Long-term variability, disturbance, and biological traits

0 downloads 0 Views 4MB Size Report
Feb 1, 1982 - and short-term effects of a small-scale prescribed fire on stream and ..... committee for making that experience relatively painless: the late .... involved one-time sampling, they usually sampled multiple sites ..... Stream width varies from 5 to 10 m, and the gradient of the study ...... of the total inertia of 0.029.
Long-term variability, disturbance, and biological traits of aquatic invertebrates in mediterranean-climate streams by Leah Anne Bêche B.A. (University of California, Berkeley) 2000

A dissertation submitted in partial satisfaction of the requirements of the degree of Doctor of Philosophy in Environmental Science, Policy & Management in the GRADUATE DIVISION of the UNIVERSITY OF CALIFORNIA, BERKELEY

Committee in charge: Professor Vincent H. Resh, Chair Professor Scott L. Stephens Professor G. Mathias Kondolf

Spring 2005

Abstract Long-term variability, disturbance, and biological traits of aquatic invertebrates in mediterranean-climate streams by Leah Anne Bêche Doctor of Philosophy in Environmental Science, Policy & Management University of California, Berkeley Professor Vincent H. Resh, Chair

Disturbance in mediterranean-climate streams can result from natural (e.g., flooding and drying) or human-induced (e.g., fire) processes. I examined long-term temporal variability of benthic macroinvertebrate communities and their biological traits in response to periods of floods and droughts, and to the impacts of a prescribed fire on aquatic and riparian communities in a Sierra Nevada mixed-conifer forest. I assessed temporal change and response to disturbance using univariate and multivariate analyses, based on three long-term datasets from northern California streams: 17-20 years at four sites in Hunting and Knoxville Creek (Lake Co.); 7 years at Big Sulphur Creek (Sonoma Co.); and 7-8 years at six sites in Blodgett Forest Research Station (El Dorado Co.). Biological traits were coded for 263 macroinvertebrate taxa. The life history of the caddisfly Heteroplectron californicum was examined using monthly data (1.5 years) from Webb Creek (Marin Co.). Biological-trait composition is relatively more stable than community composition and abundance both between seasons and among years. However, seasonal and annual changes in habitat from variable precipitation patterns are reflected in biological-trait

1

composition. Traits conferring resilience or resistance to flooding or drying are more common in both wet seasons and years, and dry seasons and years, respectively. Extreme multi-year droughts resulted in long lasting, directional change in communities. Some taxa are temporally rare; up to one-third of taxa at a site only occurred once in 7-20 years. Prescribed fire was patchy in the riparian zone and had short-term or no effects on water chemistry, sediment, large woody debris, periphyton, and benthic macroinvertebrate communities. The life history of a wood-boring caddisfly, Heteroplectron californicum, was found to be one-year with some cohort-splitting. In mediterranean-climate streams several patterns are evident: 1) long-term variability (both seasonal and annual) of aquatic macroinvertebrate communities results from interannual variability in precipitation, 2) the effects of droughts are long-lasting, 3) temporal rarity of invertebrate taxa is common in long-term datasets, 4) there are minor and short-term effects of a small-scale prescribed fire on stream and riparian communities, and 5) biological traits provide mechanistic understanding of communitylevel variation and relationships in response to habitat conditions.

2

TABLE OF CONTENTS LIST OF FIGURES …………………………………………………………………....vi LIST OF TABLES ……………………………………………………………………vii ACKNOWLEDGEMENTS ………………………………………………………….viii CHAPTER 1. General Introduction ...……………………………………...………… Introduction …………..……………………………………………………………….. How long are studies conducted in stream ecology? ...…………………………... Description of research chapters ...………………………………………………. References ...……………..………………………..…………………………….…….. CHAPTER 2. Long-term seasonal variation of benthic-macroinvertebrate biological traits in two mediterranean-climate streams in California, USA ...……… Abstract ….…………...………………………………...………………………... Introduction ………………....……………….…..………………………………. Methods …………………………………………...……………………………... Study sites ………………………………..………………………………… Precipitation and discharge …………………………...……………………. Data collection ……………………………………………………………... Macroinvertebrates ...………………………………………………... Biological traits ……………………………………………………… Analyses ……………………………………………………………………. Results …………………………………………………………………………… Taxonomic composition and abundance …………………………………… Structural measures …………………………………………………. Multivariate ordination ……………………………………………… Biological trait composition ………………………………………………... Multivariate ordination ……………………………………………… Trait frequencies …………………………………………………….. Discussion ……………………………………………………………………….. Seasonal variability of taxonomic composition and abundance …………… Seasonal variability of biological traits …………………………………….. Seasonal stability of biological traits: intra-annual ………………………… Seasonal stability of biological traits: inter-annual ………………………… Conclusion ………………………………………………………………….. References ………………………………………………………………………..

1 2 4 7 9

17 18 19 22 22 23 24 24 26 26 29 29 29 30 32 32 34 36 36 39 41 43 45 46

CHAPTER 3. Interannual variation of benthic macroinvertebrate communities and their biological traits: patterns observed over 20 years in mediterranean-climate streams ………………………………………...…………………………………….. 67 Abstract ………………………………………………………………………….. 68 i

Introduction ……………………………………………………………………… 69 Methods ………………………………………………………………………….. 72 Site descriptions …………………………………………………………….. 72 Hunting/Knoxville Creeks …………………………………………... 72 Big Sulphur Creek …………………………………………………… 73 Blodgett Forest Creeks ………………………………………………. 74 Physical Habitat …………………………………………………………….. 75 Precipitation …………………………………………………………. 75 Discharge and bed movement ……………………………………….. 76 FST-hemisphere values ……………………………………………… 77 Biological traits …………………………………………………………….. 78 Data analysis ………………………………………………………………... 79 Temporal change in communities …………………………………… 79 Wet and dry year patterns …………………………………………… 81 Stability and persistence of communities …………………………… 82 Results …………………………………………………………………………… 83 Physical habitat ……………………………………………………………... 83 Precipitation …………………………………………………………. 83 Discharge and bed movement ……………………………………….. 84 FST-hemisphere values ……………………………………………… 86 Temporal change in community structure ………………………………….. 87 Long-term trends in community summary metrics ………………….. 87 Relationships among years: community composition and abundance 89 Stability and persistence of community composition and abundance . 93 Biological traits ……………………………………………………………... 95 Long-term trends in trait modalities ………………………………… 95 Relationships among years: biological trait composition …………… 97 Stability of biological trait composition…………………………… 103 Discussion ……………………………………………………………………... 115 Long-term variation of community structure and abundance ……………. 115 Trends in community metrics ……………………………………... 115 Relationships among years ………………………………………... 120 Long-term variation of biological traits ………………………………….. 121 Trends in individual trait modalities ………………………………. 121 Relationships among years ………………………………………... 123 Environment-community relationships: what should be measured? ……... 125 Total precipitation …………………………………………………. 126 Discharge and bed movement ……………………………………... 126 FST-hemisphere values (sheer stress) …………………………….. 127 Stability and persistence of communities ………………………………… 128

ii

Stability and persistence of community composition and abundance …………………………………………………………. Stability of biological traits ……………………………………….. Instability of dry-season communities …………………………….. The influence of rare taxa on temporal trends ……………………………. Long-term effects of extreme drought ……………………………………. Conclusion ………………………………………………………………... References ……………………………………………………………………... Appendices ……………………………………………………………………..

128 131 133 134 135 138 139 183

CHAPTER 4. Rarity in long-term benthic macroinvertebrate surveys: a common pattern in undisturbed streams ………………………………………………….… Abstract ………………………………………………………………………... Introduction ……………………………………………………………………. Methods …………………………………………………………………….….. Site description and sampling methods …………………………………... 19 to 20-year dataset ………………………………….…………… 7-year dataset ………………………………….…………………... 8-year dataset ………………………………….…………………... Analysis ………………………………….……………………………….. Occurrence of rare and common taxa ……………………………... Abundance of rare and common taxa ……………………………... Biological traits of rare and common taxa ………………………… Results ………………………………….……………………………………… Occurrence of rare and unresolved taxa ………………………………….. Abundance of rare and unresolved taxa ………………………………….. Biological traits of rare and common taxa ……………………………….. Discussion ………………………………….………………………………….. References ………………………………….…………………………………..

196 197 197 200 200 200 201 202 203 203 204 204 205 205 206 207 207 212

CHAPTER 5. The effects of riparian prescribed fire on aquatic and riparian communities in a mixed-conifer forest watershed, Sierra Nevada, California, USA ………………………………………………………………………………. Abstract ………………………………….…………………………………….. Introduction ………………………………….………………………………… Methods ………………………………….…………………………………….. Study site ………………………………….……………………………… Study design ………………………………….…………………………... Prescribed fire treatment ………………………………….……………… Data collection ………………………………….………………………… Riparian vegetation ………………………………….…………….. Large woody debris (LWD) ………………………………….……

222 223 223 226 226 227 228 228 228 230 iii

Hydrology ………………………………….……………………… Sediment and substrate characterization …………………………... Channel morphology ………………………………….…………... Water chemistry ………………………………….………………... Periphyton ………………………………….……………………… Macroinvertebrates ………………………………….…………….. Data Analysis ………………………………….…………………………. Results …………………………………………………………………………. Prescribed fire behavior and fuels inventory ……………………………... Riparian vegetation ………………………………….……………………. Large woody debris ………………………………….…………………… Hydrology ………………………………….……………………………... Sediment composition ………………………………….………………… Channel morphology ………………………………….………………….. Water chemistry ………………………………….………………………. Periphyton ………………………………….…………………………….. Macroinvertebrates ………………………………….……………………. Discussion ………………………………….………………………………….. Prescribed fire in the riparian zone: fuels and vegetation ………………... Large woody debris ………………………………….…………………… Sediment composition and channel morphology ………………………… Water chemistry ………………………………….………………………. Periphyton ………………………………….…………………………….. Macroinvertebrates ………………………………….……………………. Conclusion ………………………………….…………………………….. References ………………………………….………………………………….. Appendices ………………………………….………………………………….

231 231 231 232 232 234 235 237 237 237 239 240 240 241 242 244 245 247 248 250 251 254 257 258 260 262 297

CHAPTER 6. Life history of a wood-boring caddisfly, Heteroplectron californicum (Trichoptera: Calamoceratidae) in a coastal California stream …….. Abstract ………………………………….…………………………………….. Introduction ………………………………….………………………………… Methods ………………………………….…………………………………….. Site description ………………………………….………………………... Sampling ………………………………….………………………………. Case-building behavior ………………………………….………………... Results and discussion ………………………………….……………………… Life cycle ………………………………….……………………………… Egg stage ………………………………….……………………………… Larvae ………………………………….…………………………………. Case construction ………………………………….……………………...

300 301 301 302 302 303 303 304 304 305 306 307

iv

Pupal stage ………………………………….…………………………….. Emergence and adults ………………………………….…………………. Conclusion ………………………………….…………………………….. References ………………………………….…………………………………..

308 309 309 309

CONCLUSION ………………………………….………………………………... 319

v

LIST OF FIGURES Figure 2.1 ……………………………………………………………………………. 59 Figure 2.2 ……………………………………………………………………………. 60 Figure 2.3 ……………………………………………………………………………. 61 Figure 2.4 ……………………………………………………………………………. 64 Figure 2.5 ……………………………………………………………………………. 65 Figure 2.6 ……………………………………………………………………………. 66 Figure 3.1 ………………………………………………………………………….. 165 Figure 3.2 ………………………………………………………………………….. 166 Figure 3.3 ………………………………………………………………………….. 167 Figure 3.4 ………………………………………………………………………….. 168 Figure 3.5 ………………………………………………………………………….. 171 Figure 3.6 ………………………………………………………………………….. 173 Figure 3.7 ………………………………………………………………………….. 175 Figure 3.8 ………………………………………………………………………….. 176 Figure 3.9 ………………………………………………………………………….. 179 Figure 3.10 ………………………………………………………………………… 181 Figure 4.1 ………………………………………………………………………….. 217 Figure 4.2 ………………………………………………………………………….. 218 Figure 4.3 ………………………………………………………………………….. 219 Figure 4.4 ………………………………………………………………………….. 220 Figure 4.5 ………………………………………………………………………….. 221 Figure 5.1 ………………………………………………………………………….. 285 Figure 5.2 ………………………………………………………………………….. 286 Figure 5.3 ………………………………………………………………………….. 287 Figure 5.4 ………………………………………………………………………….. 288 Figure 5.5 ………………………………………………………………………….. 289 Figure 5.6 ………………………………………………………………………….. 290 Figure 5.7 ………………………………………………………………………….. 293 Figure 5.8 ………………………………………………………………………….. 294 Figure 5.9 ………………………………………………………………………….. 295 Figure 5.10 ………………………………………………………………………… 296 Figure 6.1 ………………………………………………………………………….. 315 Figure 6.2 ………………………………………………………………………….. 316 Figure 6.3 ………………………………………………………………………….. 317 Figure 6.4 ………………………………………………………………………….. 318

vi

LIST OF TABLES Table 1.1 …………………………………………………………………………….. 14 Table 1.2 …………………………………………………………………………….. 15 Table 1.3 …………………………………………………………………………….. 16 Table 2.1 …………………………………………………………………………….. 54 Table 2.2 …………………………………………………………………………….. 55 Table 2.3 …………………………………………………………………………….. 56 Table 2.4 …………………………………………………………………………….. 57 Table 3.1 …………………………………………………………………………... 148 Table 3.2 …………………………………………………………………………... 149 Table 3.3 …………………………………………………………………………... 150 Table 3.4 …………………………………………………………………………... 151 Table 3.5 …………………………………………………………………………... 152 Table 3.6 …………………………………………………………………………... 153 Table 3.7 …………………………………………………………………………... 154 Table 3.8 …………………………………………………………………………... 155 Table 3.9 …………………………………………………………………………... 156 Table 3.10 …………………………………………………………………………. 157 Table 3.11 …………………………………………………………………………. 158 Table 3.12 …………………………………………………………………………. 159 Table 3.13 …………………………………………………………………………. 162 Table 3.14 …………………………………………………………………………. 163 Table 3.15 …………………………………………………………………………. 164 Table 4.1 …………………………………………………………………………... 215 Table 4.2 …………………………………………………………………………... 216 Table 5.1 …………………………………………………………………………... 274 Table 5.2 …………………………………………………………………………... 275 Table 5.3 …………………………………………………………………………... 276 Table 5.4 …………………………………………………………………………... 277 Table 5.5 …………………………………………………………………………... 278 Table 5.6 …………………………………………………………………………... 279 Table 5.7 …………………………………………………………………………... 280 Table 5.8 …………………………………………………………………………... 281 Table 5.9 …………………………………………………………………………... 282 Table 5.10 …………………………………………………………………………. 283 Table 5.11 …………………………………………………………………………. 284 Table 6.1 …………………………………………………………………………... 312 Table 6.2 …………………………………………………………………………... 313 Table 6.3 …………………………………………………………………………... 314

vii

Acknowledgements This dissertation would not have been possible without the assistance, cooperation, advice, and friendship from numerous individuals; I am indebted to them for all that they have done. I express my gratitude to Vince Resh, my major advisor; I cannot thank him enough for all that he has contributed to my beginning and future as a stream ecologist and my life. He has taught me that science is much more than research and publishing, and that teaching is truly a privilege and an art. Throughout my graduate career, Vince has provided encouragement, guidance, financial and emotional support, and a lot of creative editing that will not be forgotten. I aspire to be such a wonderful advisor and mentor. I thank Scott Stephens for not only being a member of my dissertation committee, but for being a wonderful collaborator and mentor. Scott has provided support and encouragement throughout my graduate career and has always been enthusiastic about my research ideas and direction. Without the efforts of Scott and his research group, the study on the effects of prescribed fire on streams would not have been possible. In particular, I am grateful to Jason Mogaddhas for planning, obtaining permits, and helping to implement the fire in Blodgett Forest. Emily Mogaddhas, Tadashi Moody, Leda Kobziar, Danny Fry, Dan Stark, Andy Amacher, and other members of the Stephens Lab, as well as other undergraduate and graduate student volunteers were instrumental in making this prescribed fire happen. In particular, employees of the U.S. Forest Service in El Dorado County, the California Department of Forestry and Fire Protection. I also appreciate the involvement that Matt Kondolf has had with my research and also in reviewing my dissertation as a member of my dissertation committee. Matt has

viii

provided a regular forum for me, as well as other stream and river researchers, to present and share ideas. I would also like to thank the three other members of my orals committee for making that experience relatively painless: the late Don Dahlsten, Joe McBride, and Alex Horne. Eric McElravy has been gracious enough to provide years of data, and has helped immensely with the analysis and editing of three chapters. I thank him for being a helpful collaborator. My experience as a graduate student would not have been bearable or complete without the friendship, support, encouragement, and occasional sympathy from the (official and honorary) members of the Resh lab. I am especially grateful to Deborah Rudnick for getting me involved in research as an undergraduate; Allison Purcell and Matt Cover for thoroughly reviewing my papers and providing advice; Igor Lacan for listening to my rants; Matt Deitch for developing flow models and answering my uneducated questions on hydrology; Andreas Hoffmann for getting me started on the Heteroplectron study and for showing me what a joy caddisflies can be; Rosalie del Rosario and Marilyn Myers for helping me to survive my first year as a graduate student; Faith Kearns for showing me that there is a world beyond the ivory tower. Finally, I am especially fortunate to have spent the last four years with Tina Mendez and Rafi Mazor, who have kept me sane during the past four years. They have been my sounding board for ideas and great friends. Several devoted and hard-working undergraduates contributed to my research effort. Liliana Aguas, Valerie Chan, and Liat Zovodivker were incredibly efficient field assistants. Bobbi Martin measured and helped to collect Heteroplectron specimens.

