Neilson & Lewin, 1974). In nature, it has been demonstrated that in polar regions, for example, algae that encounter extended periods of low irradiance due to shading

168 by snow and ice can be metabolically active, incorporating amino acids, glucose, and other exogenous carbon sources (Palmisano et al., 1985; Rivkin & Putt, 1987). Active transport appears to be the primary means by which algae acquire organic carbon substrates from the environment. Although passive diffusion of exogenous organic carbon is possible (Stadelmann, 1962), concentrations of organic sources are likely too low in most natural environments to sustain efficient passive uptake. Active transport of carbon sources has been documented in several species of diatoms (Hellebust & Lewin, 1977), and is likely the most common mechanism for sequestering these compounds. These active transport mechanisms appear to be regulated by irradiance in many species of algae. For example, when irradiance is high enough to activate photosynthesis, the uptake of glucose by Cyclotella cryptica was minimized, while in the dark, glucose uptake was maximized (Hellebust, 1971). However, photoinhibition of heterotrophic metabolism does not always occur. Phototrophic (photoassimilating) cells have been shown to incorporate a selective range of organic carbon and nitrogenous compounds at low irradiance for storage products as a means of subsidizing their photosynthetic metabolism (Allison et al., 1953; Zotina et al., 2003). However, this ‘luxury consumption’ is generally less common and uptake rates are lower than by dark-grown algal cells. In certain species, nevertheless, it may contribute a significant portion of the carbon and nitrogen budgets of algae growing under low irradiance (e.g., Zotina et al., 2003). The proximal relationship between algae and bacteria, particularly in benthic assemblages, may lead to competition between heterotrophically metabolizing algae and bacteria for organic substrates. In situations where heterotrophic algae and bacteria are competing for organic carbon, it has been assumed that the more numerically abundant bacteria with higher metabolism and a greater surface area to volume ratio would outcompete algae for most organic carbon substrates (Stewart, 1974). It is therefore possible that heterotrophically metabolizing algae utilize different substrates than bacteria in an attempt to reduce the effects of competition.

Although there are numerous examples of algal heterotrophy, the heterotrophic utilization of a large number of different organic carbon sources by algae has not been investigated under a single set of controlled conditions for an array of species. The purpose of this study was to survey the ability of freshwater benthic diatoms to sequester and utilize 95 naturally occurring organic carbon compounds. Uptake of these substrates was compared between diatoms grown in the light and those grown in total darkness to determine if uptake mechanisms were light dependent. In addition, diatom substrate utilization was also compared to that of five species of aquatic bacteria to determine if substrate use differed between these two groups of organisms, which could indicate niche separation as a mechanism by which they avoid competition.

Materials and methods Maintaining cultures Eight species of freshwater benthic diatoms were selected for study based on their diverse autecologies. Achnanthidium minutissimum (Kutz.) Czar. and Achnanthidium rostratum Ostr. are adnate, monoraphid, non-motile benthic diatoms that are commonly found at the base of periphyton mats in streams (Steinman et al., 1987; Tuchman & Stevenson, 1991). Encyonema minutum (Hilse in Rabh.) D. G. Mann, Encyonema minutum var. pseudogracilis (Choln.) Czar., and Gomphonema accuminatum Ehrenb. are biraphid, stalked benthic diatoms (Hudon & Bourget, 1981; McCormick & Stevenson, 1991). Navicula trivialis L.-Bert, Nitzschia linearis (Ag. Ex W. Sm.) W. Sm., and Nitzschia palea (Kutz) W. Sm., are biraphid and capable of motility (Bertrand, 1992). Cultures of diatoms were obtained from the live diatom herbarium of Loras College (Dubuque, IA). Even though the algal cultures were not axenic, bacterial content of these cultures prior to treatment was determined to be negligible using an assay that incorporated an antibiotic protocol (Schollett, 1998). In addition, airborne bacterial contamination was also accounted for (see below) so that heterotrophy could be assumed to be of algal origin. Each culture was divided into eight

169 125 ml Erlenmeyer flasks containing 75 ml of sterile Alga-Gro inorganic nutrient media (Carolina Biological Supply Co., Burlington, NC). Flasks were stopped with sterile gauze plugs and placed into a Percival environmental chamber on a shaker table and maintained at 15 C for 10 days under one of two illumination conditions: (1) 8 h of 250 lmol quanta m)2 s)1 and 14 h darkness (light cultures) or, (2) 24 h of 3 lmol quanta m)2 s)1 (dark cultures). Five aquatic bacterial species, Aeromonas sobria Popoff and Veron, Aquaspirillum itersoniii (Giesberger) Hyleman, Aquaspirillum serpens (Muller) Hylemon et al., Aquaspirillum sinuosum (Williams and Rittenberg) Hylemon et al., and Bacillus cereus Frankland and Frankland (Carolina Biological Supply Co., Burlington, NC), were selected for study based on their abundance in natural aquatic habitats. Individual strains were received on agar plates and transferred to 75 ml of sterile, BBL inorganic nutrient media (Benton Dickinson Microbiology Systems, Cockeysville, MD) in autoclaved 125 ml Erlenmeyer flasks (eight replicates), stopped with sterile gauze and placed in a Fisher Science Isotemp Chamber at 23 C with illumination levels of 3 lmol quanta m)2 s)1 to allow for cell growth. Cultures were manually shaken every 12 h for 10 days to circulate media and redistribute cells. Every 48 h, 50% of the culture media in each flask was pipetted from the surface and exchanged with fresh media to facilitate nutrient replenishment and waste removal. Determining metabolic activity of diatoms and bacteria To quantify the metabolism of exogenous organic carbon sources by bacteria, dark-grown diatoms, and light-grown diatoms, Biolog Microtiter plates (Biolog Inc., Haywood, CA) were inoculated with bacteria or diatom samples from each illumination treatment. Biolog Microtiter plates were used to test the ability of the microorganisms to utilize a wide range of different carbon sources that occur naturally in aquatic ecosystems. A plate consisted of 96 wells; one well contained water and served as a control for airborne bacterial contamination and the remaining 95 were pre-filled with equal molar amounts of

different organic carbon substrates (Table 1). Each well was inoculated with 150 ll of microorganism culture in saline suspension. The culture suspensions were prepared by centrifuging the cultures, decanting the inorganic nutrient media, so as not to add additional substrates into each well, and resuspending cells in 0.85% saline solution. The density of cells in these suspensions was standardized among samples by diluting with 0.85% saline solution or concentrating by centrifugation until a 55% transmittance level at 680 nm was obtained on a Spectronic GenesysTM 2 spectrophotometer. To compare biomass-specific oxidation rates among diatoms and bacteria, each individual aliquot of sample was adjusted to a uniform biovolume. In each Microtiter well, density of each species of bacteria and algae used to inoculate the well was quantified and standardized to an equivalent biovolume. Algal cell densities were determined for each species by preparing microscopic slides of cultures (densities that were spectrophotometerically quantified to 55% transmittance at 680 nm) and enumerating at least 300 cells at 1000X magnification. Biovolume of each algal species was estimated by measuring 10 random cells per slide and applying the appropriate geometric equation. Bacterial biovolumes were obtained from the literature (Pelczar et al., 1977). Observed oxidation slopes were then divided by biovolume per well to produce biovolume-adjusted oxidation rates. Following inoculation, plates were incubated for 12 days under respective light and temperature conditions in which the cultures were originally maintained. If the microorganisms oxidized the carbon source, an associated tetrazolium dye produced a color change that was quantified by a Titertek Multiskan Plus turbidimeter. Readings of each plate were made at 4, 12 and then every 24 h for 12 days, after which no further increase in oxidation was observed in any well. Oxidation rates reported here were measured at day 12. Data analysis Oxidation rates for each carbon source were calculated from the slope of each oxidation curve and corrected for values obtained in the control wells.

170 Table 1. Substrates tested in Biolog GN Microtiter Plates. Bold headings represent chemical families Carbohydrates

Carboxylic acids

Amino acids

Aromatics

Adonitol

Acetic acid

D,L-carnitine

Inosine

Alpha-D-glucose

Alpha-hydroxybutyric acid

D-alanine

Thymidine

Alpha-D-lactose Beta-methyl D-glucoside

Alpha-keto butyric acid Alpha-keto glutaric acid

D-serine

Uridine Urocanic acid

Gamma-amino butyric acid

Cellobiose

Alpha-keto valeric acid

Glycyl-L-aspartic acid

D-arabitol

Beta-hydroxybutyric acid

Glycyl-L-glutamic acid

Alcohols

D-fructose

Bromo succinic acid

Hydroxy L-proline

2,3-butanediol

D-galactose

cis-aconitic acid

L-alanine

2-amino ethanol

D-mannitol

Citric acid

L-alanyl-glycine

Glycerol

D-mannose

D,L-lactic

L-aspartic

D-melibiose

D-galactonic

acid

acid

L-asparagine

D-psicose

acid lactone D-galacturonic acid

D-raffinose

D-gluconic

L-histidine

D-sorbitol

D-glucosaminic

acid acid

Putrescine

L-leucine

D-trehalose

D-glucuronic D-saccharic

i-erythritol

Formic acid

L-proline

L-arabinose

Gamma-hydroxybutyric acid

L-pyroglutamic

L-fucose

Itaconic acid Malonic acid

L-serine

acid

Amines Phenyl ethylamine

acid

Gentiobiose

L-rhamnose

acid

L-glutamic

L-ornithine

Esters

L-phenylalanine

Methyl pyruvate Mono-methyl succinate acid

L-threonine

Phosphorylated hydrocarbons Glucose-1-phosphate D,L-Alpha

Lactulose

p-hydroxy phenylacetic acid

m-inositol

Propionic acid

Amides

Glucose-6-phosphate

Maltose

Quinic acid

Alaninamide

N-acetyl-D-galactosamine

Sebacic acid

Glucuronamide

Polymers

N-acetyl-D-glucosamine

Succinic acid

Succinamic acid

Alpha-cyclodextrin

Sucrose

Glycogen

Turanose Xylitol

Tween 40 Tween 80

This correction eliminated any plate-wide contamination effects from airborne bacteria. Mean oxidation rates and the percentage of all 95 substrates that were utilized were calculated for each chemical group and compared between diatoms (each species used as a replicate, n=8) grown in the dark and light using a series of paired t-tests (1-tailed, H0: dark £ light). Mean oxidation rates for each chemical family by dark-grown diatoms (n=8) were compared to bacteria (each species used as a replicate, n=5) using t-tests (Systat version 9, SPSS Inc., Chicago, IL). Patterns in bacterial substrate utilization were compared to dark grown diatom substrate utilization using principal components analysis (Systat version 9, SPSS Inc., Chicago, IL). All data were

glycerol phosphate

first LOG10 (x+1) transformed. The analysis was performed on the correlation matrix with no rotations. An ordination plot was created using the first two principal component axes, and t-tests were performed between bacteria and algae using component scores from each axis.

Results Light- and dark-grown diatoms Greater numbers of substrates were utilized by each diatom species when grown in the dark than the light (Fig. 1, Appendix), with an average of 68% utilized in the light to an average of 94% in

171 light dark

Relative number of substrates oxidized (%)

100

80

60

40

20

0 A.

r.

G.

a.

N.

p.

m. m.p. E. E.

N.

t.

l. N.

m. A.

Figure 1. Relative number (%) of 95 substrates utilized by 8 diatom species grown under both light and dark conditions. A.r.=Achnanthidium rostratum, G.a.=Gomphonema accuminatum, N.p.=Nitzschia palea, E.m.=Encyonema minutum, E.m.p.=Encyonema minutum var. pseudogracilis, N.t.=Navicula trivialis, N.l.=Nitzschia linearis, A.m.=Achnanthidium minutissimum.

the dark (paired t-test, p<0.001). The number of substrates used in the light ranged from 81% by Nitzschia palea to 56% by Achnanthidium minutissimum. In the dark, the number of substrates utilized ranged from 99% by Encyonema minutum var. pseudogracilis and Gomphonema accuminatum to 90% used by Encyonema minutum and Navicula trivialis. The percentage of total substrates utilized increased significantly (paired t-test, p<0.05) in the dark vs. the light treatments for each individual chemical family (Fig. 2). This increase was most pronounced in the amides, where the number of substrates oxidized increased from 0 to 100%. The rate of oxidation of most substrates by diatoms also increased in the dark (Fig. 3). Oxidation rates were significantly higher in the dark than in the light for amides (p=0.005), amines (p=0.011), amino acids (p=0.002) aromatics (p=0.002), carboxylic acids (p=0.048), esters (p=0.001), phosphorylated hydrocarbons (p=0.003) and polymers (p=0.006). They did not differ significantly between light and dark for alcohols (p=0.438) or carbohydrates (p=0.447).

Differences in bacterial and dark-grown diatom substrate utilization Bacteria and diatoms appeared to differ in their use of the different organic substrates. Unlike diatoms, each bacterial species utilized all 95 of the organic substrates. Differences between diatoms and bacteria were primarily due to oxidation rates of different compounds. At the chemical family level, only polymers were significantly different (p=0.012) between dark-grown diatoms and bacteria (Fig. 3). Bacteria oxidized polymers at twice the rate of diatoms. The results of the principal components analysis indicate that 36% of the variance in the data could be explained by the first component (PC1), which was heavily weighted by the utilization of amino acids. The second axis explained 22% of the variance. The ordination plot indicated that bacteria and diatoms were separated distinctly along PC2, but not along PC1 (Fig. 4). Principal component scores did not differ significantly between bacteria and algae with respect to PC1 (p=0.500), but did differ significantly with respect to PC2 (p<0.001).

172

100

100

80

80

60

60

40

40

20

20

0

0

100

Relative number of substrates oxidized (%)

*

Alcohols

*

Amines

100

80

80

60

60

40

40

20

20

0 100

*

100

80

80

60

60

40

40

20

20

0

80

*

100

*

Aromatics

*

Esters

*

Polymers

*

80 60

40

40

20

20

0

80

Amino Acids

0

Carboxylic Acids

60

100

*

0

Carbohydrates

100

Amides

0

Phosph. Hydro.

*

100 80

60

60

40

40

20

20

0

0

Light

Dark

Light

Dark

Figure 2. Number of substrates (mean % ±1 SE) in each chemical family oxidized by diatoms (n=8) grown in the light and dark measured over 12 days. Asterisks represent pairs of means that are significantly different (p<0.05).

The ordination plot indicates that diatoms primarily used substrates that were negatively correlated to PC2, while those that were positively correlated were primarily utilized by bacteria. Substrates that were highly correlated with PC2 (r>Œ0.6 Œ) are listed in Table 2. Tween 80 (a polymer) had the highest positive correlation with PC2, while putrescine had the highest negative correlation. Substrates that were negatively correlated to PC2 and thereby more heavily utilized by diatoms were largely made up of carbohydrates (7 out of 11). Substrates that were

positively correlated to PC2 and thereby more widely used by bacteria consisted of a more even mixture of substrate families. However, polymers (1st and 3rd highest loadings) and carboxylic acids appeared to be more heavily utilized by bacteria.

Discussion The diatom species examined in this study all inhabit benthic habitats where they may often encounter light depletion due to frequent burial in

173 25

25

Alcohols

20

20

15

15

10

10

5

5

0 25

0 25

Amines

-1 6 -3 Oxidation rate (Δ absorption units h x 10 μg )

20 15 10

10

Aromatics

0 25

*

10

10

5

5

Carboxylic Acids

0 25

20

20

15

15

10

10

5

5

Phosphorylated Hydrocarbons

0 25

15

10

10

*

0

t a ar k igh er i s -l s-d Bact m m to to D ia Dia

Esters

* Polymers

20

15

5

Carbohydrates

20 15

20

*

5

15

0 25

Amino Acids

15

*

20

0 25

*

20

5 0 25

Amides

5

*

0

t a ar k ig h er i s -l s -d Bac t m m to to Dia D ia

Figure 3. Biomass specific oxidation rates (mean ±1 SE) for each chemical family by diatoms (n=8) grown in the light and dark for 12 days vs. bacteria (n=5), after 12 days. Asterisks represent pairs of means that are significantly different (p<0.05).

the sediments, highly turbid or stained overlying water, or from high density algal mats. Under such light-limited conditions these diatoms can activate mechanisms for uptake and metabolism of organic substrates as a survival strategy. In addition, they have the ability to turn off these metabolically costly uptake mechanisms when irradiance is adequate for photosynthesis. The capacity of a

microorganism to oxidize a compound is initially dependent on two conditions; first the organism must possess a transport system for that particular compound, and second, the environmental conditions must be appropriate for activation of the transport system (Amblard, 1991). The existence of transport systems for a few compounds has been well documented in various species of

174

8

Aquaspirillum serpens Bacillus cereus

6 4

Aeromonas sobria

Aquaspirillum itersoni

PC2 (22%)

Aquaspirillum sinuosum

2 0

Encyonema minutum var. pseudogracilis

-2

Gomphonema accuminatum

Nitzschia linearis

Navicula trivialis

-4 -6

Encyonema minutum

Achnanthidium minutissimum

Nitzschia palea Achnanthidium rostratum

-8 -12 -10 -8

-6

-4

-2

0

0

0

0

0

10

PC1 (36%) Bacteria Diatoms Figure 4. Ordination of 8 dark-grown diatom species and 5 bacterial species using the first two principal component axes (PC1 and PC2).

diatoms (Hellebust & Lewin, 1977). Hellebust & Lewin (1977) demonstrated the specificity of these transport systems for an individual compound when they observed the requirement for a least three transport systems for amino acid uptake: one for acidic, one for basic, and one for neutral amino acids. The high specificity of these transport systems suggests that to use a wide range of compounds, an equally diverse series of transport systems must be operational. In addition to transport mechanisms, many environmental factors contribute to the regulation of an individual transport system, such as light, pH, temperature, and substrate concentration (Neilson & Lewin, 1974). The results of this study indicate that the eight benthic diatom species tested possess a diverse array of transport systems to allow utilization of a large number of natural substrates, and that many of these mechanisms are regulated in part by irradiance. Since Nitzschia palea was capable of oxidizing 81% of the organic substrates in the light, it is possible that, in addition to low irradiance, substrate concentrations are important for activating uptake mechanisms in this species. Nitzschia palea often thrives in habitats that are organically enriched, such as downstream of sewage outfalls (e.g.,

Michels, 1998; Gurbuz & Kivrak, 2003). It is likely that Nitszchia palea is able to supplement carbon fixation with heterotrophy in high irradiances when organic substrate concentrations are high. Although some of the compounds examined in this study do not likely occur in nature in the concentrations or chemical forms used in the experiments, the diversity of substrate oxidation demonstrated in this study suggests that diatoms under dark conditions can likely utilize similar forms of these compounds. Comparable forms of these compounds and their derivatives are present in aquatic ecosystems and may be introduced into the environment from a number of natural sources, including decomposition, allochthonous inputs, autochthonous inputs, animal or plant excretions, or UV-B photodegradation of DOC (Wetzel et al., 1995; De Lange et al., 2003). In addition to using these compounds as a source of carbon, diatoms might utilize amides, amines, and amino acids as a source of nitrogen as well. Diatoms readily take up dissolved organic nitrogen in both benthic (e.g., Nilsson & Sundba¨ck, 1996) and planktonic habitats (see reviews by Anitia et al., 1969; Flynn & Butler, 1986; Paerl, 1991). For example, Nilsson & Sundba¨ck (1996) demonstrated that a number of benthic algae,

175 Table 2. Correlations of individual substrates to the second principal component axis (PC2). Only substrates with correlations>Œ0.6 Œ were included. Positive values represent substrates that were more intensively used by bacteria and negative values represent substrates that were more intensively used by diatoms Substrate

Correlation with PC2

Chemical family

Tween 80

0.907

Polymers

L-histidine

0.870

Amino acids

Dextrin

0.840

Polymers

L-serine

0.825

Amino acids

Urocanic acid

0.783

Aromatics

D,L-latic

acid

Mono-methyl succinate Alpha-hydroxybutyric acid

0.774

Carboxylic acid

0.772 0.710

Esters Carboxylic acid

Glycyl-L-aspartic acid

0.688

Amino acids

Citric acid

0.676

Carboxylic acid

Alpha-D-glucose

0.666

Carbohydrates

L-threonine

0.645

Amino acids

D-mannose

0.637

Carbohydrates

cis-aconitic acid

0.634

Carboxylic acid

Sucrose Bromo succinic acid

0.626 0.626

Carbohydrates Carboxylic acid

D-glucosaminic

acid

0.601

Carboxylic acid

Putrescine

)0.788

Amines

D,L-carnitine

)0.773

Amino acids

D-raffinose

)0.773

Carbohydrates

Gentiobiose

)0.693

Carbohydrates

L-fucose

)0.690

Carbohydrates

D-melibiose

)0.658 )0.653

Carbohydrates Carbohydrates

m-inositol D-glucuronic

acid

Alpha-D-lactose D,L-Alpha

Lactulose

glycerol phosphate

)0.652

Carboxylic acid

)0.625

Carbohydrates

)0.613

Phosphorylated hydrocarbons

)0.611

Carbohydrates

including diatoms, could utilize free amino acids. Similarly, Liu & Hellebust (1973) reported that algal cells exhibit growth on glutamine and arginine at rates equal to that of nitrate, and ornithine, asparagine, glycine, alanine, and aspartate can also support growth, although at lower rates than that of inorganic nitrogen. Although diatom and bacterial species were tested in monocultures under conditions which do not reflect the potential competitive environment that would be encountered in nature, discernible differences in substrate utilization patterns between diatoms and bacteria may reflect specializations that have evolved allowing these two groups of organisms to coexist. For example,

diatoms and bacteria most effectively utilized different organic compounds. The three carbohydrates that were utilized more by bacteria were simple, while diatoms appeared to use more complex carbohydrates. Of the 11 carbon sources that were utilized more by diatoms (high negative correlation to PC2), seven were complex carbohydrates. The two chemical families that appeared more highly utilized by bacteria were polymers and carboxylic acids. Although a strong potential exists for competition between bacteria and heterotrophically metabolizing diatoms for organic substrates in nature, it is also possible that the interaction between these two groups of organisms is largely

176 mutualistic. Algae can exude organic compounds that can be assimilated by bacteria (e.g., Petit et al., 1999; Descy et al., 2002; Puddu et al., 2003), and bacteria can make both organic and inorganic carbon available to algae (e.g., Klug, 2005). Algae and bacteria may exchange trace organic substances such as vitamins (Cole, 1982), and bacteria may also release exoenzymes that degrade complex organic molecules (e.g., Sinsabaugh & Linkins, 1988) that are not transportable across cell membranes. Indeed, in this study we found that growing diatoms in cultures that contained bacteria was much more successful than growing axenic diatom cultures. This study demonstrates several important points: (1) that eight species of benthic diatoms have the ability to sequester and utilize a wide range of organic substrates, (2) that uptake mechanisms are largely light activated, and (3) that diatoms likely sequester a different set of organic substrates than bacteria to possibly reduce competition. These ideas need to be expanded to include organic substrate utilization under natural conditions, which would include monitoring multispecies assemblages and a range of natural organic and inorganic substrate concentrations. However, the information gained here suggests that freshwater benthic diatoms may be much more metabolically active under light-limiting conditions than was previously acknowledged, which may begin to explain, in part, how diatoms that are 10 cm below the sediment/water interface can be metabolically active.

Acknowledgements We would like to thank John Smarrelli and Martin Berg for advice and technical assistance and David Czarnecki for live diatom cultures from the Loras College herbarium. We would also like to thank Christopher Peterson for computer programming assistance and Robert Genter for advice on use of the Microtiter plates. References Allison, R. K., H. E. Skipper, M. R. Reid, W. A. Short & G. L. Hogan, 1953. Studies on the photosynthetic reaction I. The

assimilation of acetate by Nostoc muscorum. Journal of Biological Chemistry 204: 197–205. Amblard, C., 1991. Carbon heterotrophic activity of microalgae and cyanobacteria – ecological significance. Annee Biologique 30: 6–107. Anitia, N. J., J. Y. Cheng & F. J. R. Taylor, 1969. The heterotrophic growth of a marine photosynthetic cryptomonad (Chroomonas saline). Proceedings of the International Seaweed Symposium 6: 17–29. Bertrand, J, 1992. Mouvements des diatomees II -synthese des mouvements. Cryptogamie: Algologie 13: 49–71. Burkholder, J. M., R. G. Wetzel & K. L. Klomparens, 1990. Direct comparison of phosphate uptake by adnate and loosely attached microalgae within an intact biofilm matrix. Applied Environmental Microbiology 56: 2882–2890. Cole, J. J., 1982. Interactions between bacteria and algae in aquatic ecosystems. Annual Review of Ecology and Systematics 13: 291–314. De Lange, H. J., D. P. Morris & C. E. Williamson, 2003. Solar ultraviolet photodegradation of DOC may stimulate freshwater food webs. Journal of Plankton Research 25: 111–117. Descy, J. P., B. Leporcq, L. Viroux, C. Francois & P. Servais, 2002. Phytoplankton production, exudation, and bacterial reassimilation in the River Meuse (Belgium). Journal of Plankton Research 24: 161–166. Dodds, W. K., 1992. A modified fiber-optic light microprobe to measure spherically integrated photosynthetic photon flux density: Characteristics of periphyton photosynthesisirradiance patterns. Limnology and Oceanography 37: 871– 878. Flynn, K. J. & T. Butler, 1986. Nitrogen sources for growth of marine microalgae: Role of dissolved free amino acids. Marine Ecology Progress Series 34: 281–304. Gurbuz, H. & E. Kivrak, 2003. Seasonal variations of benthic algae of Kuzgun Dam Reservoir and their relationship to environmental factors. Fresenius Environmental Bulleting 12: 1025–1032. Hellebust, J. A., 1971. Glucose uptake by Cyclotella cryptica: Dark induction and light inactivation of transport system. Journal of Phycology 7: 345–349. Hellebust, J. A. & J. Lewin, 1977. Heterotrophic nutrition. In Werner, (ed.) The Biology of Diatoms. University of California Press, Berkeley, 169–197. Hudon, C. & E. Bourget, 1981. Initial colonization of artificial substrate: community development and structure studied by scanning electron microscopy. Canadian Journal of Fisheries and Aquatic Sciences 38: 1371–1384. Johnson, R. E., N. C. Tuchman & C. G. Peterson, 1997. Changes in the vertical microdistribution of diatoms within a developing periphyton mat. Journal of the North American Benthological Society 16: 503–519. Jorgenson, B. B., N. P. Revsbech & Y. Cohen, 1983. Photosynthesis and structure of benthic microbial mats: Microelectrode and SEM studies of four cyanobacterial communities. Limnology and Oceanography 28: 1075– 1093. Klug, J. L., 2005. Bacterial response to dissolved organic matter affects resource availability for algae. Canadian Journal of Fisheries and Aquatic Sciences 62: 472–481.

177 Liehr, S. K., M. T. Suidan & J. W. Eheart, 1990. A modeling study of carbon and light limitation in algal biofilms. Biotechnology and Bioengineering 35: 233–243. Liu, M. S. & J. A. Hellebust, 1973. Utilization of amino acids as nitrogen sources, and their effects on nitrate reductase in the marine diatom Cyclotella cryptica. Canadian Journal of Microbiology 20: 1119–1124. McCormick, P. V. & R. J. Stevenson, 1991. Mechanisms of benthic algal succession in lotic environments. Ecology 72: 1835–1848. Michels, A., 1998. Effects of sewage water on diatoms (Bacillariophyceae) and water quality in two tropical streams in Costa Rica. Revista de Biologı´ a Tropical 46: 153–175. Neilson, A. H. & R. A. Lewin, 1974. The uptake and utilization of organic carbon by algae; an essay in comparative biochemistry. Phycologia 13: 227–264. Nilsson, C. & K. Sundba¨ck, 1996. Amino acid uptake in natural microphytobenthic assemblages studied by microautoradiography. Hydrobiologia 332: 119–129. Paerl, H. W., 1991. Ecophysiological and trophic implications of light-stimulated amino acid utilization in marine picoplankton. Applied Environmental Microbiology 57: 473– 479. Palmisano, A. C., J. B. SooHoo, D. C. White, G. A. Smith, G. R. Staton & L. H. Burckle, 1985. Shade adapted benthic diatoms beneath Antarctica sea ice. Journal of Phycology 21: 664–667. Parker, B. C., H. C. Bold & T. R. Deason, 1961. Facultative heterotrophy in some chlorococcacean algae. Science 133: 761–763. Pelczar, M. J., R. Reid & E. C. S. Chan, 1977. Microbiology. McGraw-Hill Book Company, New York, 952 pp. Petit, M., G. P. Alves & P. Lavandier, 1999. Phytoplankton exudation, bacterial reassimilation and production for three diel cycles in different trophic conditions. Archiv fu¨r Hydrobiologie 146: 285–309. Puddu, A., A. Zoppini, S. Fazi, M. Rosati, S. Amalfitano & E. Magaletti, 2003. Bacterial uptake of DOM released from Plimited phytoplankton. FEMS Microbial Ecology 46: 257– 268. Rivkin, R. B. & M. Putt, 1987. Heterotrophy and photoheterotrophy by Antarctic microalgae: Light-dependent incor-

poration of amino acids and glucose. Journal of Phycology 23: 442–452. Schollett, M. A., 1998 Organic nutrient preferences for benthic diatoms: An approach to quantifying heterotrophic metabolism. Master’s Thesis. Loyola University Chicago, Chicago, IL, 91 pp. Sinsabaugh, R. L. & A. E. Linkins, 1988. Exoenzyme activity associated with lotic epilithon. Freshwater Biology 20: 249– 261. Stadelmann, E. J., 1962. Permeability. In Lewin, (ed.) Physiology and Biochemistry of Algae. Academic Press, New York, 493–528. Steinman, A. D., C. D. McIntire, S. V. Gregory, G. A. Lamberti & L. R. Ashkenas, 1987. Effects of herbivore type and density on taxonomic structure and physiognomy of algal assemblages in laboratory streams. Journal of the North American Benthological Society 6: 175–188. Stewart, W. D. P., 1974. Algal Physiology and Biochemistry. University of California Press, Los Angeles, 989 pp. Tuchman, N. C., 1996. The role of heterotrophy in benthic algae. In Stevenson, R. J. M. Bothwell, & R. Lowe (eds.) Algal Ecology: Freshwater Benthic Habitats. Academic Press, San Diego, 299–319. Tuchman, N. C. & R. J. Stevenson, 1991. Effects of selective grazing by snails on benthic algal succession. Journal of the North American Benthological Society 10: 430–443. Wasmund, N., 1987. Live algae in deep sediment layers. Internationale Revue der Gesamten Hydrobiologie 74: 589– 597. Wetzel, R. G., P. G. Hatcher & T. S. Bianchi, 1995. Natural photolysis by ultraviolet irradiance of recalcitrant dissolved organic matter to simple substrates for rapid bacterial metabolism. Limnology and Oceanography 40: 1369–1380. Zhang, Q., R. Gradinger & Q. S. Zhou, 2003. Competition within the marine microalgae over the polar dark period in the Greenland Sea of high Arctic. Acta Oceanologica Sinica 22: 233–242. Zotina, T., O. Koster & F. Juttner, 2003. Photoheterotrophy and light-dependent uptake of organic and organic nitrogenous compounds by Planktothrix rubescens under low irradiance. Freshwater Biology 48: 1859–1872.

Hydrobiologia (2006) 561:179–189  Springer 2006 R.J. Stevenson, Y. Pan, J.P. Kociolek & J.C. Kingston (eds), Advances in Algal Biology: A Commemoration of the Work of Rex Lowe DOI 10.1007/s10750-005-1613-3

Using diatom assemblages to assess urban stream conditions Christopher E. Walker1,2* & Yangdong Pan1 1

Environmental Sciences and Resources, Portland State University, Portland, OR, 97207, USA USGS, Columbia River Research Lab, Cook-Underwood Rd, 5501-AWA, 98605-9717, USA (*Author for correspondence: E-mail: [email protected]) 2

Key words: canonical correspondence analysis, TWINSPAN, urban-to-rural gradient, bioassessment, diatoms

Abstract We characterized changes in diatom assemblages along an urban-to-rural gradient to assess impacts of urbanization on stream conditions. Diatoms, water chemistry, and physical variables of riffles at 19 urban and 28 rural stream sites were sampled and assessed during the summer base flow period. Near stream land use was characterized using GIS. In addition, one urban and one rural site were sampled monthly throughout a year to assess temporal variation of diatom assemblages between the urban and rural stream sites. Canonical correspondence analysis (CCA) showed that the 1st ordination axis distinctly separated rural and urban sites. This axis was correlated with conductivity (r=0.75) and % near-stream commercial/ industrial land use (r=0.55). TWINSPAN classified all sites into four groups based on diatom assemblages. These diatom-based site groups were significantly different in water chemistry (e.g., conductivity, dissolved nutrients), physical habitat (e.g., % stream substrate as fines), and near-stream land use. CCA on the temporal diatom data set showed that diatom assemblages had high seasonal variation along the 2nd axis in both urban and rural sites, however, rural and urban sites were well separated along the 1st ordination axis. Our results suggest that changes in diatom assemblages respond to urban impacts on stream conditions.