ix

Kathryn Farrar, Lisa Kitchin, Jenny Oh, Liliana Aguas, and Bobbi Martin diligently and persistently sorted invertebrates from sediment for hours on end. Maren Farnum and Amattullah R’id conducted case-building studies using Heteroplectron californicum. The logistics of my prescribed fire research were made possible by numerous people at Blodgett Forest Research Station including: Bob Heald, Frieder Schurr, Sheryl Rambeau, Dave Rambeau, Russell Seiffert, Rob York, Jennifer Prentiss, Nadia Hamey, Laura Schweitzer, and Andrew Corr. Bernhard Statzner and Núria Bonada provided analytical advice and made comments on my research. The Amundsen and Casida labs were generous in allowing me to use their equipment. Sue Jennison-Baumgartner and Richard Battrick made dealing with administration seem easy. Financial support for my research was provided by a graduate fellowship from the U.S. Environmental Protection STAR (Science to Achieve Results; U-91572101), Entomology Student’s Organization, Divisional Research Assistantship, Departmental Travel Grant, an a Graduate Division Travel Grant. I am particularly honored to have received the Robert L. Usinger Memorial Award in Aquatic Entomology and the Harvey Magy Scholarship for Outstanding Graduate Students, as well as an Outstanding Graduate Student Instructor Award. I am thankful to my family for their support and love. In particular, I am fortunate for my mom, Denise, that has shown me the value of education and has inspired me to continually work towards improvement. My sisters, Madylan and Allison, have supported me through many ups and downs both before and during graduate school. I am also thankful to be able to share my accomplishments with my father, Jonathon, my grandparents, Donna, Al, and Roy; my great-grandmothers Jean and Mamy, my aunt

x

Sharon, and the rest of the Rogers, Foster, and Walkey families whose members are too numerous to mention. I am thankful for the time that I have spent with family members that are no longer here: grandma Norma, Bubie, Nana, grandpa Jerry, and Mickie. Finally, I am lucky to share my life with my husband Jean-Francois; he is my daily inspiration. Throughout this often trying process, he has been patient, provided unwavering support and good humor, and love. I thank him for making every day promising bright and the future infinitely bright.

xi

Chapter 1

General introduction

1

Introduction The importance of long-term data in understanding ecological phenomena has been emphasized in many studies (e.g., Connell and Sousa 1983, McElravy et al. 1989, Resh and Rosenberg 1989, Resh and McElravy 1993, Robinson et al. 2000). The analysis of long-term data, particularly in light of recent developments in ecological theory (e.g., Statzner et al. 1996), can potentially provide important insights into the mechanisms controlling the abundance and occurrence of organisms. Furthermore, fluctuations in the abundance and occurrence of organisms over long-time scales (e.g., in terms of persistence and stability, sensu Connell and Sousa 1983) can have important implications for the biological assessment of aquatic environments (e.g., Metzeling et al. 2002), as well as for the future, the long-term applicability of research results (e.g., Scarsbrook 2002, chapter 3). Temporal fluctuations in composition and abundance may be expected to be highest in areas with relatively high variability in climate (e.g., Boulton et al. 1992), such as the mediterranean-climate. The mediterranean-climate is characterized by predictable seasonal variation in precipitation and temperature, with hot, dry summers and cool, wet winters when at least 65% (but up to 95%) of precipitation occurs (Gasith and Resh 1999). Although the timing of the seasonal variation in precipitation is predictable, the total amount of precipitation varies considerably on an annual basis. High interannual and seasonal variability results in disturbances that range from drought to severe spates; this variability in disturbance regime will tend to increase annual variability of aquatic communities (McElravy et al. 1989). Similarly, the predictable seasonal pattern of wet

2

and dry results in marked seasonality in community composition and abundance (e.g., McElravy et al. 1989, Bonada 2003). The importance of disturbance (e.g., flooding and drying) as the primary factor influencing stream biota has been recognized for at least twenty years (e.g., Resh et al. 1988); however, most research has focused only on flooding, whereas drying disturbance as result of seasonal or multiyear drought has largely been ignored (Boulton 2003, Lake 2003). In mediterranean-climate regions (including northern California), aquatic communities are structured by both seasonal flooding and drying and interannual differences in precipitation, which lead to differences in the magnitude and frequency of these disturbances (Gasith and Resh 1999). Thus, aquatic communities in mediterraneanclimates may be expected to exhibit pronounced seasonality (e.g., Boulton and Lake 1992a,b), but high interannual variability, and thus low stability and persistence. However, long-term data are needed to determine whether communities are persistent (i.e., the tendency for the composition of a community to remain constant through time) or stable (i.e., the tendency of relative abundance of taxa, or biological traits, to remain constant through time) (Connell and Sousa 1983). Conclusions on the temporal variability of aquatic communities based on studies on spanning only a few years may be not provide underestimate natural variability because these studies generally fail to capture infrequent events (e.g., l00-year floods, fires) or long-term cycles in climate (e.g., multi-year droughts) (e.g., Connell and Sousa 1983, Eby et al. 2003). Furthermore, in assessing the impacts of management practices or other anthropogenic effects, long-term data may also be required to accurately differentiate the effects of natural variability (e.g., as a result of flooding or drought) from the impacts of

3

interest, particularly in regions where high natural variability exists (such as in the mediterranean-climate). The temporal variability of communities on a seasonal and annual time scale is related to life history strategies or behavioral characteristics, or biological traits. The adaptation of biological traits to flow conditions (i.e., to avoid floods or droughts) is dependent on the predictability of timing of seasonal flow (Lytle and Poff 2004). Despite this connection between seasonal and interannual variability in communities and life history strategies, there have been few studies explicitly examining either the seasonality (Snook and Milner 2002) or interannual variability of biological traits at the community level (Usseglio-Polatera 1997). In general, seasonality and temporal variability in benthic macroinvertebrate communities have been examined in terms of taxonomic (i.e., structural) characteristics (e.g., Robinson et al. 2000, Scarsbrook 2002). However, nontaxonomic aggregations of taxa into biological traits categories based on life history, behavior, and morphology may be more informative for investigating the relationship between physical habitat and community response (e.g., Richards et al. 1997). The use of biological traits can increase our predictive ability in ecology because traits provide a mechanistic explanation for variability (e.g., seasonal or annual) in aquatic invertebrate communities (e.g., Resh et al. 1994, Townsend and Hildrew 1994). How long are studies conducted in stream ecology? Because the importance of long-term datasets has been repeatedly emphasized, particularly over the last two decades (e.g., McElravy 1988, McElravy et al. 1989, Resh and McElravy 1993, Resh et al. 1994, Robinson et al. 2000, Scarsbrook 2002, Eby et al. 2003), I reviewed the time-scale of investigation used by aquatic ecologists working in

4

streams and rivers, as reported in the peer-reviewed literature. To do this, I examined research articles published in seven hydrobiological or ecological journals (Table 1.1) over a six year period (1999-2004). A similar literature survey was done by McElravy (1988) covering articles published in five journals from 1980 to 1987. In the present survey, I added two more journals: Ecological Applications (a sister journal to Ecology that began publishing in 1991), and I examined issues of the Journal of the North American Benthological Society from 1982-1987 and 1999-2004 for an additional comparison. Only studies examining macroinvertebrates in streams were included, and short-term experiments and experiments conducted in artificial channels or mesocosms were excluded. Given substantial year-to-year variation in populations and communities (e.g., McElravy et al. 1989, chapters 2 – 5), we cannot assume that conclusions based on one year (or less) have any predictive value (McElravy 1988). Despite this fact, the majority of freshwater (macroinvertebrate) studies (76%) published from 1999 to 2004 are oneyear or less in duration (Table 1.1). A similar result (72% of studies were one-year or less in duration) was obtained by McElravy (1988) from his literature survey from 1980 to 1987 (Table 1.1). Furthermore, the proportion of short-term studies, particularly those where sampling was only done once or was limited to < 6 months, increased substantially (21% in 1980-1987 cf. 45% in 1999-2004). The proportion of long-term studies (> 3 years duration) increased (5% in 1980-1987 cf. 9% in 1999-2004) as well, although the number of studies was still low. The majority of studies published from 1999-2004 were community studies (206, 53%); however, these studies were most likely to be short in duration (i.e., ≤ 1 year). For

5

example, only 8.2% of community studies were 3 years or longer in duration. In contrast, 12.2 (biological assessment) to 25.1% (impact monitoring) were ≥ 3 years in duration (Table 1.2). However, when only studies ≥ 3 years in duration were examined, 66% of these long-term studies were either community (34%) or impact monitoring (32%) studies; population studies accounted for 24% and biological assessment articles accounted for only 10% of long-term studies from 1999 to 2004 (Table 1.3). In contrast, McElravy (1988) found that most articles using long-term data were community studies (48%), while only 25% of long-term studies were impact assessment and 28% were population studies (Table 1.3). The high proportion of studies where sampling was only done once (Table 1.2) suggests that spatial variation in macroinvertebrate communities and populations is being studied more often than temporal variation. For example, although 24% of all studies involved one-time sampling, they usually sampled multiple sites (e.g., Kay et al. 2001, Heino et al. 2002). Furthermore, 11% of all of the studies investigated were biological assessment methods papers (e.g., comparing different methods, assessing the influence of different metrics or identification levels on assessment outcomes, developing new methods), which are usually based on large-scale surveys (with 10’s to 100’s of sites) that are only sampled once (e.g., Hawkins and Vinson 2000, Zweig and Rabeni 2001). The pattern of increasing spatial replication and decreasing study duration was also apparent in impact assessment studies. For example, only 36% of impact assessment studies lasted longer than one year, whereas from 1985-1987, 56% of impact assessment studies were longer than one year (Resh and McElravy 1993). Furthermore, Resh and McElravy (1993) noticed an increasing trend in the duration of impact assessment

6

studies from 1980-1984 (20% of studies were > 1 year) to 1985-1987 (56% of studies were > 1 year). The results of the present survey suggest that this trend is now in reversal. Although some studies have examined temporal variation on annual (e.g., Robinson et al. 2000, Scarsbroook 2002) or over decade (e.g., Bradt et al. 1999) time-scales, most investigations of temporal variation of macroinvertebrates focused on seasonality over the course of a single year or less (e.g., Burgherr et al. 2002, Robinson et al. 2001). However, the applicability of these short-term seasonal studies to other time periods may be limited; I have found considerable year-to-year variation in seasonality of communities and their biological traits that can be attributed to annual variation in precipitation (chapter 2).

Description of research chapters The purpose of the following chapters is to examine temporal variability of aquatic invertebrate communities in terms of intra-annual (seasonal) and inter-annual (year-toyear) variation and responses to disturbance over long time scales (7 to 20 years). In particular, I investigated community-level responses to annual variability in precipitation (and thus discharge and bed hydraulics) and disturbance (i.e., floods, droughts, and prescribed fire). Furthermore, I used biological traits of organisms to relate temporal variability of communities and their traits to temporal variability in habitat, following the habitat template concept of Southwood (1977, 1988) and Townsend and Hildrew (1994). In chapter 2, I examine how the seasonality of aquatic invertebrate communities and their biological traits varies over long time-scales (i.e., 19 years) in response to annual

7

differences in precipitation, and, thus discharge and near-bed hydraulic conditions, in two mediterranean-climate streams. In addition, I assess the influence of temporal habitat variability (i.e., flow permanence, intermittent cf. perennial stream) on seasonality of communities, where seasonal habitat variability is higher in the intermittent stream than in the perennial stream. In chapter 3, I investigate the annual and inter-annual variability of aquatic invertebrate communities over 7 to 20 years, using three datasets from northern California. In particular, I assess the relationship between variability of community composition, community abundance, and biological traits and annual variability of precipitation in mediterranean-climate streams. I also examine the persistence and stability of community composition, and abundance and biological traits (Connell and Sousa 1983). In chapter 4, I examine long-term patterns of taxa occurrence in mediterraneanclimate streams. For example, in examining long-term data from the previous two chapters (2, 3), I found that a high proportion (up to 35%) of aquatic invertebrate taxa were temporally rare (i.e., they occurred only once during the study period of 7 to 20 years). Thus, in this chapter I assess the patterns, potential causes, and the consequences of temporal rarity in benthic macroinvertebrate surveys. In chapter 5, I examine the effects of a watershed-scale disturbance (prescribed fire) on aquatic and riparian communities, and stream habitat, in a mixed-conifer watershed in the Sierra Nevada, California. In this chapter, a beyond-BACI (before-after-controlimpact) experimental design (e.g. Underwood 1994) is used with long-term data (3 to 8 years) to examine the effects of a moderate-intensity upland and riparian fire (surface

8

fire) on aquatic and riparian communities in a small (1st-order) watershed at Blodgett Forest Research Station in the mixed-conifer region of northern California. In chapter 6, I examine the life-history of a wood-boring caddisfly, Heteroplectron californicum (Trichoptera: Calamoceratidae), in a northern California stream with a mediterranean-climate, which is endemic to the Pacific northwest of North America, and is one of only two species known to make cases out of burrowed twigs in the world. Thus, in this chapter, I provide detailed trait information at the species-level that I not only incorporated into my other dissertation chapters, but which can also be used by others in the future. I end this dissertation with a short conclusion in which I discuss the values and limitations of long-term studies and the use of biological traits to assess temporal variability and response to disturbance of aquatic invertebrate communities.

References Bonada, N. 2003. Ecology of the macroinvertebrate communities in Mediterranean rivers at different scales and organization levels. Ph.D., University of Barcelona, Spain, 355 pages. Boulton, A.J. and Lake, P.S. 1992a. The ecology of two intermittent streams in Victoria, Australia. II. Comparisons of faunal composition between habitats, rivers, and years. Freshwater Biology 27: 99-121. Boulton, A.J. and Lake, P.S. 1992b. The ecology of two intermittent streams in Victoria, Australia. III. Temporal changes in faunal composition. Freshwater Biology 27: 99121.

9

Boulton, A.J., Peterson, C.G., Grimm, N.B., and Fisher, S.G. 1992. Stability of an aquatic macroinvertebrate community in a multiyear hydrologic disturbance regime. Ecology 73: 2192-2207. Boulton, A.J. 2003. Parallels and contrasts in the effects of drought on stream macroinvertebrate communities. Freshwater Biology 48: 1173-1185. Bradt, P., Urban, M., Goodman, N., Bissell, S., and Spiegel, I. 1999. Stability and resilience in benthic macroinvertebrate assemblages: impact of physical disturbance over twenty-five years. Hydrobiologia 403: 123–133. Burgherr P., Ward, J.V., and Robinson, C.T. 2002. Seasonal variation in zoobenthos across habitat gradients in an alpine glacial floodplain (Val Roseg, Swiss Alps). Journal of the North American Benthological Society 21: 561-575. Connell, J.H. and Sousa, W.P. 1983. On the evidence needed to judge ecological stability or persistence. The American Naturalist 121: 789-824. Eby, L.A., Fagan, W.F., and Minckley, W.L. 2003. Variability and dynamics of a desert stream community. Ecology 13: 1566-1579. Gasith, A. and Resh, V.H. 1999. Streams in mediterranean climate regions: abiotic influences and biotic responses to predictable and seasonal events. Annual Review of Ecology and Systematics 30: 51-81. Hawkins, C.P., and Vinson, M.R. 2000. Weak correspondence between landscape classifications and stream invertebrate assemblages: implications for bioassessment. Journal of the North American Benthological Society 19: 501-517. Heino, J., Muotka, T., Paavola, R., Heikkihämälä, I., and Koskenniemi, E. 2002. Correspondence between regional delineations and spatial patterns in

10

macroinvertebrate assemblages of boreal headwater streams. Journal of the North American Benthological Society 21: 397-413. Kay, W.R., Halse, S.A., Scanlon, M.D., and Smith, M.J. 2001. Distribution and environmental tolerances of aquatic macroinvertebrate families in the agricultural zone of southwestern Australia. Journal of the North American Benthological Society 20: 182–199. Lake, P.S. 2003. Ecological effects of perturbation by drought in flowing waters. Freshwater Biology 48: 1161-1172. Lytle, D.A. and Poff, N.L. 2004. Adaptation to natural flow regimes. Trends in ecology and evolution 19: 94-100. McElravy, E.P. 1988. Temporal variability in abundance of aquatic insects: a comparison of temperate and tropical environments. Ph.D., University of California, Berkeley, 206 pages. McElravy, E.P., Lamberti, G.A., and Resh, V.H. 1989. Year-to-year variation in the aquatic macroinvertebrate fauna of a northern California stream. Journal of the North American Benthological Society 8: 51-63. Metzeling, L., Robinson, D., Perriss, S., and Marchant, R. 2002. Temporal persistence of benthic invertebrate communities in south-eastern Australian streams: taxonomic resolution and implications for the use of predictive models. Marine and Freshwater Research 53: 1223-1234. Resh, V.H., Brown, A.V., Covich, A.P., Gurtz, M.E., Li, H.W., Minshall, W., Reice, S., Sheldon, A.L., Wallace, J.B., and Wissmar, R. 1988. The role of disturbance in stream ecology. Journal of the North American Benthological Society 7: 433-455.

11

Resh, V.H. and Rosenberg, D.M. 1989. Spatial-temporal variability and the study of aquatic insects. Canadian Entomologist 121: 941-963. Resh, V.H. and Jackson, J.K. 1993. Rapid assessment approaches to biomonitoring using benthic macroinvertebrates. Pages 195-233 in Resh, V.H. and Rosenberg, D.M. (Eds.), Freshwater biomonitoring and benthic macroinvertebrates. Chapman and Hall, New York, 488 pages. Resh, V.H. and McElravy, E.P. 1993. Contemporary quantitative approaches to biomonitoring using benthic macroinvertebrates. Pages 159-194 in Resh, V.H. and Rosenberg, D.M. (Eds.), Freshwater biomonitoring and benthic macroinvertebrates. Chapman and Hall, New York, 488 pages. Resh, V.H., Hildrew, A.G., Statzner, B., and Townsend, C.R. 1994. Theoretical habitat templets, species traits, and species richness: a synthesis of long-term ecological research on the Upper Rhône River in the context of concurrently developed ecological theory. Freshwater Biology 31: 539-554. Richards, C., Haro, R.J., Johnson, L.B., and Host, G.E. 1997. Catchment and reach-scale properties as indicators of macroinvertebrate species traits. Freshwater Biology 37: 219-230. Robinson, C.T., Minshall, G.W., and Royer, T.V. 2000. Inter-annual patterns in macroinvertebrate communities of wilderness streams in Idaho, U.S.A. Hydrobiologia 421: 187-198. Robinson, C.T., Uehlinger, U., and Hieber, M. 2001. Spatio-temporal variation in macroinvertebrate assemblages of glacial streams in the Swiss Alps. Freshwater Biology 46: 1663-1672.

12

Scarsbrook, M.R. 2002. Persistence and stability of lotic invertebrate communities in New Zealand. Freshwater Biology 47: 417-431. Snook D.L. and Milner A.M. 2002. Biological traits of macroinvertebrates and hydraulic conditions in a glacier-fed catchment (French Pyrénées). Archiv für Hydrobiologie 153: 245-271. Southwood, T.R.E. 1977. Habitat, the templet for ecological strategies. Journal of Animal Ecology 46: 337-365. Southwood, T.R.E. 1988. Tactics, strategies and templets. Oikos 52: 3-18. Statzner, B., Resh, V.H., and Dolédec, S. 1996. A synthesis of long-term ecological research on the Upper Rhône River in the context of concurrently developed ecological theory. Archiv für Hydrobiologie Supplement 113: 45-50. Townsend, C.R. and Hildrew, A.G. 1994. Species traits in relation to a habitat templet for river systems. Freshwater Biology 31: 265-275. Underwood, A.J. 1994. On beyond BACI: sampling designs that might reliably detect environmental disturbances. Ecological Applications. 4: 3-15. Usseglio-Polatera, P. 1997. Long-term changes in the Ephemeroptera of the River Rhône at Lyon, France assessed using a fuzzy coding approach. Pages. 227-234 in Ephemeroptera & Plecoptera: Biology-Ecology-Systematics. MTL, Fribourg. Zweig, L.D. and Rabeni, C.F. 2001. Biomonitoring for deposited sediment using benthic invertebrates: a test on 4 Missouri streams. Journal of the North American Benthological Society 20: 643–657.

13

14

1.5-2 years

3 (6%) 24 (48%) 2 (4%) 1 (9%) 5 (45%) 1 (9%) 1 (3%) 14 (48%) 5 (17%) 7 (14%) 20 (39%) 3 (6%) 13 (10%) 47 (38%) 10 (8%) 16 (30%) 10 (19%) 3 (6%) 41 (13%) 120 (38%) 24 (8%)

1 year

3 years

50 11 29 51 125 53 319

> 3 years N

6 (12%) 2 (4%) 3 (6%) 1 (9%) 0 0%) 1 (9%) 4 (14%) 1 (3%) 2 (7%) 10 (20%) 6 (12%) 2 (4%) 9 (7%) 8 (6%) 7 (6%) 2 (4%) 2 (4%) 1 (2%) 32 (10%) 19 (6%) 16 (5%)

2 years

1999-2004 Archiv für Hydro. 25 (20%) 30 (25%) 22 (18%) 20 (16%) 6 (5%) 10 (8%) 4 (3%) 5 (4%) 122 CJFAS 4 (33%) 2 (17%) 1 (8%) 0 0%) 1 (8%) 0 0%) 2 (17%) 2 (17%) 12 2 Ecol. Appl. 2 (18%) 1 (9%) 0 0%) 2 (18%) 0 0%) 0 0%) 0 0%) 6 (55%) 11 Ecology 1 (25%) 2 (50%) 0 0%) 0 0%) 0 0%) 0 0%) 0 0%) 1 (25%) 4 Freshwater Biology 25 (32%) 15 (19%) 8 (10%) 14 (18%) 4 (5%) 5 (6%) 1 (1%) 7 (9%) 79 Hydrobiologia 13 (15%) 20 (22%) 10 (11%) 21 (24%) 9 (10%) 4 (4%) 4 (4%) 8 (9%) 89 JNABS 24 (32%) 14 (19%) 8 (11%) 13 (18%) 3 (4%) 2 (3%) 5 (7%) 5 (7%) 74 All Journals 94 (24%) 84 (21%) 49 (13%) 70 (18%) 23 (6%) 21 (5%) 16 (4%) 34 (9%) 391 1 JNABS was formerly known as Freshwater Invertebrate Biology from 1982-1986; journal began publishing 1982. 2 Ecological Applications began publishing in 1991.

1980-1987 Archiv für Hydro. 8 (16%) 2 (4%) CJFAS 0 0%) 2 (18%) Ecology 0 0%) 2 (7%) Freshwater Biology 1 (2%) 2 (4%) Hydrobiologia 15 (12%) 16 (13%) 1 JNABS 7 (13%) 12 (23%) All Journals 31 (10%) 36 (11%)

One-time 6 monthssample < 6 months 1 year

20.3 2 1.8 0.7 13.2 14.8 9.3 65.2

6.3 1.4 3.6 6.4 15.6 8.8 42.0

Avg./yr

American Benthological Society; Ecol. Appl. = Ecological Applications; Archiv für Hydro. = Archiv für Hydrobiologie.

(McElravy 1988) and from 1999-2004. CJFAS = Canadian Journal of Fisheries and Aquatic Sciences; JNABS = Journal of the North

Table 1.1. Duration of selected macroinvertebrate studies reported in 7 hydrobiological or ecological journals from 1980-1987

Table 1.2. Duration of selected research topics (as percentages of total number of articles in each topic, which is given in parentheses) published in seven hydrobiological or ecological journals from 1999 to 2004.

Biological Community Impact Population assessment studies assessment studies (41) (206) (64) (80) One-time sample 51.2% 20.9% 25.0% 17.5% < 6 months 7.3% 25.2% 17.2% 22.5% 6 months -1 year 9.8% 13.1% 6.3% 17.5% 1 year 9.8% 20.4% 15.6% 17.5% 1.5-2 years 7.3% 6.8% 4.7% 3.8% 2 years 2.4% 5.3% 6.3% 6.3% 3 years 4.9% 2.4% 6.3% 6.3% > 3 years 7.3% 5.8% 18.8% 8.8%

15

Table 1.3. Proportion of long-term studies (≥ 3 years) comprised of selected research topics reported in 319 articles from 1980 to 1987 and 391 articles from 1999 to 2004. Because the research topic categories were not identical between McElravy’s (1988) survey and the present survey, percentages were rescaled from the 1980-1987 period to include only the three overlapping categories (i.e., community, impact assessment, and population studies). Furthermore, values may add up to more than 100% because of rounding.

1980 – 1987 1999 – 2004 Biological assessment -10% Community studies 48% 34% Impact assessment 25% 32% Population studies 28% 24%

16

Chapter 2

Long-term seasonal variation of benthic-macroinvertebrate biological traits in two mediterranean-climate streams in California, USA

17

Abstract I investigated the seasonal variation of biological traits and the influence of interannual rainfall variability on this pattern. Using long-term survey data (6-19 years) from an intermittent and a perennial stream in the mediterranean-climate region of northern California, I examined 16 fuzzy-coded biological traits (e.g., maximum size, life cycle duration, and respiration mode). Seasonal habitat variability is higher in the intermittent stream than in the perennial stream. During the winter and spring wet-season both streams flood; however, during the summer dry-season, the intermittent stream forms isolated pools in (occasionally drying completely) and the perennial stream maintains flow. Seasonal habitat variability influences both taxonomic and biological trait composition. Distinct taxonomic communities are present in each season, particularly in the intermittent stream. The intermittent stream also exhibits more seasonal variation of biological traits than the perennial stream. Despite statistically significant seasonal variation, trait composition is relatively stable among seasons in comparison to taxonomic composition and abundance. Taxonomic composition varies considerably between seasons because of high seasonal and interannual replacement of taxa resulting from seasonal habitat changes. The seasonality of taxonomic composition and abundance is sensitive to interannual rainfall variability. In dry years, the taxonomic composition of communities is more similar than in wet years, while trait composition is relatively insensitive to rainfall variability. The seasonal stability of biological traits in these two undisturbed streams suggests that the stream environments are so physically harsh (regardless of season) that traits in aquatic macroinvertebrates vary less than would

18

be expected over long time scales given that there is high variation in abundance and taxonomic composition.

Introduction Seasonality of aquatic macroinvertebrate communities has been demonstrated in a wide variety of climatic regions, from glacier-fed streams (Robinson et al. 2001, Burgherr et al. 2002) to the humid tropics (e.g., Wolda 1978, 1988, McElravy et al. 1982). This seasonality, as in many terrestrial communities (e.g., plants, insects, etc.), is often a result of variation in weather (e.g., Butler 1984). Seasonal patterns in climate, such as precipitation and insolation, result in within-year changes in flow and temperature in aquatic systems. Variation in these features greatly influences the timing of emergence (e.g., Wise 1980), reproduction (e.g. Sweeney and Vannote 1978, Vannote and Sweeney 1980), and growth and development (e.g., Becker 1973, Mackey 1977) in many aquatic macroinvertebrates, which in turn influences the seasonal replacement of organisms. The seasonal patterns of occurrence and abundance of invertebrates depend on their life history and behavioral characteristics (Wolda 1988), or biological traits. Biological traits of stream macroinvertebrates that are influenced by seasonality of abiotic and biotic factors include the presence and timing of diapause and emergence, adult longevity, and voltinism. In addition, the adaptation of life history strategies to flow conditions (i.e., to avoid floods or droughts) is dependent on the predictability of timing of seasonal flow (Lytle and Poff 2004). Despite this connection between seasonal patterns in communities and life history strategies, there have been few studies (e.g., Snook and Milner 2002)

19

explicitly examining seasonality of biological traits at the community level. In general, seasonality and temporal variability in benthic macroinvertebrate communities have been examined in terms of taxonomic (i.e., structural) characteristics (e.g., Flecker and Feifarek 1994). However, non-taxonomic aggregations of taxa into biological traits categories based on life history, behavior, and morphology may be more informative for investigating the relationship between physical habitat and community response (e.g., Richards et al. 1997). In most regions of the world, floods and a gradual reduction in discharge structure stream communities; however, the gradual reduction of flow is often interrupted by periods of rainfall. Dry season rainfall can increase flow and reduce the effects of drying (Gasith and Resh 1999), as when the dry periods are interrupted by monsoon rains in New Mexico, USA. In contrast, the mediterranean climate (including parts of coastal California, Chile, South Africa, Australia, and the Mediterranean Basin) is defined by seasonal precipitation and temperature patterns (Aschmann 1973) that result in hot, dry summers and cool, wet winters. Variability and seasonality in precipitation are the principal, defining attributes of the mediterranean climate (Gasith and Resh 1999), and the resulting hydrology influences the structure of stream communities. These predictable disturbances, in this case flooding and drying, may alter the prevalence of aquatic macroinvertebrates with behaviors or life history characteristics that allow them to exhibit resistance or resilience (by withstanding or escaping stresses, respectively) from one year to the next or from season to season (Boulton and Lake 1992a, b, Boulton 2003, Lake 2003, Flecker and Feifarek 1994). For example, aerial respiration permits the avoidance of oxygen stress, and emergence prior

20

to summer allows for avoidance of low food availability and high stream temperatures. However, some traits may only be useful at certain times, such as desiccation resistance, which may prevent invertebrates from drying-out during extreme dry seasons but is not useful during wet season floods. Alternatively, some biological traits may be stable among seasons, implying that they are equally advantageous in wet and dry seasons because the stream environment presents unique habitat constraints, regardless of the season (e.g., Snook and Milner 2002). The objective of our study was to examine whether the seasonality of aquatic macroinvertebrate communities in mediterranean-climate streams is reflected in the biological traits of organisms. This objective is based on the habitat templet concept (Southwood 1977 1988, Townsend and Hildrew 1994), where I examined the influence of seasonal variability of the physical habitat on biological trait composition in an intermittent and perennial stream using long-term datasets. In particular, I sought to discover (1) whether the known and well-described seasonal changes in taxonomic composition and abundance of these two streams are also reflected in the biological traits of these organisms; (2) whether flow permanence influences the seasonality of biological traits; and (3) whether the stability of the relationship between wet and dry season communities varies with respect to interannual differences in rainfall (comparing taxonomic versus biological trait composition).