Introduction Urban ecosystem dynamics are driven by natural ecological processes, human activities, and their interactions (Grimm et al., 2000). Examining changes in ecosystem function and structure along an urban-to-rural gradient may identify predictable responses to different intensities of urbanization and generate hypotheses for these changes (McDonnell & Pickett, 1990). Urbanization in watersheds alters stream flow, riparian structure, and physical habitats (Leopold, 1968; Booth & Jackson, 1997; Finkenbine et al., 2000; Bledsoe & Watson, 2001). Cumulative impacts of urbanization on stream ecosystems may be best reflected by resident biota. Benthic biota such as macroinvertebrates and periphyton can integrate

effects of multiple environmental stressors over time (Karr & Chu, 1999; Stevenson & Pan, 1999). Sonneman et al. (2001) examined changes in diatom community composition related to urban intensity and found that they related better to water quality, while macroinvertebrates were better indicators of catchment disturbance. Using diatoms as bioindicators has a long history (see review by Stevenson & Pan, 1999), but the potential for using diatoms as a bioassessment tool has not been fully realized. A recent survey showed that even though many US state agencies are interested, benthic diatoms have not been widely used by state water quality programs (Kroeger et al., 1999). The metropolitan area of Portland, Oregon is an ideal place to study urban ecosystems. Unlike

180 many metropolitan cities that have developed uncontrollably, growth in the Portland area has been managed in a way to prevent urban sprawl. An Urban Growth Boundary (UGB) around the metropolitan area was enacted in 1972 to constrain urban development within the boundary (Metro, 1997). This boundary allows the comparison of highly urbanized watersheds to rural watersheds with similar characteristics immediately outside the boundary. This study was designed to examine whether diatoms could be used to assess urban stream conditions. The first objective of this study was to examine changes in benthic diatom assemblages along an urban-to-rural gradient and relate those changes to environmental factors. Second, we wanted to assess temporal variability of diatom assemblages between an urban and a rural stream site.

derived from ancient volcanic activity and flood deposits (Swanson et al., 1993). Johnson Creek has the common characteristics of many urban streams, severe alteration by channelization, storm water inputs, removal of riparian vegetation, increases in surrounding impervious surface area, and industrial discharges. As a result, it has poor water quality, habitat quality, and dwindling fish populations (Abrams & Prescott, 1999). It is a free-flowing stream beginning in rural areas outside the UGB and continues into the suburbs, and eventually flows into commercial and industrial areas. The Clear Creek watershed has mixed land use associated with forest, agriculture, and rural residential living. Land use within the Deep Creek watershed is similar to Clear Creek. Field sampling

Materials and methods Study area Sites were sampled from Clear Creek, Deep Creek, and Johnson Creek, located in the plains and foothills of the Willamette Valley Ecoregion (Clarke et al., 1991; Fig. 1). The plain, approximately 170 km long and 70 km wide, is a trough with modest relief between the Coastal Range in the west and the Cascades in the east (Uhrich & Wentz, 1999). Land use and cover in the ecoregion are predominately agriculture with some forest and urban area (Clarke et al., 1991). The proximity of the Willamette Valley to the Pacific Ocean and the prevailing weather patterns produce a temperature regime characterized by cool wet winters and warm dry summers. Annual precipitation ranged from 102 to 127 cm from 1961 to 1990 (Uhrich & Wentz, 1999). Most of the precipitation (@75%) occurs from October through March with <5% occurring during July and August (Uhrich & Wentz, 1999). These three watersheds were selected across the UGB. Johnson Creek flows into a highly urbanized area within the UGB, both Clear Creek and Deep Creek are located immediately outside the UGB. The geologic characteristics in these three watersheds are similar. All watersheds have geology characterized by sandstone and alluvium

A total of 45 sites were sampled for physical, chemical, and biological variables during late August through early September of 1999. Of 45 sites, 25 were in Johnson Creek, 12 in Clear Creek, and 8 in Deep Creek. It was assumed that observations at these sites were spatially independent. Stevenson (1984) found that variability in species composition among adjacent sites was no different than among random sites. In addition, one downstream site in Johnson Creek located inside the UGB (‘urban’) and one in Clear Creek located outside the UGB (‘rural’) were selected for sampling throughout the year. These sites were sampled at least monthly for diatoms and water chemistry. Stream discharge was measured continuously in the lower urban Johnson Creek site by a United States Geological Survey (USGS) gaging station located just above the urban site. Riffle habitats ranging from 5 to 20 m in length were defined as the sampling unit. Five crossstream transects were set up in each riffle by dividing the riffle into four equal length intervals. Only three transects were used when the length of available riffle habitat was limited. Stream physical habitat was characterized by assessing channel morphology, substrate composition, riparian conditions, and discharge. Thalweg depth and channel width were measured at each transect. Channel gradient was measured for the entire sampling area using the Suunto PM-5/360 PC clinometer. Stream

181

Figure 1. Study area showing three watersheds (Johnson, Clear, and Deep Creek) and sampling sites. The downstream portion of the Johnson Creek Watershed is inside the UGB which is of the Portland Metropolitan Area located west of the boundary.

discharge was measured using a Swoffer flow meter. The percent of fine sediment (<2 mm) at each site was assessed by randomly placing a grid on the streambed at 20 locations and counting intersections (Torquemada & Platts, 1989). Two water samples were taken for nutrient analyses and stored on ice during sampling. One sample was unfiltered; and the other was filtered through 47 mm Millipore type HA filters (0.45 lm pore size) using a Nalgene hand pump

filtration unit and frozen until analysis. Conductivity normalized for temperature at 25 C, dissolved oxygen (DO), and stream temperature were measured using the YSI Model 85 meter. Turbidity was measured using the HACH Model 2100P Turbidimeter and pH was measured using the Orion Model 210A pH meter. Diatoms were sampled by selecting two rocks, sizes ranging from coarse gravel to cobble, at random from each transect (a total of 10 rocks per

182 site). Diatoms were then scraped from a known area of rock using a toothbrush and a rubber delimiter and combined into 1 composite sample per site.

conditions have a strong influence on biological assemblages (Carter et al., 1996; Pan et al., 1996; Richards et al., 1997). Data analysis

Laboratory analysis Diatom samples were digested with concentrated sulfuric acid and potassium dichromate, rinsed with deionized water repeatedly until pH was approximately neutral, then mounted on slides using Naphrax high resolution mounting medium. Using a Nikon Eclipse E600 microscope at 1000 magnification, transects on slides were scanned until 500 diatom valves were enumerated and identified to the species level. Patrick & Reimer (1966, 1975) and Krammer & Lange-Bertalot (1986, 1988, 1991a, b) were used as primary references for diatom taxonomy. Water samples were analyzed for nitrate and nitrite by ion chromatography and colorimetric methods (EPA methods 300.0, 1979, and 353.2, 1993). These two constituent’s values were added together for this study (NO3+NO2). Soluble reactive phosphorus (SRP) concentrations were determined using colorimetric methods (EPA method 365.1, 1993). Total phosphorus (TP) concentrations were determined using persulfate digestion and colorimetric methods (EPA method 365.1, 1993). Near-stream land use characterization Land use for each site was characterized by using a geographic information system (GIS). The program ArcView (ESRI, 1997) was used for displaying and analyzing spatial coverage of maps and their corresponding databases created by Metro (1999). Sonoda et al. (2001) used GIS analyses to characterize near-stream land use and the relationship with water chemistry in Johnson Creek. They found that nutrients correlated well with near-stream land use characterized within a 30, 91, and 152 m radius around each site. A similar approach was taken but only a 91 m radius was used for this analysis. We realize that stream ecosystems are continuous and that conditions at a particular site on a stream reflect the cumulative inputs upstream in the watershed, however studies have also shown that local environmental

Diatom assemblages and their relation to environmental variables were examined using canonical correspondence analysis (CCA). A separate CCA was performed on temporal and spatial data. All environmental variables, except land use percentage data and pH, were log10 transformed to ‘normalize’ their distribution prior to the analysis. Species proportions of assemblages were transformed by first taking the square root then their arc sine (thus arc-sine square root transformed). Monte Carlo permutation tests were used to select a set of environmental variables that relate best with species assemblages (ter Braak & Smilauer, 1998). A few variables that were not selected by this procedure, but were important to this study such as urban land use, were also included in the final analysis. Unrestricted global Monte Carlo permutation tests were used to test the significance of the first two CCA axes (999 permutations). Species with relative abundance <1% were excluded in the analysis. CCA was performed using the computer software CANOCO for windows (v. 4) (ter Braak & Smilauer, 1998). Sites were classified into site groups based on diatom assemblages using TWINSPAN (two-way indicator species analysis) (Hill et al., 1975). Univariate analyses (ANOVA, or Kruskal-Wallis, and multiple comparison Student–Newman–Keuls (SNK) test) were used to test differences among classification site groups for environmental variables and land use (Zar, 1999). A paired t-test was used to test differences between the urban and rural sites throughout the year.

Results Spatial patterns of diatom assemblages and relations to environmental variables A wide range of physical, chemical, and biological variables were observed. Average stream widths ranged from 1.3 to 17.7 m and mean thalweg depths ranged from 0.1 to 0.47 m. Conductivity

183 varied from 56 to 231 lS cm)1. Nutrient concentrations such as SRP and NO3+NO2 ranged from 0.02 to 0.37 and 0.12 to 5.68 mg l)1 respectively. Urban near-stream land use ranged from 0 to 100%. A total of 84 diatom species and varieties were identified. Diatom assemblages were dominated by Achnanthes pyrenaicum Hustedt (26%), Cocconeis placentula Ehrenburg (14%), and Rhoicosphenia abbreviate Agardh (11%). CCA showed that the 1st two ordination axes accounted for 15.1% of variation in diatom species composition among sites (Fig. 2). Collectively, eight selected environmental variables explained 49.2% of the variation in diatom species distributions captured by the 1st two axes. The species-environmental correlations for the 1st two axes were high (r=0.81 for axis I and II). Monte Carlo permutation tests showed

that both axes were statistically significant (p<0.01). The 1st CCA axis may represent an urban-torural land-use gradient. Most of the rural sites were ordinated on the left side of the 1st axis based on species composition and their relation to measured environmental variables while urban sites were on the right side of the axis (Fig. 2). Several sites from Johnson Creek also ordinated on the left with other rural sites, but nearly all these Johnson Creek sites were headwater sites located outside the UGB. This axis was positively correlated with conductivity (r=0.75) and % near-stream commercial/industrial land use (r=0.55) (Table 1). Conductivity was highly correlated with % imperviousness of ‘catchment area upstream of each site’ (r=0.84), however GIS data for this variable was only available in the Johnson Creek

Figure 2. CCA ordination diagram showing the relation between diatom assemblages at each site and selected environmental variables (Clear Creek: dark circles, Deep Creek: light circles, squares=rural sites in Johnson Creek outside UGB, triangles=urban sites in Johnson Creek within UGB, N:P=dissolved inorganic nitrogen:dissolved inorganic phosphorus, Comm/Ind=commercial and industrial land use, Res.=residential land use, Cond.=conductivity).

184 Table 1. Correlation coefficients between selected environmental variables and the first two CCA axes CCA axes Variables

1

2

N:P* % Fines

0.37 0.08

)0.22 )0.41

Conductivity*

0.75

)0.20

Stream temperature*

0.38

)0.56

Stream Discharge*

0.24

0.58

% Residential

0.24

)0.15

% Commercial/industrial*

0.55

)0.04

)0.23

0.41

% Forest (*=Significant correlation p<0.05).

watershed, and therefore, was not used in this CCA. The 2nd axis may represent a stream longitudinal gradient. This axis correlated negatively with water temperature (r=)0.56) and positively with stream discharge (r=0.58). TWINSPAN classified all sites into four site groups based on diatom species composition (Fig. 3). Multiple comparison tests showed that relative abundance of A. pyrenaicum and F. pinnata Ehrenburg in sites of Group D were significantly different from sites in Groups A, B, and C (SNK, p<0.01). Relative abundance of C. placentula in Group A was significantly different from Groups B, C, and D (SNK, p<0.01). Relative abundance of R. abbreviate was significantly different between Groups A and D (SNK, p<0.05). For this species, Group B was also significantly different than Group D, as well as Group A being different than C (SNK, p<0.025). The species Nitzschia inconspicua Grunow, Achnanthidium minutissimum (Ku¨tz.) Czarnecki, N. fonticola Grunow, and Gomphonema parvulum Ku¨tzing were also compared among groups but were not significantly different. Group A comprised mostly rural sites. In contrast, sites in Groups C and D were mostly urban sites. Group B contained a mix of sites from all three watersheds. Water quality variables were significantly different among the site groups. Conductivity was significantly lower in Groups A and B than in Groups C and D (SNK, p<0.05) (Table 2). NO3+NO2 concentrations showed a similar pattern (SNK, p<0.05) (Table 2). Percent of substrate as fine sediment was lowest in Group A. The

amount of fine sediment for Group A was statistically different from Groups B and C (SNK, p<0.05). Group A had significantly more urban residential land use (ANOVA, p<0.05, p=0.001) than all other groups (SNK, p<0.05). Land use for agriculture, commercial/industrial, parks/open spaces, rural residential, and forest was not significantly different among groups. Stream discharge, pH, total phosphorus, and SRP did not differ among groups. Temporal patterns of diatom assemblages and relations to environmental variables Physical and chemical variables showed high seasonal variability at both urban and rural sites. Mean monthly discharge for Johnson Creek averaged 1.81 m3 s)1 and varied from 0.51 to 4.79 m3 s)1. Conductivity ranged from 121.5 to 216.0 lS cm)1 averaging 166.4 lS cm)1 in the wet season (Oct.–Apr.) and 196.8 lS cm)1 during the dry season (May–Sept.). Nutrient concentrations such as SRP and NO3+NO2 varied from 0.01 to 1.03 mg l)1 and 1.41 to 5.28 mg l)1, respectively. For Clear Creek, conductivity varied from 38.2 to 71.5 lS cm)1. Average conductivity was 51.9 lS cm)1 during the wet season and 63.3 lS cm)1 during the dry season. Concentrations of SRP ranged from 0.01 to 0.02 mg l)1 and NO3+NO2 varied between 0.10 to 1.11 mg l)1. Conductivity correlated well with nutrient measures such as N:P ratios (r=0.59). Physical, chemical, and biological variables were different between the urban and rural sites. Conductivity, SRP, and NO3+NO2 were greater at the urban site than the rural site throughout the year (Paired t-test, p<0.001) (Fig. 4). The species Planothidium lanceolatum Bre´bisson ex Ku¨tzing and C. placentula were common at both the urban and rural site. However, the diatom assemblages in each site were dominated by different species on average during the year. The urban site was uniquely comprised of Fragilaria pinnata Ehrenberg (20%) and the rural site was uniquely comprised of A. pyrenaicum (30%). The 1st two CCA axes explained about 28% of the variation for diatom assemblages at the urban and rural site throughout the year and showed distinguishable spatial and temporal differences. The species-environmental correlations for CCA

185

Figure 3. Common species, (a), and near-stream land use (b) that were significantly different (p<0.05) for groups classified based on diatom assemblages using TWINSPAN.

axes I and II were high (r=0.91 and 0.80, respectively). Collectively, the selected environmental variables explained 55.1% of the variation of diatom species composition captured by the 1st two axes. Monte Carlo permutation tests showed that both axes were statistically significant (p<0.01). The 1st axis may represent an urban-to-rural land use gradient. The rural sites with different sampling dates were ordinated on the left side of the 1st axis, while the urban sites were on the right side of the axis. This axis was correlated with conductivity (r=0.89), SRP (r=0.79), and NO3+NO2 (r=0.81). The 2nd axis was weakly correlated with stream temperature (r=0.39) and turbidity (r=)0.37). This axis may represent seasonal changes in stream conditions. Samples from

both urban and rural sites varied along the 2nd axis.

Discussion Spatial patterns of diatom assemblages along the urban-to-rural gradient Changes in diatom species distributions correlated well with the urban-to-rural gradient. Ordination analysis showed that diatom assemblages changed along the 1st CCA axis, which correlated most strongly with conductivity and % near-stream commercial/industrial land use. Our results are consistent with a previous study in the region (Carpenter & Waite, 2000) where conductivity

186 Table 2. Comparison of environmental variables (n=45) for TWINSPAN groups (mean and SD) Stream groups Variables

A

B

C

D

NO3+NO2* (mg l)1)

0.78 (1.13)

0.58 (0.49)

2.23 (1.81)

2.23 (2.20)

SRP (mg l)1)

0.06 (0.11)

0.03 (0.04)

0.05 (0.04)

0.05 (0.03)

Total P (mg l)1) Conductivity* (mS cm)1)

0.08 (0.12) 78.0 (30.5)

0.07 (0.04) 102.9 (36.5)

0.22 (0.46) 136.1 (52.1)

0.09 (0.02) 152.8 (52.1)

Temperature* (C)

13.3 (2.0)

pH Discharge % Fines*

15.1 (2.2)

15.4 (1.7)

15.2 (1.6)

7.37 (0.35)

7.35 (0.29)

7.32 (0.44)

7.44 (0.33)

0.36 (0.23)

0.41 (0.45)

0.09 (0.10)

0.45 (0.40)

24 (22)

24 (17)

19 (16)

7 (7)

(*=Significant difference, p<0.05, ANOVA).

distinguished urban and agricultural streams from forested sites in a study examining 25 streams in Oregon’s Willamette Valley. Conductivity, a measure of total dissolved ions in water, is largely a function of basin biogeochemistry and land use. Distinguishing effects of natural and anthropogenic sources of variability in stream conductivity is important for determining effects of the urban-to-rural gradient on diatom species composition. Increases in conductivity may be accompanied by elevated dissolved nutrients in streams. In this study, conductivity and nutrient measures correlated well. Leland (1995) also found that conductivity and nutrients correlated well with algal species distributions (r=)0.78) in Columbia Plateau streams of the Yakima Basin. Welch et al. (1998) suggested that conductivity might be a surrogate of urban development in the Pacific Northwest. Bryant (1995) found that base flow conductivity related well to impervious surface area in Puget Sound lowland streams, Washington (r2=0.83). Our data showed that conductivity correlated well with near-stream urban land use. Classification analysis was consistent with CCA in reflecting urban and rural conditions. For example, Group D included sites characterized by the highest average conductivity and urban land use and the diatom assemblage was dominated by R. abbreviate, a species classified as halophilous taxon (Lowe, 1974). Leland (1995) reported that this species was associated with waters of relatively higher ionic strength and occurred at levels of

higher conductivity in Yakima Basin streams (@391 lS cm)1 at 25 C). This species is also known to prefer stable habitats with high nutrient supply (Biggs et al., 1998). Biggs (1995) observed this species at sites in nutrient-rich watersheds during base flow conditions. Stream sites in Group D were characterized by the highest average conductivity and urban land use. In comparison, sites in Group A had the lowest conductivity and had more land cover as forest. The diatom assemblages in Group A sites were co-dominated by A. pyrenaicum and C. placentula, two species that both prefer waters of lower ionic strength (Lowe, 1974). A. pyrenaicum is thought to prefer low-nutrient conditions and was shown to decrease in abundance at stream sites highly polluted by effluent input from a sewage treatment plant (Klotz et al., 1976). Temporal patterns of diatom assemblages between the urban and rural sites Changes in diatom assemblages may result from both natural and anthropogenic factors (Pan et al., 1996). Diatom assemblages can change spatially and temporally. In this study, the 2nd CCA axis correlated well with discharge and temperature indicating longitudinal changes of diatom species. Species change may be related to stream size or other factors that change longitudinally from headwaters to mouth. Molloy (1992) found differences in algal biomass accumulation rates and benthic algal species assemblages longitudinally in

(a)

+3.0

187

Rural

Urban 1

1 2 8

Stream Temperature 5 4 3 6 7 SRP 2 16 Cond.

15 3 4

11 13 12 9 16 7 14 5 6 10

13 Turbidity

14

15 NO3 +NO2

9

8

10

11

-3.0

12

Date

A. pyrenaicum

20 8/31/00

6/30/00

0 4/30/00

8/26/00

6/26/00

4/26/00

2/26/00

12/26/99

10/26/99

0

C. placentula

40

2/29/00

20

P. lanceolatum

60

12/31/99

40

80

10/31/99

P. lanceolatum C. placentula F. pinnata

60

100

8/31/99

80

Relative abundance (%)

+3.0

100

8/26/99

(b)

Relative abundance (%)

-3.0

Date

Figure 4. CCA ordination diagram (a) showing the relation between diatom assemblages in Johnson and Clear Creek and selected environmental variables. The numbers correspond to sampling date and sequence (Urban: right, Rural: left, closed circles=wet season, open squares=dry season). Seasonal variation (b) of relative abundance (%) for dominant diatom species in Johnson and Clear Creek (Johnson Creek:right, Clear Creek:left).

streams. Seasonal variation of diatoms or recovery of diatom assemblages after flood disturbance has been well documented (Biggs, 1996). In bioassessment, it is crucial to distinguish the variability in biota due to anthropogenic factors (signal) and the variability due to natural variation (noise). A high signal-to-noise ratio is often considered as one of the most important criteria for selecting bioindicators. In this study, diatom species and environmental variables varied seasonally in both urban and rural streams. We expected seasonal patterns in diatom assemblages to be similar between the two sites due to their proximity. Our data showed that changes in diatom spe-

cies composition had similar patterns in both urban and rural sites throughout the year. Despite the high seasonal variation, differences in overall diatom assemblages between the urban site and rural site remained distinct. CCA showed that the urban site and rural site were separated along the 1st axis regardless of sampling dates. This axis was highly correlated with conductivity and dissolved nutrients, two environmental variables which co-varied with diatom assemblages along the urban-to-rural gradient during the summer base flow period. It is tempting to attribute differences among diatom assemblages to watershed land use alone. However, stream sites may have different charac-

188 teristics such as geology, channel geomorphology, and riparian conditions. In this study, these watersheds were chosen specifically to alleviate confounding factors by selecting watersheds with similar characteristics, including geology, gradient, and climate. Also, due to the proximity of all watersheds to each other and the dispersal ability of diatoms, it is likely they share a similar species pool. In summary, changes in diatom assemblages correlated well with conductivity along the urban-to-rural gradient. Also, despite the high seasonal variation, diatom assemblages were consistently different between the urban and rural sites, even though the urban and rural watersheds exhibited many similar characteristics. This study provides evidence that diatom assemblages can be used as a biotic indicator of urban stream conditions.

Acknowledgements This study was funded by the City of Portland, Bureau of Environmental Services. We thank Chris Prescott and Mary Abrams for supporting this project and C. Walker’s graduate program. Kazuhiro Sonoda and Julie Berry were of great help with both field and lab work. Ray Hoy, Scott Rollins, Brian Bowder, and Kristi DeWerff helped in field sampling. Christine Weilhoefer critically reviewed the manuscript. We also want to thank the residents in the Johnson, Clear, and Deep creek watersheds who let us access some of the study sites.

References Abrams, M. M. & C. Prescott, 1999. The endangered species act enters the urban landscape: can Portland streams sustain salmon? In Sakrison, R. & Sturtevant (eds), Watershed Management to Protect Declining Species. American Resources Association, Seattle Washington: 95–98. Biggs, B. J. F., 1995. The contribution of flood disturbance, catchment geology and land use to the habitat template of periphyton in stream ecosystems. Freshwater Biology 33: 419–438. Biggs, B. J. F., 1996. Patterns in benthic algae of streams. In Stevenson, R. J. M. L. Bothwell, & R. L. Lowe (eds), Algal Ecology: Freshwater Benthic Ecosystems. Academic Press, San Diego: 31–56.

Biggs, B. J. F., R. J. Stevenson & R. L. Lowe, 1998. A habitat matrix conceptual model for stream periphyton. Archive fu¨r Hydrobiologia 143: 21–56. Bledsoe, B. P. & C. C. Watson, 2001. Effects of urbanization on channel instability. Journal of the American Water Resources Association 37: 255–270. Booth, D. B. & C. R. Jackson, 1997. Urbanization of aquatic systems: degradation thresholds, stormwater detection, and the limits of mitigation. Journal of the American Water Resources Association 33: 1077–1090. Bryant, J., 1995. The Effects of Urbanization of Water Quality in Puget Sound Lowland Streams. Masters Thesis, University of Washington, Seattle. Carter, J. L., S. V. Fend & S. S. Kennelly, 1996. The relationships among three habitat scalesand stream benthic invertebrate community structure. Freshwater Biology 35: 109–124. Carpenter, K. D. & I. R. Waite, 2000. Relations of habitatspecific algal assemblages to land use and water chemistry in the Willamette Basin, Oregon. Environmental Monitoring and Assessment 64: 247–257. Clarke, S. E., D. White & A. L. Schaedel, 1991. Oregon, USA, ecological regions and subregions for water quality management. Environmental Management 16: 847–856. Environmental Protection Agency, 1993. Methods for the determination of inorganic substances in environmental samples. Cincinnati, Ohio. ESRI, 1997. Environmental System Research Institute Inc., Redlands, CA. Finkenbine, J. F., J. W. Atwater & D. S. Mavinic, 2000. Stream health after urbanization. Journal of the American Water Resources Association 36: 1149–1160. Grimm, N. B., J. M. Grove, S. T. A. Pickett & C. L. Redman, 2000. Integrated approaches to long-term studies of urban ecological systems. BioScience 50: 571–584. Hill, M. O., R. G. H. Bunce & M. W. Shaw, 1975. Indicator species analysis, a divisive polythetic method of classification and its application to a survey of native pinewoods in Scotland. Journal of Ecology 63: 597–613. Karr, J. R. & E. W. Chu, 1999. Restoring Life in Running Waters: Better Biological Monitoring. Island Press, Washington, D.C. Klotz, R. L., J. R. Cain & F. R. Trainor, 1976. Algal competition in an epilithic river flora. Journal of Phycology 12: 363–368. Krammer, K. & H. Lange-Bertalot, 1986. Bacilariophyceae, Teil 1. Naviculaceae. VEB Gustav Fisher Verlag, Jena. Krammer, K. & H. Lange-Bertalot, 1988. Tiel 2. Epithemiaceae, Bacilariophyceae Surirellaceae. VEB Gustav Fisher Verlag, Jena. Krammer, K. & H. Lange-Bertalot, 1991a. Bacilariophyceae, Tiel 3. Centrales, Fragilariaceae, Eunotiaceae, Achnanthaceae. VEB Gustav Fisher Verlag, Jena. Krammer, K. & H. Lange-Bertalot, 1991b. Bacilariophyceae, Tiel 4. Achnanthaceae, Kritische Erganzungen zu Navicula (Lineolate) und Gomphonema. VEB Gustav Fisher Verlag, Jena. Kroeger, S., E. Fensin, K. Lynch & M. Vander Borgh, 1999. United States Water Quality Programs that Use Algae as a

189 Biological Assessment Tool. North Carolina Division of Water Quality, Raleigh, NC. Leland, H. V., 1995. Distribution of phytobenthos in the Yakima River basin, Washington, in relation to geology, land use, and other environmental factors. Canadian Journal of Fisheries and Aquatic Sciences 52: 1108–1129. Leopold, 1968. Hydrology for urban land planning? a guidebook on the hydrologic effects of urban land use. USGS Circular 554. Lowe, R. L., 1974. Environmental requirements and pollution tolerance of freshwater diatoms. United States Environmental Protection Agency, Cincinnati, Ohio. EPA-670/4– 74–005. McDonnell, M. J. & S. T. A. Pickett, 1990. Ecosystem structure and function along urban-rural gradients: an unexploited opportunity for ecologys. Ecology 71: 1232–1237. Metro, 1997. Clackamas River Watershed Atlas. Portland, Oregon. Metro., 1999. Regional Land Information System (RLIS) Lite (CD rom). Metro’s Data Resource Center, Portland, Oregon. Molloy, J. M., 1992. Diatom communities along stream longitudinal gradients. Freshwater Biology 28: 59–69. Pan, Y., R. J. Stevenson, B. H. Hill, A. T. Herlihy & G. B. Collins, 1996. Using diatoms as indicators of ecological conditions in lotic systems: a regional assessment. Journal of North American Benthological Society 15: 481–495. Patrick, R. & C. W. Reimer, 1966. The Diatoms of the United States. Vol. 1. Monographs of the Academy of Natural Sciences of Philadelphia No.13. Patrick, R. & C. W. Reimer, 1975. The Diatoms of the United States. Vol. 2. Part 1 Monographs of the Academy of Natural Sciences of Philadelphia No.13. Richards, C., R. J. Haro, L. B. Johnson & G. E. Host, 1997. Catchment and reach-scale properties as indicators of macroinvertebrate species traits. Freshwater Biology 37: 219–230. Sonneman, J. A., C. J. Walsh, P. F. Breen & A. K. Sharpe, 2001. Effects of urbanization on streams of the Melbourne

region, Victoria, Australia. II. Benthic diatom communities. Freshwater Biology 46: 553–565. Sonoda, K., J. A. Yeakley & C. E. Walker, 2001. Near-stream landuse effects on streamwater nutrient distribution in an urbanizing watershed. Journal of the American Water Resources Association 37(6): 1517–1532. Stevenson, R. J, 1984. Epilithic and epipelic diatoms in the Sandusky River, with emphasis on species diversity and water pollution. Hydrobiologia 114: 161–174. Stevenson, R. J. & Y. Pan, 1999. Assessing ecological conditions in streams and rivers with diatoms. In Stoermer, E. & J. P. Smol (eds), The Diatoms: Applications to Environmental and Earth Sciences. Cambridge University Press, Cambridge, U.K: 11–40. Swanson, R. D., W. D. McFarland, J. B. Gonthier & J. M. Wilkinson, 1993. A description of hydrologic units in the Portland basin, Oregon and Washington. U.S. Geological Survey, Water-Resources Investigations Report 90–4196, 56 pp. ter Braak, C. J. F. & P. Smilauer, 1998. CANOCO for Windows: Software for Canonical Community Ordination (version 4). Microcomputer Power, Ithaca, New York. Torquemada, R. & W. Platts, 1989. A comparison of sediment monitoring techniques of potential use in sediment/fish population relationships. Boise, Idaho, Dept. Fish and Game. Uhrich, M. A. & D. A. Wentz, 1999. Environmental setting of the Willamette Basin, Oregon. U.S. Dept. of the Interior, U.S. Geological Survey, Water-Resources Investigations Report 97–4082-A. Welch, E. B., J. M. Jacoby & C. W. May, 1998. Stream quality. In Naiman, R.J. & R.E. Bilby (eds), River Ecology and Management. Springer-Verlag, New York: 69– 94. Zar, J. H., 1999. Biostatistical Analysis (4th ed.). Prentice-Hall, Upper Saddle River, New Jersey.

Hydrobiologia (2006) 561:191–206  Springer 2006 R.J. Stevenson, Y. Pan, J.P. Kociolek & J.C. Kingston (eds), Advances in Algal Biology: A Commemoration of the Work of Rex Lowe DOI 10.1007/s10750-005-1614-2

Developing and testing diatom indicators for wetlands in the Casco Bay watershed, Maine, USA Yi-Kuang Wang1,4, R. Jan Stevenson1,*, P. Roger Sweets2 & Jeanne DiFranco3 1

Department of Zoology, Michigan State University, East Lansing, MI, 48824, USA Biology Department, University of Indianapolis, Indianapolis, IN, 46227, USA 3 Maine Department of Environmental Protection, 312 Canco Rd., Portland, ME, 04103, USA 4 General Education Center, Nan-Hua University, Taiwan, R.O.C. (*Author for correspondence: E-mail: [email protected]) 2

Key words: ecological assessment framework, wetlands, diatoms, index of ecological integrity, stressor identification, human disturbance

Abstract Diatom indicators of wetland condition were developed and tested by assessing human disturbance, water chemistry, and species composition of benthic, epiphytic, and planktonic diatoms from 20 wetlands sampled for 2 years. One sample from each site was randomly selected to form a development data set, while the rest were used as the test data set. Human disturbance indicated substantial differences among wetlands in hydrologic modification, impervious surface, and potential for non-point source contamination. These landscape alterations were related to increases in pH, non-nutrient ions, and nutrients and decreases in dissolved organic carbon and water color. Pre-existing diatom indicators, calculated with autecological information from lakes and aquatic habitats, correlated highly to relevant water chemistry and human disturbance scores. Weighted average models (WAM) of Cl), conductivity, pH, and alkalinity derived with the Maine development data set correlated to relevant water chemistry and human disturbance of the test wetlands. Diatom assemblage attributes that correlated with human disturbance were selected to combine into a multimetric index of biotic condition (IBC). IBCs and WAMs from benthic and epiphytic diatoms were usually more precisely related to relevant environmental factors than planktonic diatoms. These results showed that human disturbance alkalized wetlands, enriched them with nutrients, and diatom assemblages responded to these changes. Indicator development protocols for streams can be readily adapted for use in wetlands.