21

Methods Study sites Hunting Creek, a 1st order stream that flows through the Morgan Valley of the California Coast Range in the University of California McLaughlin Natural Reserve in Lake County, California, was sampled biannually from 1985 to 2003. Hunting Creek is intermittent in summer and in most years isolated pools form; however, in 2 of the 19 years studied the streambed dried completely (2002 and 2003). Hunting Creek will be hereafter referred to as the “intermittent” site or stream. The 4.4 km2 watershed is characterized by steep slopes, ranging from 10 to >40%, with a mixture of serpentine and volcanic soils (Boucher and Waddell 2003). Benthic macroinvertebrate samples were collected from a single headwater reach approximately 100 m in length (38°51’56”N, 122°26’15”W, 634 m elevation). Within the sample reach, the stream habitat is characterized by riffles and shallow runs, and fine sediment and small gravel are the dominant substrates. Stream width varies from 5 to 10 m, and the gradient of the study reach is 1.5%. Riparian and aquatic vegetation consists of sedges (Scirpus and Carex spp.), rushes (Juncus spp.) and oaks (Quercus spp.), with heavy encroachment by cattail (Typha spp.) in the summer months when flow decreases. Big Sulphur Creek, a 3rd order stream that flows through The Geysers Known Geothermal Area in north eastern Sonoma County, California, was sampled biannually from 1977 to 1983. The drainage basin is formed by the steep-walled valley of the Mayacmas Mountains in the California Coast Range. Big Sulphur Creek is perennial and never dried during the study period; this site will be hereafter referred to as the “perennial” site or stream. The 35 km2 watershed is characterized by steep slopes, with

22

70-75% of the stream basin having slopes >30% with unstable and erosive terrain (McElravy et al. 1989). Benthic macroinvertebrate samples were collected from a single reach, approximately 100 m in length (38°48’3”N, 122°48’36”W, elevation 415 m), 13 km below the headwaters (McElravy et al. 1989). Within the sample reach, there are several habitat types, ranging from pools to fast riffles, and substrates ranging from large boulders (>1 m diameter) to fine sand or silt. Stream width varies from 10 to 25 m and the gradient is 2.5%. Riparian vegetation on the north-facing bank of the stream consists of blackberry (Rubus vitifolius Cham. and Schlecht), sedges (Carex spp.), and oaks (Quercus spp.), while vegetation is absent along the steep, south-facing canyon wall (McElravy et al. 1989).

Precipitation and discharge At both sites, nearly all precipitation events occur during the 7-month period from November to May (Fig. 2.1a). However, precipitation (which occurs as rainfall) is extremely variable between years, despite the highly predictable seasonal patterns within a year (Fig. 2.1a). Precipitation data are grouped by rainfall year, which is defined as October 1 to April 15 (the sample date for benthic macroinvertebrates, see below) for the intermittent site and October 1 to May 15 (the sample date) for the perennial site. Wet season rainfall at the intermittent site varied from 294 mm in 1990 to 1062 mm in 1995 (mean ± SD = 605 ± 232 mm, CV = 38.4%, 1985-2003), representing values 48 to 176% of the mean (Fig. 2.1b). The perennial site receives 43-67% more rainfall than the intermittent site, and the rainfall patterns are slightly more variable (coefficient of variation, CV = 50.6%), ranging from 442 mm in 1977 to 2505 mm in 1983, (mean ± SD

23

= 1499 ± 752 mm, 1977-1983), which is 30 to 167% of the mean (Fig.1b). Despite the differences in the amount of total rainfall between the two sites, there were no climatic differences between the two time periods of the study within a site (i.e., I compared rainfall 1977-1983 cf. 1985-2003 at each site: intermittent site, p = 0.07; perennial site, p = 0.97, Mann-Whitney U-test). Discharge also shows a predictable seasonal pattern, even though total amounts per year vary widely. Storms during the wet season commonly increase discharge 20-200 times above base flow. In addition, the timing of the wet season rainfall greatly influences flow at the time of sampling. For example, in some years (e.g., 1993), the majority of the wet season rainfall fell in December and January, while in other years, most of the rainfall occurred much later, in March or April (e.g., 1991). In summer, the perennial stream usually maintains surface flow (connected pools), whereas the intermittent stream, as noted above, has disconnected pools; in extremely dry years, the stream dries completely (e.g., 2002 and 2003).

Data collection Macroinvertebrates Aquatic macroinvertebrates were sampled at a single 100 m reach within the intermittent site at the end of the wet season (April 15 ± 3 days), and at the end of the dry season (August 15 ± 3 days), from 1985-2003. In 2002 and 2003, because of the total drying of the stream in summer, only wet season (April) samples were taken. After stratifying by uniform substrate size, five random Surber samples (0.093 m2, mesh 0.5 mm) were collected on each sampling date in each year.

24

Aquatic macroinvertebrates were sampled at a single 100 m reach within the perennial site at the end of the wet season (May 15 ± 5 days), and at the end of the dry season (August 15 ± 5 days), from 1977-1982; in 1983, only wet season (May) samples were taken because the study was terminated. After stratifying by uniform substrate size, four random Surber samples (0.093 m2, mesh 0.5 mm) were collected on each sampling date in each year. At both sites, benthic macroinvertebrate samples were taken to coincide with the end of the wet season (late spring sampling), which is prior to the emergence period of most aquatic insects at each site, and at the end of the dry season (late summer sampling). Stratification by substrate size biases our sample towards a specific habitat; however, this habitat usually comprises > 90% of all habitats at both sites. Most samples were processed without sub-sampling, except for samples with large amounts of organic debris, or > 2500 individuals (intermittent: three dry season samples; perennial: ~1/3 of samples), and ¼ to ½ of the sample was processed. All specimens were identified to the lowest taxonomic level possible (usually genus, with species distinguished as OTUs, i.e. operational taxonomic units sensu Resh (1979)); in most cases, the OTUs represent a single species. Diptera and non-insects were often identified to the family or sub-family levels with lower taxonomic units identified as OTUs, and usually represent species. Both datasets have minimized operator inconsistency, a common problem in benthic macroinvertebrate studies (Needham and Usinger 1956). The same persons performed all of the identifications, sampling, and sample processing for the duration of both studies (Intermittent site: E.P McElravy and V.H. Resh; Perennial site: G. A. Lamberti and E.P McElravy).

25

Biological traits I used 16 biological traits (73 modalities) for taxa in each stream (e.g., Resh et al. 1994, Statzner et al. 1997, Tachet et al. 2000, Usseglio-Polatera et al. 2000, Gayraud et al. 2003) (Table 2.1). Information was collected from a variety of literature sources that ranged from published life histories to regional fauna guides, particularly from mediterranean-climate regions. Biological traits were coded at the level of identification, which was usually genus. Trait information was collected for 149 (of 152) taxa from the intermittent stream and for 109 (of 109) taxa from the perennial stream (185 total taxa, with 73 taxa being shared between both sites). Early instars of Diptera, Trichoptera, and Plecoptera (3 taxa identified only to order) were not included to reduce ambiguity in coding. Fuzzy coding was used to record the biological traits, which allows for variation within a taxon for a specific trait (Chevenet et al. 1994). Each taxon is given a score (between 0 and 3, or 0 and 5), corresponding to its affinity to the modality (Chevenet et al. 1994. For taxa that were identified at the family or sub-family level (e.g., early instars of Chloroperlidae), the average of the taxa belonging to that group was used. Traits were rescaled as proportions (sum = 1), such that values ranged from 0 (no affinity among individuals for the modality) to 1 (all individuals had exclusive affinity for the modality).

Analyses

Data from the intermittent and perennial sites were analyzed separately. Samples from each site for a given season and year were pooled (intermittent site = 5 samples,

26

perennial site = 4) to create a composite sample, representing a site-season-year (hereafter referred to as a ‘sample’). These data were then log10 (x + 1) transformed to reduce the effect of extremely abundant taxa present during the dry season. Traits were weighted by the transformed abundance by multiplying the two matrices and rescaling the data (sum =1 within each trait). Density (numbers/m2), taxa richness (number of taxa), and Shannon diversity (H’) were calculated using raw abundance. The coefficient of variation of each statistic was also calculated (CV = standard deviation/mean). Comparisons were made using nonparametric Wilcoxon III paired tests to compare differences between wet and dry seasons within a year at a given site. To examine changes in community structure and biological traits between seasons over time, I used three different analysis techniques. First, I examined taxonomic composition, abundance, and trait composition between seasons using multivariate ordination. Each site was ordinated separately using non-metric multidimensional scaling (NMS; Kruskal 1964) with Sørensen (Bray-Curtis) distance for abundance or traits. NMS was performed using PC-ORD 4.27 (McCune and Mefford 1999). NMS is non-parametric distance-based ordination method that maximizes the correspondence between measured dissimilarities and distance between data points within a predefined number of spatial dimensions (Legendre and Legendre 1998). An initial ordination was conducted with 2000 runs of real data and randomized starting coordinates. The number of dimensions to be retained was evaluated after inspecting the stress (goodness of fit) of solutions with dimensions 1 through 6. Stress measures the departure from monotonicity in the relationship between the distance in the original p-dimensional space and distance

27

in the reduced k-dimensional ordination space, and values range from 0 to 100, with values close to 0 being a good fit of the data. A subsequent NMS was run (2000 runs real data, 999 runs randomized data) with the specified dimensions and using the first solution as starting coordinates. The final ordination helps to ensure that a stable solution is reached and that the ordination is based on global minima and not local minima (McCune et al. 2002). Significance was assessed by conducting Monte Carlo tests using 999 runs of randomized data in the final ordination. A p-value is calculated as a function of the number randomized runs that resulted in a stress less than or equal to the observed stress (McCune and Mefford 1999). The following NMS analyses were performed: (1) NMS of taxonomic composition by season-year for each site separately (intermittent, perennial); and (2) NMS of trait composition by season-year for each site (intermittent, perennial) separately. I examined the traits that were significantly correlated with explanatory axes representing time (year) and rainfall using a non-parametric Spearman rank correlation (rs). Furthermore, to examine temporal variation of communities, I plotted the axis that explained the most variation against year for each ordination. Second, I compared trait proportions at a given site between seasons and years. Paired-comparisons of trait proportions between seasons (for each site) were conducted using a non-parametric paired Wilcoxin III test, with a Bonferroni correction for multiple comparisons. A non-parametric test was chosen because the data (trait proportions) were not normally distributed. The alpha value (α = 0.05) was divided by the number of trait modalities within a trait group (e.g., six modalities in “maximum body size”) to obtain the corrected alpha value (e.g., α = 0.008).

28

Third, I compared how taxonomic composition, abundance, and trait composition varied between seasons in relationship to annual rainfall variability (a surrogate for flow data that were not available for both sites). That is, are the differences between taxonomic and trait composition between seasons constant through time, and do they respond to the habitat templet concept as related to flow variability? To do this, I examined the dissimilarity between wet and dry season communities by calculating the Sørensen distance between seasons (based on log10-transformed abundance and trait data, separately) for each year. The distance values were plotted against total rainfall (mm) in each year for both sites separately. Values for dry-season modalities in each year were regressed against values for wetseason trait modalities values. If the linear regression is not significantly different from y = x (as determined by a t-test for differences in slope and intercept; Zar 1999), then the traits do not vary between seasons. Linear regressions were performed using log10transformed abundance data for all taxa and for each site separately.

Results Taxonomic composition and abundance Structural measures The intermittent and perennial streams are similar in faunal composition and structure. Many taxa (73) are common to both streams, comprising 48.0% of total taxa from the intermittent stream and 67.0% of total taxa from the perennial stream. However, the wet and dry season communities at both sites are relatively discrete taxonomically. The intermittent stream has more taxa “unique” to each season than the

29

perennial stream. For example, at the intermittent site, 70 taxa (46.1%) are present in both seasons, 42 (27.6%) only in the wet season and 40 (26.3%) only in the dry season. At the perennial site, 65 taxa (59.6%) are common to both seasons, only 28 (25.6%) are unique to the wet season, and even fewer (16, 14.7%) taxa are unique to the dry season. At the intermittent site, density and diversity differ between seasons (density: dry > wet season, p = 0.05; diversity, H’: wet > dry season, p = 0.003), whereas taxa richness is not significantly different (p = 0.07). At the perennial site, only density differs significantly between seasons (dry > wet season, p = 0.03). Although some structural measures vary substantially between years in wet and dry seasons, the stability of these measures is not consistent (Table 2.2). For example, the CV of density (no. of individuals/m2) indicates low stability (ranging from 70.8% in perennial dry season to 117.4% in intermittent dry season), whereas the CV of diversity (H’) indicates stability of this metric (ranges 15.4% in intermittent wet season to 25.4% in intermittent dry season) (Table 2.2).

Multivariate ordination Similar to patterns in structural measures, NMS ordination of log10 transformed abundance from the intermittent stream indicates that there are distinct wet and dry season communities. For example, the NMS ordination (stress = 11.2, p = 0.001) represents 92.3% of the variation in the dataset, with 53.8% on axis 1 and 38.5% on axis 2 (Fig. 2.2a, Table 2.3). The points appear to be separated by season on axis 1. Year and total rainfall are significantly correlated with axis 2 (rs = 0.55, and rs = 0.36, respectively) (Table 2.3). Three years (1985-87) appear to be distinct from other years (separated

30

along axis 1); in these years, the total abundance of organisms is less than 35% than the average of the other years. In addition, in 1997 the stream was almost dry at the time of sampling and the very few taxa that were found were in low abundance (separated along axis 3). Despite these irregular years, the summer communities are generally more variable between years (variance of position on axis 2 = 0.49) than are wet season communities (axis 2 variance = 0.08) (Figs. 2a, 4a). Furthermore, the wet season community appears to cycle in response to periods of drought (e.g., 1987-1990), whereas the dry season community does not cycle in response to rainfall (Fig. 2.4a). The NMS ordination of log10 transformed abundance and taxonomic composition from the perennial stream is similar to the intermittent stream in terms of differences between seasons. However, the year of sampling and total rainfall also affected the community structure (Fig. 2.2b). The NMS ordination (stress = 10.0, p = 0.001) represents 90.4% of the variation in the dataset, with 13.9% on axis 1, and 76.5% on axis 2 (Fig. 2.2b, Table 2.3). The points appear to be separated by season along axis 2. When both seasons are considered together, year is not significantly correlated with either axis, but total rainfall is significantly correlated with axis 2 (rs = 0.57) (Table 2.3). The distance between wet and dry season samples is less in years where rainfall is very low (lower discharge), and more when rainfall is high (higher discharge) (Fig. 2.4b); this is also apparent in analyses of distance and stability (below). In comparison to the results from the intermittent site, wet season communities of the perennial site are more variable between years (variance on axis 2 = 0.50) than are dry season communities (variance on axis 2 = 0.13) (Fig. 2.2b, 2.4b). Furthermore, the high correlation between axis 2 and year (not statistically significant)

31

suggests that there is directional change in the community, which is more evident in the wet season than in the dry season (Fig. 2.4b).

Biological trait composition Multivariate ordination The NMS ordination of seasonal trait composition from the intermittent stream (stress = 6.4, p = 0.001) represents 96.5% of the variation in the dataset in a 3-axis solution (Fig. 2.3a-c). Axis 1 represents 16.4% of the variance in the data matrix, axis 2 represents 62.0%, and axis 3 represents 18.2% of the variance (Fig. 2.3a-c, Table 2.3). Season appears to separate the points along axis 2. Year is significantly correlated with axis 1 (rs = -0.36), and rainfall is not correlated with any axis (Table 2.3). Similar to the ordination of taxonomic composition and abundance (Fig. 2.2a), there is more variation in trait composition in the dry season trait than in the wet season, with several years standing out as unique (1985-87, 1997) on axis 1 or 3. When axis 2 (62% of the variation) is plotted against year (Fig. 2.4c), it is evident that both wet and dry seasons cycle in response to periods (i.e., 1987-1990) or years of drought (i.e., 1994). The weighted average score of traits indicates that dry season communities are dominated by traits such as desiccation resistance (h1), terrestrial eggs (l6), moderate (b2) to strong body armoring (b3), and very small maximum body size (< 2.5 mm, a1), among other traits (negative side of axis 2, Fig. 2.3b-c). Wet season communities are dominated by traits such as CPOM consumption (o2), fixed clutches (l3), no body armoring (b1), and short-lived adults (≤ 1d, g1 and 1-10 days, g2), among other traits (positive side of axis 2, Fig. 2.3b-c).

32

The NMS ordination of seasonal trait composition from the perennial stream (stress = 5.6, p = 0.001) represents 96.6% of the variance, with 92.6% on axis 1 and only 4.0% on axis 2 (Fig. 2.3d, Table 2.3). The points appear to be separated by season along axis 1. Year is significantly correlated with axis 1 (rs = 0.70), but total rainfall is not significantly correlated with either axis (Table 2.3). The gradient of dry to wet season samples is more pronounced than in the ordination of taxonomic composition and abundance from the same site (Fig. 2.3d cf. Fig. 2.2b). There is also a clear temporal gradient (on axis 1) that is weakly correlated with rainfall (rs = 0.46, p = 0.11). However, upon further examination of this temporal trend, it appears that there is linear, directional change in both wet and dry season communities (NMS axis 1 vs. year, Fig. 2.4d) after a severe drought year (1977; Fig. 2.1b). The weighted average score of traits indicates that dry season communities are dominated by traits such as a very small maximum size (< 2.5 mm, a1), life cycle duration > 1 year (d2), aquatic adults (f3), adult diapause (i3), and ovoviparity (l7), among other traits (negative side of axis 1, Fig. 2.3f). Wet season communities are dominated by traits such as short-lived adults (≤ 1d, g1; 1-10 days, g2; and > 10-30 days, g3), diapause in egg (i2), larva or pupa (i2), fixed, single eggs (l1), and burrowing (n5) or endobenthic (n6) locomotion among other traits (positive side of axis 1, Fig. 2.3f). Fuzzy-correspondence analysis (FCA; Chevenet et al. 1994), whose use has been advocated for the analysis of fuzzy-coded trait data, produced similar results to NMS for both sites (intermittent and perennial) in terms of the season-year and trait scores, as well as relationships of the primary axes to year and total rainfall.

33

Trait frequencies Seasonally, there are significant differences among biological traits in both streams, although the differences are greater in the intermittent stream than the perennial stream. At the intermittent site, 39 of the 73 biological-trait modalities differ between seasons (p ≤ 0.05 after Bonferroni correction) (Table 2.4). In contrast, none of biological-trait modalities differ significantly between seasons at the perennial stream after a Bonferroni correction, although 10 traits differ at p < 0.10 with the Bonferroni correction (= nearly significant, “n” in Table 2.4). The number of paired samples (i.e., wet-dry seasons) that would be required to detect a difference with the adjusted p-values is > 6, and only size years (i.e., season pairs) were available for analysis at the perennial site. Furthermore, when data from both sites are combined, there are 36 statistically significant modalities between the seasons, and these are not always the same as those from the intermittent site (I+P, Table 2.4). Overall, however, trait composition of the perennial stream does not vary seasonally as much as at the intermittent stream, which is consistent with the multivariate analyses. In fact, trait composition of the perennial stream in the dry season is generally more similar to trait composition of the intermittent stream in the wet season than in the dry season (graph not shown).

Stability of taxonomic composition, abundance, and biological traits Trait composition is more similar between seasons than is taxonomic composition and abundance (Fig. 2.5), even though the trait composition of communities at both sites does differ (Fig. 2.3, Table 2.4). For example, the seasonal distance (based on Sørensen distance) for taxonomic composition and abundance at the intermittent site is 0.62 ± 0.09

34

(n = 17 years ± SD), and 0.48 ± 0.09 (n = 6 years ± SD) at the perennial site. In contrast, the distance for trait composition between seasons at the intermittent site is 0.13 ± 0.02 and 0.08 ± 0.02 at the perennial site. Similar to the results from both the NMS ordinations and frequency analyses (Fig. 2.2 and 3), the seasonal difference between communities is greater at the intermittent site than the perennial site, regardless of the data type (Mann-Whitney U-tests, intermittent vs. perennial: trait composition, p 40 Life span b1 ≤ 1 yr b2 > 1 yr No. generations/yr c1 1 Aquatic stage d1 egg d2 larvae d3 adult Reproduction e1 sexual e2 asexual e3 parthenogenic Type of egg(s) f1 single, free f2 single, attached f3 clutches, free f4 clutches, attached f5 endophytic f6 terrestrial f7 ovoviparious

Dispersal h1 aquatic passive h2 aquatic active h3 aerial passive h4 aerial active Resistance i1 desiccation i2 diapause i3 none Respiration j1 cutaneous j2 gills j3 plastron j4 aerial respiration j5 respiratory pigments Locomotion k1 flight k2 swimmer surface k3 swimmer k4 crawling k5 burrower (epibenthic) k6 endobenthic (interstitial) Food eaten l1 FPOM 1mm l3 periphyton/algae l4 living macrophyte l5 dead animals > 1 mm l6 microinvertebrates l7 macroinvertebrates l8 vertebrates

Feeding habits m1 absorber m2 deposit feeder m3 shredder m4 scraper m5 filterer m6 piercer (plant/animal) m7 predator m8 parasite Adult longevity n1 ≤ 1 d n2 > 1-10 d n3 > 10-30 d n4 > 30d to 365 d n5 > 365 d Diapause stage p1 egg p2 larvae/pupae p3 adult p4 none Body armoring q1 none (soft-bodied) q2 moderate (sclerotized) q3 strong (snails/stone cased) Body shape s1 flattened s2 cylindrical s3 spherical Female dispersal t1 < 1 km t2 > 1 km

149

Table 3.3. Spearman rank correlation (rs) among all physical variables examined. Significant correlations (p ≤ 0.05) are indicated in bold.

Year Hunting/Knoxville Creeks Year -Total precip. 0.45 1-month precip. 0.19 2-week precip. 0.41 Bed movement 0.27 FST-hemisphere 0.15

Total precip.

1-month precip.

2-week precip.

Bed movement

-0.54 0.54 0.90 0.33

-0.46 0.44 0.27

-0.26 0.38

-0.26

Big Sulphur Creek Year Total precip 1-month precip. Bed movement

-0.54 0.43 0.58

-0.82 0.99

-0.79

--

--

Blodgett Forest Year Total precip. Bed movement

--0.86 -0.70

-0.82

--

--

--

150

Table 3.4. Wet-season bed movement (number of times bed moved) and sediment characteristics for each site. The years in which the maximum and minimum amount of bed movement occurred are noted. “%50 finer” refers to the sediment size that ~50% of all pebbles sampled were finer than. “% Embedded” refers to the % of sediment particles measured that were embedded.

1D

Bed movement 5.7 ± 7.0

Max 25 (1995)

2D

1.3 ± 2.1

6 (1995, 1997)

1P 2P BSC BC DA DE GC MC

28.2 ± 26.1 10.0 ± 11.9 19.1 ± 14.7 --4.9 ± 7.1 22.8 ± 23.8 --

76 (1998) 40 (1995) 39 (1983) --60 (1997) 20 (1997) --

Min 0 (1988-90, 1992, 1994, 1999) 0 (1984-85, 1987-92, 1994, 1996, 19992003) 2 (1990) 0 (1988, 1990, 1998) 0 (1977) --0 (2001-03) 0 (2001) --

50% finer (mm) 32

% embedded 14

64

27

8 32 - 1mm* periphyton* macrophytes dead animals > 1mm microinvertebrates* macroinvertebrates vertebrates absorber* shredder* scraper filterer* piercer predator parasite* 1-10 d* >10-30 d* >30-365 d* >365 d*

HC/KC D W (40) (32) 3 -1 -1 -2 --2 1 --3 -- -1 -2 -1 -2 --2 -1 2 --1 2 --3 -3 -3 -- -4 --- --1 -3 2 --3 1 -1 -1 -2 1 -4 1 -2 -1 -1 1

BSC D W (15) (16) -1 -- -1 --1 1 --1 1 --- -1 --1 -- -1 --1 1 --- --1 -1 1 -1 --- -1 --1 -- --- --- --- --- --- --- --- --- --- --- --1 -- --- --

BFRS D W (36) (32) 3 --- -3 -2 --3 3 --4 -1 -2 4 -2 -1 2 4 --5 4 -2 -3 --3 -- --2 -- -4 --4 -- -5 -5 1 -- --1 -- -1 -3 --3 3 1 -2 1 1 1 --

% of sites D W (58) (58) 60 10 10 -50 -40 10 10 50 40 10 10 70 -- 10 20 20 60 10 30 -40 20 40 30 10 60 60 -20 20 50 10 10 60 10 30 -- 50 10 -80 10 -- 40 -- 10 50 30 70 10 -- 30 10 10 10 -20 -50 10 -- 70 40 10 20 30 20 10 20 10

160

Trait category Diapause

Modality

p1 p2 p3 p4 Body armoring q1 q2 q3 Body shape s1 s2 s3 Female dispersal t1 t2

egg larva/pupa* adult none* none moderate strong* flattened cylindrical* spherical low (< 1km) * high (>1km) *

HC/KC D W (40) (32) -2 3 --- --- --3 3 -1 2 -1 1 -2 --2 2 --

BSC D W (15) (16) -- --1 -- -1 --- --- --- --- --1 -- -1 --1

BFRS D W (36) (32) -- -4 --2 1 3 -2 2 --- --- --- -1 --3 3 --

% of sites D W (58) (58) -- 20 70 10 -- 20 20 30 -- 50 50 -10 20 -- 10 10 10 30 -10 50 50 10

161

Table 3.13. Number of trait modalities (total = 75) that were significantly correlated (Spearman rank correlation, p ≤ 0.05) with each physical variable at each site. For site 1P only wet-season data was included. The total number of modalities correlated with one or more physical variables are indicated. Traits correlated with one or more variable were counted only once to determine totals.

Total precipitation 1-month precipitation 2-week precipitation Bed movement FST-hemispheres Total

1D 2D 1P 2P BSC BC DA DE GC MC 23 29 23 7 18 34 50 24 38 38 16 14 4 7 27 -----8 29 21 4 ------7 17 21 8 18 --- 23 22 -30 1 32 7 ------43 40 39 18 31 34 50 25 40 38

162

Table 3.14. Summary of community change analyses for the composition and abundance of taxa at each site. Long-term persistence and stability values are given (i.e., between first and last year of sampling), as well as whether this value is greater than (>), less than (
0.51 BC 8 0.34 = 0.90 DA 8 0.30 < 0.82 DE 7 0.30 < 0.88 GC 8 0.37 < 0.80 MC 8 0.40 > 0.84 1 TLR was nearly significant (p = 0.06).

= > < > >
), less than ( 0.19 < 0.25 > 0.10 > 0.20 > 0.21 >

TLR no change no change directional no change no change directional directional directional directional directional directional

Ordination directional then convergence convergence directional, convergence directional, then no change directional then convergence directional then no change directional directional directional then convergence directional then no change directional then convergence

164

Figure 3.1. Total precipitation for each water year: (a) the Hunting and Knoxville Creek area (October 1-April 15, 1984-2003); (b) Big Sulphur Creek (October 1-May 15, 19761983); and (c) Blodgett Forest (October 1-September 30, 1994-2003). Long-term average precipitation is indicated by the solid line and the dashed lines represent mean ± 1 SD. Hunting/Knoxville Creek: 653 ± 251 mm (1946-2003); Big Sulphur Creek: 872 ± 286 mm (1896-2003); Blodgett Forest: 1579 ± 543 mm (1962-2003).

Total precipitation (mm)

1200

1600

(a)

1000

1200 800

1000

600

800 600

400

400 200

200

0

0 1985

3000

Total precipitation (mm)

(b)

1400

1990

1995

2000

1976

1978

1980

1982

(c)

2500 2000 1500 1000 500 0 1994

1996

1998

2000

2002

Water year

165

Figure 3.2. Average daily discharge for site: (a) 2P (modelled; Oct. 1, 1984 – Apr. 30, 2002); (b) BSC; (c) DA. 140

(a)

Flow (m3/s)

120 100 80 60 40 20 0 1985

140

1987

1989

1991

1993

1995

1997

1999

2001

2003

(b)

Flow (m3/s)

120 100 80 60 40 20 0 1981

1.2

1982

1983

(c)

Flow (m3/s)

1.0 0.8 0.6 0.4 0.2 0.0 1995

1996

1997

1998

1999

2000

2001

2002

2003

166

Figure 3.3. Average (± SD) FST-hemisphere at each Hunting/Knoxville Creek site from 1989 to 1998. 1st order sites (1D, 1P) are represented by circles, 2nd-order sites are represented by squares. Temporary-flow sites (1D, 2D) are represented by open symbols with dashed lines and permanent-flow sites (1P, 2P) are represented by filled symbols with solid lines.

16

FST-hemisphere

14 12 10 8 1D 2D 1P 2P

6 4 2 0 1990

1992

1994

1996

1998

Year

167

168

-1.5

NMS Axis 1

-2

03 (end)

(a) 1D

02

NMS Axis 3

88

94

91 87

96 99 98 93 92 (start) 86 97 8584 01

95

00

90

89

2

2

-2

NMS Axis 2

1.5

-2

03 (end)

(b) 2D

01 94

87

91 92 93 85 88

96

86

84 (start) 95

NMS Axis 1

0297

98

99 00

90 89

2

1.5

-1.5 -1.5

95

99 0389 (end)

91

NMS Axis 3

94

88 97 92 90

93 9896

0100

02

(c) 1P (wet)

is arbitrary). Each point is represented by the sample year and the first (“start”) and last year (“end”).

87

85 (start)

86

1.5

2P; (f) BSC; (g) all BFRS; (h) BC; (i) DA; (j) DE; (k) GC; (l) MC. The two axes explaining the most variance are shown (axis order

Figure 3.4. Non-metric multidimensional scaling analysis (NMS) plots for each site: (a) 1D; (b) 2D; (c) 1P; (d) 1P (dry-season); (e)

NMS Axis 2

169

2.0

-2

NMS Axis 2

2

-2.0

97

94

92

93 00 98 91 90 88 01 (end)

NMS Axis 1

89

96

99

-2

BC DA DE GC MC

NMS Axis 1

(h) all Blodgett Forest

-1.5

87 8685 (start)

(d) 1P (dry)

NMS Axis 2

95

1.5

2

2

-2

2

-2

NMS Axis 2 -2

93

-2

(i) BC

(e) 2P

94

03 (end)

02

87

96 97 95 (start)

99

84 (start)

NMS Axis 1

00 03 (end) 01 98

02

NMS Axis 1

95

89 88 96 01 00 97 85 91 86 98

92

90

2

2

0

2

4

6

8

-2

-1.5 -1.5

1.5

Rank

81

79

(j) DA

77 (start)

(f) BSC

NMS Axis 1

98 95 (start)

97 96

00 0301 (end)

02

NMS Axis 1

78

80

82

1.5

2

83 (end)

total rain, 1-mo, 2-wk, FST

170

2

-2 -1.5

NMS Axis 2

02 01 03 (end) 00

(k) DE

NMS Axis 1

9796 98(start)

1.5

-2

2

-2

02 01 00(end) 03

(l) GC

NMS Axis 1

95 (start)

98 97 96

2

-2

2

-2

(m) MC

9796 95 (start)

NMS Axis 1

98

00 03 (end)

02 01

2

171

2000

2000

-2

1995

-2

1990

-1

-1

1

2

0

1985

(e) 2P

0

1

2

-2

1995

-2

1990

-1

-1

1

2

0

1985

(a) 1D

0

1

2

(k) MC.