Introduction Wetlands are important elements of landscapes and provide natural capital and many ecosystem services, such as migratory bird habitat, biodiversity hotspots, water retention, groundwater recharge, flood reduction, and carbon, sediment, and nutrient sequestration (Gale et al., 1993; Drewien et al., 1996; Mitsch & Gosselink, 2000; Rapalee et al., 2001; Bendjoudi et al., 2002; Severo et al., 2002; Leibowitz, 2003). A historic lack of understanding of wetland ecosystems has led to mass destruction of wetlands for economic develop-

ment. Recent public awareness has stimulated protection of wetlands by government agencies. In 1998, however, ecological condition had only been assessed by state agencies for 9 % of the wetlands in the conterminous U.S. (U.S. EPA, 1998a). Sedimentation, filling and draining, and habitat alteration were the major threats to wetland integrity (U.S. EPA, 1998a). Facing increasing pressure from human development, effective assessment tools are needed for consistent evaluation of the condition and stressors of wetland resources and to provide information for solving problems.

192 The goals of ecological assessment are to evaluate biological condition (sensu Cairns, 1977; Karr, 1991) and other valued ecological attributes, and to assess human disturbance, contaminants, and habitat alterations that could impair valued ecological attributes (Smith et al., 1997; D’Elia et al., 2003; Stevenson et al., 2004). Then, water quality criteria and stressor–response relationships (Roux et al., 1999; Stevenson & Hauer, 2002; Yuan & Norton, 2003) can be used to synthesize all information for management purposes (Stevenson, 1998; U.S. EPA, 1998b; Jackson et al., 2000; Wang & Stevenson, 2002; Stevenson et al., 2004). Diatom indicators can be used to evaluate both biological condition and pollutants affecting biological condition. Most diatom indicators have been developed as weighted average models (WAMs) to infer chemical or physical conditions of streams and lakes. Recently diatom indicators were developed to assess biological condition independently from inferences of pollution (Hill et al., 2000; Wang et al., 2005). Existing wetland studies have evaluated WAMs of water chemistry in Kentucky and Michigan wetlands (Pan et al., 1996; Stevenson et al., 1999). In the Everglades, changes in algal species composition and biomass have been related to environmental factors (McCormick & O’Dell, 1996; Slate & Stevenson, 2000; Pan et al., 2000; Stevenson et al., 2002a) However, no wetland studies of algae have developed multiple metrics of biological condition and tested multimetric indices of biological condition (Stevenson et al., 2002b). In addition, no wetland studies have evaluated which habitat in wetlands is best for sampling and assessing algal assemblages in wetlands. The objectives of this study were to develop and test a set of diatom indicators of wetland conditions and to determine which habitats were best to sample. We developed and tested separate indicators of biological condition and indicators of physicochemical condition. Biological condition in this study was defined as the condition of algal assemblages in wetlands with low human disturbance. First we related changes in water chemistry of wetlands to human activities around them. We then tested metrics calculated with species autecological information from lake and stream literature by relating them to changes in human activities and related water chemistry of wetlands.

WAMs were developed and tested with Maine wetlands data to characterize species environmental optima and tolerances in wetlands and to compare to metrics calculated with pre-existing information. Finally, we developed and tested metrics and multimetric indices of biological condition in wetlands with benthic, epiphytic, and planktonic diatoms.

Materials and methods Wetland sampling and sample analysis Twenty wetlands were sampled in 1998 and 1999 in the Casco Bay watershed of Maine (Fig. 1), with two different wetlands sampled in each year. Sites were selected along an urban gradient, with landscapes conditions around wetlands ranging from natural to suburban and urban. This watershed was glaciated and had sand, gravel, and silt deposited over granite bedrock. About two-thirds of sites were riverine wetlands, while one-third were isolated marshes. Because of the small sample size, different wetland types were not analyzed separately, which could improve precision of indicators. Eighteen wetlands were sampled during 2 years to assess annual variability in plants, invertebrates, and water chemistry, which are not reported in this paper. Based on results from studies in streams (Catherine Riseng et al., University of Michigan, unpublished data), we assumed that repeated measures of algal species composition and water chemistry in the same wetland across years were independent, because annual variability in these factors was as great among wetlands of similar disturbance regimes as within each wetland. Human disturbance was characterized using a field assessment of the wetland landscape developed by Maine Department of Environmental Protection. Similar approaches have been adopted by state agencies, such as Washington Department of Ecology (Washington DE, 1993) and Ohio EPA (Mack, 2001). Field assessments were calibrated between people in charge of the assessments for consistency. The assessment was conducted by walking around the wetlands. Five categories of disturbance were characterized: vegetation modi-

193

Figure 1. Locations of sampling sites in Maine and USA.

fications (VM), hydrological modifications (HM), the evidence of chemical pollutants (CP), impervious surface (IS), and potential for non-point source pollution (NPS). Each category had five questions and each question was answered with a ranking from 0 to 5 (none observed (0), minimal (1), moderate (3), to severe (5)). A maximum total score of 25 could be assigned to each category. The total human disturbance score (THDS) was the sum of scores from these five categories. Water grab samples were collected from multiple locations in each wetland site and composted for chemical analysis. Ca+2, Mg+2, K+, Na+, Si, Cl), NO3)N, NH4)N, total N, SO4, PO4)P, total P, and dissolved organic carbon were measured in

the lab with standard methods (APHA, 1998). A grab sample of 500 ml was acidified with HNO3 and stored in a refrigerator for Ca+2, Mg+2, K+, and Na+ analyses, while a 250- and 500-ml grab samples were stored on ice in the field and kept in a refrigerator for nutrients (NO3)N, NH4)N, total N, PO4)P, total P) analyses in the lab respectively. All metals were analyzed by inductively coupled plasma-atomic emission spectrometry, while nutrients were analyzed by automated or semiautomated colorimetry. Conductivity, temperature (Hanna hand-held meter, model HI 9635), and dissolved oxygen (Hanna hand-held meter, model HI 9142) were measured in the field. pH and alkalinity were analyzed in the lab.

194 Diatoms were sampled qualitatively from three habitats (water depth <1 m): the water column, plants, and sediments. Composite algae samples were collected randomly. Phytoplankton was sampled by collecting water from several locations in the wetland and composting them in a 1-l bottle for each site. Epiphytic algae were sampled by randomly selecting plants from the wetland, cutting stems, placing underwater sections of plant stems in a Whirl-Pak bag with distilled water, rubbing plants together to remove epiphyton, and then removing the stems from the bag. Sediment algae were sampled by using a turkey baster, which is like a large pipette with a 0.5-cm opening. Diatoms were cleaned with boiling nitric acid catalyzed by potassium dichromate, neutralized with de-ionized water, and mounted on slides with Naphrax. Diatom valves were identified and counted until a minimum of 600 valves and at least 10 valves for 10 different taxa were observed. Diatom taxonomy was based on Krammer and Lange-Bertalot (1986, 1988, 1991a, b) and Patrick and Reimer (1966, 1975). Data analysis First, we separated the data set into development and test data sets. One of two samples from each site was randomly selected to form a development data set so that year of sampling was not a confounding variable during testing. The rest of samples were used in a test data set. Differences in water chemistry between data sets were tested with a Mann-Whitney U test. We tested the hypothesis that differences in water chemistry among wetlands were related to human activities around wetlands. Pearson correlation coefficients were calculated to characterize relationships between individual and combined scores of human disturbance categories with each water chemistry variable measured. Separate analyses were performed for 1998 and 1999 because water levels were substantially higher in 1999. Diatom indices of water chemistry were calculated with existing autecological information from two major sources to evaluate transferability of metrics across regions and habitat types, to evaluate generic indicators, and to develop and test WAMs as diagnostic indices. These indices were calculated with diatom relative abundances in

planktonic samples only, to limit the length of results to report in this paper. Planktonic samples were chosen for these analyses over epiphytic and benthic samples because we assumed plankton would most directly reflect measured water chemistry conditions. WAMs were calculated using species environmental optima and their relative abundances (Dixit & Smol, 1994). van Dam’s indices (1994) were calculated for N, salinity, pH, trophic state, and saprobity using the categorical autecological classification of diatoms, which were derived from multiple habitat types in Netherlands and adjacent countries. Dixit’s WAMs were calculated for TP, pH, and Cl) using species optima from Dixit et al. (1999), which were derived from a large number of lakes in northeastern U.S. WAMs are capable of inferring quantitative values of water quality, while biotic indices based on categorical autecological ranks simply infer relative ranks of water quality. Due to different taxonomic systems used in different studies, species names in this study were matched with species names used in van Dam’s indices and Dixit’s WAMs before calculations. Generic indicators were calculated with relative abundance of genera in which most species favor acidic environments, are motile, or require high nutrients to survive (eutraphentic), according to the classification of genera in Wang et al. (2005). Transferability of metrics across regions and habitat types was evaluated by comparing the magnitude of correlations between diatom indicators and human disturbance scores and relevant environmental factors (i.e., those for which the indicator was developed, such as Cl), TN, and TP concentrations and pH). Thus correlations between Dixit’s WAMs or van Dam’s indices and relevant environmental factors were compared to correlations between the same environmental factors and respective Maine wetland WAMs. If correlations between environmental factors and Dixit’s WAMs or van Dam’s indices were similar to correlations with Maine wetland WAMs, then transferability was considered high. WAMs were calculated for environmental variables in Maine wetlands that were highly correlated to THDS by using the development data set and CALIBRATE (Juggins & ter Braak, 1992). WAMs were calculated with and without downweighting the importance of species with high tolerances (ter Braak & van Dam, 1989) and re-

195 ferred to as WA and WATol models, respectively. The Maine WAMs were tested in three ways. First, WAMs were tested using bootstrapped techniques with the development data set. Second, WAMs derived with the development data set were then tested with the test data set by inferring conditions at test sites with the species relative abundances and measured conditions at test sites. Third, WAM-inferred conditions were compared with human disturbance scores at test sites. WAMs were calculated for all three habitat types. Differences in precision of WAMs among habitats were compared by correlations between inferred conditions and relevant environmental factors. Multimetric indices of biotic condition (IBCs) were developed, tested, and compared among habitats. The first step in IBC development was to select a set of diatom attributes that could respond to human disturbance. Both biotic condition and diagnostic metrics were included in IBC (Barbour et al., 1999). Six categories of attributes were used in index development: diversity, biotic indices inferring stressors, similarity to reference condition, sensitive and tolerant species, growth forms, and genus-level community structure. In the analyses, reference sites were defined as those with THDS smaller than 7, the 25th percentile of THDS among sites. Five diversity indices were assessed. Shannon diversity (Shannon & Weaver, 1949), Hurlbert’s evenness, and species and generic richness indices were expected to decrease with increasing human disturbance, while Simpson’s dominance index was expected to increase with increasing human disturbance (Odum, 1985). Similarity measures have been recognized as a powerful tool to assess human disturbance on aquatic communities (Boyle et al., 1990). Average similarity between species composition of the assemblage in a test site and assemblages in reference sites is an overall measurement of community structural change resulting from human disturbance (Sheehan, 1984). We used the Bray– Curtis dissimilarity index to evaluate similarity of assemblages between a test site and reference sites. Similarity was calculated by subtracting the dissimilarity value from 1. We expected that average similarities with reference sites would decrease with increasing human disturbance.

Ecosystems lose sensitive species and gain tolerant species under increasing stress (Schindler, 1987). Sensitive and tolerant species were distinguished in the development data set based on the presence or absence at reference or impaired sites, respectively, with indicator species analysis (Dufrene & Legendre, 1997). The number and percentage of sensitive species were expected to decrease with increasing human stressor levels, while the number and percentage of tolerant species were expected to increase with increasing stressor levels. Diatoms are sensitive to nutrients and take up nutrients as primary producers in stream ecosystems. Therefore, we categorized diatom functional groups with their preferences to nutrient conditions. The index values usually increase with relevant water chemistry levels. van Dam’s indices and generic groups were expected to increase with human disturbance. The relative abundance of diatoms with different growth forms indicates assemblage developmental status (Peterson, 1996). Prostrate diatoms grow below the boundary layer and can avoid scouring, and are usually dominant during early successional stages or heavy grazing. Erect and stalked diatoms are capable of overgrowing prostrate diatoms and appear during mid and late succession. Unattached diatoms are planktonic and accumulate in slow moving water. Motile diatoms are capable of moving through sediments, which may indicate the level of sedimentation and bank erosion. Shifts in assemblage composition at family and genus levels of organization are commonly used in the assessment of biological condition. Some diatom genera are also capable of indicating specific environmental conditions. For example, Eunotia and Stenopterobia prefer acidic waters; Frustulia, Pinnularia, and Tabellaria prefer soft waters (Round, 1990). Most Nitzschia, Planothidium, Amphora, Anomoeneis, and Stauroneis are high-trophic status indicators, whereas most Brachysira, Cymbella, Eunotia, Frustulia, and Stenopterobia are low-nutrient indicators (Kelly, 1998). Relative abundances of Achnanthidium and Cymbella are less abundant in streams with high human disturbance in a region studied by Wang et al. (2005). Therefore, relative abundances of each

196 genus were evaluated as potential metrics of biological condition. Metrics of the IBC were selected according to the following procedures. Diatom attributes were selected as metrics if they were among the three highest Pearson correlation coefficients with each human disturbance categories and if they were significantly correlated with THDS (p < 0.05). Correlations among selected metrics were evaluated to reduce redundant metrics. Each metric was scaled to a 0–10 scoring system. If the metric decreased with human disturbance, the score was scaled by dividing the 95th percentile of metric values, and then multiplied by 10. If the metric increased with disturbance rank, the score was calculated by dividing the metric value by the 95th percentile, subtracting the quotient from 1, and then multiplying that difference by 10. The total IBC score was a summation of values of selected metrics. IBCs were developed with the development data set and tested on both data sets for three habitat specific assemblages. They were evaluated and compared with Pearson correlation analysis to assess their relationships with three indicators of human disturbance: THDSs for each site; site scores on the first principal component axis of a principal component analysis (PCA) of human disturbance scores; and site scores on the first principal component axis of a PCA of water chemistry data. PCA was used to summarize

multivariate environmental gradients into a single latent variable. The first principal component axis usually explains most of the variance, hence can represent the overall gradient in conditions at the sites. Statistical analyses were performed with SPSS software.

Results Human disturbance and physicochemical attributes THDS ranged from 1 to 25 with a mean of 8.7 and had a distribution skewed toward low scores (Fig. 2). Variation was relatively great for hydrologic modification with scores ranging from 0 to 6, impervious surfaces ranging from 0 to 9, and nonpoint source pollution ranging 0 to 9. Less than half of the sites had vegetation modification scores greater than 0 (maximum = 3); only three sites had chemical pollution scores greater than 0 (maximum = 5). Because of low variation in the vegetation and chemical disturbance scores, they were not included in most analyses. Physicochemical variables varied greatly among wetlands and correlated with the THDS (Tables 1 and 2). Indicators of ionic content of waters were usually low with medians of 3.87 mg Cl l)1, 29 lS cm)1 of conductivity, and 6.7 pH (Table 1). Indicators of nutrient concen-

Figure 2. Distribution of scores for different categories of human activities near Maine wetlands: HM, hydrologic modification; VM, vegetation modification; CP, chemical pollution; IS, impervious surface; NPS, non-point source pollution; SumHDS, sum of human disturbance ranks for all categories.

197 Table 1. Summary statistics for the development and test data sets (n = 20) Cl)

Statistic

NO3–N )1

(mg l )

)1

(mg l )

NH4–N )1

(mg l )

TN )1

(mg l )

PO)2 4

TP

Cond

SiO2–Si

(lg l)1)

(lg l)1)

(lS cm)1)

(mg l)1)

pH

Development data set Minimum Percentiles

25th

0.7

0.01

0.01

0.23

1

11

15

0.11

6.2

2.33

0.01

0.02

0.56

2

22

19.8

1.68

6.4

50th

3.87

0.03

0.05

0.74

3

42

29

3.11

6.7

75th

10.95

0.04

0.05

0.93

5

51

74.8

4.37

6.9

1.34

0.17

22

12

300

1345

5.4

8.1

Maximum

357

Test data set Minimum Percentiles

25th

0.52 3.04

0.01 0.01

0.01 0.02

0.4 0.49

1 2

12 24

14 27.6

0.6 1.72

6.2 6.6

50th

5.34

0.03

0.04

0.71

3

55

39.1

2.67

6.8

75th

7.87

0.03

0.05

1.06

5

71

45

4.26

7

0.89

0.32

2.7

14

120

1820

5.2

7.6

Maximum

388

Table 2. Pearson correlation coefficients characterizing relationships between physiochemical variables and human disturbance rank scores Pchem variable

HM 1998

IS 1999

1998

NPS 1999

1998

THDS 1999

1998

1999

Ca+2

0.51

0.46

0.79

0.80

0.51

0.56

0.65

0.67

Mg+2

0.54

0.49

0.77

0.79

0.57

0.58

0.70

0.68

K+

0.10

0.50

0.68

0.72

0.24

0.54

0.37

0.66

Na+

0.53

0.39

0.66

0.74

0.56

0.52

0.65

0.60

Cl)

0.56

0.36

0.61

0.73

0.52

0.49

0.60

0.57

Cond

0.53

0.50

0.72

0.79

0.54

0.52

0.67

0.66

Alk SO)2 4

0.45 0.31

0.50 0.45

0.79 0.31

0.69 0.48

0.59 )0.11

0.69 0.44

0.70 0.12

0.72 0.54

NH4)N

0.21

0.12

0.72

0.10

0.37

0.40

0.49

0.36

NO3)N

0.07

)0.31

0.37

)0.07

0.02

)0.29

0.08

)0.25

Total N

0.04

0.26

0.30

0.50

0.36

0.34

0.28

0.41

PO4)P

0.11

)0.08

0.08

0.34

0.07

0.28

0.09

0.22

Total P

0.31

0.48

0.29

0.35

0.60

0.56

0.47

0.54

Si

0.28

0.32

0.23

0.28

0.29

0.73

0.29

0.60

0.22 )0.32

0.57 )0.28

0.10 0.01

0.58 )0.13

0.40 0.17

0.57 )0.24

0.31 )0.04

0.66 )0.28

Chl a DOC pH Color

0.31

0.32

0.80

0.62

0.36

0.54

0.55

0.59

)0.46

)0.48

)0.56

)0.38

0.03

)0.40

)0.30

)0.49

Bold coefficients indicate p < 0.05 (n = 20).

trations were also low, with median NO3, NH4, and TN of 0.03, 0.05, and 0.74 mg N l)1, a median Si of 3 mg l)1, and median PO4 and TP of 3 and 42 lg P l)1, respectively. No significant differences were observed in physicochemical

variables between the development and test data sets (Mann-Whitney U tests, p > 0.05) (Table 1). For example, non-nutrient ionic factors (Ca+2, Mg+2, K+, Na+, Cl), SO)2 4 , alkalinity, and conductivity) significantly correlated to scores of

198 disturbance categories and to sums of disturbance scores in more than 90% of the combinations of disturbance and physicochemical variables. Variables related to nutrient enrichment (NO3)N, NH4)N, TN, PO4, TP, Si, and chl a) were positively correlated (p < 0.05, n = 20) to human disturbance categories in less than 30% of the variable combinations (15 of 56), but results of correlation analyses were two times as likely to be significant during 1999 vs. 1998. DOC was not related to human disturbance, but pH was often positively related and color was often negatively related to human disturbance variables. Functional groups, WAMs, and biotic indices Indices based on pre-existing autecological information were highly correlated with both physicochemical and human disturbance variables. Diatom functional groups (van Dam’s indices) were significantly correlated with most relevant disturbance and water chemistry variables (Table 3). Scores of van Dam’s saprobity and nitrogen indices were correlated to most physicochemical and human disturbance variables. Scores of van Dam’s indices were always correlated with Cl, pH, and IS scores. All of Dixit’s WAMs were also significantly correlated with relevant variables and more than 50% of variables in Table 3.

Two of the three genus-level indices were correlated with most relevant variables (Table 3). Percentages of diatoms in eutraphentic genera did not correlate with any relevant variables. Percentages of diatoms in acidic genera negatively correlated to 6 of 8 relevant variables. Percentages of motile diatoms was significantly correlated with 7 of the 8 relevant variables. WAMs derived from the Maine development data set were correlated with many relevant variables for assemblages in all three habitats when tested by bootstrapping within the development data set, applied to the test data set, and related to THDS (Table 4). For example, the benthic pH WAM correlated with pH in the development data set, applied to the test data set, and with THDS (r = 0.58, p = 0.007; r = 0.58, p = 0.007; r = 0.69, p < 0.001; respectively). Benthos had more significant relationships than epiphyton (p < 0.05 for 26 and 21 of 60, respectively). Plankton had only 16 significant correlations of the 60 possible tests. More significant correlations were observed between WAMS and THDS than between WAMS and relevant variables with the test data set (32/60 vs. 19/60, respectively). Few correlations were found between WAMS and measured environmental variables when bootstrapped with the development data set (12/60). WAMs derived from the Maine development data set for ionic variables (Cl, conductivity,

Table 3. The correlations of planktonic functional groups and indices with relevant disturbance scores and water chemistry for the plankton samples of the test data set VD

VD

trophic

saprobity

VD N

% Eutrophentic

% Acidic

% Motile

Dixit’s

Dixit’s

Dixit’s

VD

diatom

diatom

diatom

TP

pH

Cl

Salinity

VD pH

Cl)

0.64

0.75

0.72

)0.24

)0.64

0.64

0.56

0.68

0.54

0.56

0.11

TN TP

0.36 0.51

0.76 0.70

0.74 0.70

0.18 0.08

)0.38 )0.48

0.80 0.67

0.61 0.54

0.37 0.54

0.56 0.52

0.29 0.40

)0.03 0.20 0.06

)0.13

0.04

)0.02

0.27

0.18

0.02

0.20

)0.09

0.24

0.01

pH

0.63

0.60

0.63

)0.04

)0.71

0.26

0.35

0.77

0.41

0.69

0.31

Chl a

0.16

0.50

0.50

0.14

)0.25

0.77

0.49

0.27

0.50

0.09

)0.21

DOC

HM

0.19

0.30

0.26

)0.30

)0.31

0.60

0.30

0.26

0.29

0.13

)0.11

CP

0.50

0.83

0.80

)0.01

)0.47

0.86

0.68

0.48

0.61

0.36

0.02

IS

0.51

0.68

0.62

)0.27

)0.62

0.51

0.55

0.59

0.54

0.55

0.00

NPS THDS

0.38 0.49

0.49 0.66

0.50 0.64

0.11 )0.07

)0.46 )0.59

0.57 0.70

0.28 0.49

0.42 0.55

0.31 0.49

0.31 0.44

0.11 0.04

Bold face marks significant correlation. IS – impervious surface rank; NPS – non-point source pollution rank; ns – non-significance (n = 20).

199 Table 4. Correlation statistics for WAMs developed with and without down-weighting species for their environmental tolerance (WA and WATol, respectively) WAM variable

Habitat

Dev data set

Test data set

WA

WATol

WA

THDS WATol

WA

WATol

Cl) Cond

Benthos Benthos

0.01 0.20

0.36 0.53

Alk

Benthos

0.66

0.64

0.79

0.95

0.69

0.71

pH

Benthos

0.58

0.31

0.58

0.49

0.69

0.65

0.86 0.90

0.46 0.92

0.59 0.63

0.67 0.71

SO4

Benthos

0.00

0.02

0.57

)0.04

0.45

)0.02

Si

Benthos

0.03

0.10

0.21

0.26

0.12

0.25

NH4

Benthos

0.25

0.01

0.03

0.03

0.68

0.70

NO3

Benthos

0.00

0.00

)0.01

)0.03

0.47

0.39

TN TP

Benthos Benthos

0.00 0.11

0.07 0.11

)0.05 0.10

0.16 0.03

)0.13 0.40

0.41 0.31 0.20

Cl)

Epiphyton

0.02

0.27

0.22

)0.05

0.59

Cond

Epiphyton

0.13

0.30

0.24

)0.02

0.62

0.36

Alk

Epiphyton

0.77

0.60

0.62

0.56

0.71

0.64

PH

Epiphyton

0.54

0.43

0.70

0.57

0.60

0.53

SO4

Epiphyton

0.00

0.00

0.26

)0.01

0.51

0.08

Si

Epiphyton

0.03

0.03

0.30

0.14

0.53

0.26

NH4 NO3

Epiphyton Epiphyton

0.22 0.00

0.00 0.01

0.26 )0.01

)0.06 )0.06

0.75 0.52

0.56 )0.08

TN

Epiphyton

0.00

0.01

)0.05

0.05

0.05

0.46

TP

Epiphyton

0.02

0.01

0.16

)0.04

0.65

0.13

Cl)

Plankton

0.01

0.04

0.49

0.29

0.39

0.60

Cond

Plankton

0.12

0.19

0.55

0.10

0.40

0.42

Alk

Plankton

0.66

0.46

0.19

0.86

0.44

0.60

PH

Plankton

0.55

0.37

0.44

0.35

0.46

0.57

SO4 Si

Plankton Plankton

0.01 0.00

0.05 0.01

0.42 0.45

0.10 0.37

0.35 0.72

0.03 0.68 0.47

NH4

Plankton

0.15

0.02

)0.03

)0.06

0.34

NO3

Plankton

0.00

0.07

)0.02

)0.05

0.35

0.36

TN

Plankton

0.04

0.06

)0.04

)0.04

)0.09

)0.35

TP

Plankton

0.07

0.00

0.09

0.05

0.35

0.22

Development (Dev) data set coefficients relate measured and inferred conditions with the development data set based using bootstrapping. Test data set coefficients relate measured physicochemical conditions of the test data set and conditions inferred for test data sites using species optima from the development data set and species relative abundances from the test data set. Human dist rank coefficients relate human disturbance ranks of test sites with conditions inferred for test data sites using species optima from the development data set and species relative abundances from the test data set. Coefficients in bold indicate that p < 0.05 (n = 20 for epiphyton and benthos, n = 19 for plankton).

alkalinity, and pH, excluding the SO4 WAM) were related to relevant variables in 20 out of the 24, 14 of the 24, and 12 of the 24 tests of benthos, epiphyton, and plankton, respectively (Table 4). The WAMs for SO4 and nutrients were significantly correlated in less than 10% of cases with the development or test data sets for any assemblage, however they were

usually related to THDS. Nutrient WAMs were significantly related to THDS at test sites for only 4, 6, and 2 of 10 possible relationships (WA and WATol for 5 nutrient parameters) for benthos, epiphyton, and plankton, respectively. WAMs for Cl, pH, and nutrients using autecological information from the Maine wetlands

200 were not always better related to relevant environmental variables than WAMs using existing autecological indices. Correlation coefficients of Cl and pH WAMs derived from Maine wetlands with relevant variables were higher when assessing the test data set and were lower when assessing the development data set with bootstrapping (Table 4). Correlation coefficients for Cl and pH WAMs were usually higher with THDS at test sites when using autecologies derived from the Maine development data set than when using existing autecologies. However, correlations for nutrient WAMs were lower in all situations when using autecological information from Maine wetlands than when using existing autecologies. Biotic condition Six metrics of benthic assemblages were selected for the benthic IBC according to correlations with human disturbance scores (Table 5). Selected metrics were: % motile diatom, % Synedra, no. sensitive species, average similarity with reference sites, % Neidium, and % Eunotia (Table 5). Percent Nitzschia and Surirella both had high correlations with % motile diatom (r = 0.94, p < 0.001; r = 0.70, p < 0.001), so both were not included in the IBC. Eight metrics were selected for the epiphytic IBC according to correlations with human disturbance scores (Table 5). The 8 metrics were: % motile diatom, % Eunotia, no. sensi-

tive species, average similarity with reference sites, no. tolerant species, van Dam’s saprobity index, van Dam salinity index, and van Dam’s nitrogen index (Table 5). Percent Nitzschia was not included in the IBS due to high correlations with % motile diatom (r = 0.97, p < 0.001). Eight metrics were selected for the planktonic IBC according to correlations with human disturbance scores (Table 5). The 8 metrics were: % prostrate diatoms, % motile diatoms, no. sensitive species, van Dam’s saprobity index, % Eunotia, % Neidium, % Pinnularia, and average similarity with reference sites. Percent sensitive species was highly correlated with no. of sensitive species, hence it was not included in the planktonic IBC. Multimetric biotic indices from each of the three habitats significantly correlated with three evaluation criteria in the development data set (Tables 6–8). In the PCA results of water physicalchemistry variables, the 1st axis explained 57.4% variance, which was around five times the variance explained by the 2nd axis (Table 6). In the PCA results of human disturbance variables, the 1st axis explained 51% variance, which was about twice the variance of the 2nd axis (Table 7). In the development data set, the epiphytic IBC had the highest average correlation with evaluation criteria (r = )0.87, p < 0.001; r = )0.98, p < 0.001; r = )0.81, p < 0.001) and the benthic IBC was the second (r = )0.84, p < 0.001; r = )0.95, p < 0.001; r = )0.79, p < 0.001) (Table 8). In the

Table 5. Selected metrics for each habitat Habitat

Disturbance categories HM

Benthos

Epiphyton

Plankton

CP

IS

NPS

% Nitzschia (0.60)

% Synedra (0.91)

No. sensitive species ()0.84)

% Neidium ()0.57)

% Motile (0.59) % Synedra (0.56)

% Motile (0.80) % Surirella (0.72)

Average similarity ()0.82) % Synedra (0.74)

% Eunotia ()0.57) No. sensitive species

% Motile (0.6)

% Surirella (0.74)

No. sensitive species ()0.83)

VD saprobity (0.65)

% Nitzschia (0.59)

No. sensitive species ()0.73)

Average similarity ()0.76)

VD N (0.62) VD salinity (0.63)

% Eunotia ()0.57)

VD saprobity (0.70)

No. tolerant species (0.76)

% Prostrate (0.58)

% Nitzschia (0.83)

No. sensitive species ()0.82)

% Neidium ()0.57)

% Motile (0.55)

% Motile (0.82)

Average similarity (0.73)

% Pinnularia ()0.53)

% Sensitive species ()0.55)

VD saprobity (0.80)

% Eunotia ()0.69)

% Nitzschia (0.51)

Attributes with the highest three correlation coefficients with each disturbance rank category were selected as metrics. ( ) encloses Pearson correlation coefficient (n = 20 for plankton and epiphyton, n = 19 for benthos).

201 Table 6. Results of PCA on water chemistry variables

% Variance explained

PCA axis 1

PCA axis 2

Table 8. Correlation between IBC from development data set and IBC from test data set for each habitat and evaluation criteria

57.35

12.40

Ca+2

0.98

0.01

Mg+2

0.98

0.05

+

K

0.95

0.11

Na+

0.96

0.08

)0.84

)0.95

)0.79

Si

0.52

)0.15

(n = 19)

(<0.001*)

(<0.001*)

(<0.001*)

Cl) NH4

0.93 0.53

0.02 )0.46

)0.79

)0.69

)0.68

(n = 20)

(<0.001*)

(0.001*)

(0.001*)

NO3)N

0.03

)0.50

Epiphyte

TN

0.68

0.53

PO4

0.23

)0.27

TP

0.73

0.43

SO4

0.70 )0.17

DOC Cond Alkalinity PH Color

Habitat

DR PCA1

Water PCA1

Human DR

Benthos

)0.87

)0.98

)0.81

(n = 20) Test

(<0.001*) )0.61

(<0.001*) )0.67

(<0.001*) )0.62

0.25

(n = 20)

(0.005*)

(0.001*)

(0.004*)

0.78

Plankton

0.97 0.96

)0.03 )0.02

(n = 20)

0.87

)0.25

Test

)0.71

)0.73

)0.64

)0.61

0.58

(n = 19)

(<0.001*)

(<0.001*)

(<0.001*)

Variance explained and correlations of PCA axes with water chemistry variables were shown in the table.