1990

1978

(f) BSC

1985

(b) 2D

1980

1995

1982

2000

0.0

0.5

1.0

1.5

2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

1996

(g) BC

1985

1998

1990

(c) 1P

2000

1995

2002

2000

-1.0

-0.5

0.0

0.5

1.0

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

1996

(h) DA

1985

1998

1990

(d) 1P (dry)

2000

1995

2002

2000

(explaining the most variation) over time: (a) 1D; (b) 2D; (c) 1P; (d) 1P (dry-season); (e) 2P; (f) BSC; (g) BC; (h) DA; (i) DE; (j) GC;

Figure 3.5. Temporal change of community composition and abundance at each site as shown by position along the NMS axis

NMS Axis

NMS Axis

172

NMS Axis

-2.0 2002

-1.5 2000

-1.5

-1.0

1998

-1.0

-0.5

0.0

0.5

-0.5

1996

(i) DE

0.0

0.5

1.0

1996

(j) GC

1998

2000

2002

-1.0

-0.5

0.0

0.5

1.0

1996

(k) MC

1998

2000

2002

173

2

4

6

8

10

2

4

6

8

10

2

3

1

3

Timelag (sqrt)

2

R2 = 0.06 (p = 0.0007)

(e) 2P

1

R2 = 0.10 (p 1 km is 71% compared to 39% of rare taxa). However, even though rare taxa do not disperse very far (< 1 km), they were more likely to disperse via flight than common taxa (50% compared to 25%; Table 4.2). The predominance of flight dispersal in the rare taxa indicates that these taxa are rare because of an infrequent expansion in spatial distribution. Common taxa were also more likely to be of intermediate size (5-10 mm), whereas rare taxa came from a range of sizes (2.5-40 mm), a pattern which has also been found in studies of taxonomically similar animals (Gaston and Kunin 1997). I also found that the pattern of trait differences between rare and common taxa differed among sites, with only a few traits (e.g., dispersal traits) showing similar patterns among different study areas (Table 4.2). Is the preponderance of rare taxa that I found in these northern California streams a pattern that is likely to be found in long-term datasets from other regions? The mediterranean-climate of this region results in high predictability in terms of timing of the wet and dry season, but low predictability in terms of the amount of precipitation (Fig. 4.1), which may influence temporal rarity patterns. Although our objective in this study was not a long-term analysis of others’ data, I chose a 10-yearr study from The Netherlands (Pillot 2003), a region with a completely different region and climate, to

209

examine whether a similar trend was apparent. In that study, 22.9% of the 201 benthic macroinvertebrate taxa collected only occurred one year, and 9.0% occurred each of 10 years; these results are comparable to the 7–8 year datasets examined for Big Sulfur Creek and Blodgett Forest (Fig. 4.3), and indicate that the pattern observed is not specific to our study streams. Unresolved taxa will continue to be a problem for interpretation of benthic macroinvertebrate studies in the immediate future. Unresolved taxa, which in our study comprised 2-3% of all taxa, may eventually become identifiable through (1) the development of better keys for difficult-to-identify life history stages (e.g., early instars), (2) the use of DNA technology (e.g., bar-coding) to associate life history stages of the same species, or (3) the incorporation of rearing techniques into macroinvertebrate survey programs. However, all these inclusions add cost to survey studies and the tendency in current benthic macroinvertebrate surveys is to develop approaches that expend less effort (and consequently cost). Some protocols even consider dropping unresolved taxa from datasets in regions that are attempting standardization among multiple monitoring programs. With this approach, unresolved taxa will surely disappear from future examination and analyses. There has been a great deal of discussion about the inclusion of rare taxa in the analysis of benthic surveys (e.g., Cao et al. 1998, 2001, Cao and Williams 1999, Marchant 1999, 2002). Rare species are often excluded from studies of community persistence because they are assumed to be chance occurrences that are not reflective of overall community change (e.g., Scarsbrook 2002). This study, although limited to northern California streams, demonstrates that rare species are in fact temporally

210

common in their occurrence but low in their densities. Even if all rare and unresolved taxa collected during a single year were combined, their densities would be a fraction of the total fauna collected. Arguably, these rare species may be superfluous or redundant to the community in terms of ecosystem function or biomass contributions.

However,

rare taxa may also be important members of communities in undisturbed streams (e.g., Robinson et al. 2000), or they may provide valuable information on community change (e.g., Woodward et al. 2002). Despite the different approaches to dealing with rare taxa in long-term studies, they greatly influence measures of community persistence and, thus, interpretations of the effects of environmental change on community assemblages. In conclusion, the assumption that long-term benthic macroinvertebrate surveys in streams can provide useful baseline data for biomonitoring has 2 major underpinnings: constancy of metrics and communities over time in the absence of impact, and negligible effects of inter-operator differences. Even though taxa and summary metrics are assumed to be fairly constant over time, few long-term datasets have been collected to test whether this is correct. Moreover, most long-term surveys that have been done have involved many different collectors, and the patterns observed could be attributable to inter-operator differences (e.g., Needham and Usinger 1956). Because a huge investment has been made in starting long-term biomonitoring with benthic macroinvertebrates at selected sites in much of the United States (Carter and Resh 2001) and in other developed countries (Bonada et al. 2006), we want to assume that long-term data will yield significant benefits for management decisions. This study is not intended to diminish the need for, or value of, long-term benthic macroinvertebrate surveys; however, the results

211

do emphasize that we need to develop new ways of looking at temporal survey data collected in reaching conclusions that have management implications.

References Bonada, N., Prat, N., Resh, V.H., Rieradevall, M., and Statzner, B. 2006. Developments in aquatic insect biomonitoring: a comparative analysis of recent approaches. Annual Review of Entomology. In Press. Carter, J.L. and Resh, V.H. 2001. After site selection and before data analysis: sampling, sorting, and laboratory procedures used in stream benthic macroinvertebrate monitoring programs by USA state agencies. Journal of the North American Benthological Society 20:658-682. Chevenet, F., Dolédec, S., and Chessel, D. 1994. A fuzzy coding approach for the analysis of long-term ecological data. Freshwater Biology 31:295-309. Cao, Y., Williams, D.D., and Williams, N.E. 1998. How important are rare species in aquatic community ecology and bioassessment? Limnology and Oceanography 43:1403-1409. Cao, Y. and Williams, D.D. 1999. Rare species are important for bioassessment—reply to Marchant’s comments. Limnology and Oceanography 44:1841-1842. Cao, Y., Larsen, D.P., and St-J. Thorne, R. 2001. Rare species in multivariate analysis for bioassessment: some considerations. Journal of the North American Benthological Society 20:144-153. Connell, J.H. and Sousa, W.P. 1983. On the evidence needed to judge ecological stability or persistence. The American Naturalist 121:789-824.

212

del Rosario, R.B. 2000. Aquatic insect responses to hydrologic intermittency and organic enrichment in headwater streams. Ph.D. Dissertation, University of California, Berkeley, California. 140 pages. Gasith, A. and Resh, V.H. 1999. Streams in mediterranean climate regions: abiotic influences and biotic responses to predictable seasonal events. Annual Review of Ecology and Systematics 30:51-81. Gaston, K.J. 1994. Rarity. Chapman & Hall, New York. Gaston, K.J. and Kunin, W.E. 1997. Rare-common differences: an overview. Pages 1229 in Kunin, W.E. and Gaston, K.J. (Eds.), The biology of rarity: causes and consequences of rare-common differences. Chapman & Hall, New York. Marchant, R. 1999. How important are rare species in aquatic community ecology and bioassessment? A comment on the conclusions of Cao et al. 1999. Limnology and Oceanography 44:1840-1841. Marchant, R. 2002. Do rare species have any place in multivariate analysis for bioassessment? Journal of the North American Benthological Society 21:311-313. McElravy, E.P., Lamberti, G.A., and Resh, V.H. 1989. Year-to-year variation in the aquatic macroinvertebrate fauna of a northern California stream. Journal of the North American Benthological Society 8:51-63. Needham, P.R. and Usinger, R.L. 1956. Variability in the macrofauna of a single riffle in Prosser Creek, California, as indicated by the Surber sampler. Hilgardia 24:383-409. Nijboer, R.C. and Verdonschot, P.F.M. 2004. Rare and common macroinvertebrates: definition of distribution classes and their boundaries. Archiv für Hydrobiologie 161:45-64.

213

Pillot, H.M. 2003. Hoe waterdieren zich handhaven in een dynmische wereld. 10 jaar onderzoek in de Roodloop, een bovenloopje van der Reusel in Noord-Brabant. Stichting het Noordbrabants Landschap. Haaren, The Netherlands. Resh, V.H. 1979. Biomonitoring, species diversity indices, and taxonomy. Pages 241-253 in Patil, G.P., Smith, W.K. and Taillie, C. (Eds.), Ecological Diversity in Theory and Practice. J.F. Grassle, International Cooperative Publishing House, Burtonsville. Robinson, C.T., Minshall, G.W. and Royer, T.V. 2000. Inter-annual patterns in macroinvertebrate communities of wilderness streams in Idaho, U.S.A. Hydrobiologia 421:187-198. Scarsbrook, M.R. 2002. Persistence and stability of lotic invertebrate communities in New Zealand. Freshwater Biology 47:417-431. Woodward, G., Jones, J.I., and Hildrew, A.G. 2002. Community persistence in Broadstone Stream (U.K.) over three decades. Freshwater Biology 47:1419-1435.

214

Table 4.1. Summary of physical characteristics of the study sites. Stream width is based on post wet-season measurements.

Latitude (N)

Longitude (W)

Elevation (m)

Stream order

Watershed area (km2)

Stream width (m)

38°51’56” 38°49’45” 38°48’30”

122°24’54” 122°22’45” 122°22’36”

634 402 348

1 2 2

4.4 22.1 29.3

5-10 10-15 10-15

Knoxville Creek (1984-2003) 1D 38°47’56”

122°18’53”

390

1

2.1

3-8

Big Sulphur (1977-1983) BSC

38°48’3”

122°48’36”

415

3

35.0

10-25

Blodgett Forest (1995-1998, 2000-2003) DA 38°54’21” BC 38°54’44” MC 38°54’17” DE 38°55’13” GC 38°52’12”

120°38’48” 120°39’25” 120°38’39” 120°40’14” 120°38’31”

1452 1380 1310 1280 1243

1 1 1 1 2

1.3 1.0 2.4 1.4 4.9

0.8-2.1 0.5-2.4 0.9-4.6 1.5-5 2.1-7.4

Study site (study years) Hunting Creek (1984-2003) 1P 2D 2P

215

216

Dispersal

Passive aquatic dispersal (drift) Active aerial dispersal Female dispersal potential > 1 km > 1 generation/year Life cycle Aquatic adults Aquatic larvae Cutaneous Respiration Aerial respiration FPOM (< 1 mm) Food eaten CPOM (> 1mm) Reproduction Synchronous emergence Maximum size 5-10 mm Morphology Soft-bodied Cylindrical body shape Spherical body shape Larval or pupal diapause Resistance

Trait

and post dry-season as separate sites).

common 0.42 0.25 0.71 0.59 0.01 0.77 0.65 0.03 0.38 0.31 0.23 0.49 0.82 0.86 0 0.16

rare 0.17 0.50 0.39 0.24 0.14 0.51 0.33 0.37 0.13 0.10 0.07 0.20 0.37 0.55 0.14 0.04

Proportion with trait

# sites with common-rare difference (p < 0.05) Hunting/ Big Sulphur Blodgett Knoxville (5) (2) (5) 4 -2 5 --3 -1 5 1 -2 -1 2 --4 --4 --5 ----2 2 1 0 1 -1 --4 3 --2 -1 4 ---

U-test), and the number of sites in each dataset where the difference was significantly different are listed (treating 1P post wet-season

Table 4.2. The proportion of rare and common taxa having a trait where there was a significant difference (p < 0.05, Mann-Whitney

Figure 4.1. Total rainfall for each wet season (October 1-June 1, 1984-2003) for Hunting and Knoxville Creek. Rainfall average for the study duration are indicated by the solid horizontal bar (587 ± 224 mm, average ± SD).

1200

1000

Rainfall (mm)

800

600

400

200

0 1985

1990

1995

2000

Wet season

217

Figure 4.2. Percentage of taxa (identified, gray; unresolved, black) versus the number of years of occurrence for: a-d) Hunting Creek (1P, 2P, 2D post wet-season, 1P post dryseason); e) Knoxville Creek (1D); and f) all post wet-season Hunting Creek (All HC) sites combined.

35

a) 1P (post wet-season)

b) 2P

% of total taxa

30

identifiable unresolved

25 20 15 10 5 0 35

c) 2D

d) 1P (post dry-season)

% of total taxa

30 25 20 15 10 5 0 35

e) 1D

f) all HC

% of total taxa

30 25 20 15 10 5 0 0

5

10

15

20

Number of years of occurrence

0

5

10

15

20

Number of years of occurrence

218

Figure 4.3. Percentage of taxa (identified, gray; unresolved, black) versus the number of years of occurrence for: a) Big Sulphur Creek (post wet-season); and b-f) Blodgett Forest streams (BC, DA, DE, MC, GC). a) BSC (post wet-season)

b) BC

35 30

identifiable unresolved

% of taxa

25 20 15 10 5 0 35

c) DA

d) DE

e) MC

f) GC

30

% of taxa

25 20 15 10 5 0 35 30

% of taxa

25 20 15 10 5 0 1

2

3

4

5

6

7

8

1

2

3

4

5

6

7

8

Number of years of occurrence Number of years of occurrence

219

Figure 4.4. Average log10 density (no./0.093 m2) of each taxon (identified, clear; unresolved, black) versus the number of years of occurrence for: a-d) Hunting Creek (1P, 2P, 2D post wet-season, 1P post dry-season); e) Knoxville Creek (1D); and f) all post wet-season Hunting Creek (All HC) sites combined. Regressions (p < 0.01) do not include unresolved taxa. identifiable unresolved

3.0

Log10 density

2.5

b) 2P R2 = 0.53

a) 1P (post wet-season) 2

R = 0.72

regression

2.0 1.5 1.0 0.5 0.0

3.0

Log10 density

2.5

d) 1P (post dry-season)

c) 2D R = 0.61

R2 = 0.50

e) 1D

f) All HC

R2 = 0.53

R2 = 0.60

2

2.0 1.5 1.0 0.5 0.0 3.0

Log10 density

2.5 2.0 1.5 1.0 0.5 0.0 0

5

10

15

20

Number of years of occurrence

0

5

10

15

20

Number of years of occurrence

220

Figure 4.5. Average log10 density (no./0.093 m2) of each taxon (identified, clear; unresolved, black) versus the number of years of occurrence for: a) Big Sulphur Creek (post wet-season); and b-f) Blodgett Forest streams (BC, DA, DE, MC, GC). Regressions (p < 0.01) do not include unresolved taxa.

Log10 density

3.0

a) BSC 2 2.5 R = 0.48

3.0

b) BC

2.5

R2 = 0.40

2.0

2.0

1.5

1.5

1.0

1.0

0.5

0.5

0.0

0.0

3.0

Log10 density

2.5

identifiable unresolved linear regression

d) DE

c) DA R = 0.53

R2 = 0.38

f) GC

e) MC

2

2.0 1.5 1.0 0.5 0.0 3.0

Log10 density

2.5

2

2

R = 0.51

R = 0.36

2.0 1.5 1.0 0.5 0.0 1

2

3

4

5

6

7

8

Number of years of occurrence

1

2

3

4

5

6

7

8

Number of years of occurrence

221

Chapter 5

The effects of riparian prescribed fire on aquatic and riparian communities in a mixedconifer forest watershed, Sierra Nevada, California, USA

222

Abstract Concerns about the effects of fire on ecologically sensitive habitats have limited the use of prescribed fire in riparian management. Using a beyond-BACI (before-aftercontrol-impact) experimental design, the effects of a 26-ha prescribed fire that burned upland and riparian areas of a 1st order watershed was examined in one burn and six control sites for up to seven years pre-fire and one-year post-fire. I monitored pre- and post-fire water chemistry, riparian vegetation, periphyton, large woody debris, sediment, and aquatic macroinvertebrates. The fire consumed 79% of the pre-fire fuel in the riparian zone, with 34% of total surface fuel consumed and 90% of the total ground fuel consumed. There were increases in some water chemistry parameters (SO4- , total phosphorous, Ca2+, and Mg2+), a decrease in periphyton biomass, and an increase in fine sediment composition post-fire. Each of these changes were short-term, and recovery was documented in less than one year post-fire. Aquatic macroinvertebrate density, diversity and composition did not change post-fire. Likewise, there were no changes in LWD volume, recruitment, or movement. These results indicate that this carefully planned prescribed fire in and near the riparian zone of a mixed-conifer watershed did not adversely affect stream habitat or biota.

Introduction Fire is one of the most important natural disturbances influencing the heterogeneity and diversity of terrestrial landscapes (e.g., Romme 1982, Agee 1993) and aquatic ecosystems (e.g., Resh et al. 1988). Fire can directly and indirectly affect aquatic and riparian communities at spatial scales ranging from microhabitats to entire watersheds,

223

and temporal scales ranging from days to decades (e.g., Minshall et al. 1989, Minshall 2003). Because of the role that wildfire plays in terrestrial and aquatic ecosystem dynamics, there is a growing impetus to restore fire to degraded forested landscapes through the use of prescribed burning (e.g., Agee 1993, Stephens 1998, Miller and Urban 2000). For example, in mixed-conifer forests in the Sierra Nevada mountains of California, low-intensity surface fires were frequent over the past millennium (e.g. Kilgore and Taylor 1979, Stephens and Collins 2004), and prescribed burning been successfully used to restore mixed-conifer terrestrial habitats (e.g., Keifer et al. 2000). Fire-history research suggests that fire may have played an important role in riparian zones (e.g., Everett et al. 2003, Skinner 2003). Although the importance of fire to riparian ecosystem dynamics in the Sierra Nevada mountain ranges is not clear (e.g., Russell and McBride 2001), recent research indicates that fire usually occurred less frequently in California and the western U.S. riparian zones than in adjacent upland areas (Arno 1996, Russell and McBride 2001, Everett et al. 2003, Skinner 2003). However, moisture regimes affect the frequency of riparian fires. For example, in dry climates, such as the inland forests of Oregon, fires occurred with similar frequency in both riparian and adjacent upland areas (Oslon 2000). Similarly, on north facing aspects (with higher moisture and cooler temperatures), the fire return interval is more similar between upland and riparian forests than on west facing slopes (Everett et al. 2003). However, the use of prescribed fire in riparian zones (e.g., Kattelmann and Embury 1996) poses a complex management problem. Current management practices avoid the use of prescribed burning near aquatic ecosystems (e.g. DeBano and Neary 1996, Erman 1996) to prevent potentially negative impacts (e.g., increased erosion or altered

224

hydrographs) on physical habitat and aquatic biota that have been found to occur after wildfires (e.g., Roby and Azuma 1995, Rinne 1996, Minshall et al. 1995, 1997, 2001a,b,c). Prescribed fires may have less severe impacts on aquatic ecosystems than wildfires because of their low to moderate intensity, which may cause little mortality to mature trees (e.g., Gresswell 1999). However, few studies have been conducted on the effects of prescribed fire on riparian and aquatic communities in the western U.S. (cf. Chan 1998) and the applicability of research conducted in other ecosystems (e.g., Elliott et al. 1999, Townsend and Douglas 2000, Britton 1990, 1991a,b) to fire management and potential responses to fire in western conifer forests is not clear. For example, Britton (1991b) found no effect of prescribed fire on macroinvertebrate communities in South Africa, whereas Chan (1998) found decreases in the diversity and abundance of aquatic macroinvertebrates post-fire in mixed-conifer forests of Sequoia National Park, CA, USA, possibly caused by an increase in fine sediment deposition. In this study, a prescribed fire treatment was implemented that was consistent with protocols of actual management burns (i.e., size of the fire, burning conditions, and burn objectives). The exception was that it was also allowed to burn into and was actively ignited in the riparian zone. The objectives of this study were to determine: 1) the immediate and direct effects of the fire on riparian plant communities; water chemistry and physical water quality; 2) the delayed (but short-term, ≤ 1 year) impacts on the physical stream habitat (sediment composition, channel morphology, hydrology, large woody debris); 3) the effects of habitat changes on periphyton and aquatic

225

macroinvertebrates; and 4) the time to recovery, if possible, for both the physical habitat and periphyton and aquatic macroinvertebrates.

Methods Study site Blodgett Forest Research Station (BFRS) is located in El Dorado Co., CA, USA (38 55’N, 120 40’W). The streams in this forest research station drain watersheds of mixedconifer forests with well-documented land use histories (e.g., timber harvest, seasonal cattle grazing), and all of the streams have well-protected buffer zones (at least 50 m). The average annual precipitation (1961-2003) is 155.6 ± 53.8 cm (mean ± SD), 15-20 % of which occurs as snow. Prior to the prescribed burn, the streams at BFRS can be considered unimpaired because of the well-managed watershed and intact buffer zones. The six study sites were located in four 1st order streams (two sites separated by 400 m in Bacon, B1 and B2, and one site in each of Dark Canyon, D1, Deep Canyon, D2, and Mutton Creeks, M1), and one 2nd order stream (Gaddis Creek, G2) (Table 5.1, Fig. 5.1): The last fire occurred in this watershed in 1905, although the historical median fire return interval (FRI) is 5 years (mean FRI is 6.8 years), based on a 15 ha sample (from 1649 to 1921, Stephens and Collins 2004). Tree species in the mixed-conifer forest of BFRS include sugar pine (Pinus lambertiana Dougl.), ponderosa pine (Pinus ponderosa P.& C. Lawson), white fir (Abies concolor (Gord. and Glend.) Lindl. ex Hildebr), incense-cedar (Calocedrus decurrens (Torr.) Florin), Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco), California black oak (Quercus kelloggii Newb.), tan oak (Lithocarpus densiflorus (Hook. & Arn.) Rehder),

226

bush chinquapin (Chrysolepis sempervirens (Kell.) Hjelmg.), and Pacific madrone (Arbutus menziezii Pursh). The riparian forests also include white alder (Alnus rhombifolia Nutt.), Pacific yew (Taxus brevifolia Nutt.), Pacific dogwood (Cornus nuttallii Audubon ex Torr. & Gray), and willows (Salix L. sp.). All of the riparian areas studied are dominated by incense-cedar (Table 1). For example, at D1 incense-cedar makes up 52.1% of all trees and 26.0% of the basal area. This differs from the upland vegetation in the watershed, where white fir is dominant (Stephens and Collins 2004), making up 32% of the basal area (incense cedar is 17% of the upland basal area). The common understory plants in the riparian areas include western azalea (Rhododendron occidentalis (Torr. & Gray ex Torr.) Gray), twin flower (Linnaea borealis L.), sierra-laurel (Leucothoe davisiae Torr. ex Gray), western bracken fern (Pteridium aquilinum (L.) Kuhn), star flower (Trientalis latifolia Hook), false Solomon’s seal (Smilacina racemosa (L.) Desf.), western prince’s pine (Chimaphila umbellata (L.) W. Bart), blue elderberry (Sambucus nigra L. ssp. cerulea (Raf.) R. Bolli), smooth yellow violet (Viola glabella Nutt.), and drops of gold (Disporum hookeri (Torr.) Nichols. var. trachyandrum (Torr.) Q. Jones).

Study design I used a Beyond BACI (Before-After-Control-Impact) experimental design (Underwood 1994) to examine the effects of a prescribed burn on stream structure at BFRS. Beyond BACI experimental designs are asymmetrical, with multiple controls and one or more impacted sites that are monitored before and after the impact. Thus, I sampled before and after treatment at one burned and five control sites, all located within

227

four watersheds (Fig. 5.1). Changes in (1) water chemistry, (2) periphyton biomass, (3) riparian forest communities, (4) large woody debris, (5) hydrology, (6) channel morphology and sediment, and (7) macroinvertebrate communities were documented in both control and impacted streams (1st to 2nd order) (Table 5.3).

Prescribed fire treatment On October 21-23, 2002, a prescribed fire was burned in a 26 ha plot within the Dark Canyon Creek watershed (Fig. 5.1). Ignition patterns were manipulated to control fireline intensity (Martin and Dell 1978) in the riparian areas using strip-headfires. Average flame lengths were estimated visually during the prescribed fires. Fireline intensity, the heat produced during by flaming combustion (kW m-1), was estimated using the Byram fireline intensity equation and average flame lengths (Byram 1959). No attempt was made to exclude the prescribed fires from riparian areas or the stream, and there was active ignition in the riparian zone.

Data collection Riparian vegetation I surveyed permanent riparian vegetation plots to determine the amount of fuel consumption, and the direct (e.g., mortality) and indirect (e.g., reduced regeneration) effects of the prescribed fire on riparian vegetation. Permanent riparian vegetation sites were surveyed pre-fire (1997, 2001) in both the burned (n = 5 sites) and control (n = 6) watersheds. Specifically, there were five sites established on Dark Canyon Creek, and two sites established on each of Mutton, Bacon, and Gaddis Creeks (total of six sites in control streams). Each site was re-surveyed in summer 2003 (post-burn). At each site,

228

one large plot (400 m2) and two nested sub-plots (40 m2) were established (therefore, five large plots and 10 sub-plots were surveyed in both the burned stream and the control streams). Within each large (400 m2) plot, trees with a diameter at breast height (DBH) > 11.5 cm were identified and measured (DBH, height, and height to live-crown base). In both large and nested sub-plots, the relative cover of shrubs and herbaceous plants was determined (visual estimation of percent cover for each plant species). Canopy cover was measured in all plots using a sighting tube on a set of grid points. Regeneration was determined in subplots by counting the number of small trees in 14 size categories, ranging from < 0.3 m to 6-7.5 m. Fuel characteristics (duff and litter depth, distribution of 1, 10, 100, and 1000-hr fuel classes) were measured according to the procedures in Brown (1974) in each large plot (2 transects per plot) at each sampling date (therefore, n = 10 transects in burned and n =10 for control streams per sample date). In addition, fuel was surveyed 2 weeks post-fire (Nov 11, 2002) at a sub-sample of the plots in only the burned watershed (n = 5 transects). Fuel consumption was calculated by subtracting post-burn fuel loads from pre-burn fuel loads. We used regression equations for Sierra Nevada forests (from van Wagtendonk et al. 1998) to calculate duff and litter fuel loads, and we modified Brown’s (1974) equations with parameters from van Wagtendonk et al. (1996) to calculate surface fuel loads. Coefficients required to calculate surface and ground fuel loads were arithmetically weighted by basal area fraction to produce accurate and precise estimates of fuel loads (Stephens 2001). Summary statistics and the coefficient of variation (CV = standard deviation/mean) were calculated for the percent cover of surface vegetation for each sample period,

229

including total % cover (% of sampled area covered by vegetation), taxa richness (number of taxa), and Simpson’s diversity (D).

Large woody debris (LWD) To determine whether prescribed fires result in an increase in LWD recruitment and movement (e.g., Young 1994), I used standard methods for the inventory and monitoring of instream LWD (e.g., Lienkamper and Swanson, 1987; Platts et al., 1987). Pre-burn surveys of LWD were conducted in summer 2001 and 2002. During each initial survey, 50 pieces of LWD were measured by hand with a meter stick and tape and mapped at each site. LWD was defined by having a length ≥ 1 m or a diameter ≥ 10 cm. The orientation of the wood in relation to the channel (oblique, parallel, or perpendicular) and the position (driftwood, bridge over stream or ramp along bank) was documented, in addition to the longitudinal position in the stream (distance from the start point). Wood pieces were tagged with numbered aluminum markers. Upon subsequent survey in 2003, I documented the movement of previously tagged pieces of wood, and I measured and tagged new pieces of LWD. The volume of each piece of wood was calculated using the following equation from Lienkaemper and Swanson (1987): Volume =

[π (D

2 1

)]

+ D22 L 8

Where D1 and D2 are the end diameters (m) and L is the length (m) of a given piece of wood.

230

Hydrology To determine whether prescribed fire results in an increase in runoff and a change in precipitation-runoff relationships, data from stream gauges and weather stations (operated by BFRS) were used to construct precipitation-runoff relationships for each stream before and after the fire disturbance. Stream discharge, temperature, and precipitation were measured four times per hour at station-operated stream gauges and weather stations, respectively. These gauges have collected data continuously since 1995 for Bacon and Dark Canyon Creeks, and since 1997 for Gaddis Creek.

Sediment and substrate characterization To detect changes in fine sediment deposition as a result of the fire, I sampled the substrate surface layer using the random pebble count method (Wolman 1954, Kondolf and Li 1992) in June and October 2001-2003. The intermediate diameters of 100 randomly selected pebbles (first pebble touched at boot tip with eyes averted) were measured along a 10 m transect in riffle/run habitat at each site (except for sandybottomed Bacon Creek sites, where all substrate is < 4mm in diameter). Additionally, I determined the amount of fine sediment in pools by calculating the residual volume of the pool comprised of fine sediment, V* (Hilton and Lisle 1993). Fine sediment in pools was measured in each pool within 30m adjacent to a riparian vegetation plot.

Channel morphology To determine whether the prescribed fire resulted in changes in channel morphology, I surveyed channel cross-sections following the methods in Platts et al. (1987) in June

231

and October 2001-2003, using a telescoping level and rod. In addition, I surveyed the longitudinal profile (200 m) and three cross-sections of Dark Canyon Creek October 2001 (pre-burn) and October 2003 (post-burn) using a telescopic level and a survey rod.

Water chemistry To determine if ash-deposition or runoff in burned areas resulted in changes in water chemistry, water was sampled at each site monthly from June 2001 to October 2003, except immediately after the fire, when samples were taken every 24 hours for one week post-fire. I measured in-stream temperature, pH, dissolved oxygen, and conductivity using portable meters in the field; Nitrate (NO3-), ammonia (NH4+), total nitrogen (TKN), total phosphorous (tot P), soluble phosphorous (sol P), calcium (Ca2+), magnesium (Mg2+), potassium (K+), and sulfate (SO4-) concentrations were determined from water samples which were stored on ice (4°C), and processed in accordance with APHA standard methods for the analysis of water (Clesceri et al. 1998) by the University of California Division of Agriculture and Natural Resources Analytical Laboratory in Davis, CA.

Periphyton I sampled periphyton biomass to determine if an increase in nutrients, reduced shading and/or increases in fine sediment deposition caused by the prescribed fire affected periphyton. Standing crop of benthic periphyton was estimated by determining the ash-free dry mass (AFDM). I began sampling periphyton in September 2001, and continued to take samples approximately monthly through October 2003.

232

Periphyton was collected from colonized artificial substrates (i.e., clay tiles), because only two sites (G2, D2) have large enough mineral substrate to sample periphyton. Clay tiles have been shown to be indistinguishable from natural substrate in terms of periphyton biomass and community composition (Lamberti and Resh 1985). A total of 10 tiles (each 161 cm2) were placed at each site, stratified by habitat type (pool, riffle/run) and by light regime (shaded and open areas). At each site and sampling date, two samples were taken from five tiles (10 samples per site) approximately monthly, by removing periphyton with a razor and toothbrush from an 8 cm2 area. The scrubbed sample was rinsed with water into a 50-mL vial, stored on ice (4°C) and filtered within 24 hours of collection using Whatman® Glass Microfibre Filters (GF/C). The filtered samples were dried at 60°C for 24 hours, placed in a dessicator for 24 hours, weighed (initial dry mass), combusted in a muffle furnace at 550°C for 2 hours, and then reweighed (Clesceri et al. 1998). The difference between the ash residue and the initial dry mass provides the AFDM (Steinman and Lamberti 1996). Biomass data often contain many small decimal fractions, and a log (x + 1) transformation would mask differences between small values. Thus, I used a [log(x + d) – c] transformation, where c = an order of magnitude constant = integer value of: log(smallest non-zero value in data), and d = log-1(c), resulting in a transformation equation of: log(x + 0.01) – (-2). This transformation tends to preserve the original order of magnitudes in the data and results in values of zero when the initial value was zero (McCune et al. 2002).

233

Macroinvertebrates To determine what indirect effects prescribed fire has on aquatic macroinvertebrates at the population and community level, benthic samples were collected on October 1-3 of each year. At each site, five D-frame samples were collected once per year in riffle and run habitats within a 50 m reach at each site (each sample separated by 10 m). For each sample, a 1m2 area of the streambed was disturbed, including stream banks, by kicking the substrate, dragging a D-frame net, and scouring gravel by hand. A D-frame net (1 mm mesh) was then swept along the disturbed area to collect dislodged invertebrates. Macroinvertebrates were sorted without sub-sampling and identified to genus, family (for some Diptera), or order (for some non-insect taxa). Samples were taken in the same location during base-flow conditions every October 1 (+/- 3 days) at all six sites from 1995-1998 (del Rosario 2000) and 2000-2003. Additionally, using the same collecting methods, 5 samples were taken from D1 2 days post-fire (Oct. 25), and 3 samples from each from D1 and M1 were taken 10 and 19 days post-fire (November 2, and 11, 2002). Samples from Nov. 11 (19 days post-fire) were taken 12 hours after a 4-day storm which resulted in 18.7 cm of rain, and an increase in discharge of 0.038 m3/s (from 0.007 m3/s to a peak of 0.045 m3/s, discharge at time of sampling was 0.01 m3/s). Benthic macroinvertebrate samples were preserved in 90% ethanol. Each sample was sorted entirely and all invertebrates were identified to genus (except some Diptera, such as Chironomidae, which were identified to the family level, and non-insects such as Oligochaeta, which were identified to order).