Table 7. Results of PCA on human disturbance variables PCA axis 1

PCA axis 2

51

24.7

HM

0.72

)0.11

VM CP

0.09 0.81

0.97 )0.35

IS

0.81

)0.07

NPS

0.84

0.39

)0.71

)0.83

)0.74

(<0.001*)

(<0.001*)

(<0.001*)

Multihabitat (n = 20)

% Variance explained

Evaluation criteria

)0.88

)0.95

)0.85

(<0.001*)

(<0.001*)

(<0.001*)

Test

)0.79

)0.72

)0.70

(n = 20)

(<0.001*)

(<0.001*)

(0.001*)

DR PCA1 denotes disturbance rank principal component analysis axis 1. Water PCA1 represents water chemistry principal component analysis axis 1. Human DR is total human disturbance rank score. Values in the table are Pearson correlation coefficient. Sample sizes were 20. Values in ( ) denote probability; * indicates statistical significance.

Variance explained and correlations of PCA axes with human disturbance variables were shown in the table.

Discussion

test data set, the benthic IBC had the highest average correlation with evaluation criteria (r = )0.79, p < 0.001; r = )0.69, p < 0.001; r = )0.68, p < 0.001) and the planktonic IBC was the second (r = )0.71, p < 0.001; r = )0.73, p < 0.001; r = )0.64, p < 0.001). The epiphytic IBC had the largest decrease in Pearson correlation coefficients from the development data set to the test data set. The multiple-habitat IBC had higher correlations with evaluation criteria in both development and test data sets than IBCs from the three specific habitats.

Many attributes of diatom assemblages were related to physicochemistry and human disturbance of the Casco Bay watershed wetlands in both development and test sets of data. These attributes included deviations in taxonomic composition from reference condition and new and existing indicators of water chemistry. Correlations among human disturbance scores, measured water chemistry, and diatom indices of water chemistry showed that human activities alkalized and enriched these wetlands. Diatom assemblages changed from naturally occurring taxa that tolerate low pH and low nutrient conditions to those

202 that require circumneutral to alkaline pH and elevated nutrients. No studies have developed and tested simple and multimetric indices of water chemistry and biological condition in wetlands as this study for the Casco Bay watershed wetlands. Although our sample size was relatively small, we did test metrics with a relatively ‘‘naı¨ ve’’ data set and demonstrated the reliability of indices at spatial and temporal scales that better simulated sources of variability than the bootstrapping approach commonly used to test WAMs. Our results show that existing metrics are transferable across regions and habitats. Ten of the 12 indices based on existing autecological information and species composition of plankton were correlated significantly with THDS. Almost all species level indicators based on Dixit’s and van Dam’s autecological information were significantly correlated with relevant environmental variables. In many cases, index precision was as good or better when calculated with autecological information gathered from existing literature for lakes and streams as when calculated with species autecologies derived in this study. Nineteen of the 30 metrics based on new autecological information were correlated with THDS. Less than 50% of Maine-derived WAMs were statistically significant. Even though indices based on existing autecological data were correlated with relevant environmental factors and THDS as well or better than indices based on autecologies developed in this study, we argue that regional refinement of indicators is very important. Regional differences in population phenotypes and cryptic species (sensu, Sa´ez and Lozano, 2005) likely reduce accuracy of transferring species autecological information among regions. The small sample size of wetlands in this study probably constrained development of precise autecological information. Constraining indicator development to specific types of wetlands would probably increase precision of metrics. More precise measure of environmental factors with relatively high-temporal variability would likely increase precision of metrics. For example, the WAMs developed in this study were more precisely related for Cl and pH than to nutrient concentrations. Nutrient concentrations in shallow waters such as wetlands vary greatly at daily and diurnal temporal scales due to exogenous weather

factors and endogenous metabolic regulation of uptake and release from periphyton, macrophytes, and sediments. Cl probably varies little diurnally. Both Cl and pH probably vary less than nutrients with weather related factors. Weather related temporal variability in nutrient relationships with human disturbance was indicated by higher correlations during 1999 vs. 1998, which was probably due to higher water levels in 1999. Thus, more thorough assessment of water chemistry with repeated visits of wetlands and characterizing biota and chemistry of more wetlands should improve precision and accuracy of WAMs and regionally refined indicators. Metrics from diatom assemblages in all three habitats provided reliable assessments of ecological conditions in wetlands. WAMs and IBCs calculated with species composition of benthic and epiphytic diatoms were more precise than with planktonic diatoms. No justification for using multiple assemblages was detected, such as differing responses of assemblages to contaminants or categories of human activities. Thus benthic or epiphytic diatom assemblages should provide similar characterizations of wetland conditions, and plankton could be used if a slight increase in ease of sampling was more important than the loss in precision of assessments. Down-weighting tolerant species in WAMs and combining multiple metrics into a single multimetric index may be valuable, respectively, for improving precision of some metrics and providing a summary statistic for characterizing wetland health. However, application of these indicator modification techniques should be chosen with appropriate justification. The lack of pattern in whether WA or WATol models were more precise for different types of chemical variables (e.g., ions or nutrients), and more precise in one habitat or another, showed that differences in indicator performance were random. Therefore, down-weighting tolerant species in WAMs may just add another source of variability and may not be worthwhile. In this study, we emphasized the development and testing of indices that were more directly related to measures of biological condition than WAMs of stressor conditions. Metrics such as number of sensitive and tolerant species, similarity of species composition to reference condition, and

203 % individuals of specific genera should be included in multimetric indices of biological condition (Stevenson and Smol, 2002; Wang et al., 2005). They more independently measure biological condition than WAMS of stressor condition. Maintaining independence in measures of biological conditions and measures of pollutants and human activity is important in ecological assessment (Stevenson et al., 2004). If the goals of assessment are to evaluate valued ecological attributes and to diagnose stressors and human activities causing problems, then measures of valued attributes and stressors must be as independent as possible to avoid circularity when arguing cause-effect relationships. Many indicators of biological condition were highly correlated to human disturbance, such as those mentioned above and % Synedra, Nitzschia, Neidium, and Eunotia. The high performance of many metrics of biological condition show that this approach can be used with diatoms as with other organisms. In conclusion, diatom indicators of biological condition and water chemistry of wetlands provide reliable methods for assessing wetland condition and diagnosing potential stressors. This study demonstrated procedures and measures involved in developing a set of tools that assess ecological condition and diagnose potential causes of impairment. Classifying the great diversity of wetland types according to their expected condition and response to stressors should improve precision of assessments. These additional steps can be taken to refine indicator as more data on biological condition, contamination, habitat alteration, and human activities are accumulated for wetlands.

Acknowledgements We thank Kalina Manoylov for providing help on algal taxonomy. Vanessa Lougheed critically reviewed an earlier version of this manuscript. The reviewers, editors and Thomas Danielson provided helpful comments. This research was supported by a contract to process algal samples from the Maine Department of Environmental Protection and to analyze data from the United States Environmental Protection Agency.

References American Public Health Association (APHA), 1998 Standard Methods for the Examination of Water and Wastewater 20th edn. American Public Health Association, Washington, D.C. Bahls, L. L., 1993. Periphyton Bioassessment Methods for Montana Streams. Montana Department of Health and Environmental Sciences, Helena, MT, USA. Barbour, M. T., J. Gerritsen, B. D. Snyder & J. B. Stribling, 1999. Rapid bioassessment protocols for use in wadeable streams and rivers. EPA 841-B-99-002. Office of Water, US Environmental Protection Agency, Washington, D.C., USA. Bendjoudi, H., P. Weng, R. Guerin & J. F. Pastre, 2002. Riparian wetlands of the middle reach of the Seine river (France): historical development, investigation and present hydrologic functioning. A case study. Journal of Hydrology 263: 131–155. Boyle, T. P., G. M. Smillie, G. M. Anderson & D. R. Beeson, 1990. A sensitivity analysis of nine diversity and seven similarity indices. Journal of the Water Pollution Control Federation 62: 749–762. Cairns, J. Jr., 1977. Quantification of biological integrity. In Ballentine, R. K. & L. J. Guarria (eds) . Integrity of Water. US Environmental Protection Agency, Washington, D.C., USA: 171–187. D’Elia, C. F., W. R. Boynton & J. G. Sanders, 2003. A watershed perspective on nutrient enrichment, science, and policy in the Patuxent River, Maryland: 1960–2000. Estuaries 26: 171–185. Dixit, S. S. & J. P. Smol, 1994. Diatoms as indicators in the Environmental Monitoring and Assessment program – surface water (EMAP-SW). Environmental Monitoring and Assessment 31: 275–306. Dixit, S. S., J. P. Smol, D. F. Charles, R. M. Hughes, S. G. Paulsen & G. B. Collins, 1999. Assessing water quality changes in the lakes of the Northeastern United States using sediment diatoms. Canadian Journal of Fisheries and Aquatic Sciences 56: 131–152. Dodds, W. K., 2003. The role of periphyton in phosphorus retention in shallow freshwater aquatic systems. Journal of Phycology 39: 840–849. Drewien, R. C., W. M. Brown & D. S. Benning, 1996. Distribution and abundance of sandhill cranes in Mexico. Journal of Wildlife Management 60: 270–285. Dufrene, M. & P. Legendre, 1997. Species assemblages and indicator species: the need for a flexible asymmetrical approach. Ecological Monographs 67: 345–366. Gale, P. M., K. R. Reddy & D. A. Graetz, 1993. Nitrogen removal from reclaimed water applied to constructed and natural wetland microcosms. Water Environment Research 65: 162–168. Hill, B. H., A. T. Herlihy, P. R. Kaufmann, R. J. Stevenson F. H. McCormick & C. B. Johnson, 2000. Use of periphyton assemblage data as an index of biotic integrity. Journal of the North American Benthological Society 19: 50–67. Jackson, L. E., J. C. Kurtz & W. S. Fisher, 2000. Evaluation guidelines for ecological indicators. EPA/620/R-99/005. Of-

204 fice of Research and Development, U.S. EPA, Research Triangle Park, NC, USA. Juggins S. & C. J. F. ter Braak, 1992. CALIBRATE – a program for species-environment calibration by [weightedaveraging] partial least squares regression. Unpublished computer program, Environmental Research Centre, University College, London. Karr, J. R., 1991. Biological integrity: A long-neglected aspect of water resource management. Ecological Applications 1: 66–84. Kelly, M. G., 1998. Use of the trophic diatom index to monitor eutrophication in rivers. Water Research 32: 236–242. Krammer, K. & H. Lange-Bertalot, 1986. Bacillariophyceae, Teil 1: Naviculaceae. Gustav Fischer Verlag, Stuttgart, Germany. Krammer, K. & H. Lange-Bertalot, 1988. Bacillariophyceae, Teil 2: Bacillariaceae, Epithemiaceae, and Surirellaceae. Gustav Fischer Verlag, Stuttgart, Germany. Krammer, K. & H. Lange-Bertalot, 1991a. Bacillariophyceae, Teil 3: Centrales, Fragilariaceae, and Eunotiaceae. Gustav Fischer Verlag, Stuttgart, Germany. Krammer, K. & H. Lange-Bertalot, 1991b. Bacillariophyceae, Teil 4: Achnanthaceae. Gustav Fischer Verlag, Stuttgart, Germany. Leibowitz, S. G., 2003. Isolated wetlands and their functions: An ecological perspective. Wetlands 23: 517–531. McCormick, P. V. & M. B. O’Dell, 1996. Quantifying periphyton responses to phosphorus in the Florida Everglades: a synoptic-experimental approach. Journal of North American Benthological Society 15: 450–468. Mitsch, W. J. & J. G. Gosselink, 2000. The value of wetlands: importance of scale and landscape setting. Ecological Economics 35S: 25–33. Odum, E. P., 1985. Trends expected in stressed ecosystems. BioScience 35: 419–422. Mack, J. J., 2001. Ohio rapid assessment method for wetlands, manual for using version 5.0. Ohio EPA Technical Bulletin Wetland/2001-1-1. Ohio Environmental Protection Agency, Division of Surface Water, 401 Wetland Ecology Unit, Columbus, OH, USA, 72 pp. Pan, Y. D. & R. J. Stevenson, 1996. Gradient analysis of diatom assemblages in western Kentucky wetlands. Journal of Phycology 32: 222–232. Pan, Y., R. J. Stevenson, P. Vaithiyanathan, J. Slate & C. J. Richardson, 2000. Changes in algal assemblages along observed and experimental phosphorus gradients in a subtropical wetland, U.S.A. Freshwater Biology 43: 339 353. Patrick, R. & C. W. Reimer, 1966. The diatoms of the United States, Volume 1. The Academy of Natural Sciences of Philadelphia, Philadelphia, PA, USA. Patrick, R. & C. W. Reimer, 1975. The Diatoms of the United States, Volume 2. The Academy of Natural Sciences of Philadelphia, Philadelphia, PA, USA. Peterson, C. G., 1996. Response of benthic algal communities to natural physical disturbance. In Stevenson, R. J. M. L. Bothwell & R. L. Lowe (eds), Algal Ecology: Freshwater Benthic Ecosystems. Academic Press, New York, USA: 375–403.

Potapova, M. G., D. F. Charles, K. C. Ponader & D. M. Winter, 2004. Quantifying species indicator values for trophic diatom indices: a comparison of approaches. Hydrobiologia 517: 25–41. Rapalee, G., S. E. Trumbore, E. A. Davidson, J. W. Harden & H. Veldhuis, 2001. Soil carbon stocks and their rates of accumulation and loss in a boreal forest landscape. Global Biogeochemical Cycles 12: 687–701. Round, E. F., 1990. The Diatoms: Biology and Morphology of the Genera. Cambridge University Press, Cambridge, UK. Roux, D. J., P. L. Kempster, C. J. Kleynhans, H. R. Van Vliet & H. H. Du Preez, 1999. Integrating stressor and response monitoring into a resource-based water-quality assessment framework. Environmental Management 23: 15–30. Severo, J. B., E. E. Lopez-Lopez & K. A. B. Stanley, 2002. Spatial and temporal variation patterns of a waterfowl community in a reservoir system of the Central Plateau, Mexico. Hydrobiologia 467: 123–131. Schindler, D. W., 1987. Detecting ecosystem responses to anthropogenic stress. Canadian Journal of Fisheries and Aquatic Sciences 44(Suppl. 1): 6–25. Saez, A. G. & E. Lozano, 2005. Body doubles. Nature 433: 111. Sheehan, P. J., 1984. Effects on community and ecosystem structure and dynamics. In Sheehan, P. J. (ed.), Effects of pollutants at the ecosystem level. Wiley, Chichester, UK: 23–145. Shannon, C. E. & W. Weaver, 1949. The Mathematical Theory of Communication. University of Illinois Press, Urbana, IL, USA. Slate, J. E. & R. J. Stevenson, 2000. Recent and abrupt environmental change in the Florida everglades indicated from siliceous microfossils. Wetlands 20: 346–356. Smith, R. A., G. E. Schwarz & R. B. Alexander, 1997. Regional interpretation of water-quality monitoring data. Water Resources Research 33: 2781–2798. Stevenson, R. J., 1998. Diatom indicators of stream and wetland stressors in a risk management framework. Environmental Monitoring and Assessment 51: 107–118. Stevenson, R. J., P. R. Sweets, Y. Pan & R. E. Schultz, 1999. Algal community patterns in wetlands and their use as indicators of ecological conditions. In McComb, A. J. & J. A. Davis (eds), Proceedings of INTECOL’s Vth International Wetland Conference. Gleneagles Press, Adelaide, Australia: 517–527. Stevenson, R. J. & F. R. Hauer, 2002. Integrating hydrogeomorphic and index of biotic integrity approaches for environmental assessment of wetlands. Journal of the North American Benthological Society 21: 502–513. Stevenson, R. J., Y. Pan & P. Vaithiyanathan, 2002a. Ecological assessment and indicator development in wetlands: the case of algae in the Everglades, USA. Verhandlungen Internationale Vereinigung fu¨r Theoretische und Andgewandte Limnologie 28: 1248–1252. Stevenson, R. J., P. V. McCormick & R. Frydenborg, 2002b. Methods for evaluating wetland condition. #11 Using algae to assess environmental conditions in wetlands. EPA-822-R02–021. U.S. Environmental Protection Agency, Washington, D.C.

205 Stevenson, R. J., B. C. Bailey, M. C. Harass, C. P. Hawkins, J. ALBA-TERCEDOR , C. Couch, S. Dyer, F. A. Fulk, J. M. Harrington, C. T. Hunsaker & R. K. Johnson, 2004. Interpreting results of ecological assessments. In Barbour, M. T. S. B. Norton, H. R. Preston & K. W. Thornton (eds), Ecological Assessment of Aquatic Resources: Linking Science to Decision-making. Society of Environmental Toxicology and Chemistry, Pensacola, FL: 85–111. ter Braak, C. J. F. & H. van Dam, 1989. Inferring pH from diatoms: a comparison of old and new calibration methods. Hydrobiologia 178: 209–223. U.S. EPA (US Environmental Protection Agency), 1998a. 1998 National water quality inventory report to congress. EPA841-F-00-006. Office of Water, US EPA, Washington, D.C., USA 633 pp. U.S. EPA., 1998b. Guidelines for ecological risk assessment. EPA/630/R-95/002F. Office of Research and Development, Risk Assessment Forum, U.S. EPA, Washington, D.C, USA 188 pp. U.S. EPA., 2002. Methods for evaluating wetland condition: developing metrics and indexes of biological integrity. EPA-

822-R-02-016. Office of Water, US EPA, Washington, D.C., USA 45 pp. van Dam, H., A. Mertenes & J. Sinkeldam, 1994. A coded checklist and ecological indicator values of freshwater diatom in the Netherlands. Netherlands Journal of Aquatic Ecology 28: 117–133. Wang, Y.-K. & R. J. Stevenson, 2002. An ecological assessment framework for watershed management. Journal of the Chinese Institute of Environmental Engineering 12: 209–216. Wang Y.-K., R. J. Stevenson & L. Metzmeier, 2005. Development and evaluation of a diatom-based index of biotic integrity for the Interior Plateau Ecoregion, Kentucky, USA. Journal of the North American Benthological Society 24: 990–1008. Washington Department of Ecology (DE), 1993. Washington State Wetlands Rating System, Western Washington, 2nd edn. Publication No. 93–74. Washington DE, Olympia, WA, USA, 69 pp. Yuan, L. L. & S. B. Norton, 2003. Comparing responses of macroinvertebrate metrics to increasing stress. Journal of the North American Benthological Society 22: 308–322.

206 Appendix 1. Correlation coefficients between human disturbance scores of test sites and conditions inferred for test data sites using species optima from the

Habitat

IndVar

HM WA

VM WATol

development data set and species relative abundances from the test data set. Coefficients in bold indicate that p < 0.05 (n = 20 for Epiphyton and Benthos, n = 19 for Plankton).

CP

WA

WATol

WA

IS WATol

WA

NPS WATol

WA

WATol

Benthos

Cl)

0.43

0.38

)0.21

)0.09

0.60

0.94

0.79

0.63

0.31

0.63

Benthos

Cond

0.43

0.39

)0.20

)0.19

0.68

0.85

0.81

0.89

0.81

0.89

Benthos

Alk

0.42

0.33

)0.16

)0.15

0.81

0.94

0.83

0.85

0.83

0.85

Benthos

pH

0.39

0.35

)0.07

)0.11

0.74

0.66

0.78

0.84

0.78

0.84

Benthos Benthos

SO)2 4 )S Si

0.39 0.01

0.02 0.10

)0.21 0.48

)0.15 0.48

0.39 )0.22

)0.02 )0.04

0.67 )0.18

0.06 )0.07

0.67 )0.18

0.06 )0.07

Benthos

NH+ 4 )N

0.39

0.35

)0.18

)0.18

0.82

0.95

0.85

0.84

0.85

0.84

Benthos

NO)3 )N

0.40

0.13

)0.15

0.14

0.36

0.27

0.67

0.55

0.67

0.55

Benthos

TN

)0.04

0.29

)0.22

)0.13

)0.01

0.45

0.05

0.56

0.05

0.56

Benthos

TP

0.33

0.15

)0.21

0.17

0.46

0.33

0.53

0.47

0.53

0.47

Epiphyton

Cl)

0.33

0.43

0.15

)0.20

0.59

0.17

0.44

)0.07

0.44

)0.07

Epiphyton

Cond

0.33

0.46

0.15

)0.06

0.63

0.33

0.49

0.09

0.49

0.09

Epiphyton Epiphyton

Alk pH

0.38 0.26

0.42 0.20

0.13 0.18

)0.08 0.02

0.73 0.60

0.72 0.69

0.62 0.58

0.59 0.59

0.62 0.58

0.59 0.59

Epiphyton

SO)2 4 )S

0.22

0.05

0.16

)0.02

0.55

0.13

0.44

0.08

0.44

0.08

Epiphyton

Si

0.43

0.20

0.37

0.38

0.21

0.04

0.19

)0.11

0.19

)0.11

Epiphyton

NH+ 4 )N

0.40

0.35

0.14

)0.10

0.75

0.40

0.63

0.79

0.63

0.79

Epiphyton

NO)3 )N

0.21

)0.39

0.24

0.49

0.53

)0.19

0.41

0.08

0.41

0.08

Epiphyton

TN

0.05

0.33

0.19

0.23

)0.05

0.31

)0.06

0.28

)0.06

0.28

Epiphyton

TP

0.48

0.16

0.31

)0.01

0.42

0.06

0.41

0.02

0.41

0.02

Plankton Plankton

Cl Cond

0.22 0.22

0.40 0.14

)0.02 0.00

)0.18 )0.19

0.32 0.32

0.84 0.85

0.59 0.59

0.47 0.38

0.59 0.59

0.47 0.38

Plankton

Alk

0.23

0.14

0.07

)0.11

0.34

0.91

0.57

0.76

0.57

0.76

Plankton

pH

0.18

0.25

0.19

)0.13

0.29

0.61

0.58

0.83

0.58

0.83

Plankton

SO)2 4 )S

0.20

)0.28

)0.05

0.04

0.28

0.25

0.58

0.28

0.58

0.28

Plankton

Si

0.52

0.47

0.01

0.18

0.58

0.47

0.58

0.47

0.58

0.47

Plankton

NH+ 4 )N

0.13

0.28

0.12

)0.18

0.23

0.39

0.47

0.79

0.47

0.79

Plankton

NO)3 )N

0.19

0.21

)0.04

)0.08

0.29

0.19

0.60

0.65

0.60

0.65

Plankton Plankton

TN TP

0.08 0.37

)0.17 0.18

0.07 )0.06

0.19 )0.15

0.01 0.35

)0.31 0.35

)0.17 0.29

)0.37 0.26

)0.17 0.29

)0.37 0.26

Hydrobiologia (2006) 561:207–219  Springer 2006 R.J. Stevenson, Y. Pan, J.P. Kociolek & J.C. Kingston (eds), Advances in Algal Biology: A Commemoration of the Work of Rex Lowe DOI 10.1007/s10750-005-1615-1

Diatom assemblages and their associations with environmental variables in Oregon Coast Range streams, USA Christine L. Weilhoefer* & Yangdong Pan Environmental Sciences and Resources, Portland State University, Portland, Oregon, USA (*Author for correspondence: E-mail: [email protected])

Key words: periphyton, nonmetric multidimensional scaling (NMDS), cluster analysis, TWINSPAN, bioassessment, Oregon Coast Range

Abstract The Oregon Coast Range, rich in natural resources, is under increasing pressure from rapid development. The purpose of this study was to examine diatom species patterns in relation to environmental variables in streams of this region. Diatoms, water quality, physical habitat and watershed characteristics were assessed for 33 randomly selected stream sites. Watershed size, elevation, geology, vegetation and stream morphology varied substantially among sites. Streams were characterized by dilute water chemistry and a low percent of fine substrate. A total of 80 diatom taxa were identified. Taxa richness was low throughout the region (median 15, range 10–26). Assemblages were dominated by two adnate species, Achnanthidium minutissimum and Achnanthes pyrenaicum. Diatoms sensitive to organic pollution dominated the assemblages at all sites (median 85%). Non-metric multidimensional scaling (NMDS) and correlational analysis showed quantitative relationships between diatom assemblages and environmental variables. NMDS axes were significantly correlated with watershed area, watershed geology, conductivity, total nitrogen, total solids and stream width. Diatom-based site classification (Two-way Indicators Species Analysis, (TWINSPAN)) yielded 4 discrete groups that displayed weak correlations with environmental variables. When stream sites were classified by dominant watershed geology, overall diatom assemblages between groups were significantly different (Analysis of Similarity (ANOSIM) global R = 0.19, p < 0.05). Our results suggest that streams in the coastal region are in relatively good condition. High natural variability in stream conditions in the Oregon Coast Range ecoregion may obscure quantitative relationships between environmental variables and diatom assemblages. A bioassessment protocol that classifies sites by major landscape variables and selects streams along the major human disturbance gradient might allow for detection of early signs of human disturbance in environmentally heterogeneous regions, such as the Pacific Northwest.

Introduction Streams are a distinctive feature in the Oregon Coast Range ecoregion, with a typical density of 1–2 km of perennial streams per square km in the mountains (Omernik & Gallant, 1986). The declining status of stream biota, particularly salmonid fish, has been of interest in recent years,

resulting in the implementation of several federal, state and local programs to assess the status of these streams (Hansmann & Phinney, 1973; Fore et al., 1996; Ford & Rose, 2000). One aspect of many stream assessment plans is the use of aquatic biota as indicators of stream conditions. In Oregon, stream bioassessment is commonly performed using macroinvertebrate assemblages and fish, but

208 the utility of stream periphyton assemblages is only beginning to be explored (e.g., Carpenter & Waite, 2000). Periphyton assemblages have frequently been used in stream bioassessment because they respond rapidly to changes in stream conditions (Leland, 1995; Pan et al., 1996; 2000). Streams in the Coast Range ecoregion have naturally dilute water chemistry, with low alkalinity and nutrient levels, due to regional bedrock (Welch et al., 1998). Consequently, their periphyton may be highly sensitive to anthropogenic disturbances and serve as a good bioindicator of stream conditions. For example, periphyton respond positively to nutrient enrichment in mountain streams of British Columbia (Perrin et al., 1987). The major anthropogenic disturbances in the Coast Range are activities associated with forest management practices (McClain et al., 1998). However, the watersheds vary widely in elevation, size, geology and vegetation cover, leading to a high natural variability in water chemistry and physical habitat conditions. This high natural variability in stream conditions may interfere with the use of diatoms in bioassessment by obscuring relationships between pollution or habitat alterations and species assemblages. The main objective of this study was to characterize diatom species composition in streams of the Oregon Coast Range ecoregion. The diatom assemblages of these streams were then related to environmental variables, including water chemistry, physical habitat, watershed vegetation and watershed geology, in an attempt to explain the observed patterns. A better understanding of the role of natural landscape variability in influencing diatom assemblages may improve the utility of diatom-based bioassessment in the region.

Methods

annual precipitation ranges between 140 and 320 cm (Wigington et al., 1998), the heaviest typically occurring from October through March. The natural vegetation of the area is a mixture of Sitka spruce (Picea sitchensis Carr.), occurring below 150 m in elevation, and western hemlock (Tsuga heterophylla Sargent), occurring at higher elevations (Frenkel, 1993). The bedrock geology of the study area is a mixture of marine tuffaceous sandstones and shales and basaltic volcanic rocks with unrelated intrusives (Rosenfeld, 1993). Land use/land cover in the area is predominately forested in the uplands, much of which is managed commercially for timber harvest, and a mixture of dairy farmland, cropland and orchards in the lowlands. Much of the population residing in the Coast Range is confined to small towns and a few larger urban areas on the coast. Data were collected as part of Oregon DEQ’s Oregon Plan for Salmon and Watersheds. A stratified, random probability design was used to select 1st–3rd order, wadeable streams throughout Oregon (Herlihy et al., 2000). For this study, the 35 sites that fell within the range of the Coastal Landscape Analysis and Modeling Study (CLAMS; Oregon State University College of Forestry, U.S. Forest Service and Oregon Department of Forestry joint study) were utilized. These sites were selected so that watershed geology and land use coverages for a geographic information system (GIS) could be generated. After preliminary analysis of diatom species, two sites were excluded from the study because their assemblages were dominated by one species, Achnanthes oblongella Oestrup, which was not present at any of the other sites (n = 33; Figure 1). Nine sites were designated as ‘‘reference sites’’ a priori by Oregon DEQ. Reference sites were defined as streams with minimally impacted watersheds (Mrazik, 1999). Criteria used to select reference sites included professional opinion and the percentage of roads in the stream’s watershed.

Study area and site selection Field sampling The Oregon Coast Range ecoregion (Level III Ecoregion, Omernik & Gallant, 1986) extends along the coast from the Pacific Ocean to the Willamette Valley on the east side of the Coast Range and is characterized by rugged mountains inland and flat plains along the coast. Average

All sites were sampled once for biological, chemical and physical variables during June through September of 1999. The sampling unit was the stream reach, defined as 40 times the average wetted width measured at 3 locations within 10 m

209

Figure 1. Sampling locations and streams (5th order or higher) in the Oregon Coast Range ecoregion.

of the stream coordinates selected by the random probability design. The reach was divided into 11 equally spaced transects parallel to the ends of the reach, including 1 at either end of the reach. Stream physical habitat characterization included channel morphology, substrate composition, riparian area condition and discharge. Thalweg depth and wetted width were measured at each transect and the mean of these values was used to represent the depth and width of the reach. Instantaneous stream discharge was measured using a flow meter (Marsh-McBirney ‘‘Flo-mate 2000’’). Fine substrate (<2 mm) at each site was assessed by placing a grid on the streambed at 2

random locations and counting the number of intersections covered by fine sediment particles. Qualitative assessment of riparian and stream conditions were made according to the EPA’s Rapid Bioassessment Protocol (Barbour et al., 1999). Within each reach, conductivity, dissolved oxygen and stream temperature were measured with a YSI Model 85 m. Turbidity was measured using a HACH Model 2100P Turbidimeter and pH was measured using an Orion Model 210A meter. Water samples were collected in the middle of the stream for nutrient analyses. Two water samples were taken at each site, 1 filtered on site

210 (47 mm Millipore type HA filters, 0.45 lm pore size), and the other left unfiltered and stored on ice until returned to the laboratory for nutrient analysis. Water samples were analyzed for concentrations of total and dissolved nutrients according to standard EPA methods (Clesceri et al., 1998). Periphyton were sampled by randomly selecting 10 cobbles within riffle habitats of the stream sampling reach. A known area (7.1 cm2) of each rock was scraped using a toothbrush and delimiter and combined into 1 composite sample per site. The composite sample was homogenized and split into 3 subsamples for identification and enumeration of diatom species and other assays. Samples for diatom identification and enumeration were preserved with formalin for a final formalin concentration of 4%.

Sample and data analysis For diatom slide preparation, samples were rinsed repeatedly with deionized water to remove excess formalin. Samples were homogenized and digested using concentrated sulfuric acid and potassium dichromate for 12 h. Samples were rinsed repeatedly with deionized water until the pH was approximately neutral and then mounted on slides with Naphrax, a high resolution mounting medium. Slides were scanned using transects until at least 500 diatom valves were identified and enumerated to the species level using a Nikon Eclipse E600 microscope at 1000 magnification. The primary references for diatom taxonomy were Krammer & Lange-Bertalot (1986; 1988; 1991a; 1991b; 2000) and Patrick & Reimer (1966, 1975). Diatom autecological indices were calculated for each site based on the pollution tolerance values of Lange-Bertalot (1979) and Bahls (1993). Geologic categories and vegetative cover within the watershed upstream of each sample reach were quantified using a GIS with ArcInfo version 8 and ArcView version 3.2 software. The spatial analyst extension of ArcInfo was used to delineate the watershed upstream from each sampling reach. Ten-meter digital elevation models (10 m resolution USGS; CLAMS, 1996) were used and sampling site coordinates were used as the outlet point for each watershed. Vegetation within each watershed was quantified from the 1996 Gradient

Nearest Neighbor Vegetation Class developed by CLAMS (25 m grid, TM satellite imagery and plot data; CLAMS, 1996). Vegetation data were grouped into three major forest classes: broadleaf, conifer forest and mixed forest (broadleaf and conifer). Major lithologic categories within each watershed were quantified from the 1991 Geologic Map of Oregon (1:500,000 coverage; Walker & MacLeod, 1991) using overlay analysis in ArcInfo. Geologic types were grouped into major lithologic categories (i.e., volcanics, sedimentary, alluvium) using guidelines in Johnson & Raines (1995). Diatom assemblages were represented as the proportion of total species at each site (relative abundances). In an effort to reduce the influence of ‘‘rare’’ species in the data set on the data analyses, only species occurring with greater than 1% relative abundance at three or more sites were included in data analyses (n = 28). Stream sites were ordered based on diatom species composition using non-metric multidimensional scaling techniques (NMDS). NMDS was performed using PC-ORD v.4 (Bray-Curtis distance measure, 40 real runs and 400 maximum iterations). The Monte-Carlo permutation procedure was used to determine if the axes extracted by NMDS explained more variation than by chance alone (PC-ORD; 50 randomized runs). An Analysis of Similarity (ANOSIM, PRIMER-4, Bray–Curtis distance measure, 999 permutations; Clarke & Green, 1988) was used to test differences in diatom species composition between reference and randomly selected sites. Measured environmental variables were correlated to the axes of the NMDS to identify a subset of environmental variables that co-vary with changes in diatom species composition among streams. Environmental variables that were not normally distributed were log10 or arcsine-square root transformed (proportional data) to produce near-normal distributions (Zar, 1999). Relations of environmental variables to diatom species composition were examined by correlating variables with NMDS axes scores. To examine the effects of classifying sites based on watershed geology on the relationships between diatom assemblages and environmental variables, sites were divided into 2 groups based on their dominant watershed geology. The watersheds of 21 sites had greater than 50% sedimentary rocks and these

211 sites were considered to have sedimentary rock dominated watersheds for analysis while the remaining 12 sites were considered to have nonsedimentary dominated watersheds. Rock types other than sedimentary and volcanic (primarily alluvial/unconsolidated material) were found at 6 sites, with the relative abundance ranging from <1 to 9%. To simplify data analysis, these rock types were grouped with volcanic-dominated sites. Average % sedimentary rock in the sedimentarydominated group was 91% while average % volcanic rock in the volcanic-dominated group was 78%. The % sedimentary rock and % volcanic/ other rock varied significantly between the two groups (ANOVA p < 0.001, in both cases). An ANOSIM was used to detect difference in diatom assemblages between the 2 geology-based groups (ANOSIM, PRIMER-4, Bray–Curtis distance measure, 999 permutations). Differences in environmental variables and relative abundances of dominant diatom species between geology-based site groups were examined with t-tests. Two-way indicator species analysis (TWINSPAN, PC-ORD), a divisive cluster method using a top–down approach, was used to classify stream sites into several groups based on diatom relative abundance data (Hill et al., 1975). To determine the number of TWINSPAN groups to retain, differences in species composition between the groups were tested using ANOSIM (PRIMER-4; Bray–Curtis similarity measure, 999 permutations). A separate ANOSIM was performed on the groups created by the first 4 TWINSPAN breaks. Groups were retained if the probability of sites being members of a TWINSPAN specified group was greater than the probability of sites being members of randomly created groups. A set of indicator species for each of the retained TWINSPAN groups was characterized using Indicator Species Analysis using PC-ORD (Dufrene & Legendre, 1997). Statistical significance of each species indicator value was tested using a MonteCarlo permutation test (999 permutations, p<0.05). Differences in environmental variables and diatom indices between TWINSPAN groups were detected using an Analysis of Variance (ANOVA, SigmaStat v. 1.00, Jandel Scientific). Multiple comparisons between TWINSPAN groups were performed with a Student–Newman– Keuls test using a Bonferroni-corrected alpha.

Results Environmental factors Watershed area and sampling point elevation varied widely among sites. Watershed area ranged between 1 and 169 km2 (median = 15 km2; Table 1). Elevation ranged between 3 and 707 m (median = 171 m). Sedimentary rocks were present in the watersheds of all but 3 sites and 13 watersheds were composed entirely of sedimentary rock. Rock of volcanic origin was present in the watersheds of 18 sites and only 3 watersheds were composed entirely of volcanic rocks. Sites with volcanic-dominated watersheds tended to be in the northern part of the Coast Range ecoregion. Conifer forest was the most common vegetation type within the watersheds (median coverage = 63%; Table 1). Mixed forest covered between 1 and 36% of the watersheds, while coverage by broadleaf vegetation ranged between 0 and 25%. Spatial vegetation coverages were similar between the sedimentary and volcanicdominated watersheds, however, the age structure for conifer forest is unknown. Correlations between watershed vegetation type (conifer, mixed and broadleaf forest) and water quality, physical habitat and watershed geology variables were weak overall (|r| < 0.5). Stream morphology varied substantially among sites. Mean bankful width varied between 3 and 52 m, with a median of 9 m. Mean thalweg depth ranged between 7 and 37 cm. Channel slope was low overall all, but varied between 0 and 29% of mean reach length (Table 1). Riparian canopy cover ranged between 13 and 88%. In general, the percent of fine substrate was low at most sites (median 8%); only 2 sites had greater than 50% percent fine substrate. Percent of fine substrate was the only environmental variable that was significantly different between sedimentary- and volcanic-dominated watersheds (median = 11 c.f. 2%; p = 0.003). The streams in the study had low stream ionic strength and nutrient levels (Table 1). Median total nitrogen (TN) concentration was 0.4 mg l)1 and median dissolved inorganic nitrogen (DIN) was 0.2 mg l)1. Both total phosphorus (TP) and soluble reactive phosphorus (SRP) concentrations ranged from below detection limits (0.01 mg l)1)

212 Table 1. Minimum, median and maximum values for landscape, water chemistry and physical habitat variables for Oregon Coast Range streams (n=33) Minimum

Median

Maximum

Watershed area (km2) Elevation (m)

1 3

15 171

169 707

Sedimentary rock in watershed (%)

0

75

100

Volcanic rock in watershed (%)

0

25

100

Conifer forest in watershed (%)

23

63

98

Broadleaf forest in watershed (%)

0

9

25

Mixed forest in watershed (%)

1

12

36

Alkalinity (mg CaCO3/l)

6

17

45

Conductivity (lS/cm) Total Nitrogen (mg/l)

38 0.2

66 0.4

146 0.8

Nitrate + Nitrite (mg/l)

0.0

0.2

0.6

Total Phosphorus (mg/l)

0.01

0.02

0.10

Soluble Reactive Phosphorus (mg/l)

0.01

0.01

0.02

Total Solids (mg/l)

41

56

93

Bankful width (m)

3

9

52

Thalweg depth (cm)

7

15

37

Slope (% of mean reach) Fine substrate (%)

0 0

2 8

29 69

Riparian canopy coverage (%)

13

51

88

to 0.1 mg l)1 (TP) and 0.02 mg l)1 (SRP). Alkalinity, conductivity and total solids varied throughout the study area, but were low overall. Alkalinity ranged between 6 and 45 mg CaCO3 l)1, conductivity ranged between 38 and 146 lS cm)1 and total solids ranged between 41 and 93 mg l)1. Conductivity, alkalinity and total solids were positively correlated for all sites (r > 0.6 for all). Diatom assemblages A total of 80 diatom species, from 22 different genera, were identified from the 33 sites. Nitzschia, Navicula and Achnanthes/Achnanthidium were the most common genera with 13, 10 and 9 species, respectively. Overall species richness was low, ranging between 10 and 26 species per site (median = 15; Table 2). Shannon diversity index values ranged from 0.9–2.4 (median = 1.8; Table 2). Over 50% of the total diatoms counted were from the genus Achnanthes/Achnanthidium. Achnanthidium minutissimum (Ku¨tz.) Czarnecki was present at all sites and was the most abundant species throughout the study area (median relative abun-

dance = 19%, range = 3–64%; Table 2). Achnanthes pyrenaicum (Hustedt) Kobayasi was also common. Diatoms sensitive to organic pollution dominated the assemblages at all sites (median = 85%). Overall siltation index was low; however, there were a few sites with high siltation index values (median = 8%, range = 0–82). A three-dimensional solution was obtained for the NMDS ordination that explained 87% of the variance in the diatom distance matrix (Figure 2). Seventy percent of the variance was partitioned between the second (r2 = 0.28) and third (r2 = 0.42) axes. Axis 1 explained 17% of the variance in the diatom data. The axes explained significantly more variance than would be expected by chance based on Monte-Carlo permutation tests (p = 0.03). Axis 3 was primarily driven by relative abundance of A. pyrenaicum (r = )0.69; Table 3) and secondarily by N. inconspicua Grunow (r = 0.55) and Rhoicosphenia abbreviata (Agardh) Lange-Bertalot (r = 0.61). Axis 2 was driven primarily by the relative abundances of A. minutissimum (r = 0.79) and Cocconeis placentula Ehrenberg (r = )0.56). Axis 1 was driven by the relative abundance of Cocconeis

213 Table 2. Relative abundance of the most common diatom species and diatom autecological indices values (median and ranges) for all sites and sites grouped by dominant watershed geology. Bold values indicate significant differences between the two watershed geology groupings. Pollution sensitivity values are from Lange-Bertalot (1979). Siltation index values are from Bahls (1993) All Sites (n=33)

Sedimentary-dominated (n=21)

Volcanic-dominated (n=12)

Achnanthidium minutissimum

19 (3–64)

15 (4–64)

38 (3–64)

Achnanthidium pyrenaicum

13 (0–64)

21 (0–64)

5 (0–47)

Cocconeis placentula

6 (0–52)

11 (0–46)

1 (0–52)

Nitzschia inconspicua

3 (0–34)

2 (0–19)

10 (0–34)

Rhoicosphenia abbreviata

2 (0–32)

2 (0–32)

3 (0–32)

Shannon diversity

1.8 (0.9–2.4)

1.8 (0.9–2.4)

1.5 (1.2–2.0)

Species richness % Pollution sensitive taxa

15 (10–26) 85 (24–99)

17 (10–25) 85 (28–99)

14 (10–26) 84 (24–97)

Siltation Index

8 (0–82)

4 (0–37)

15 (0–82)

width positively correlated to axis 1 (r = 0.52 and 0.42, respectively; p<0.05 in both cases; Table 3). Conductivity and total solids positively correlated to the 3rd axis (conductivity, r = 0.52; total solids, r = 0.50). Total nitrogen was correlated to axes 1 and 2 (r = )0.39 and 0.41, respectively). Watershed vegetation was not significantly correlated to any of the ordination axes. Watershed geology correlated to the 2nd ordination axis, with % volcanic rock correlating positively (r = 0.58; Figure 2b) and % sedimentary rock correlating negatively (r = )0.57). The ANOSIM performed

placentula (r = 0.59) and Planothidium lanceolatum (Bre´bisson ex Ku¨tzing) Lange-Bertalot (r = )0.60). Diatom assemblages in reference and randomly selected sites did not separate along the NMDS axis (Figure 2a). In addition, diatom assemblages were not significantly different between reference and randomly selected sites (ANOSIM global R = 0.13, p = 0.09). Correlations of environmental variables with site locations along the ordination axes revealed gradients associated with changes in species composition. Stream watershed area and bankful

Table 3. Pearson’s correlation coefficients (r) for environmental variables and common diatom species with significant correlations to NMDS axes. Bold values indicate significant correlations (a = 0.05) Axis 1

Axis 2

Axis 3

Watershed area (km2)

0.52

0.08

)0.31

Elevation (m)

0.00

)0.35

0.24

Sedimentary rock in watershed (%) Volcanic rock in watershed (%)

)0.07 0.05

)0.57 0.58

0.00 0.03

Conductivity (lS/cm)

)0.26

)0.15

0.52

Total Nitrogen (mg/l)

)0.39

0.41

0.17

Nitrate + Nitrite (mg/l)

)0.34

0.52

0.08

Total Solids (mg/l)

)0.24

)0.21

0.50

Bankful width (m)

0.42

0.11

)0.30

Slope (% of mean reach)

)0.38

)0.04

0.44

Achnanthidium pyrenaicum Achnanthidium minutissimum

0.37 )0.12

)0.48 0.79

)0.69 )0.40 0.25

Cocconeis placentula

0.59

)0.56

Nitzschia inconspicua

)0.22

0.46

0.55

Planothidium lanceolata

)0.60

)0.35

0.37

Rhoicosphenia abbreviata

)0.44

)0.02

0.61

(a)

Axis 3

214

A

Group II :

Achnanthidium minutissimum Nitzschia fonticola Nitzschia frustulum

Group IIIB: Achnanthidium pyrenaicum

Group IVAC:

Achnanthes bioretti Eunotia pectinalis Navicula cryptocephala Planothidium lanceolata

Group IC:

Amphora pediculus Rhoicosphenia abbreviata

Axis 2

Axis 3

(b)

Figure 3. Dendrogram illustrating results for TWINSPAN classification of sites based on diatom species assemblages. Groups were retained if ANOSIM results were significant. Superscripts indicate significant differences in species assemblages among TWINSPAN groups. Indicator species for each TWINSPAN group are the result of the Indicator Species Analysis.

Axis 2

Figure 2. Non-metric multidimensional scaling ordination plot of sites based on diatom species assemblages (sites in species space). (a). Open circles are randomly selected sites and closed circles are reference sites. (b) Open circles are sedimentary-dominated watersheds and closed circles are volcanic-dominated watersheds.

N. frustulum (Ku¨tzing) Grunow. Group III was characterized by a single species, A. pyrenaicum. Group IV was characterized by Achnanthes bioretti Germain, Planothidium lanceolata, Eunotia pectinalis (Dillwyn) Rabenhorst, and Navicula cryptocephala Ku¨tzing. Relationships between landscape level environmental variables and TWINSPAN groups were not clear (Table 4). Group III sites had significantly lower total nitrogen levels (p = 0.005). Groups IV sites had the highest percent of fine substrate (p = 0.004). There were no significant differences in diatom metrics among groups.

Discussion on diatom assemblages between the 2 geologic groups was significant (global R = 0.19, p <0.05). The relative abundance of A. minutissimum was significantly greater at sites with volcanic-dominated watersheds ( p = 0.01), while that of A. pyrenaicum was significantly lower ( p = 0.05). TWINSPAN produced four groups with statistically significant site membership based on ANOSIM results (Figure 3). Indicator species analysis showed that Group I was characterized by Amphora pediculus (Ku¨tzing) Grunow and R. abbreviata. Group II was characterized by A. minutissimum, Nitzschia fonticola Grunow and

The strength of utilizing a random probability sampling design is that it allows for extrapolation to regional patterns of stream condition. Diatom species assemblages and autecological metrics suggest overall the Coast Range streams have relatively high water quality. The median percent of pollution sensitive species was 85%. Assemblages were not significantly distinct between reference sites and randomly selected sites (ANOSIM, p = 0.09). Diatom richness was low throughout the region, similar to findings in other low-impacted basins of the Coast Range (Naymik et al., 2005). The dominance of stream assemblages by Achnanthidium minutissimum and

215 Table 4. Landscape, water quality, physical habitat variables and diatom metrics for diatom-based TWINSPAN groups for Oregon Coast Range streams. Superscripts indicate significant differences among groups (ANOVA, alpha = 0.05). Pollution sensitivity values are from Lange-Bertalot (1979) Variables

TWINSPAN Group I (n=6)

II (n=6)

III (n=18)

IV (n=3)

Watershed area (km2)

11

6

17

33

Elevation (m)

27a

320b

175b

207b

Watershed sedimentary rock (%)

100

28

83

100

Watershed volcanic rock (%)

0

72

15

0

Alkalinity (mg CaCO3/l)

28a

14ab

17b

14ab

Conductivity (lS/cm) Nitrate + Nitrite (mg/l)

82 0.3

66 0.2

63 0.1

64 0.5

Total Nitrogen (mg/l)

0.5ab

0.4ab

0.3a

0.7b

Total Solids (mg/l)

59

55

53

62

Bankful width (m)

7ab

15a

10a

4

Thalweg depth (cm)

9ab

23a

16ab

10b

Slope (% of mean reach)

4a

1b

1b

1ab

Fine substrate (%)

5a

2a

10a

45b

Riparian canopy coverage (%) Species richness

49 18

50 15

56 14

18 20

Pollution sensitive taxa (%)

84

84

87

76

A. pyrenaicum may indicate waters with dilute water chemistry (Van Dam et al., 1994; Potapova & Charles, 2003). In addition, the widespread distribution of these adnate species may indicate the importance of the physical environment in controlling biotic assemblages. Achnanthes/Achnanthidium species are often small in size and associated with shaded headwater streams (Steinman & McIntire, 1986). Although diatoms have been shown to respond to environmental factors over several scales, quantitative relationships between diatom assemblages and environmental variables were relatively weak in this study. Random probability sampling reflects regional complexity, however it may not capture the variance of human disturbance adequately. Diatom-based assessment has proven successful in areas with high variance of human disturbance. Strong relationships between environmental variables and diatom assemblages were found in streams covering a gradient of acid mine drainage impacts (Verb & Vis, 2000). Diatom assemblages were different between reference streams, streams in golf courses under construction and streams in operational golf courses

b

spanning a nutrient enrichment gradient (Winter et al., 2003). Diatom-based bioassessment in minimally impacted areas, such as the Oregon Coast Range, may be improved with careful site selection that ensures the coverage of the entire human disturbance gradient of interest. The impacts of logging activities on stream periphyton were detected in Oregon Coast Range streams selected to cover the entire gradient of logging activities (Naymik et al., 2005). The effectiveness of a biodindicator is based on its ability to separate the human disturbance signal from natural environmental variability. In regions where high natural variability may obscure the human disturbance signal, diatom-based bioassessment may require a more sophisticated approach. One possible approach is to first classify sites into relatively homogeneous groups to minimize confounding factors not of interest in the assessment. There are two main classification methods employed in bioassessment: (1) landscape classification (e.g., ecoregions) and (2) biota-based classification (e.g., River Invertebrate Prediction and Classification System (RIVPACS); Wright et al., 1993). The ecoregion approach defines areas

216 of relatively homogeneous ecosystems based on landscape factors (soils, vegetation, climate, geology and physiography; Omernik & Gallant, 1986). The use of ecoregions to classify sites has produced mixed results. Invertebrate communities of New Zealand streams differed significantly among ecoregions (Harding et al., 1997). In Oregon, fish and macroinvertebrates differed significantly only between mountain and valley ecoregions (Whittier et al., 1988). Pan et al. (2000) also found significant differences in diatom assemblages only between montane and valley ecoregions. The weak relationships between diatom assemblages and environmental variables found in this study point to the fact that Level III ecoregion classification might be too coarse to detect the human disturbance signal in this moderately impacted area. The Oregon Coast Range ecoregion encompasses both the coastal mountain range and the flat, coastal lowlands with a mixture of geology, soil, vegetation and landuse/land cover (Clarke et al., 1991). In this study, when sites were stratified by dominant watershed geology, relationships between diatom assemblages and environmental variables became stronger. Diatom assemblages were significantly different between sedimentary and volcanic-dominated watersheds. In addition, the relative abundance of dominant species, A. minutissimum and A. pyrenaicum were significantly different between the two groups. A major criticism of the ecoregion classification approach is that it places too much emphasis on landscape features. In summarizing several studies on the relationship between landscape classifications and stream biota, Hawkins et al. (2000a) found that the amount of variance explained by landscape variables was low. The ecoregion approach is non-hierarchical and therefore smaller-scale variation (e.g., reach-level) is overlooked. In a study of stream macroinvertebrates, reachscale physical properties were more predictive of species assemblages than were landscape-scale properties (Richards et al., 1997). In regions with high environmental heterogeneity, biota-based stream classification may delineate discrete stream types with more clear relationships between biota and environmental conditions. The RIVPACS assessment approach first classifies sites based on biota and then uses a predictive model based on environmental variables to provide a list of taxa to

be expected in the absence of human disturbance (Wright et al., 1993). The ratio of observed taxa to expected is an estimate of the stream condition. The RIVPACS bioassessment approach using macroinvertebrates has been successful in montane streams in California (Hawkins et al., 2000b). A RIVPACS model using diatoms in Australian streams detected differences in predicted values at test sites (Chessman et al., 1999). The RIVPACS approach is data intensive and the sample size in our study was too small to employ this approach. In an attempt to explore the utility of a biotabased classification in the Oregon Coast Range ecoregion, we used TWINSPAN to classify sites into 4 discrete groups. Although diatom assemblages were significantly different between groups, relationships with landscape level variables were not apparent. It may be that land cover and geology characterized at the watershed level are at a coarser scale than stream biota respond. To improve diatom-based bioassessment in environmental heterogeneous regions more sophisticated in-stream sampling and analytical approaches may be required. The sampling design used in this study randomly selects and composites periphyton samples from the stream reach. Several studies have demonstrated that benthic algal assemblages are spatially structured, varying with habitat features, such as substrate and current velocity (Stevenson & Hashim, 1989; Sabater et al., 1998; Passy 2001). A sampling design that targets dominant stream habitats might maximize the signal to noise ratio in the data. The statistical treatment of cosmopolitan and rare taxa might also improve the detection of patterns. Hawkins & Vinson (2000) partially attributed the weak relationships between invertebrates and environmental variables to cosmopolitan taxa masking real differences among sites and rare taxa contributing little ecological information. Four of the five most common taxa in this study were among the most commonly occurring taxa in U.S. rivers (Potapova & Charles, 2002). The dominance of Coast Range diatom assemblages by cosmopolitan taxa might mask disturbance patterns. The treatment of rare taxa also has implications for bioassessment. In reviewing the effects of rare taxa in bioassessment, Cao et al. (2001) concluded that the inclusion/ exclusion of rare taxa appeared to relate to the sensitivity of the bioassessment, particularly in

217 studies covering small spatial extents. Hawkins et al. (2000b) found that rare taxa responded positively to impacts and their inclusion in data analysis improved precision and sensitivity of the bioassessment. The Bray–Curtis distance measure used in our study is insensitive to rare taxa. However, we did not include rare taxa in our analysis due to taxonomic uncertainty of taxa encountered only a few times. This resulted in the exclusion of 35% of the taxa encountered in the study. Conventional fixed-count methods (counting 500–600 diatom valves per sample) may not be adequate to characterize rare species. Employment of a stratified counting method that stops counting dominant taxa after their abundance stabilizes but continues to count rarer taxa until taxonomic precision is established is one possible approach to adequately identify and enumerate rare species. In summary, diatom-based assessment suggests that Coast Range ecoregion streams are in good condition overall. The high natural variability in stream condition and short human disturbance gradients in this region might obscure the patterns between diatoms and environmental variables. In minimally impacted, complex landscapes, diatombased bioassessment might benefit from: (1) classifying sites by either major landscape variables (i.e., geology) or stream biota, (2) selecting sites that cover the entire human disturbance gradient of interest, (3) sampling dominant in-stream habitat and (4) careful statistical treatment of rare and cosmopolitan taxa. Acknowledgements Stream sampling was conducted by the Oregon Department of Environmental Quality as part of the Salmon Plan for Oregon Watersheds. We thank Rick Hafele, Mike Mulvey, Doug Drake, Paul Gill and Shannon Hubler for their assistance. Watershed geology and landscape calculations would not have been possible without the CLAMS dataset. Jesse Naymik and Brian Bowder provided considerable assistance with GIS calculations. An EPA STAR Graduate Fellowship provided funding for C. L. Weilhoefer during the writing of this manuscript. The comments of Jan Stevenson and an anonymous reviewer greatly improved the quality of the manuscript.

References Bahls, L. L., 1993. Periphyton bioassessment methods for Montana streams. Water Quality Bureau, Department of Health and Environmental Sciences, Helena, Montana., 23 pp. Barbour, M. T., J. Gerritsen, B. D. Snyder & J. B. Stribling, 1999. Rapid Bioassessment Protocols for Use in Streams and Wadeable Rivers: Periphyton, Benthic Macroinvertebrates and Fish, 2nd edn. Report EPA 841-B-99-002. U. S. Environmental Protection Agency, Office of Water, Washington, DC. Cao, Y., D. P. Larsen & R. St-J. Thorne, 2001. Rare species in multivariate analysis for bioassessment: some considerations. Journal of the North American Benthological Society 20: 144–153. Carpenter, K. D. & I. R. Waite, 2000. Relations of habitatspecific algal assemblages to land use and water chemistry in the Willamette Basin, Oregon. Environmental Monitoring and Assessment 64: 247–257. Chessman, B., I. Growns, J. Currey & N. Plunkett-Cole, 1999. Predicting diatom communities at the genus level for the rapid biological assessment of rivers. Freshwater Biology 41: 317–331. CLAMS, 1996. Coastal Landscape Analysis and Modeling Study. www.fsl.orst.edu/clams. Clarke, K. R. & R. H. Greene, 1988. Statistical design and analysis for a ‘biological effects’ study. Marine Ecology Progress Series 46: 213–222. Clarke, S. E., D. White & A. L. Schaedel, 1991. Oregon, USA, ecological regions and subregions for water quality management. Environmental Management 15: 847–856. Clesceri, L. S., A. E. Greenberg & A. D. Eaton, 1998. Standard Methods for the Examination of Water and Wastewater. American Public Heath Association, Washington, DC 1325 pp. Dufrene, M. P. & P. Legendre, 1997. Species assemblages and indicator species: the need for a flexible asymmetrical approach. Ecological Monographs 67: 345–366. Ford, J. & C. E. Rose, 2000. Characterizing small subbasins: a case study from coastal Oregon. Environmental Monitoring and Assessment 64: 359–377. Fore, L. S., J. R. Karr & R. W. Wisseman, 1996. Assessing invertebrate responses to human activities: evaluating alternative approaches. Journal of the North American Benthological Society 15: 212–231. Frenkel, R. E., 1993. Vegetation. In Jackson, P. L. & A. J. Kimmerling (eds), Atlas of the Pacific Northwest. Oregon State University Press: 58–65. Hansmann, E. W. & H. K. Phinney, 1973. Effects of logging on periphyton in coastal streams of Oregon. Ecology 54: 194–199. Harding, J. S., M. J. Winterbourn & W. F. McDiffett, 1997. Stream faunas and ecoregions in South Island, New Zealand: do they correspond?. Archiv fu¨r Hydrobiologie 140: 289–307. Hawkins, C. P. & M. R. Vinson, 2000. Weak correspondence between landscape classifications and stream invertebrate assemblages: Implications for bioassessment. Journal of the North American Benthological Society 19: 501–517.

218 Hawkins, C. P., R. H. Norris, J. Gerritsen, R. M. Hughes, S. K. Jackson, R. K. Johnson & R. J. Stevenson, 2000a. Evaluation of the use of landscape classifications for the prediction of freshwater biota: Synthesis and recommendations. Journal of the North American Benthological Society 19: 518–540. Hawkins, C. P., R. H. Norris, J. H. Hogue & J. W. Feminella, 2000b. Development and evaluation of predictive models for measuring the biological integrity of streams. Ecological Applications 10: 1456–1477. Herlihy, A. T., D. P. Larsen, S. G. Paulsen, N. S. Urquhart & B. J. Rosenbaum, 2000. Designing a spatially balanced, randomized site selection process for regional stream surveys: The EMAP Mid-Atlantic pilot study. Environmental Monitoring and Assessment 63: 95–113. Hill, M. O., R. G. H. Bunce & M. W. Shaw, 1975. Indicator species analysis, a divisive polythetic method of classification and its application to a survey of native pinewoods in Scotland. Journal of Ecology 63: 597–613. Johnson, B. R. & G. L. Raines, 1995. Digital map of major bedrock lithologic units for the Pacific Northwest: A Contribution to the Interior Columbia River Basin Ecosystem Management Project. U. S. Geological Survey, Open File Report 95–680. Krammer, K. & H. Lange-Bertalot, 1986. Bacillariophyceae, Teil 1. Naviculaceae . VEB Gustav Fisher Verlag, Jena, 876 pp. Krammer, K. & H. Lange-Bertalot, 1988. Bacillariophyceae, Teil 2. Epithemiaceae, Bacillariophyceae, Surirellaceae. VEB Gustav Fisher Verlag, Jena 596 pp. Krammer, K. & H. Lange-Bertalot, 1991a. Bacillariophyceae, Teil 3. Centrales, Fragilariaceae, Eunotiaceae, Achnanthaceae. VEB Gustav Fisher Verlag, Jena, 576 pp. Krammer, K. & H. Lange-Bertalot, 1991b. Bacillariophyceae, Teil 4. Achnanthaceae, Kritische Erganzungen zu Navicula (Lineolate) und Gomphonema. VEB Gustav Fisher Verlag, Jena, 437 pp. Krammer, K. & H. Lange-Bertalot, 2000. Bacillariophyceae, Part 5. English and French Translation of the Keys. VEB Gustav Fisher Verlag, Jena, 311 pp. Lange-Bertalot, H., 1979. Pollution tolerance of diatoms as a criterion for water quality estimation. Nova Hedwigia 64: 285–305. Leland, H. V., 1995. Distribution of phytobenthos in the Yakima River Basin, Washington, in relation to geology, land use, and other environmental factors. Canadian Journal of Fisheries and Aquatic Sciences 52: 1108–1129. McClain, M. E., R. E. Bilby & F. J. Triska, 1998. Nutrient cycles and responses to disturbance. In Naiman, R. J. & R. E. Bilby (eds), River Ecology and Management: Lessons from the Pacific Coastal Ecoregion. Springer-Verlag, New York: 347–372. Mrazik, S., 1999. Reference Site Selection: A Six-Step Approach for Selecting Reference Sites for Biomonitoring and Stream Evaluation Studies. Oregon Department of Environmental Quality Technical Report BIO99-03. 14 pp. Naymik, J., Y. Pan & J. Ford, 2005. Diatom assemblages as indicators of timber harvest effects in coastal Oregon streams. Journal of North American Benthological Society: accepted.

Omernik, J. M. & A. L. Gallant, 1986. Ecoregions of the Pacific Northwest. Map scale 1:2,500,000. Report EPA/600/3–86/ 033. U. S. Environmental Protection Agency, Corvallis, Oregon, 39 pp. Pan, Y., R. J. Stevenson, B. H. Hill, A. T. Herlihy & G. B. Collins, 1996. Using diatoms as indicators of ecological conditions in lotic systems: a regional assessment. Journal of the North American Benthological Society 15: 481–495. Pan, Y., R. J. Stevenson, B. H. Hill & A. T. Herlihy, 2000. Ecoregions and benthic diatom assemblages in Mid-Atlantic Highlands streams, USA. Journal of the North American Benthological Society 19: 518–540. Passy, S. I., 2001. Spatial paradigms of lotic diatom distribution: a landscape ecology perspective. Journal of Phycology: 370–378. Patrick, R. & C. W. Reimer, 1966. The Diatoms of the United States. Vol. 1: Monographs of the Academy of Natural Sciences of Philadelphia No. 13, 688 pp. Patrick, R. & C. W. Reimer, 1975. The Diatoms of the United States. Vol. 2: Part 1. Monographs of the Academy of Natural Sciences of Philadelphia No. 13, 213 pp. Perrin, C. J., M. L. Bothwell & P. A. Slaney, 1987. Experimental enrichment of a coast stream in British Columbia: effects of organic and inorganic additions on autotrophic periphyton production. Canadian Journal of Fisheries and Aquatic Sciences 44: 1247–1256. Potapova, M. G. & D. F. Charles, 2002. Benthic diatoms in USA rivers: distributions along spatial and environmental gradients. Journal of Biogeography 29: 167–187. Potapova, M. & D. F. Charles, 2003. Distribution of benthic diatoms in U.S. rivers in relation to conductivity and ionic composition. Freshwater Biology 48: 1311–1328. Richards, C., R. J. Haro, L. B. Johnson & G. E. Host, 1997. Catchment and reach-scale properties as indicators of macroinvertebrate species traits. Freshwater Biology 37: 219–230. Rosenfeld, C. L., 1993. Landforms and geology. In Jackson, P. L. & A. J. Kimerling (eds), Atlas of the Pacific Northwest. Oregon State University Press, Corvallis, OR: 40–47. Sabater, S., S. V. Gregory & J. R. Sedell, 1998. Community dynamics and metabolism of benthic algae colonizing wood and rock substrate in a forest stream. Journal of Phycology 34: 561–567. Steinman, A. D. & C. D. McIntire, 1986. Effects of current velocity and light energy on the structure of periphyton assemblages in laboratory streams. Journal of Phycology 22: 352–361. Stevenson, R. J. & S. Hashim, 1989. Variation in diatom community structure among habitats in sandy stream. Journal of Phycology 25: 678–686. Van Dam, H., A. Mertens & J. Sinkeldam, 1994. A coded checklist and ecological indicator values of freshwater diatoms from the Netherlands. Netherlands Journal of Aquatic Ecology 28: 117–133. Verb, R. G. & M. L. Vis, 2000. Comparison of benthic diatom assemblages from streams draining abandoned and reclaimed coal mines and nonimpacted sites. Journal of the North American Benthological Society 19: 274–288. Walker, G. W. & N. S. Macleod, 1991. Geologic Map of Oregon, scale 1:500,000. United States Geologic Survey.

219 Welch, E. B., J. M. Jacoby & C. W. May, 1998. Stream quality. In Naiman, R. & R. E. Bilby (eds), River Ecology and Management: Lessons from the Pacific Coastal Ecoregion. Springer-Verlag, New York, NY: 69–85. Whittier, T. R., R. M. Hughes & D. P. Larsen, 1988. Correspondence between ecoregions and spatial patterns in stream ecosystems of Oregon. Canadian Journal of Fisheries and Aquatic Sciences 45: 1264–1278. Wigington, P. J. Jr., M. R. Church, T. C. Strickland, K. N. Eshleman & J. Van Sickle, 1998. Autumn chemistry of Oregon Coast Range Streams. Journal of the American Water Resources Association 34: 1035–1049.

Winter, J. G., P. J. Dillon, C. Paterson, R. A. Reid & K. M. Somers, 2003. Impacts of golf course construction and operation on headwater streams: bioassessment using diatoms. Canadian Journal of Botany 81: 848–858. Wright, J. F., M. T. Furse & P. D. Armitage, 1993. RIVPACSa technique for evaluating the biological quality of rivers in the UK. European Water Quality Control 3: 15–25. Zar, J. H., 1999. Biostatistical Analysis. Prentice-Hall, Upper Saddle River, New Jersey, 663 pp.

Hydrobiologia (2006) 561:221–238  Springer 2006 R.J. Stevenson, Y. Pan, J.P. Kociolek & J.C. Kingston (eds), Advances in Algal Biology: A Commemoration of the Work of Rex Lowe DOI 10.1007/s10750-005-1616-0

Algal assemblages in multiple habitats of restored and extant wetlands Lei Zheng1,2,* & R. Jan Stevenson1 1

Department of Zoology, Michigan State University, East Lansing, MI 48824, USA Tetra Tech, Inc., Suite 200, Owings Mills, MD 21117, USA (*Author for correspondence: E-mail: [email protected]) 2

Key words: epiphyton, epipelon, phytoplankton, habitats, restored wetlands, extant wetlands

Abstract The functional and structural attributes of algal assemblages were studied in 25 restored and 20 extant depressional wetlands in southern central Michigan. Environmental conditions and algal assemblages were compared between restored and extant wetlands and among habitats within wetlands. Restored marshes generally had lower shading by macrophytes, nutrient concentrations, and sediment organic matter. Relative biovolume of non-diatom algae was significantly different among plankton, macrophyte and sediment habitats in restored wetlands, but did not differ between macrophytes and sediments in extant wetlands. Species composition of diatom assemblages was not significantly different between plants and sediments in both restored and extant wetlands. The observed differences in non-diatom algae could not be attributed to any measured environmental variable; however, diatom assemblage differences between habitats increased with light irradiance. Differences in sediment diatom assemblages were observed between restored and extant wetlands and were related to differences in nutrients, pH, and canopy cover. Differences were also observed between epiphytic diatom assemblages in restored and extant wetlands and they were related to light and dissolved oxygen. In summary, differences in light and nutrient availability were the main environmental factors differentiating algal communities in wetlands.

Introduction Wetland restoration has been increasingly used to mitigate loss and anthropogenic impact to wetland systems (McDonnell & Pickett, 1990; Holland et al., 1995; Detenbeck et al., 1999; Zedler, 2000). Beginning in the early 1990s, a large effort was initiated in Michigan (USA) to compensate for wetland losses by establishing wetland restoration programs (MDEQ, 2001a). Hundreds of wetland restoration projects have been carried out since then. Most of the wetland restoration projects involve simple techniques, such as plugging agricultural ditches or breaking field tiles to restore the hydrology of an area, with the assumption that this will restore the biological

community. Restored wetlands support aquatic and terrestrial habitats and sustain many plant, invertebrate, and wildlife species (NRC, 1995; USEPA, 2000). Ecological development after wetland restoration has been studied extensively (Whigham et al., 1990; Zedler, 1993; Keddy, 1999; Middleton, 1999), and many environmental variables change during transition from one successional state to another (Gosz, 1992). Nutrient accumulation is one of the most noticeable changes in biogeochemical process affecting the biological community in wetlands. In human impacted landscapes, nutrient enrichment from human activities affects the wetland restoration process, and its detrimental effects on wetland systems have been frequently

222 reported (Koerselman & Beltman, 1988; Blair, 1996; Magee et al., 1999). Nutrient accumulation and storage from agricultural and urban runoff affect wetland productivity, species diversity, water quality, as well as the restoration rate of wetlands (Rader & Richardson, 1992; Craft & Casey, 2000). A number of questions regarding how nutrient accumulation shapes the biological communities remain to be answered (Stolt et al., 2000). Although algae are one of the most important groups of primary producers in wetlands and play crucial roles in wetland biogeochemical cycles (Vymazal, 1995), algal ecological responses in restored wetlands present a major knowledge gap in wetland ecology and restoration ecology. Wetland restoration attempts to mimic one of the most complex examples of ecosystem development and is therefore difficult to monitor and predict. Algae, especially diatoms, are effective indicators of ecological conditions (van Dam et al., 1994; Adamus, 1996; Cottingham & Carpenter, 1998; Detenbeck, 2001; Stevenson et al., 2001) and have been used as indicators of ecological conditions in wetlands (Pan & Stevenson, 1996; Stevenson et al., 1999, 2001; Pan et al., 2000). Algae are also sensitive to nutrient concentration change (Tilman, 1981). Soil organic C, N and P accumulation can affect the biomass and species composition of wetland algal communities (McCormick et al., 1996; Pan et al., 2000). Therefore, algal assemblages may differ between restored and extant wetlands and could be valuable indicators of restoration success because algal species composition and diversity would differ in low- and high-nutrient wetlands (John, 1993; Mayer & Galatowitsch, 1999). Habitat differences, along with a number of other environmental factors (such as substratum type, substratum stability, and nutrient availability), contribute to spatial and temporal variability of algal assemblages in aquatic ecosystems (Leland, 1995; Burkholder, 1996; Stevenson, 1996, 1997). A number of studies have reported that different habitats support different algal assemblages in aquatic systems (Stevenson & Hashim, 1989; Pringle, 1990; Burkholder, 1996; McCormick et al., 1998; Hillebrand & Kahlert, 2001; Lim et al. 2001). It has been hypothesized that the extent and conditions in four major algal habitats: phytoplankton, epiphyton, epipelon, and metaphyton, vary temporally at different stages in wetland cycles

(Goldsborough & Robinson, 1996; Wu & Mitsch, 1998), but the underlying mechanisms have not been determined. Wetland restoration provides a unique system to examine how changes in environmental conditions of algal habitats following creation or restoration affect changes in species composition among habitats. We expect that the lack of substratum should distinguish the plankton from the benthic communities, but heterogeneity between planktonic and benthic habitats could decrease if more benthic algae were suspended in the water during either early or late stages of development. We also expect that epiphyton and epipelon become more similar with the increase of wetland age for two reasons. First, wetland bottoms will be covered by increasing aquatic plants and macroalgae in late stages so the benthic habitats will be more similar. Second, nutrient availability may become more similar for the epiphytic and sediment habitats in late stages. Nutrient supply from sediments directly to epipelon may be available earlier during wetland development than the indirect supply from sediments, through plants, to epiphyton. Similarities between epiphyton and epipelon probably increase with wetland development as nutrient availability increases and differences between habitats decrease. To evaluate the success of wetland restoration and investigate the impact of human disturbance to wetland systems, we selected 45 (25 restored, 20 extant) freshwater depressional wetlands in the Maple and Upper Grand River watersheds in southern central Michigan. The environmental characteristics and algal assemblages of phytoplankton, epiphyton, and epipelon in different habitats in these wetlands were examined. We hypothesized that (1) environmental variables would differ between restored and extant wetlands; (2) algal assemblages would differ among habitats within wetlands and between restored and extant wetlands; (3) differences in algal assemblages among habitats within wetlands and between restored and extant wetlands would decrease with increasing age of restored wetlands, and that decrease would be due to increasing similarity of environmental characteristics. Therefore, we expected to identify environmental factors that determine the distribution of algal assemblages and biodiversity among habitats and between restored and extant wetlands.

223 Materials and methods Site description Forty-five (25 restored, 20 extant wetlands) depressional wetlands were selected in the Maple and Upper Grand River watersheds of Southern Michigan (Fig. 1). All restored wetlands were reestablished within the past 15 years as part of wetland mitigation program by U.S. Fish and Wildlife Service (USFWS) and the U.S. Department of Agriculture, Natural Resources Conservation Service (NRCS). These wetlands were built in areas of hydric soils that had been drained or manipulated for agricultural purposes, and they were located in landscapes dominated by agricultural land. Extant wetlands were selected based on similar geomorphology and site location with the restored wetlands. These wetlands were isolated semi-permanent to permanent wetlands (Tiner,

2003), less than 2 acres in size and 2 m in depth. The sampled wetlands were mostly Typha or Nuphar dominated marshes, but also included several forest marshes. Duckweed (Lemmaceae) and other floating plants appeared in at least 2/3 of the sampled sites during the sampling period. Watershed land use data were obtained from the Michigan Department of Natural Resource (DNR) spatial data library to determine land use around these wetlands. Wetland boundaries for each site were visually determined based on shifts from wetland to upland vegetation and changes in slope between the wetlands and the adjacent upland. Sampling The selected wetlands were sampled and assessed in July 2000. Water temperature and dissolved oxygen (DO) were measured with a YSI DO

Figure 1. Wetland sampling locations in southern Michigan.

224 meter (YSI Inc., Yellow Springs, Ohio, USA) at the beginning and at the end of the sampling period in the mornings, allowing at least 1 h between measurements. Because of time limitations, we used the DO change in the morning to estimate net primary production in the wetlands. Wetland net production was calculated based on DO change over 1 h times the estimated average depth of the wetland. Conductivity and pH were measured using a YSI model 33 SCT meter with a conductivity and pH electrode (Denver Instrument Company). Canopy cover by emergent vegetation was estimated with a spherical canopy densiometer held at water level and by averaging canopy cover from four directions. Canopy readings were recorded from the locations where sediment samples were taken. Duckweed coverage was estimated based on the proportional coverage of the wetland surface. Irradiance at the water surface (I0) and 30 cm below the water surface (Iz) were recorded with a LICOR light meter and light extinction coefficients (LEC) were calculated as ln(I0))ln(Iz)/0.3 m (Wetzel, 2001). Water samples from open water area in each wetland were collected in two 125-ml acid-washed polyethylene bottles, stored on ice, transported to the laboratory, and frozen until chemical analysis. Phytoplankton samples from 10 cm below the water surface were collected in two 1-l polyethylene bottles for chlorophyll a (chl a) and algal cell counts. Sediment samples were collected randomly from eight different locations in each wetland using hard plastic tubes (internal diameter = 3.8 cm, length = 20 cm) and the top 1 cm of all eight sediment samples from a wetland were composited for epipelon community determination. Epiphytic algae were collected by cutting 4–5 macrophyte stems, brushing epiphytes from each macrophyte surface, and putting them in a Whirl-pak bag. Metaphyton were collected for qualitative identification. Algae were preserved with M3 (APHA, 1998) after subsampling for chl a analysis. Sample analysis One 125-ml water sample from each wetland was analyzed for total nitrogen (TN) and total phosphorus (TP). Another bottle was filtered through Coleman glass fiber filters (0.45 lm pore size) and the filtrate was analyzed for nitrate-nitrite

(NO)3 + NO)2 –N) and soluble reactive phosphorus (SRP) using a Spectronic Genesys 2 Spectrophotometer. TP and SRP were measured using the ascorbic acid method, while TN and nitrate and nitrite (Nox–N) were measured using cadium reduction methods (APHA 1998). Ammonia (NH4–N) was analyzed using standard methods and a Wuick-Chem 8000 autoanalyzer. Dissolved inorganic nitrogen (DIN) was calculated as the sum of ammonia and nitrate–nitrite concentrations. Sediment samples were rinsed with 200 ml of deionized water and homogenized using a Biospec homogenizer and were then digested according to standard methods (APHA, 1998). Appropriate dilution was made for sediment samples to measure sediment TN and TP concentration following the same methods as the water samples. Chl a in water samples was assayed by extracting in 90% buffered acetone overnight at 4 C, reading absorbance on a Spectronic Genesys 2 Spectrophotometer, and calculating pheophytin-corrected chl a concentration (APHA, 1998). Dry mass (DM) and ash free dry mass (AFDM) of algae were determined according to standard methods (APHA, 1998). Organic matter proportion in sediment (OMP) was the AFDM/DM ratio of surface sediments. Phytoplankton, epiphytic algae, and epipelon were diluted or concentrated as necessary before counting. Algal densities and non-diatom species composition were determined using a Leica microscope and a Palmer-Maloney (0.1 ml) counting chamber under 400 and by counting 300 natural units (cells or colonies naturally attached under microscope were defined as one natural unit). Algal taxa were identified to the lowest possible taxonomic level. Biovolume was estimated for at least 15 cells of each taxon by assigning them a geometric shape and measuring appropriate cell dimensions (Hillebrand et al., 1999; Charles et al., 2002). Diatom taxa and relative abundance were determined from permanent Naphrax mounts of acid-cleaned diatoms (Stosch & Reimann, 1970). At least 500 diatom valves were counted. Alpha diversity (species richness per locality), gamma diversity (total species diversity within landscape, all wetlands sampled), and beta diversity (variation in species composition between two communities or habitats) were calculated for non-diatom algae and diatoms (Beilman, 2001).

225 Statistical analysis To compare environmental variables between restored and extant wetlands, both individual variables and axis scores from principle component analyses (PCA) were used. Environmental variables were log-transformed to evenly distribute the variance before analysis, as necessary. The student t-test was used to compare individual variables between extant and restored wetlands. PCA based on correlation coeffiecents was used to summarize major environmental patterns in this region. Jackson’s broken stick rule (Jackson, 1993) was used to determine the significance of PCA axes. The scores of the first three PCA axes were used as composite variables for multiple analysis of variance (MANOVA) to test the overall differences between restored and extant wetlands. All statistics were performed using SYSTAT v.10. Attributes of non-diatom algae and diatoms were analyzed separately. The relative biovolume of non-diatom algae and relative abundance of diatom species were calculated and used for comparisons. Only diatom species with a relative abundance ‡1% in a minimum of three sites or ‡5% in at least one site were included in the analysis. Non-metric multidimensional scaling (NMDS), a multivariate ordination technique (Clarke, 1993), was used to summarize and illustrate patterns of non-diatom and diatom distribution among different habitats and between restored and extant wetlands. Stress values and dimensionality were selected using criterion defined in PC-Ord autopilot. Normally, a stress value less than 20 was considered an acceptable NMDS solution. Because taxnomical composition of non-diatom algae exhibited large variation among wetlands and between habitats, and a high stress value was always expected, so we did not use NMDS for making decisions on group differences. Instead, multiresponse permutation procedure (MRPP), a non-parametric procedure for testing the hypothesis of no difference between two or more groups (Biondini et al., 1985), was used to assess differences in taxonomic composition of non-diatom and diatom assemblages among different habitats and between restored and extant wetlands. A p-value was provided by the MRPP to characterize the significance of between group differences. An A value (0–1), the chance-corrected within-group

agreement, characterizes within-group homogeneity compared to the random expectation. When all items are identical within groups, then A=1; if heterogeneity within groups equals the expectation by chance, then A=0. Sorensen (Bray–Curtis) similarity index was used as the distance measurem in both MRPP and NMDS. Both MRPP and NMDS were performed using PC-ORD v.4.0 (McCune & Mefford, 1997). To relate differences in taxonomic composition to environmental variables, Bray–Curtis similarity matrices (similarity in relative abundances or biovolumes) were calculated for species composition of non-diatom and diatom assemblages among different habitats and between restored and extant wetlands. Stepwise automatic linear regressions were used to explore the relationship between nondiatom or diatom assemblages and environmental variables. This was done by using similarity of non-diatom algae or diatom assemblages between habitats as separate dependent variables and a set of 19 uncorrelated environmental variables as independent variables. Using these same procedures, similarities of species composition between a restored wetland and all extant wetlands were calculated and expressed as a function of differences in environmental variables between wetlands. We expected that the increase in similarity of species composition between a restored wetland and extant wetlands would increase when the differences in their environmental characteristics decreased. The dependent variables were the average Bray–Curtis similarities in species composition of a restored wetland with all the extant wetlands. The Euclidean distance matrix for each environmental variable among different wetlands was calculated to indicate environmental differences between wetlands and the average Euclidean distances of individual environmental variables between a restored wetland with all extant wetlands were used as independent variables. Those variables that had the highest correlations (r>0.4, p<0.05) with other variables were screened out, so only the most uncorrelated environmental variables were included for the regression analyses. For example, the distance of conductivity between restored and extant wetlands was positively correlated with the distance of sediment TP (r=0.424, p=0.035), negatively correlated with pH distance (r=0.430,

226 availability (duckweed cover) (loadings=0.631) and light extinction coefficient (loadings =0.69). The second axis was negatively correlated with N and P content in sediments (loadings=)0.792 and )0.691, respectively) and shading by vegetation (loadings =)0.61), and positively correlated with conductivity (loadings=0.48). The third axis was positively correlated with SRP (loadings =0.641) and DO (loadings =0.751). Overall, the scores of the first three axes were significantly different between restored and extant wetlands (MANOVA, Wilks’s lamba F3,42=5.938, p=0.002). PCA axis 1 scores were significantly different between restored and extant wetlands (ANOVA, p=0.012), but the scores of PCA axis 2 and axis 3 were not different (p>0.1). Many variables did not differ between the two types of wetlands (p>0.1). Water temperature varied from 18.1 to 32.2 C among wetlands during the sampling period. Conductivity was highly variable (from 149 to 1003, 499±34 lmho cm)1 (mean±S.E.)), as well as pH (from 6.81 to 9.62, 7.92±0.93). DO in the morning could be as high as 11 mg l)1 and as low as 0.70 mg/l in some of the wetlands (4.57±0.38 mg l)1). Wetland net

p=0.032), and related to several other variables. Therefore, it was excluded from the model. All these analyses were performed using SYSTAT v. 10.

Results Environmental variables Land cover in the region was dominated by agriculture (96%) with smaller portions of urban (0.9%), forest (1.6%), and other land types. Most of the restored wetlands were Typha-dominated marshes. In contrast, extant marshes in this area were seldom dominated by one macrophyte species. Phragmites, Typha, Nuphar, Ceratophyllum demersum, Myriophyllum, and Scirpus were co-dominant in many of the extant marshes. Three PCA axes accounted for 42% of environmental variance in the data set (Table 1). The first axis was positively correlated with major nutrients, such as TP (loadings=0.621), ammonium (loadings=0.784), and DIN (loadings=0.675) in the water column, as well as light

Table 1. Correlation matrix of measured environmental variables and principal component analysis (PCA) loadings of environmental variables in the 45 wetlands Shading LEC

Duckweed DO

TN

TP

DIN

NH+ SRP Chl a 4

OMP

Sed. N Sed. P Age

Shading

1.00

.

.

.

.

.

.

.

.

.

.

.

.

LEC Duckweed

0.05 0.16

1.00 0.65

. 1.00

. .

. .

. .

. .

. .

. .

. .

. .

. .

. .

DO

)0.21

.

.

.

.

.

.

.

.

.

TN

0.05

)0.37 )0.54 0.00

0.07

)0.01

1.00

.

.

.

.

.

.

.

.

TP

0.18

0.48

0.27

)0.14

0.08

1.00

.

.

.

.

.

.

.

DIN

0.40

0.43

0.28

)0.29

0.24

0.25

1.00

.

.

.

.

.

.

.

1.00

NH+ 4

0.44

0.51

0.37

)0.31

0.31

0.35

0.96

1.00

.

.

.

.

SRP

0.10

0.21

0.11

0.19

0.37

0.46

0.23

0.30 1.00

.

.

.

.

Chl a OMP

0.03 0.17

0.38 0.29

0.38 0.31

0.02 )0.15

0.11 0.04

0.35 0.25

0.19 0.32

0.26 0.29 0.39 0.28

1.00 0.49

. 1.00

. .

. .

Sed. N

0.28

-0.03

0.16

)0.08 )0.13 )0.07

0.01 )0.05 0.00

)0.05

0.29

1.00

.

Sed. P

0.26

-0.02

0.16

)0.07 )0.04

0.04

0.16

0.11 0.11

)0.04

0.12

0.76

1.00

Age

0.22

0.12

0.07

)0.07 )0.19

0.03 )0.14 )0.21 0.02

0.02

0.02

0.00

0.04

1.00

PC1

0.24

0.69

0.63

)0.33

0.62

0.68

0.78 0.45

0.65

0.66

0.14

0.08

)0.22

PC2

)0.61

0.17

-0.16

0.14

0.24

0.01

0.14

0.18 0.13

0.18

0.29 )0.79

)0.69

)0.24

PC3

0.04

-0.21 )0.36

0.75

0.26

0.38 )0.12 )0.22 0.64

0.03 )0.13 )0.44

)0.08

0.37

0.38

When a correlation coefficient r > 0.30, Dunn–Sidak p < 0.05; when r>0.39, p < 0.01. Correlation between wetland age and environmental variables had 24 degrees of freedom and no significant correlations.

40

100

20

30

75

LEC

100 80 60 40

Duckweeds(%))

Coverage(%)) Canopy Coverage

227

50

20

0

0

0

1100

10

40

9

pH

700

8

500

30

20

7

100

6

10

12

6

10

5

90 60

-1 -1

8 6 4 2 0

R

E

OMP(%)

300

DΔO (mg·L ·h )

DO(mg/L)

900

Temperature(°C) C)

10

Conductivity(μhos/cm)

25

4 3

30 10

2 1 0

R

E

R

E

Figure 2. Wetland physical and chemical characteristics and comparison between restored and reference wetlands.

production based on DO change ranged from )1.00 to 3.17 g m)2 h)1, representing heterotrophic to autotrophic wetland types. Most restored wetlands had relatively low shading by vegetation (26.2±3.8%), which was significantly less than in extant sites (p=0.031) (Fig. 2). Duckweed coverage was much higher in extant wetlands than restored ones (p=0.003). LEC at 0.3 m depth in open water (if available) of restored wetlands (1.21±0.12 lmol m)3 s)1) was much lower than in extant wetlands (3.17±0.78 lmol m)3 s)1) (p=0.010). Nutrient concentrations in these two types of wetlands were significantly different (Fig. 3). TP (1.86±0.31 mg l)1) and SRP (0.41±0.10 mg l)1)

were high in most of the wetlands, and had significantly higher concentrations in extant wetlands than in restored wetlands (p=0.007 and 0.027, respectively). TN (0.78±0.07 mg l)1 overall) in extant wetlands was not different from restored wetlands; the major form of inorganic N was ) NH+ 4 , and NO3 was almost depleted in most of the wetlands (<0.010 mg l)1). DIN (0.054±0.007) was significantly higher in extant wetlands (p=0.023) than restored wetlands. N:P and DIN:SRP ratios were generally lower than 17. N:P ratios were higher in restored wetlands than in extant wetlands (p=0.002), but DIN:SRP ratios were not significantly different between these two types of wetlands. Most of wetlands had mineral

228

101

10-2

10-1

2.0 1.5 1.0

10-2 101 N:P

0.5

100

100

101 100 10-1

10-1

N:P in sediment

P:DM(MG/G)

10-1

N:DM(mg/g)

100

10-1

10-2

TP(mg/l)

TN(mg/L)

10-3

100

DIN:SRP

SRP(mg/L)

DIN(mg/L)

10-1

100

101

10-1

100

10-2 R

E

10-1 R

E

R

E

Figure 3. Wetland nutrient concentration and comparison between restored and extant wetlands.

soils with an average OMP 17.4% (±2.3), which was higher in extant wetlands than in restored sites (p=0.035). TN (3.91±1.11 mg l)1) and TP content (0.76±0.19 mg g)1) in surface sediments was higher in extant wetlands than in restored wetlands (p=0.025 and 0.058, respectively), but N:P ratios (3.78±1.86) in sediment were not different. Many of the environmental variables related to trophic condition correlated with each other (Table 1). LEC at 30 cm depth positively correlated with duckweed cover, chl a, ammonia, water column TP concentration, and sediment organic matter, and negatively correlated with DO. DO in the water column negatively correlated with duckweed cover, LEC, and ammonia concentration. Most nutrient factors were also correlated with each other. Water column TN positively

correlated with SRP, NH+ 4 , and sediment TN. Water column TP positively correlated with SRP, ammonia, and chl a. Sediment TN and TP content were positively correlated with each other. Age of restored marshes was not a significant factor in differentiating the first 3 PCA axes. Environmental conditions between marshes less than 5-year-old and 5–10 year-old were not significantly different (MANOVA, Wilks’s lamba F3,25=0.833, p=0.48) based on the three PCA axes. None of the environmental variables correlated with age of the wetlands. Algal assemblages Of the 311 epiphytic algal taxa identified, diatoms had the highest taxa richness (157), followed by

229 Table 2. Mean alpha diversity (number of species per site), gamma (landscape) diversity, and beta diversity between habitats within a wetland and between restored and extant wetlands Alpha (average + s.e.)

Gamma

Beta

Epiphytic Sediment Planktonic Epiphytic Sediment Planktonic Epiphytic Sediment Planktonic Restored Diatoms 21.6±1.3 28.2±1.8 9.3±0.76 Non-diatom 8.7±1.2 3.2±0.3 Extant

165 26

All algae

30.0±1.9 31.3±1.9

250

191

Diatom

17.6±1.9 29.7±2.4 7.9±0.65

100

129

89

28

Non-diatom All site

132 118

9.7±1.4

2.9±2.0

Algae

25.8±2.7 32.6±2.4

189

157

Diatom

20.5±1.1 29.5±1.3 8.7±0.52

157

199

3.0±0.2

154

28.7±1.5 31.2±1.6

311

Non-diatom Algae

9.1±0.9

Chlorophyta (99) and cyanobacteria (34). Other groups (Euglenophyta, Dinophyta, and Chrysophyta) appeared, but with less richness. Chlorophyta were the most abundant epiphytic algae, with an average of 48.2% (±6.9% SE) of the total cell density and 46.1% (±5.0%) of total biovolume. Stigeoclonium, Mougeotia, and Oedogonium were the most abundant genera. Diatoms were the next most abundant epiphytic group, with 30.3% (±4.5%) of total density and 26.8% (±4.3%) of total biovolume. Eolimna minimum, Cocconeis placentula, Achnanthidium minutissimum, and Navicula cryptocephala were the most dominant diatom taxa. Cyanobacteria, Euglenophyta, Dinophyta, and Chrysophyta appeared but were low in density in epiphytic samples. The total number of algal taxa (gamma diversity) and average taxa richness (alpha diversity) were higher in restored wetlands (250 and 30.0, respectively) than in extant wetlands (189 and 25.8, respectively) (Table 2). Beta diversity of diatoms was also higher in restored wetlands than in extant wetlands. Of the 235 epipelon taxa identified, diatoms had both the highest taxa richness (199) and abundance with an average of 54.4 % (±6.6%) of total cell density and 48.8% (±6.5%) of total biovolume. Eolimna minima, A. minutissimum, N. cryptocephala, and Nitzschia palea dominated diatom communities. Chlorophyta was the next dominant group, with 19 taxa and 22.3 (±4.3%) of total density and 12.9% (±4.1%) of total biovolume. This was due to the dominance of Scenedesmus in some of the wetlands. Only 8 cyanobacteria taxa were found, but they remained in low density

111

6.1 13.6

5.9 8.1

8.3

6.1

90

5.7

4.3

9.2

9.6

7.3

4.8

7.7

6.7

38

16.9

12.7

235

10.8

7.5

153

12.0

11.4

17.6

(13±2.3%) and biovolume (16.4±4.7%). Other groups (Euglenophyta, Dinophyta and Chrysophyta) appeared but with relatively low abundance. Although the alpha diversity of non-diatom algae and diatoms was similar in restored (3.2 and 28.2, respectively) and extant wetlands (2.9 and 29.7, respectively), the gamma diversity of diatom taxa in restored wetlands (165) was much higher than in extant wetlands (129) and resulted in higher beta diversity in restored wetlands (Table 2). Phytoplankton biomass was low in most sampled wetlands (0.030±0.009 mg chl a l)1) and was not significantly different between restored and extant wetlands (p = 0.113). The direct correlations of phytoplankton biomass with major nutrient gradients were weak (r = 0.368 with TP, r=0.126 with TN), but was significantly related to OMP in sediments (r = 0.499, p = 0.025). Phytoplankton taxa richness was relatively low compared to other habitats. A total of 153 nondiatom taxa were found in phytoplankton samples, while diatom density was very low in most wetlands. Most of the phytoplankton were Chlorophyta (42.3%) and cyanobacteria (33.1%) on the basis of cell density, only 15.2 (±3.3%) of total algal abundance was due to diatoms. The average relative biovolume was dominated by dinoflagellates (37.9±4.7%), Euglenophyta (26.5±3.7%), Chrysophyta (10.5±4.4%) and green algae (9.0±2.9%). Cyanobacteria (3.5±1.6%) and diatom biovolume (2.3±0.7%) were very low. Restored wetlands had higher alpha, beta, and gamma diversity than extant wetlands (Table 2). Planktonic diatoms were not used for quantitative

230 analysis because of their extremely low density in half of the wetland sites. Metaphyton appeared in the majority of the wetlands and completely covered the basin of at least one-quarter of wetlands. The most commonly found macroalgae were Stigeoclonium, Oedogonium, and Mougeotia. Most of the time, metaphyton were mixed with other aquatic plants. Because of the difficulty of quantifying their biomass, metaphyton were not used for further analysis. Non-diatom and diatom diversity among different habitats were quite different. Alpha, beta, and gamma diversity of non-diatom epiphyton and phytoplankton were greater than epipelon in all wetlands. Sediment diatoms had higher alpha diversity, higher gamma diversity, but lower beta diversity than epiphytic diatoms (Table 2).

was larger than that of sediment diatoms, which represented larger variation in diatom assemblages among different wetlands on macrophytes than in sediments and indicated a higher beta diversity for epiphytic diatom assemblages than sediment diatoms. Differences in species composition of algal assemblages between restored and extant wetlands revealed habitat specific patterns (Table 4, Fig. 5). Both epiphytic non-diatom algae and diatom assemblages in restored wetlands were significantly different from those in extant wetlands (Table 4, Fig. 5). However, sediment algae and phytoplankton were not different between these two types of wetlands.

Differences in taxonomic composition

Multiple linear regression of non-diatom algal similarities among habitats indicated that little variance could be explained by environmental variables (Table 5). Water column P and sediment TP had very weak effects on non-diatom assemblages among habitats. None of the regression models were significant. Bray–Curtis similarity between epiphytic and sediment diatoms, however, revealed a significant regression model with shading by vegetation, LEC, and DO as predictors (R2=0.346, p=0.001). With the increase of plant shading, LEC, and DO, similarity in diatom species composition between epiphytes and epipelon increased. Similar to inter-habitat comparisons, similarity between restored vs. extant wetlands of non-

Differences in algal assemblages were commonly observed among habitats within the wetlands. Stress value for the first three axes of NMDS based on non-diatom algae was 23.9 and was 18.2 based on diatoms. Most comparisons of non-diatom algae among habitats were different in both restored and extant wetlands, except that the difference between epiphyton and epipelon assemblages was not significantly different in extant wetlands (p=0.113) (Table 3, Fig. 4). Similarly, epiphytic diatom assemblages and sediment diatom assemblages were not significantly different in restored or extant wetlands (Table 3). The sample ellipse of epiphytic diatom assemblages (Fig. 4)

Algal assemblages and environmental variables

Table 3. MRPP comparisons of species composition in non-diatom algae and diatoms between different habitats in restored and extant wetlands Comparisons

Wetland type

t statistic

A

p

Restored

)3.940

0.0145

0.002

Extant

)1.241

0.0051

0.113

Phytoplankton vs. Epiphytes

Restored

)5.333

0.0174

0.000

Phytoplankton vs. Sediments

Extant Restored

)4.876 )8.660

0.0276 0.0371

0.001 0.000

Extant

)6.944

0.0442

0.000

Restored

)0.138

0.0006

0.389

Extant

)1.118

0.0057

0.133

Based on non-diatom algal biovolume Epiphytes vs. Sediments

Based on diatom relative abundance Epiphytes vs. Sediments

231

Non-diatomalgae NMDS3

Restored wetlands

Extant wetlands

1.0

1.0

0.5

0.5

0.0

0.0

-0.5

-0.5

-1.0 -1.0

-0.5

0.0

0.5

1.0

Diatoms NMDS2

1.0

-0.5

0.0

0.5

1.0

1.5

2.0 1.0

0.0

0.0

-1.0

-2.0 -1.0

-1.0 -1.0

Epiphytes Phytoplankton Sediments

-1.0 -2.0

-0.5

0.0

0.5

1.0

1.5

-1

NMDS1

0

1

2

NMDS1

Figure 4. Comparisons of algal species composition in different habitats using non-metric multidimentional scaling (NMDS). The circles are the sample ellipse which is centered on the sample means of the x and y variables. The size of the ellipse is specified a probability value of 0.6827.

Table 4. MRPP comparisons of species composition of non-diatom algae and diatoms between restored and extant wetlands

Non-diatom algae

Diatoms

Habitats

t statistic

Epiphytes

)2.219

0.0070

0.023

Sediments

0.760

)0.0042

0.762

Phytoplankton

0.466

)0.0033

0.594

Epiphytes

)2.322

0.0100

0.026

Sediment

)0.279

0.0010

0.352

A

p

232

epiphytes

sediment P=0.762

P=0.023 1.0

1.0

0.5

0.5

0.0

0.0

-0.5

-0.5

NMDS2

Non-diatom algae

phytoplankton 1.0

1.5

1.5

0.5 0.0 -0.5 P=0.594

-1.0 -2

-1

2

0

1

2

-1.0 -2

-1

1.0

P=0.026

0

1

-1.0 -1.0

P=0.352

NMDS2

Diatoms

0.0

0.5

1.0

Restored

1

Extant

0.5 0

-0.5

1 0.0

-1 -2

-1

0 1 NMDS1

2

-0.5 -0.5

0.0 0.5 NMDS1

1.0

Figure 5. Comparisons of algal species composition in restored and extant wetlands using non-metric multidimensional scaling (NMDS). The circles are the sample ellipse which is centered on the sample means of the x and y variables. The size of the ellipse is specified a probability value of 0.6827.

Table 5. Regression by forward stepwise selection of independent variables Species composition

Habitats

Variables

Non-diatom algae

Epiphytes vs. Phytoplankton Epiphytes vs. Sediments Phytoplankton vs. Sediments

Epiphytes vs. sediments

R2

p (2 tail)

)0.954

88.464

0.073

0.073

0.254

)1.619

0.065

0.114

)0.007

0.540

0.037

0.637

)0.015 )0.001

1.206 0.089 0.374

0.000

Coefficient

SRP

45

P in sediment

45

pH

45

Conductivity DDO Diatoms

t)value

N

Constant

0.475

39.308

)0.016

1.327

0.192

DO

0.027

)2.030

0.049

Canopy coverage

0.050

)3.955

0.000

LEC

0.034

)2.567

DIN:SRP

Regression

45

0.014 0.001

Dependent variable is average Bray–Curtis similarity among epiphytic, phytoplankton, and sediment non-diatom algae, and dissimilarity between epiphytic and sediment diatom assemblages, independent variables are uncorrelated environmental variables.

233 Table 6. Linear regression by forward stepwise selection of independent variables Species composition

Habitats

Non-diatom algae

Epiphytes plankton

Diatoms

Sediments Sediments

Variables

N

Canopy Coverage

22

t-value

25

R square

p

0 0.021

)1.569

25

0.110

0.132

0

Constant

0.239

9.944

pH

0.015

)1.894

0.074

TN

)0.011

1.681

0.109

DIN

)0.024

2.559

0.019

0.021

)2.515

0.021

)0.108

3.211

0.005

0.038 )0.086

12.821 )1.637

0.020

2.037

Canopy Coverage Sediment TP Regression Epiphytes

coefficient

25

Constant LEC Canopy Coverage Regression

25

0.540

0.000

0.007 0.000 0.054 0.114 0.196

0.091

Dependent variable is average Bray–Curtis similarity of algal species compositions in a restored wetland with extant wetlands. Independent variables are average Euclidian distances of a variable between a restored wetland with all extant wetlands.

diatom algae could not be as easily explained by environmental differences as that of diatom assemblages (Table 6). Only similarity of planktonic non-diatom algae between restored and extant wetlands responded slightly positively, to canopy coverage (p=0.132). Similarity in diatom species composition between restored and extant wetlands in many habitats was related to many environmental factors (Table 6). Similarity of epiphytic diatom species composition between restored and extant wetlands was negatively related to LEC difference and positively to differences in plant shading. Similarity in sediment diatom species composition was related to a number of factors, such as differences in pH, plant shading, and nutrient differences between restored and extant wetlands (R2=0.540, p=0.007). With the decrease in nutrient difference between restored and extant wetlands, similarity in diatom species composition increased.

Discussion Distinct ecological differences were evident between the restored and extant wetlands in southern central Michigan. In addition to differences in shading by vegetation and duckweed on the water surface, nutrient concentrations

and soil organic matter proportion also differed between these two types of wetlands. Differences in algal species composition between the two types of wetlands varied with habitat type. Epiphytic assemblages were more different between restored and extant wetlands than sediment and phytoplankton assemblages. Abiotic environmental variables were related to many of the differences in algal assemblages among habitats in wetlands and between restored and extant wetlands. Age of a restored site was not a determinant of algal characteristics in these wetlands. Wetland restoration and environmental change Although the initial goal of restoration is usually to restore the original physical and chemical characteristics and biological communities of ecosystems, the actual consequences of wetland restoration and construction are variable (Mitsch & Wilson, 1996; Kaiser, 2001). Wetland restoration can recover some aspects of the lost function and structure of pristine wetlands (USEPA, 2002), but most evidence (van der Valk, 1981; Galatowitsch & van der Valk, 1996; Magee et al., 1999) indicates that restored wetlands differ from extant wetlands in a number of ways. Restored wetlands generally have different hydrogeomor-

234 phology, more flooded area with expansive areas of open water (Gwin et al., 1999), narrow borders of vegetation (Kentula et al., 1992; Bedford, 1996), low floristic diversity, and different floristic composition (Galatowitsch & van der Valk, 1996; Magee et al., 1999) compared to natural wetlands. In this study, the extant wetlands in southern Michigan generally had complete zones of emergent, submergent, and floating plants. Vegetation in these wetlands was more diverse, and vegetation shading was higher than restored wetlands. Restored wetlands were mainly composed of emergent Typha and submergent Ceratophyllum with little sedge meadow. Restored vegetation was composed of one or a few species, and canopy cover was lower. The higher canopy and duckweed cover in extant wetlands suggested that light limitation may regulate algal growth and species composition. A number of studies have shown that extant wetlands in human impacted landscapes have a much higher nutrient concentration and sediment organic matter than restored wetlands, which has been related to long-term accumulation and nutrient retention (Craft et al., 1991; Holland et al., 1995; Craft & Richardson, 1997). Studies also found reference wetlands had higher C and N content in sediments than in constructed wetlands (Bishel-Machung et al., 1996; Craft & Casey, 2000). Continuous saturation near the soil surface should limit decomposition of organic matter and retain N in natural wetlands (Stolt et al., 2000). Our results also showed that nutrient concentration in the extant wetlands were well above average level (MDEQ, 2001b) in southern Michigan, but restored marshes were at the beginning of succession and had relatively low nutrient concentration and organic matter accumulation. Algal species composition between restored and extant wetlands Presumably, algal assemblages in restored wetlands would gradually increase their similarity to extant wetlands due to an increasing similarity of environmental characteristics. However, the starting point of a restored wetland may be different from its original state. Furthermore, the physical, chemical, and biological conditions during succession could differ from natural wetland succes-

sion. Human perturbation during succession could also significantly change the path of succession (Detenbeck et al., 1999). Thus, a restored wetland may not necessarily develop to a state similar to mature natural wetlands (Mitsch & Wilson, 1996; Lockwood & Pimm, 1999; Kaiser, 2001). Although high variability in community composition among extant wetlands suggested a high degree of environmental heterogeneity in this study, both environmental conditions and epiphyton in restored wetlands still were significantly different from those in extant wetlands. None of the environmental variables or algal assemblage changes correlated with age of restored wetlands in this study. Previous studies (Bishel-Machung et al., 1996) also indicated that many environmental variables did not reach restoration goals within 10 years of initiating restoration. Some authors have suggested that wetland nutrient content might recover after more than 15 years of restoration (Craft & Richardson, 1998). According to MDEQ (2001a), only 29% of the mitigation projects in Michigan were considered biologically successful, and 26% of the mitigation sites in southern Michigan were observed to have poor water clarity after 10 year restoration. Mitsch & Wilson (1996) suggested that mitigation projects should be given as much as 15–20 years rather than 5 years before judging their success. Generally, restored wetlands had higher diatom gamma diversity, beta diversity, and alpha diversity on epiphytes than observed in extant wetlands. Other studies have also observed that restored wetlands did not necessarily have less biodiversity of plant and invertebrates than reference sites (Scatolini & Zedler, 1996; Bedford et al., 1999; Magee et al., 1999; Knutson, 1999; Williams & Zedler, 1999), and no evidence indicated a lower diatom species diversity in restored wetlands (Mayor et al., 1999). Three reasons might have caused low diversity in extant wetlands. First, eutrophication decreases algal species diversity and evenness (Detenbeck et al., 1999; Hillebrand & Sommer, 2000). The significantly higher nutrient levels in extant wetlands may be the cause of lower alpha diversity. Shading in extant wetlands by dense macrophyte canopies and duckweed also may have constrained algal diversity. The third possible explanation was that higher spatial heterogeneity

235 among restored wetlands at different stages of succession provided greater habitat variety for different algal species. The lower beta diversity of diatoms in extant wetlands suggested that diatom species dispersed evenly in extant wetlands, while the higher beta diversity suggested heterogeneous distribution of diatom assemblages among habitats in restored wetlands. The patterns in similarity of diatom species composition between restored and extant wetlands were related to environmental factors that may regulate biological differences during ecological succession. Differences in algal species composition on macrophytes between these two types of wetlands were related to light variables, but the model only explained 20% of the variance of diatom similarity among wetlands. On the other hand, the similarity in sediment diatom species composition between extant and restored wetlands could be explained by difference in pH, nutrients, and light effects. This model was much stronger (explaining 54% of the variance in diatom similarity) and indicated the importance of nutrient accumulation and light regime, particularly for regulating algal assemblages during wetland succession. The apparent differences in biogeochemical conditions between restored and extant wetlands make diatom assemblages ideal indicators of wetland conditions. Algal species composition among different habitats Phytoplankton communities were very different from benthic algal communities. As expected, phytoplankton and benthic algal communities had different growth forms which were adapted to planktonic and benthic habitats (Stevenson, 1996). Phytoplankton communities contained few diatoms, despite their abundance in the benthos. The majority of southern Michigan wetlands in this study were in open and shelter states (Goldsbough & Robinson, 1996), which should be dominated either by epiphyton or metaphyton. Phytoplankton density was generally very low. Duckweed and macrophyte shading not only caused light limitation for phytoplankton, but may also have competed with phytoplankton for nutrients, as suggested in other studies (Herbst & Hartman, 1981). Phytoplankton had high beta diversity, and the structural features of algal assemblages lacked

a common pattern among sites. Grazing, allelopathy, light limitation, and alternating self-organization processes in these wetlands make phytoplankton dynamics more complex than in lakes due to the large number of interacting factors (Rojo et al., 2000a, b). The composition of non-diatom algae differed much more between benthic habitats than did diatom species composition. The difference in non-diatom algae between macrophytes and sediments in restored wetlands was probably due to the very complicated interactions of algae and their substrata, involving both physical and chemical processes (Burkholder, 1996). The physical and chemical habitats for benthic algae on organic substrata are very different than on the firm substrata of macrophytes. Lower light due to denser macrophytes and duckweed cover may have constrained the development of differences between non-diatom epipelon and epiphyton growth in extant wetlands, the only comparison of non-diatom algae that was not significantly different among habitats. In addition, high nutrients in extant wetlands may have reduced nutrient supply differences between epipelon and epiphyton. Diatom assemblages on macrophytes and in sediments, however, were not different in restored and extant wetlands. Lim et al. (2001) claimed that sediment diatom communities reflect changes in the diatom communities on epiphytic and rock substrates in response to limnological conditions. This study also demonstrated that sediment diatom assemblages contained the majority of epiphytic diatom taxa within a wetland, but included a wider range of species than epiphyton. Although epiphytic and sediment diatoms did not differ in restored and extant wetlands, the regression model indicated that differences between these two habitats were a function of several different factors including light, DO, and nutrient ratio. In summary, the physical, chemical, and biological features of restored wetlands differed from extant wetlands in a number of ways. In particular, shading and nutrient concentrations, were distinctly greater in extant wetlands and probably contributed to differences in algal communities among habitats within wetlands and between restored and extant wetlands. We found no direct evidence for increasing similarity in algal assemblages between restored and extant wetlands with

236 increasing age of restored wetlands. However, response of algal assemblages to nutrient concentrations and light, two factors distinguishing restored and extant wetlands, indicated indirectly that successional processes in restored wetlands were regulating algal assemblages. Restored wetlands <10 years old may need more time to develop mature qualities.

References Adamus, P. R., 1996. Bioindicators for Assessing Ecological Integrity of Prairie Wetlands. US Environmental Protection Agency, Western ecological division, Corvallis, OR. USA. EPA/600/R-96/082. American Public Health Association (APHA), 1998. In Clececeri, L. S., A. E. Greenberg & A. D. Eaton (eds), Standard Methods for the Examination of water and wastewater. 20th edn., Washington D.C., USA. Bedford, B., 1996. The need to define hydrologic equivalence at the landscape scale for freshwater wetland mitigation. Ecological Applications 6: 57–68. Bedford, B. L., M. R. Walbridge & A. Aldous, 1999. Pattern in nutrient availability and plant diversity of temperate North American Wetland. Ecological Applications 80: 2151–2169. Beilman, D. W., 2001. Plant community and diversity change due to localized permafrost dynamics of bogs in western Canada. Canadian Journal of Botany 79: 983–993. Bishel-Machung, L., R. A. Brooks, S. S. Yates & K. L. Hoover, 1996. Soil Properties of reference wetlands and wetland creation projects in Pennsyvania. Wetlands 16: 532–541. Biondini, M. E., C. D. Bonham & E. F. Redente, 1985. Secondary successional patterns in a sagebrush (Artemisia tridentata) community as they relate to soil disturbance and soil biological activity. Vegetatio 60: 25–36. Burkholder, J. M., 1996. Interactions of benthic algae with their substrata. In Stevenson, R. J., M. L. Bothwell & R. L. Lowe (eds), Algal Ecology: Freshwater Benthic Ecosystems. Academic Press, San Diego, pp 3–30. Charles, D. F., C. Knowles & R. S. Davis, 2002. Protocols for the analysis of algal samples collected as part of the U.S. Geological Survey National water-Quality Assessment Program. Report No. 02–06: 215/299–1000. The Academy of Natural Sciences, Patrick Center for Environmental Research, Phycology Section, Philadephia, PA. Clarke, K. R., 1993. Non-parametric multivariate analyses of changes in community structure. Australian Journal of Ecology 18: 117–143. Cottingham, K. L. & S. R. Carpenter, 1998. Population, community, and ecosystem variates as ecological indicators: Phytoplankton responses to whole lake enrichment. Ecological Applications 8: 508–530. Craft, C. B. & C. J. Richardson, 1997. Relationships between soil nutrients and plant species composition in everglades peatlands. Journal of Environmental Quality 26: 224–232.

Craft, C. B. & C. J. Richardson, 1998. Recent and long-term organic soil accretion and nutrient accumulation in the everglades. Soil Science Society of America Journal 62: 834–843. Craft, C. B., E. D. Seneca & S. W. Broome, 1991. Porewater chemistry of natural and created marsh soils. Journal of Experimental Marine Biology and Ecology 152: 187–200. Craft, C. B. & W. P. Casey, 2000. Sediment and nutrient accumulation in floodplain and depressional freshwater wetlands of Georgia, USA. Wetlands 20: 323–332. Detenbeck, N. E., 2001. Methods for Evaluating Wetland Condition: Wetlands Classification. EPA 822-R-01–007g. U.S. Environmental Protection Agency; Office of Water; Washington, DC. Detenbeck, N. E., S. M. Galatowitsch, J. Atkinson & H. Ball, 1999. Evaluating perturbations and developing restoration strategies for inland wetlands In the Great Lakes Basin. Wetlands 19: 789–820. Galatowitsch, S. M. & A. G. van der Valk, 1996. The vegetation of restored and natural prairie wetlands. Ecological Applications 6: 102–112. Goldsborough, L. G. & G. G. C. Robinson, 1996. Pattern in wetlands. In Stevenson, R. J., M. L. Bothwell & R. L. Lowe (eds), Algal ecology: freshwater benthic ecosystems. Academic Press, San Diego, pp 121–148. Gosz, J. R., 1992. Gradient analysis of ecological changes in time and space: implications for forest management. Ecological Applications 2: 248–261. Gwin, S. E., M. E. Kentula & P. W. Shaffer, 1999. Evaluating the effects of wetland regulation through hydrogeomorphic classification and landscape profiles. Wetlands 19: 477–489. Herbst, R. P. & R. T. Hartman, 1981. Phytoplankton distribution of a duckweed covered pond. Journal of Freshwater Ecology 1: 97–111. Hillebrand, H., C. D. Durselen, D. Kirschtel, U. Pollingher & T. Zohary, 1999. Biovolume calculation for pelagic and benthic microalgae. Journal of Phycology 35: 403–424. Hillebrand, H. & U. Sommer, 2000. Diversity of benthic microalgae in response to colonization time and eutrophication. Aquatic Botany 67: 221–236. Hillebrand, H. & M. Kahlert, 2001. Effect of grazing and nutrient supply on periphyton biomass and nutrient stoichiometry in habitats of different productivity. Limnology and Oceanography 46: 1881–1898. Holland, C. C., J. E. Honea, S. E. Gwin & M. E. Kentula, 1995. Wetland degradation and loss in the rapidly urbanizing area of Portland, Oregon. Wetlands 15: 336–345. Jackson, D. A., 1993. Stopping rules in principal components analysis: A comparison of heuristical and statistical approaches. Ecology 74: 2204–2214. John, J., 1993. The use of diatoms in monitoring the development of created wetlands at a sandmining site in western Australian. Hydrobiologia 269/270: 427–436. Kaiser, J., 2001. Recreated wetlands no match for original. Science 293: 25. Keddy, P., 1999. Wetland restoration: The potential for assembly rules in the service of conservation. Wetlands 19: 716–732.

237 Kentula, M. E., R. P. Brooks, S. E. Gwin, C. C. Holland, A. D. Sherman & J. C. Sifness, 1992. Wetlands: An Approach to Improving Decision Making in Wetland Restoration and Creation. Island Press, Washington, DC. 151 pp. Knutson, M. G., J. R. Sauer, D. A. Olsen, M. J. Mossman, L. M. Hemesath & M. J. Lannoo, 1999. Effects of landscape composition and wetland fragmentation on frog and toad abundance and species richness in Iowa and Wisconsin, U.S.A. Conservation Biology 13: 1437–1446. Koerselman, W. & B. Beltman, 1988. Evapotranspiration from fens in relation to penmans potential free-water evaporation (e0) and pan evaporation. Aquatic Botany 31: 307–320. Leland, H. V., 1995. Distribution of Phytobenthos in the Yakima River Basin, Washington, in Relation to Geology, land-use, and other environmental-factors. Canadian Journal of Fisheries and Aquatic Sciences 52: 1108–1129. Lim, D. S. S., C. Kwan & M. S. V. Douglas, 2001. Periphytic diatom assemblages from Bathurst Island, Nunavut, Canadian High Arctic: an examination of community relationships and habitat preferences. Journal of Phycology. 37: 379–392. Lockwood, J. L. & Pimm S. L., 1999. When does restoration succeed? In Weiher, E. & P. Keddy (eds), Ecological Assembly Rules: Perspectives, Advances, Retreats. Cambridge University Press, Cambridge, UK, 363–392. Magee, T. K., T. L. Ernst, M. E. Kentula & K. A. Dwire, 1999. Floristic comparison of freshwater wetlands in an urbanizing environment. Wetlands 19: 517–534. Mayer, P. M. & S. M. Galatowitsch, 1999. Diatom communities as ecological indicators of recovery in restored prairie wetlands. Wetlands 19: 765–774. McCormick, P. V., P. S. Rawlik, K. Lurding, E. P. Smith & F. H. Sklar, 1996. Periphyton-water quality relationships along a nutrient gradient in the northern Florida Everglades. Journal of the North American Benthological Society. 15: 433–449. McCormick, P. V., R. B. E. Shuford, J. G. Backus & W. C. Kennedy, 1998. Spatial and seasonal patterns of periphyton biomass and productivity in the northern Everglades, Florida, USA. Hydrobiologia 362: 185–208. McCune, B. & J. Mefford, 1997. PC-ORD. Multivariate Analysis Of Ecological Data. Version 3.0 for Windows. Mjm Software Design, Gleneden Beach, OR, USA. McDonnell, M. J. & S. T. A. Pickett, 1990. Ecosystem structure and function along urban rural gradients – an unexploited opportunity for Ecology. Ecology 71: 1232–1237. Michigan Department of Environmental Quality (MDEQ), 2001a. Michigan Wetland Mitigation and Permit Compliance Study, Final Report. Land and water management division, Michigan Department of environmental Quality, Lansing, MI. Michigan Department of Environmental Quality (MDEQ), 2001b. MDEQ Wetland Identification Manual: A Technical Manual for Identifying Wetlands in Michigan, Land and Water Management Division, Michigan Department Of Environmental Quality. Middleton, B., 1999. Wetland Restoration: Flood Pulsing and Disturbance Dynamics. John Wiley & Sons, Inc, New York, 6–7 pp.

Mitsch, W. J. & B. F. Wilson, 1996. Improving the success of wetland creation and restoration with know-how, time, and self-design. Ecological applications 6: 77–83. National Research Council (NRC), 1995. Wetlands: Characteristics and Boundaries. Nation Research Council. National Academy Press, Washington, DC, USA. Pan, Y. D. & R. J. Stevenson, 1996. Gradient analysis of diatom assemblages in Western Kentucky wetlands. Journal of Phycology 32: 222–232. Pan, Y., R. J. Stevenson, P. Vaithiyanathan, J. Slate & C. J. Richardson, 2000. Changes in algal assemblages along observed and experimental phosphorus gradients in a subtropical wetland, USA. Freshwater Biology 44: 339–353. Pringle, C. M., 1990. Nutrient spatial heterogeneity – Effects on community structure, physiognomy, and diversity of stream algae. Ecology 71: 905–920. Rader, R. B. & C. J. Richardson, 1992. The effects of nutrient enrichment on algae and macroinvertebrates in the everglades: a review. Wetlands 12: 121–135. Rojo, C., E. Ortega-Mayagoitia & M. Alvarez-Cobelas, 2000a. Lack of pattern among phytoplankton assemblages. Or, what does the exception to the rule mean? Hydrobiologia 424: 133–139. Rojo, C. E., E. Ortega-Mayagoitia, M. A. Rodrigo & M. Alvarez-Cobelas, 2000b. Phytoplankton structure and dynamics in a semiarid wetland, the National Park ‘‘Las Tablas de Daimiel’’ (Spain). Archiv Fur Hydrobiologie 148: 397–419. Scatolini, S. R. & J. B. Zedler, 1996. Epibenthic invertebrates of natural and constructed marshes of San Diego Bay. Wetlands 16: 24–37. Stevenson, R. J., 1996. An introduction to algal ecology in freshwater benthic habitats. In Stevenson, R. J., M. L. Bothwell & R. L. Lowe (eds), Algal Ecology: Freshwater Benthic Ecosystems. Academic Press, San Diego, pp 3–30. Stevenson, R. J. & S. Hashim, 1989. Variation in Diatom Community Structure among Habitats in Sandy Streams. Journal of Phycology 25: 678–686. Stevenson, R. J., 1997. Scale-Dependent Determinants and Consequences of Benthic Algal Heterogeneity. Journal of the North American Benthological Society 16: 248–262. Stevenson, R. J., R. P. Sweets, Y. Pan & R. E. Schultz, 1999. Diatom community patterns in wetlands of Michigan (USA) and their use as indicators of ecological conditions. In McComb, A. J. & J. A. Davis (eds), Proceedings of INTECOL’s Vth International Wetland Conference. Gleneagles Press, Adelaaide, Australia, pp 517–527. Stevenson, R. J., P. V. McCormick & R. Frydenborg, 2001. Methods for Evaluating Wetland Condition: Using Algae to Assess Environmental Conditions in Wetlands. EPA 843-B00–002k. U.S. Environmental Protection Agency, Office of Water Washington, DC. Stolt, M. H., M. H. Genthner, W. L. Daniels, V. A. Groover, S. Nagle & K. C. Haering, 2000. Comparison of soil and other environmental conditions in constructed and adjacent palustrine reference wetlands. Wetlands 20: 671–683. von Stosch, H. A. & B. E. F. Reimann, 1970. Subsilicea fragilarioide gen. Et spec. nov., eine Diatomee (Fragilariaceae)

238 mit vorwiegend organischer membran. Nova Hedwigia Beih. 31: 1–36. Tilman, D., 1981. Tests of resource competition theory using four species of lake Michigan algae. Ecology 62: 802–815. Tiner, R. W., 2003. Geographically isolated wetlands of the United States. Wetlands 23: 494–516. USEPA, 2000. Principles for the Ecological Restoration of Aquatic Resources. EPA841-F-00–003, Office of Water (4501F), United States Environmental Protection Agency, Washington, DC. 4 pp. van Der Valk, A. G., 1981. Succession in wetlands: a gleasonian approach. Ecology 62: 688–696. Vymazal, J., 1995. Algae and Element Cycling in Wetlands. Lewis Publishers, Boca, Raton, FL. 689 pp. Wetzel, R. G., 2001. Limnology: Lake and River Ecosystems. Academic Press, San Diego, USA.

Whigham, D. F., R. E. Good & J. Kvet, 1990. Wetland Ecology and Management: Case Studies. Kluwer Academic Publishers, Netherlands. Williams, G. D. & J. B. Zedler, 1999. Fish assemblage composition in constructed and natural tidal marshes of San Diego Bay: Relative influence of channel morphology and restoration history. Estuaries 22: 702–716. Wu, X. Y. & W. J. Mitsch, 1998. Spatial and temporal patterns of algae in newly constructed freshwater wetlands. Wetlands 18: 9–20. Zedler, J. B., 1993. Canopy architecture of natural and planted Cordgrass marshes: selecting habitat evaluation criteria. Ecological Applications 3: 123–138. Zedler, J. B., 2000. Progress in wetland restoration ecology. Trends in Ecology & Evolution 15: 402–407.

Hydrobiologia (2006) 561:239–249  Springer 2006 R.J. Stevenson, Y. Pan, J.P. Kociolek & J.C. Kingston (eds), Advances in Algal Biology: A Commemoration of the Work of Rex Lowe DOI 10.1007/s10750-005-1617-z

Ecology and assessment of the benthic diatom communities of four Lake Erie estuaries using Lange-Bertalot tolerance values Gerald V. Sgro1, Michael E. Ketterer2 & Jeffrey R. Johansen1,* 1

John Carroll University, University Heights, Ohio, USA Northern Arizona University, Flagstaff, Arizona, USA (*Author for correspondence: E-mail: [email protected]) 2

Key words: biomonitoring, diatoms, estuary, Lake Erie, Lange-Bertalot Index, water quality

Abstract Diatom composition of four Lake Erie estuaries was related to seasonal factors, year, location within the estuaries, and water quality parameters including nutrient and metals concentrations. Canonical correspondence analysis (CCA) revealed seasonality as the most important factor determining variability in diatom species composition among sites and dates. Alkalinity, pH, silicate, orthophosphate, and nitrite concentrations were water chemistry parameters correlated with diatom community composition. Eigenvalues for the first two CCA axes of nutrient/physical data and species data were higher than the first two CCA axes of metals data and species data. In addition, the water quality of these estuaries was evaluated using an index composed of Lange-Bertalot pollution tolerance values. The Lange-Bertalot index scores indicated that the Ashtabula estuary had the best water quality of the study sites. LangeBertalot index scores were highly correlated with a gradient of disturbance represented by the first axis of a principle components analysis of sites and nutrient data (Spearman q=0.7). The Lange-Bertalot tolerance values could be useful for discriminating ‘good’ sites from ‘‘bad’’ sites among the Lake Erie estuaries.

Introduction Anthropogenic changes in the ecology of the Laurentian Great Lakes have been occurring at least during the last century. These physical, chemical, and biological changes have resulted in a loss of biologic integrity in offshore and near shore habitats such as estuarine areas throughout the Lakes. Changes in the biological communities of these habitats have included long term alteration of diatom community structure (Davis, 1964; Stoermer et al., 1985; Herdendorf, 1989). Attempts to restore the quality of the estuarine habitats along Lake Erie’s southern shore have included regulation of point source loadings, riparian improvements to abate non-point source

loadings, and habitat restoration (OEPA, 1991; Ohio Department of Natural Resources, 1997). Measurements of the extent of degradation and restoration of the environment in these habitats often involves monitoring the biota supported by the habitat and employing indices such as the Index of Biotic Integrety-Lacustrine (IBI-L) (Thoma, 1998) and the Invertebrate Community Index (ICI) (OEPA, 1987) which relate attributes of biota at study sites with reference sites. An index of biologic integrity based on algae does not yet exist for these estuarine habitats. Phytoplankton has been used to assess offshore areas of the lakes (Makarewitcz, 1985); however, the efficacy of phytoplankton as an assessment tool in near shore habitats such as the estuarine

240 areas of Lake Erie’s tributaries is questionable because of mixing with lake and river water, water currents, and patchiness of the plankton. Benthic diatoms would be better for assessment because they attach to substrates and integrate conditions in that local area for the life of the organism (Stevenson & Lowe, 1986; Porter et al., 1993). Many indicators have been developed to infer environmental conditions using benthic diatoms including community composition and autecological indices which evaluate changes in the diatom assemblages due to sensitivity and tolerance of the species to environmental variables (see review in Sgro & Johansen, 1995). Precise indices can be constructed using the weighted average inference methods of ter Braak & van Dam (1989). However, no attempts have yet been made to use benthic diatoms to assess the relative health of Lake Erie’s estuaries. Nor have the benthic diatoms in most of Lake Erie’s estuaries been studied. How similar are the benthic diatom communities in these systems to each other? What natural seasonal and habitat factors regulate the species composition of diatom communities in these estuaries? How do environmental impacts affect the integrity of these benthic communities? The objectives of this study were to investigate the relationship between diatom composition and the impacts of environmental variables at four Lake Erie estuaries and to test the ability of the LBI (Lange-Bertalot Index, see Lange-Bertalot, 1979) to comparatively assess the quality of the sites. We used canonical correspondence analysis (CCA) of species and two categories of environmental variables to detect which of the measured variables were most important in impacting the integrity of the diatom communities. Fifty-seven samples were collected over a 2 year period from upstream and downstream sites in these estuaries were scored using the LBI. Principal components analysis (PCA) of sites and variables enabled a construction of a gradient of disturbance for these sites. We then evaluated the LBI by comparing its assessment of the habitat quality of the sites with the gradient of disturbance established from the PCA. This paper will demonstrate the merit of using the LBI, an index based on European rivers, for assessing impacted estuary sites in Lake Erie.

Material and methods Site description Benthic diatoms were assessed in four estuaries of Lake Erie, the Cuyahoga, Black, and Ashtabula Rivers and Old Woman Creek. The Cuyahoga River has the largest drainage basin and Old Woman Creek drains the smallest area (Table 1). Two sampling sites in each estuary were established to characterize diatom composition within the estuary. One was chosen upstream where lake water makes furthest encroachment, the other downstream near the river mouth. Downstream sites may be subjected to greater pollution impacts in the estuaries where urban land use occurs between the upstream and downstream sites. Lake water might have a greater influence on river mouth sites, especially during seasons of low river flow (Trebitz et al., 2002) Agriculture is the predominant land use in the Old Woman Creek watershed (Brant & Herdendorf, 1972). Silts and clays have accumulated in the estuary as agricultural activities have increased over the past 100 years (Buchanan, 1983). Flow through the estuary is controlled by storm events as well as the presence or absence of a barrier beach. The presence of the beach creates lentic conditions with water residence time measured in weeks; the absence of the beach creates lotic conditions with water residence time measured in hours (Klarer, 1988). Land use of the Cuyahoga River basin varies from predominantly agricultural in the upper basin to a densely populated urban and industrial area in the lower basin. Habitat has been severely modified in the navigation channel. The morphology of the navigation channel contributes greatly to the extended residence time of water in the estuary. Discharges from combined sewer Table 1. Size of estuaries (Brant & Herdendorf, 1972) Estuary

Length

Area

Drainage

(km)

(km2)

(km2)

Old Woman Creek

2.1

0.3

69

Black River

6.6

0.9

1217

Cuyahoga River

7.2

1.0

2095

Ashtabula River

2.8

0.3

355

241 overflows (CSO) and sanitary sewer overflows (SSO) are major sources of sediment oxygen demand. Metals and inorganic compounds reach the river from non-point and point sources. Sewage treatment plants and CSO/SSO’s are the major sources of nitrogen and phosphorous loadings (Ohio Environmental Protection Agency, 1994). The Cuyahoga River estuary is exposed to chronic pollution and is designated an Area of Concern (AOC) by the International Joint Commission (Cuyahoga River Community Planning Organization, 1992, unpublished). The Black River Basin is largely agricultural. Soil erosion from agricultural land is believed to be a major source of sediment in the estuary. Only 10% of the basin is urban. Major industrial areas are located between river miles 5.4 and 3.3. The dredged shipping channel reduces dissolved oxygen (DO) in much of the estuary, while both point and non-point sources of pollution contribute loadings of heavy metals, toxic organics, conventional pollutants and contaminated sediments. The Black River is a designated AOC (Black River Remedial Action Plan Coordinating Committee, 1994, unpublished). The Ashtabula River basin is 87% rural and agricultural and much of this activity is confined to the upper reaches of the river. The shipping channel (originally dredged to 6 m) was not dredged for 30 years due to the high PCB content of the sediment. The sources of organic contaminants are the abandoned hazardous waste sites in the Strong Brook and Fields Brook tributaries. Fields Brook is a Superfund site, due to the fact that the sediments are so PCB rich that they qualify as toxic waste. This estuary is also in an AOC (Ohio Environmental Protection Agency, 1991). Sampling and analysis Benthic diatoms on artificial substrates were collected quarterly from November, 1992 through September, 1994. Artificial substrates were used to reduce variability in microhabitat features among sites sampled and suitable, stable natural substrates for diatom colonization were rare. A floating diatometer containing six glass slides was placed at each site with slides submerged just below the water surface. The diatometers were positioned in similar conditions off stream banks

and remained in position for approximately two weeks before sampling. Periphyton on three selected diatometer slides from each site was processed independently to make three permanent diatom slides for counting. Diatometer slides that were under colonized or uneven in colonization were eliminated, and the three slides processed were chosen at random from those remaining. Permanent diatom slides were prepared by cleaning with nitric acid and potassium dichromate and mounting in Naphrax diatom mountant (Sgro & Johansen, 1995). Diatoms were examined at 1250 using an Olympus BH2 photomicroscope with Nomarski DIC optics. Nine hundred frustules were identified and counted for each site and date (three hundred from each processed slide). No attempt was made to distinguish living from dead organisms, though proportions of dead algae are probably small on diatometer slides (Stevenson & Lowe, 1986). Seven diatometers were lost and consequently no data is available for the following site/date combinations: November 1992, Black and Cuyahoga downstream; April 1993, Old Woman Creek downstream and Cuyahoga upstream; September 1993, Cuyahoga downstream; April 1994, Cuyahoga downstream; September 1994, Cuyahoga upstream. This resulted in 57 characterizations of diatom assemblages for the 2 years, 4 seasons, 2 locations per estuary, and 4 estuaries. At all sites temperature and pH were measured in the field with a Barnant 30 pH meter and DO was measured with a YSI 51B DO meter. Water samples were collected in polypropylene bottles rinsed with DI water for analysis of the following parameters: silicate, nitrate–nitrite, nitrite, ammonia, orthophosphate, sulfate, alkalinity, and conductivity. Standard methods were used (American Public Health Administration, 1989) for all the above analyses, except for a modification of the ammonia assay first recommended by Zaborojny et al. (1973). A Unicam UV 2-100 photospectrometer and a Radiometer TTT 80 autotitrator were used in these analyses. Additional water samples were collected in polypropylene bottles and filtered in the field with a 0.45 lm filter and treated with metals-grade nitric acid to pH <2. From these samples cation and heavy metal concentrations were determined using several techniques depending upon time and

242 availability of instruments. Concentrations of Fe, Ca, Mg, Cr, and Ti were determined in the samples from November 1992 to April 1993 by ICP emission spectrometry using a Jarrel-Ash 61 Spectrophotometer. Concentrations of Cu, Se, Cd, and Pb in these samples were determined using an ELAN 250 inductively coupled plasma mass spectrometer (ICP/MS). Concentrations of Mg, Ca, Fe, Cr, T, Cu, Se, Cd, Pb, and Zn in samples collected from June 1993 through April 1994 were determined using ICP/MS as described above. Concentrations of Mg and Fe were determined by flame atomic absorption analysis in samples from June to September 1994 using an Instrumentation Laboratory Smith-Hieftje Model 11AA/AE Spectrophotometer. Concentrations of Ca, Ti, Cr, Zn, Se, Cd, and Pb in these samples were determined using ICP/ MS as described above. The analyte and internal standard spectral lines monitored in this work were as follows: Mg 24, Ca 44, Ti 47, Cr 52, Fe 56, Cu 65, Zn 68, Se 82, Cd 111, Pb 206, Pb 207, and Pb 208. Hg was not detected in any sample by the methods of sample collection and analysis used in this study (Sgro & Johansen, 1998). Detrended correspondence analysis (DCA) (ter Braak, 1987, 1990) was used to determine unimodal character of species responses. Gradient lengths for the first four axes of DCA with species/ nutrient data ranged from 2.7 to 3.7 and with species/metals data from 2.8 to 3.6, thus we selected a CCA ordination model (Jongman et al., 1995) using CANOCO v3.0 (ter Braak, 1987, 1990) to analyze algal community response to water chemistry gradients. CCA ordinates both sites and taxa on axes that are linear combinations of environmental variables. Two different sets of environmental variables were used for two separate analyses. One set of environmental variables included temperature, pH, DO, conductivity, alkalinity, and drainage basin size, plus the six nutrient parameters collected for the study (Sgro & Johansen, 1998). The other set was of metals data (Sgro & Johansen, 1998). We wanted metals to have equal importance as the nutrient variables in examining pollution impact on diatom distribution in these sites and not analyzed simply as co-variables with nutrients. Therefore, we chose to keep the data sets separate for these analyses. The data were screened using CALIBRATE v.86 (Juggins & ter Braak, 1992) and CANOCO

v3.0 (ter Braak, 1987, 1990) to eliminate species and taxa which would not provide information in the analysis, for example, rare species and variables that are correlated and thus redundant. There were 57 diatometer samples used in these analyses. Taxa which were not equal to or greater than 1% relative abundance in at least three samples were eliminated. There were 73 species retained for the analysis. We did not perform square root transformations or downweighting of rare species in this analysis. Forward selection was used to remove redundant environmental variables and determine which variables accounted for the greatest amount of variance in the diatom distribution (ter Braak, 1987, 1990). Monte Carlo permutation tests (999 permutations) were used to test the significance of each variable added by this procedure. Variables were eliminated which provided approximately less than 0.11 extra fit in the analysis. Seven variables from the nutrient data set were retained for the analysis (temperature, pH, alkalinity, orthophosphate, nitrite, silicate, and drainage). Five variables from the metals data set were retained (Se Ti, Ca, Mg, and Cd). Histograms of variable values were generated with the program CALIBRATE v.86 (Juggins & ter Braak, 1992) to identify variables with skewed distributions. The variables with skewed distributions were either log or square root transformed to satisfy the statistical assumptions of ordination (Jongman et al., 1995). A LBI was calculated for each sample (LangeBertalot, 1979). The LBI was chosen because it is used to assess rivers in at least two other states in the USA (Bahls et al., 1992; Kentucky Department of Environmental Protection, 1993). Species tolerance values for the LBI were determined from observations in the Main and Rhine Rivers and based on species tolerance to organic pollution (Lange-Bertalot, 1979). An indicator value of 3 is assigned to the least pollution tolerant species, a value of 2 is assigned to more tolerant species and a value of 1 is assigned to the most tolerant species. The index was calculated by multiplying the indicator value assigned to a species by its relative density in the sample. Therefore, theoretically, a calculated index score of 3 for a sample would indicate excellent water quality and a score of 1 would indicate poor water quality.

243 In the present study, about a third of the 270 species was assigned an indicator value and was used in calculating the pollution index for these samples. The remaining species (Sgro & Johansen, 1998) in this study were not included in the LBI and were not used for calculating the pollution index for these samples. To compare LBI to a gradient of human disturbance characterized with PCA of nutrient data measured in the study (ammonia, orthophosphate, nitrate, nitrite, and alkalinity). Statistical software (Manugistics Corp., 1992) was used for PCA. A separate analysis was performed on cool season (November/December and April) and warm season (June and September) samples. An additional warm season analysis was performed on the metals parameters retained in the CCA (Ca, Cd, Mg, Se, and Ti). The influence of outliers was removed prior to PCA by calculating the natural logarithm for the variables and in the case of orthophosphate by adding 2 to the values before calculating the natural logarithm. One was added to the natural log of the metals. All data were standardized. PCA constructs theoretical axes which are linear combinations of the variables (Jongman et al. 1995). The first axis explains the greatest amount of variance within the variables and for our purpose represents a gradient of disturbance estimated from the variables for the sites. A cool season gradient for the metals was not determined due to the large number of undetected or unreported values for Se and Ti in the cool seasons. Ti concentrations were mostly below detection levels in the cool seasons by the methods used. Se concentrations were also below detection levels in cool seasons, while in December, 1993 and April, 1994 samples were not reported due to an R2 value <0.995 for the calibration curve. Spearman’s rank correlation coefficient (Spearman’s q) was determine the relationship between LBI and the gradients of disturbance represented by the component scores along the first PCA.

Results We identified 270 diatom taxa from all 57 diatometer samples (Sgro & Johansen, 1998). Navicula lanceolata (Ag.) Ehr., Achnanthidium minutissimum Ku¨tz., and Gomphonema parvulum

Ku¨tz., were the most common pennate species in the samples. Melosira varians Ag. was the most common centric diatom in these samples. In the species-nutrient CCA analysis (Fig. 1) eigenvalues for axes 1–4 were 0.509, 0.349, 0.208, and 0.167, respectively. The species environmental correlations were high, being 0.897 and 0.828 for axes 1 and 2, respectively. The first two axes explained 57.0% of the variance in the data. We interpreted the first ordination axis to be a temperature gradient with warm weather samples on the right and cool weather samples on the left. In this analysis seasonality (temperature) was the most important factor influencing the diatom community structure in these estuaries. Relatively low nutrient enrichment samples occurred above the horizontal axis and relatively high enrichment samples fell below the horizontal axis (Fig. 1). In our study, the Ashtabula River was consistently low in nutrients, while the Cuyahoga and Black Rivers were consistently high in enrichment. Old Woman Creek varied over time, and had both high and low nutrient samples, although the majority was high enrichment. In the CANOCO analysis of diatometer samples using the metals analysis data, the first ordination axis delineated an environmental gradient with samples associated with high concentrations of Se and Ti on the right (Fig. 2). The second, or vertical axis, depicted a gradient with samples associated with Ca and Ti above the horizontal axis and higher Mg, Se, and Cd concentrations below the horizontal axis. However, the eigenvalues for this analysis were low (0.378, 0.281, 0.140, 0.071, axes 1–4 respectively), indicating that the diatom assemblages do not appear, in this analysis, to be strongly impacted by metals concentrations. The highest LBI scores among all samples were obtained from the Ashtabula samples on all sampling dates except in April when Old Woman Creek had the highest score (Table 2). Old Woman Creek had generally higher scores than the Black and Cuyahoga systems except in summer. Generally, highest scores were obtained among all samples in the November/December sampling period. The least separation of sites by index scores occurred during the cool seasons. Over the 2 year period the cool-season scores ranged from 1.97 for the April, 1993, Black River upstream sample to

244

CCA 2 AJ’ AN

AS AJ’

AD’ AN AA’ AS’ OS’ AJ

AA’ AA

OJ’

AS AJ

OS’

AA ON

AD’

AS’

CCA 1

OD’ OA’ OD’ BD’

BA

OS

CN

OS

CJ OJ

BN OA’ CD’ CA BD’CD BA’ OA ON SIL BA’ CA’

Temp

OJ CJ BJ BS CS’

BA

CS

pH

CJ’ BS’ BJ’ BJ’

BJ

PHOS

OJ’

CJ’

BS

BS’

Alk

DRAIN

NIT

Figure 1. Canonical correspondence analysis (CANOCO) of diatometer data and selected environmental parameters. In CANOCO, the axes on which the species and sites are simultaneously ordinated are constrained by the environmental variables. Environmental parameters include: silicate (Sil), nitrite (Nit), orthophosphate (Phos), temperature (Temp), pH, alkalinity (Alk), and size of drainage basin (Drain). Sites are represented by circles, solid circles represent downstream sites, hollow circles represent upstream sites. Twoletter site codes represent estuary (A=Ashtabula, B=Black, C=Cuyahoga, O=Old Woman Creek) and month of collection (N=November, D=December, A=April, J=June, S=September). An apostrophe (¢) means second year of sampling.

2.46 for the November, 1992 Old Woman Creek downstream sample (Table 2). Based on the pollution index scores, the Black and Cuyahoga had a more impaired algal flora than the Ashtabula in these cool winter periods, but were less different from the Ashtabula system in April than in November/December periods. The LBI scores for the Black and Cuyahoga dropped in the warm seasons, while the Ashtabula scores tended to improve. The range of warm season scores was from 1.13 for the June 1993 Old Woman Creek downstream site to 2.99 for the June 1994 Ashtabula upstream site. The first axis of the PCA for nutrients in cool seasons accounted for 72.8% of the variance and was negatively correlated to the LBI scores (Spearman q=)0.51, p=0.009). The Spearman correlation coefficient calculated between the LBI

scores and the PCA nutrient gradient during warm season, which accounted for 55.3% of variance, was )0.71 (p=0.0001). The scatter plot of warm season LBI scores with first axis PCA scores reveals a non-linear threshold response of the diatom assemblages (Fig. 3). No sample with a score greater than )1 on the first PCA axis has a LBI score greater than 1.8. The correlation between LBI scores and the warm season metals disturbance gradient (the first axis of the metals PCA, accounting for 42.2% of the variance) was weaker (Spearman q=)0.56, p=0.0024).

Discussion Seasonality was the most important factor, among those measured in this study, in determining

245 Table 2. LBI pollution index values for diatometer samples from the four estuaries in the study (OWC=Old Woman Creek, BR=Black River, CR=Cuyahoga River, AR=Ashtabula River, U=up, D=down). Absence of index value signifies loss of diatometer for that sample Site

Nov./Dec.

Apr.

June

Sep.

OWC-U

2.12

2.02

1.51

1.84

OWC-D

2.46

1.13

1.66

BR-U

2.07

1.97

1.74

1.77

2.02

1.59

1.18

1.27

1.63

1992–93

BR-D CR-U

2.02

CR-D AR-U

2.38

2.02 2.17

1.26 2.94

2.94

AR-D

2.24

2.12

2.77

2.52

OWC-U

2.11

2.22

1.47

2.30

OWC-D

2.04

2.35

1.95

2.18

BR-U

2.0

2.05

1.52

1.40

BR-D

2.10

2.13

1.38

1.57

CR-U CR-D

2.16 2.12

2.03

1.32 1.58

1.68

AR-U

2.28

2.05

2.99

2.81

AR-D

2.32

2.18

2.59

2.70

1993–94

variability in diatom species composition among sites and dates in these Lake Erie estuaries. Distinct cool season and warm season diatom assemblages were observed in these habitats, as in other studies of changes of community structure with temperature (Klarer & Hickman, 1975; Squires et al. 1979; Vinson & Rushforth, 1989). Variability among diatom communities was less during the cool seasons (November/December and April) than warm seasons (June, September). This may reflect effects of higher flows from watersheds during cool seasons that homogenize conditions among sites and estuaries. During low flow periods (typically warm seasons), local factors may be more important determinants of water quality and increase variability among sites and estuaries. Alkalinity, pH, silicate, orthophosphate, and nitrite concentrations also were correlated with diatom community composition. Phosphorous, nitrogen, and pH are usually the most important determinants of algal growth and community composition in streams (Pan et al., 1999). This study revealed a relationship between drainage basin size and orthophosphate, alkalinity, and nitrite. Other studies have shown that water chemistry variables in rivers may be functions of

drainage basin size (Hynes, 1960; Harrel & Doris, 1968) and land use (Omernik, 1976; Osborne & Wiley, 1988) which may regulate the loading of nutrients, along with riparian zone conditions (Johnson et al., 1997). This has important management implications for setting nutrient criteria. For example, small disturbed estuaries, such as Old Woman Creek in this study, may show more variability over time and space than large disturbed estuaries which may be more modulated due to greater flow. Also, what might be considered a healthy algal assemblage or index score for a large estuary may be different from a healthy algal assemblage or index score for a small estuary. The metals variables examined in this study were not strong factors in structuring the diatom assemblages based on low eigenvalues of the CCA axes. Metals may, however have other effects on diatom assemblages not examined in this study such as lowering algal productivity. Stevenson & Stoermer (1982) and Stevenson & Lowe (1986) discuss the possible role of toxic loads on species diversity. Gradients of disturbance based on nutrients and metals in both warm and cool seasons were correlated to LBI scores. The best agreement was

246

CCA 2 AA

AJ’

AN AA’

AJ’

AS’ AJ BS’ BJ CJ

AD’

Ca

AA CA CN

CCA 1

AS’

AJ AS AS

AA’ AN

BS AD’ OJ’ CA’ OD’ON BJOS’ OJ’ OS OA’ OA BS OJ OA’ BA’ OS ON BA OJ CD’ OD’ BA

BN

CJ’

CS’

Ti

OS’ CS CJ’ CJ

BD’

BA’ BJ’

Se

BJ’

CD’

Cd

BS’

Mg

Figure 2. CANOCO of diatometer data and selected element concentrations as determined by ICP mass spectrophotometry. In CANOCO, the axes on which the species and sites are simultaneously ordinated are constrained by the environmental variables. Environmental parameters include: calcium, titanium, selenium, cadmium, and magnesium. Sites are represented by circles, solid circles represent downstream sites, hollow circles represent upstream sites. Two-letter site codes represent estuary (A=Ashtabula, B=Black, C=Cuyahoga, O=Old Woman Creek) and month of collection (N=November, D=December, A=April, J=June, S=September). An apostrophe (¢) means second year of sampling.

between the LBI scores and the gradient of disturbance for nutrient pollution in warm seasons which were moderately well correlated based on the rank correlation test. All sites where dominated by Navicula lanceolata in cool seasons which reduced the range in LBI scores. Consequently, the LBI correlated less well with a gradient of disturbance in cool seasons. The implication is that the LBI would be best used in warm seasons to assess these systems. The LBI may be useful in separating ‘good’ sites from ‘bad’ sites with regard to nutrient impacts in these estuaries. Based on the LBI, the Ashtabula estuary supported the largest percentage of pollution sensitive diatoms and thus ranked as the least impaired of the estuaries in the study. The Old Woman Creek estuary was somewhat impaired and the Black and Cuyahoga estuaries were most impaired by this measure.

Diatom assemblages in these estuaries may have responded in a threshold manner to these environmental impacts, rather than with a more clearly linear response. The threshold response has been observed for species of other organisms. Effects of urbanization on fish populations (Klein, 1979; Schuler & Gali, 1992) showed a precipitous decrease in Index of Biotic Integrity scores for sites with less than approximately 90% pervious surfaces in their watersheds. Clean water fish taxa are not supported below the 90% threshold. A threshold response by the diatom assemblage to environmental impacts suggests an assimilative capacity by the diatom assemblage. The community may be resilient to slight changes in nutrient concentrations, but greater changes cause a dramatic change in dominant species. LBI scores would tend to remain roughly the same until

247 3.5

LBI Scores

3 2.5 2 1.5 1 0.5 0

-3

-2

-1

0

1

2

3

4

Site Scores on First PCA Axis Figure 3. Scatter plot of LBI scores against site scores along the first axis of the warm season, nutrient PCA.

dominant species that are either more pollution tolerant or sensitive shift dramatically within an assemblage. The LBI may lose statistical precision because there is not a clear linear relationship between the disturbance gradient and index scores. Although the LBI was developed for other ecosystems and specific objectives, it worked remarkably well in Lake Erie estuaries. The values assigned to the species from Lange-Bertalot (1979) are based on empirical observations of effects of organic pollution on diatoms in European rivers, not Lake Erie estuaries. Organic pollution is most related to BOD which was not directly measured in this study. The values were not developed to measure impacts due to metals loadings. Furthermore, it is likely that some species will have different pollution tolerances in Lake Erie estuaries than in European rivers, even though evidence for global similarities in diatom autecology’s are indicated in the literature (Lowe, 1974; Beaver, 1981; van Dam et al., 1994). Many procedures using diatoms have been developed to reconstruct or assess environmental conditions past and present, including multiple linear regression (see Charles & Smol, 1988), weighted average regression and calibration (see Dixit et al., 1993; Dixit & Smol, 1994), weighted average-partial least squares and non-parametric

smoothing techniques (see Potapova et al., 2004). These techniques are more statistically concise than the empirically derived LBI. All of these procedures are designed to infer a particular variable such as chloride or pH from the diatoms. All the researchers above report correlations between diatom inferred and observed variables equal to or greater than the correlations between LBI scores and pollution we report in this study. Yet, the LBI can be used to assess impacts of pollution on the biota in the absence of a more rigorous mathematical model and where assessing the impact of specific nutrients is less desirable than a general pollution indicator. We found the LBI was very effective in differentiating estuaries of varying pollution impacts, and gave a metric that was easier to interpret than CCA and clustering. Future studies should refine metrics like the LBI for application in Great Lakes estuaries by characterizing autecologies of regional populations and their responses to common stressors.

Acknowledgements Collection of samples, microscopy, and initial data analysis were funded by a grant from the Lake Erie Protection Fund. Manuscript preparation and publication costs were covered by a grant

248 from the Committee on Research and Faculty Development, John Carroll University. The Old Woman Creek National Estuarine Research Reserve assisted in water chemistry analysis. Further data analysis was supported by a United States Environmental Protection Agency’s Science to Achieve Results (STAR) grant through the University of Minnesota Natural Recourses Research Institute’s Great Lakes Environmental Indicator Project. Although the research described in this article has been funded wholly or in part by the United States Environmental Protection Agency’s STAR program through cooperative agreement R8286750 to the University of Minnesota, it has not been subjected to the agency’s required peer and policy review and therefore does not necessarily reflect the views of the agency and no official endorsement should be inferred.

References American Public Health Administration (APHA), 1989. Standard Methods for the Examination of Water and Wastewater (17th ed.). American Public Health Administration, Washington, D.C. Bahls, L. L., R. Bukantis & S. Tralles, 1992. Benchmark Biology of Montana Reference Streams. Water Quality Bureau Helena, Montana, 107 pp . Beaver, J., 1981. Apparent Ecological Characteristics of Some Common Freshwater Diatoms. Ontario Ministry of the Environment, 517 pp. Brant, R. A. & C. E. Herdendorf, 1972. Delineation of Great Lakes estuaries. Proceedings 15th Conference Great Lakes Research, International Association Great Lakes Research 15: 710–718. Buchanan, D. B., 1983. Transport and deposition of sediment in Old Woman Creek Estuary of Lake Erie. Ohio State Univ., Ohio Sea Grant Tech. Bull. 10, 198 pp. Charles, D. F. & J. P. Smol, 1988. New methods for using diatoms and chrysophytes to infer past pH of low-alkalinity lakes. Limnology Oceanography 33(6, part 2): 1451–1462. Davis, C. C., 1964. Evidence for the eutrophication of Lake Erie from phytoplankton records. Limnology Oceanography 9: 275–283. Dixit, S. S., B. F. Cumming, H. J. B. Birks, J. P. Smol, J. C. Kingston, A. J. Uutala, D. F. Charles & K. E. Camburn, 1993. Diatom assemblages from Adirondack lakes (New York, USA) and the development of inference models for retrospective environmental assessment. Journal of Paleolimnology 8: 27–47. Dixit, C. C. & J. P. Smol, 1994. Diatoms as indicators in the environmental monitoring and assessment program-surface waters (EMAP-SW). Environmental Monitoring and Assessment 31: 275–306.

Harrel, R. C. & T. C. Dorris, 1968. Stream order, morphometry, physico-chemical conditions, and community structure of benthic macro invertebrates in an intermittent stream system. American Naturalist 80: 221–251. Herdendorf, C. E., 1989. Paleogeology and geomorphology. In Lake Erie Estuarine Systems: Issues, Resources, Status and Management. U.S. Dept. of Commerce, NOAA, Estuary-ofthe-Month Seminar Series No.14, 35–70. Hynes, H. B. N., 1960. The Biology of Polluted Waters . Liverpool University Press, Liverpool. Johnson, L. B., C. Richards, G. E. Host & J. W. Arthur, 1997. Landscape influences on water chemistry in Midwestern stream ecosystems. Freshwater Biology 37: 193–208. Jongman, R. H. G., C. J. F. ter Braak & O. F. R. van Tongeren, 1995. Data Analysis in Community and Landscape Ecology. Pudoc, Wageningen, the Netherlands 299 pp. Juggins, S. & C. J. F. ter Braak, 1992. CALIBRATE – a Program for Species-Environment Calibration by [Weighted Averaging] Partial Least Squares Regression. Environmental Change Research Center, University College, London. Klarer, D. M., 1988. The role of a freshwater estuary in mitigating storm water inflow. OWC Technical Report No. 5, Ohio Dept. Nat. Res., Div. Nat. Areas and Preserves. Klarer, D. M. & M. Hickman, 1975. The effect of thermal effluent upon the standing crop of an epiphytic algal community. Hydrobiology 60: 17–62. Klein, R., 1979. Urbanization and stream quality impairment. American Water Resources Assoc. Water Resources Bulletin 15(4). Kentucky Department for Environmental Protection, 1993. Methods for Investigating Biological Integrity of Surface Waters. Kentucky Department for Environmental Protection, Frankfort, Kentucky 140 pp. Lange-Bertalot, H., 1979. Pollution tolerance of diatoms as a criterion for water quality estimation. Nova Hedwigia, Beih. 64: 285–304. Lowe, R. L., 1974. Environmental Requirements and Pollution Tolerance of Freshwater Diatoms. Environmental Monitoring Series: EPA-670/4–74–005, U.S. Environmental Protection Agency Cincinnati, Ohio. Makarewitcz, J.C., 1985. Phytoplankton and Zooplankton Composition, Abundance and Distribution: Lake Erie, Lake Huron and Lake Michigan – 1983. EPA-905/2–87–002 U.S. Environmental Protection Agency Chicago, Illinois, 44 pp. Manugistics Corporation, 1992. Statgraphics Plus Statistical Software. Rockville, MD. Ohio Department of Natural Resources, 1997. Combined Coastal Management Program and Final Environmental Impact Statement for the State of Ohio. Vol I Office of Real Estate and Land Management, Columbus, Ohio. Ohio Environmental Protection Agency, 1987. Biological Criteria for the Protection of Aquatic Life. Vol II: The Role of Biological Data in Water Quality Assessment. Ecological Assessment Section, Division of Water Quality Planning and Assessment, Columbus, Ohio. Ohio Environmental Protection Agency, 1991. Ashtabula River Remedial Action Plan. Division of Water Quality Planning and Assessment, Columbus, Ohio 171 pp.

249 Ohio Environmental Protection Agency, 1994. Biological and Water Quality Study of the Cuyahoga River and Selected Tributaries. Ecological Assessment Section, Division of Surface Water, Columbus, Ohio 128 pp. Omernik, J. M., 1976. The Influence of Land Use on Stream Nutrient Levels. EPA-600/3–76–014. U.S. Environmental Protection Agency, Washington, D.C. Osborne, L. L. & M. J. Wiley, 1988. Empirical relationships between land use/cover patterns and stream water quality in an agricultural catchment. Journal of Environmental Management 26: 9–27. Pan, Y., R. J. Stevenson, B. Hill, P. R. Kaufmann & A. T. Herlihy, 1999. Spatial patterns and ecological determinants of benthic algal assemblages in Mid-Atlantic Highland streams. Journal of Phycology 35: 460–468. Porter, S. D., T. F. Cuffney, M. E. Gurtz & M. R. Meader, 1993. Methods for Collecting Algal Samples as Part of the National Water-Quality Assessment Program. U.S. Geological Survey, Earth Science Information Center, Denver, CO 39 pp. Potapova, M. G., D. F. Charles, K. C. Ponader & D. M. Winter, 2004. Quantifying species indicator values for trophic diatom indices: a comparison of approaches. Hydrobiology 517: 25–41. Schuler, T. R. & J. Galli, 1992. Environmental impacts of storm water ponds. In Watershed restoration sourcebook, Anacostia Restoration Team. Metropolitan Council of Governments. Washington D.C. 242 pp. Sgro, G. V. & J. R. Johansen, 1995. Rapid bioassessment of algal periphyton in freshwater streams. In: Butterworth F. M., et al., (eds), Biomonitors and Biomarkers as Indicators of Environmental Change, Plenum Press, New York: 291– 311. Sgro, G. V., J. R. Johansen, 1998. Ecology and Assessment of the Algal Communities of Four Lake Erie Estuaries. Ph.D. dissertation. Cleveland State University, Cleveland, Ohio. Squires, L. E., S. R. Rushforth & D. J. Brotherson, 1979. Algal response to a thermal effluent: Study of a power station on the Provo River, Utah, USA. Hydrobiology 63: 1011–1017. Stevenson, R. J. & E. F. Stoermer, 1982. Abundance patterns of diatoms on Cladophora in Lake Huron with respect to point source of wastewater treatment plant effluent. Journal of Great Lakes Research 8: 184–195.

Stevenson, R. J. & R. L. Lowe, 1986. Sampling and interpretation of algal patterns for water quality assessments. In: Isom B. G. (ed.), Rationale for Sampling and Interpretation of Ecological Data in the Assessment of Freshwater Ecosystems, ASTM STP American Society for Testing and Materials, Philadelphia, 894:118–149. Stoermer, E. F., J. A. Wolin, C. L. Schelski & J. D. Conley, 1985. An assessment of the ecological changes during the recent history of Lake Ontario on siliceous algal microfossils preserved in sediments. Journal of Phycology 21: 257– 276. ter Braak, C. J. F., 1987. CANOCO a FORTRAN Program for Canonical Community Ordination by [Partial] [Detrended] [Canonical] Correspondence Analysis, Principal Components Analysis, and Redundancy Analysis (version 2.1). Agricultural Mathematics Group, Wageningen, the Netherlands, 95 pp. ter Braak, C.J.F., 1990. Update Notes: CANOCO version 3.10. Agricultural Mathematics Group, Wageningen, the Netherlands, 95 pp. ter Braak, C. J. F. & van Dam, 1989. Inferring pH from diatoms: A comparison of old and new calibration methods. Hydrobiology 178: 209–223. Thoma, R. F., 1998. Biological monitoring and an index of biotic integrity for Lake Erie’s near shore waters. In Simon, T.P. (ed.) Assessing the Sustainability and Biological Integrity of Water Resources Using Fish Communities. CRC Press, 417461. Trebitz, A. S., J. A. Morrice & A. M. Cotter, 2002. Relative role of lake and tributary in hydrology of Lake Superior coastal wetlands. Journal of Great Lakes Research 28(2): 212–227. van Dam, H., A. Mertenes & J. Sinkeldam, 1994. A coded checklist and ecological indicator values of freshwater diatoms from the Netherlands. Netherlands Journal of Aquatic Ecology 28: 117–133. Vinson, D. K. & S. R. Rushforth, 1989. Diatom species composition along a thermal gradient in the Portneuf River, Idaho, USA. Hydrobiology 185: 41–54. Zaborojny, C., S. Saxon & R. Finger, 1973. Spectrophotometric determination of ammonia. Journal of the Water Pollution Control Federation 45: 905–912.