234

Summary statistics were calculated using raw abundance data for each site and season, including density (numbers/m2), taxa richness (number of taxa), and Shannon diversity (H’) based on log10 (x + 1) transformed data. The coefficient of variation of each statistic was also calculated (CV = standard deviation/mean). Comparisons between before/after sampling periods and between control/prescribed fire sites were made graphically.

Data analysis Because this study is based on a beyond-BACI experimental design, an asymmetrical ANOVA was used to analyze the effects of prescribed burning on periphyton biomass, riparian vegetation measures, and sediment composition changes (Underwood 1994). This type of analysis accounts for multiple sampling events at the same sites over time, avoiding temporal pseudo-replication. The following parameters were included in the ANOVA models: before/after, control/fire, site(control/fire), sample date(before/after), before/after*control/fire, before/after*site(control/fire), and control/impact*sample date(before/after). Other measured parameters (LWD, hydrology, and channel morphology) were examined graphically. To examine changes in macroinvertebrate community structure over time and among sites, I used two different analytical techniques. First, I examined community composition between sampling periods (before/after) and among sites (control/impact) using NMS ordinations (Kruskal 1964) based on the sample-year x taxa abundance matrix using PC-ORD 4.27 (McCune and Mefford 1999). NMS was chosen as the ordination method because it preserves distances in ordination space without distortion,

235

and makes no a priori assumptions about the data responses (Legendre and Legendre 1998). In addition, NMS is flexible with respect to the distance measure and format of the data (McCune and Mefford 1999). The following NMS analyses were performed: (1) NMS of community composition by site-year for all six sites together; (2) NMS of community composition by site-year for only M1 (control) and D1 (prescribed fire) and all years. M1 and D1 are adjacent watersheds, with similar vegetation, morphology, taxonomic composition, and sediment characteristics. Each NMS was performed using the quantitative version of Sørensen’s distance measure (which is equivalent to Bray-Curtis distance). An initial ordination was conducted first with 2000 runs of real data, and a Monte-Carlo test based on 900 runs of randomised data and randomised starting coordinates. Significance was assessed by conducting Monte-Carlo tests using 500 runs of randomised data and comparing the stress on each axis in real versus randomised data. A p-value was then calculated as follows: p = (1 + n)/(1 + N), where n = number of randomised runs with final stress ≤ the observed minimum stress, and N is the number of randomised runs. In this preliminary ordination, the appropriate dimensionality was assessed by examining the decrease in stress with increasing dimensionality. I used the starting coordinates from this preliminary ordination to begin a subsequent ordination using 500 runs of real data, with no step-down in dimensionality, and no Monte-Carlo tests. This final ordination helps to ensure that the ordination is based on global minima and not local minima (McCune et al. 2002).

236

Results Prescribed fire behavior and fuels inventory Strip-headfires fires were ignited to produce flame lengths of approximately 0.75-2 m; average flame lengths were approximately 1.5 m. Fireline intensity varied from 1381165 kW m-1 in the riparian area (average was approximately 625 kW m-1). Higher intensity fire (as measured by flame lengths) occurred in areas of high localized fuel loads, where slopes were steeper, and fuels were drier. Fuel consumption was similar to other upland fall prescribed burns in the Sierra Nevada (Stephens et al. 2004, Stephens and Finney 2002). Pre-fire fuel loads in the riparian area of Dark Canyon Creek were high (13.79 kg/m2, 137.9 Mg/ha). Ground fuels, which include only the duff and litter layers, contributed 57% of the total load. Large surface fuels (large woody debris > 7.6 cm diameter) accounted for 33% of the total fuel load. Surface and ground fuel consumption ranged from 6 to 100%, depending on the fuel size, condition, and location within the riparian zone (Table 5.4). Wind and rain caused branches, litter, and several trees to fall in the riparian zone within one year of the fire, which increased the fuel load (Table 5.4).

Riparian vegetation Because the intensity of the fire and the burn patterns were patchy along the stream, the effects on riparian vegetation were heterogeneous. Litter and woody debris covered an average of 43.4% (± 25.1, standard deviation) of the area surveyed prior to fire, and was reduced to 20.83% (± 25.4) after the fire. Within the measured plots, the amount of bare ground increased five fold (pre: 4.9 ± 8.1%, post-fire: 29.4 ± 21.9%). Vegetation

237

cover decreased from 63.3 ± 32.7% prior to the fire to 15.3 ± 5.2 % after the fire, with no corresponding decrease in the control plots (Table 5.5). Riparian canopy cover did not change (pre-fire: 83.1 ± 9.5, post-fire: 80.8 ± 18.6, controls: 78.8 ± 14.0, average ± standard deviation). Ground cover taxa richness decreased in the burn plots (pre-fire vs. post-fire, p = 0.015, paired t-test), although richness is not significantly different from the control plots (p = 0.45) (Table 5.5). Similarly, Shannon-Weiner (H’) diversity in the burned areas decreased in the post-fire plots relative to pre-fire measurements (p = 0.016, paired t-test), but post-fire diversity is not significantly different from the control plots (p = 0.50). Regeneration decreased in the burn units compared to pre-fire values. For example, the number of small trees (< 7.6 m) was 10.0 ± 12.6 trees/m2 (mean ± standard deviation) one-year after the fire (2003), compared to 55.5 ± 79.1 in 2001, and 57.7 ± 85.8 in 1997. This decrease in regeneration was statistically significant based on an asymmetrical ANOVA (control/fire * before/after interaction, F = 6.01, p = 0.018). Reduced regeneration counts were caused largely by a reduction in trees < 0.3 m height (CF*BA, F = 5.87, p = 0.019), based on separate ANOVAs for each size class, where all other size classes (0.3 – 7.6 m) did not change (CF*BA, p > 0.15). The fire resulted in the mortality of only 8 trees (4.4%), which ranged in size from 11.7-40.4 cm DBH (22.6 ± 11.8 cm, mean ± SD). Ten snags fell over in the riparian area after being partially burnt, while 4 snags were fully consumed during the fire over 0.16 ha. For example, 49.4% of tagged trees exhibited scorch of 0.10 m or higher, while 47.6% were not scorched. The average scorch height on tagged trees is 2.47 ± 2.25 m

238

(mean ± standard deviation), with a maximum scorch height of 7.6 m. Snags and live trees were scorched in approximately equal proportions (45.6 and 54.4%, respectively).

Large woody debris The prescribed fire did not change the amount or movement of LWD in Dark Canyon Creek relative to control streams (Table 5.6). Although the characteristics of LWD (average diameter, length, function, and total volume) was similar across all streams (Table 5.7), the length of stream containing 50 pieces of LWD varied among streams. M1 has 0.85 pieces of LWD per meter, while lower G2 has only 0.47 pieces/m. D1 has 0.74 pieces/m. The form and orientation of the LWD did not differ greatly between the sites. On average, 40.5% of the 390 pieces of tagged wood are ramps, 34.9% are drift, and 24.6% are bridges. The orientation of the wood was influenced by its form, with most ramps being oblique to the stream channel (38%), most bridges being parallel to the stream channel (46.9%), and most drift being oblique (52.9%). There were few new pieces of LWD in subsequent surveys (2002, 2003), with the exception of M1, where several rotten and decaying snags fell over the stream in 2002. Slightly different results were found when LWD was measured in two ways. First, an intensive survey was conducted to tag, measure, and map 50 pieces of LWD at each site twice before and once after the fire at all sites. Based on this survey, there was no increase in LWD amount or movement (Table 5.6). However, I also counted the number of pieces of LWD in the stream channel adjacent to each riparian vegetation plot. In two plots along Dark Canyon Creek (#3 and #6), there were 7 snags that fell over the channel

239

and 12 trees or snags that fell in the riparian zone but have not yet moved into or near the stream channel.

Hydrology The period between 1995-1998 was much wetter (average precipitation = 219.3 ± 28.0 cm, which is > 1 standard deviation higher than the 41 year average of 155.6 ± 53.8 cm) than the period between 2000-2003 (average precipitation was below the long-term average 115.2 ± 44.7 cm). The 1995-98 wet period produced larger winter floods and higher peaks during spring snowmelt, which regularly exceeded 0.14 m3/s, as compared to the relative dry period (2000-2003) where no floods exceeded 0.14 m3/s (Fig. 5.2). In addition, the rain-on-snow events during the 1997-98 El Niño Southern Oscillation (ENSO) produced floods that exceeded 1.68 m3/s, and throughout this winter, floods regularly exceeded 0.56 m3/s (Fig. 5.2). Although the ENSO resulted in dramatic flooding, the total amount of precipitation was not as great as during the 1994-95 water year (214.8 in 1997-98 vs. 256.2 cm 1994-95).

Sediment composition Sediment composition, as measured by pebble counts, did not significantly change post-fire in Dark Canyon Creek. Fine sediment (≤ 4mm) made up 47-65% of the substrate in riffles in Dark Canyon pre-fire, and 48-61% post-fire (Table 5.8, Fig. 5.3). However, in the adjacent control stream, Mutton Creek, fine sediment also varied and was 35-57% of the riffle substrate. In contrast, Deep Canyon and Gaddis Creeks have much coarser substrate, with only 0-24% of the substrate composed of fine sediment

240

(Table 5.8), and with a higher percentage of large (> 32 mm) cobbles. The fraction of pool volume filled with fine sediment, as measured by V*, did not significantly change post-fire in Dark Canyon Creek, as shown by an asymmetrical ANOVA, p = 0.10, (Table 5.9).

Channel morphology The channel surveys revealed little change post-fire (2001 v. 2003), however the surveys are not conclusive because there were no surveys conducted in 2002 or in control streams. The measured stream reach gradient at the study site on Dark Canyon Creek did not change between the periods and was measured as 2.5% in 2001 (Minear and Minick, unpublished data) and 3.1% in 2003. For the first 60 m of the study reach (80 m total), there was little difference between the 2001 and 2003 longitudinal profiles (Fig. 5.4). Although from 0 to 60 m the average change was –0.05 m between 2001 and 2003, between 60-80 m erosion occurred, with an average elevation difference (2001-2003) of 0.67 m (Fig. 5.4). However, with the exception of changes in the distribution of some leaf-litter and small instream debris, there is little field evidence to support such a large observed change in longitudinal profiles between the surveys. Thus, operator inconsistency is most likely responsible for the observed changes in the longitudinal profile. The three cross-sections revealed differing patterns of fire effects on the stream channel. Cross-section 1, at 5.7 m on the longitudinal profile, revealed no major change in instream elevation, even though the right bank elevation decreased after the fire (Fig. 5.5a). The second cross-section (at 47 m) also did not show any change in channel

241

elevation after the fire, except that the channel appears to have narrowed by 0.2 m on the right bank (Fig. 5.5b). At cross-section 3 (at 69.5 m), however, aggradation occurred between 2001 and 2003. The instream portion of the channel (5.2 – 6.0 m) increased in elevation by 0.11 m (Fig. 5.5c). There is potentially large error (potential operator inconsistency) associated with these measurements. However, the same benchmarks were used for the three cross-section measurements and there is field evidence to support the changes in elevation observed in the cross-sections. For example, high fuel consumption reduced litter and large woody debris cover in the riparian zone, which could be the cause of these changes in elevation.

Water chemistry There is no evidence of post-fire change in Dark Canyon for dissolved oxygen, conductivity or pH relative to control sites. Dissolved oxygen was generally > 7 mg/L for all sample periods in all streams, except for the August 2002 sample period (Fig. 5.6a). Conductivity fluctuated similarly between controls streams and Dark Canyon (Fig. 5.6b). Post-fire nitrate did not differ in Dark Canyon relative to the control streams (Fig. 5.6c). In addition, nitrate levels were much higher in Mutton Creek than in any other stream for the duration of the study (average during the study period = 0.13 ± 0.12 mg L1

, compared to an average below the detection limit < 0.05 ± 0.03 mg L-1 for the other

streams). Before the prescribed fire, ammonium levels were mostly below the detection limit of 0.05 mg L-1. However, post-fire, all streams (control and burn) exhibited a periodic increase in ammonia (Fig. 5.6d) that was reduced to below detection limits by October 2003.

242

Total Kjedahl Nitrogen (TKN) varied seasonally before the prescribed fire in all streams, ranging from 0.5 to 3 mg/L every three months, approximately. In general, all of the streams had similar TKN concentrations before the fire at each sampling period. One-day post-fire, TKN increased 1.35 mg/L (from 0.70 mg/L on 10/3/02 to 2.05 mg/L on 10/23/02). This increase was not seen in the control streams. Post-fire TKN concentrations dropped control stream levels within 19 days post-fire (Fig. 5.6e). SO4- remained at or below the detection limit (0.10 mg/L) prior to the fire in control and burn streams. SO4- remained at or below 0.10 mg/L in Dark Canyon until the end of the first post-fire rainstorm (Nov 11, 2002), when SO4- spiked to 0.30 ± 0.12 mg/L. During this sampling period, SO4- in the control streams did not increase. SO4concentrations returned to ≤ 0.10 mg/L within one month after the post-fire flood (Dec 12, 2002, Fig. 5.6f). Soluble phosphorous (sol P) concentrations were low in all of the study streams, usually remaining at or below the detection limit of 0.05 mg/L. However, pre-fire sol P increased to 0.09 mg/L on Oct 3, 2002, but dropped back to 0.05 mg/L by Oct 17, 2002. After the post-fire rain event (Nov 11, 2002), sol P increased to 0.08 mg/L, and returned to below the detection limit within 2 months (Fig. 5.6g). Similar to sol P, total phosphorous (total P) concentrations were generally low in all of the study streams (≤ 0.10 mg/L, the detection limit). However, on May 20, 2002, total P increased in all of the streams to 0.20-0.30 mg/L (Fig. 5.3h). One week post-fire, total P increased in Dark Canyon to 0.20 mg/L, while the control streams remained at or below 0.10 mg/L. Total P in Dark Canyon returned to pre-fire levels within 2 weeks (Nov 11), despite the post-fire flood.

243

Soluble calcium (Ca2+) and magnesium (Mg2+) concentrations were low prior to the fire (equal to the detection limit of 0.10 μeq./L), and increased after the spring snowmelt. Soluble potassium (K+) concentrations, however, were similar between controls and the burn stream for the duration of the study (Fig. 5.6k), and exhibited no change post-fire relative to the controls. After the spring snowmelt in June 2003, which was 7 months after the fire, Ca2+ and Mg2+ increased in Dark Canyon to 0.3 μeq./L and 1.6 μeq./L, respectively, while control streams maintained concentrations ≤ 0.10 μeq./L (Fig. 5.6i,j).

Periphyton The prescribed fire had a short-term impact on periphyton, by decreasing biomass within two months post-fire, and recovering within 1 year post-fire. Periphyton biomass in Dark Canyon Creek prior to the fire was generally equal to or higher than biomass in the control streams during most sample periods (except March 2001, where Dark Canyon Creek had lower periphyton biomass than control streams) (Fig. 5.7). Three days postfire, periphyton biomass was not substantially different in the burn stream relative to samples taken 4 days before the fire, (3.26 ± 3.55 mg/cm2 cf. 1.20 ± 0.57 mg/cm2, average ± standard deviation), even though one sample had a very high abundance (9.56 mg/cm2). By 1 week post-fire, periphyton biomass from the burn stream was equal to biomass in control streams (1.50 ± 1.16 mg/cm2 cf. 1.03 ± 0.95 mg/cm2). Beginning with the December 12, 2002 sample date (7 weeks post-fire) and continuing until June 3, 2003, periphyton biomass was lower in the burn stream than in the control streams. By the October 1, 2003 sample date, periphyton biomass in the burn stream was higher than in the control streams, which is consistent with pre-fire patterns (Fig. 5.7).

244

An asymmetrical ANOVA revealed a significant Before/After x Fire/Control interaction (F = 25.64, p 1 SD above the long-term average, resulting in peak flows of 65 m3/s on date and of > 20 m3/s on several dates). In particular, the adjacent streams, M1 (control) and D1 (fire) are the most similar pre- and post-fire (again, with the exception of 1998). Relative to the controls and pre-fire data, there was no immediate (10 and 19 days PF) or delayed (1-year PF) effect of the fire on macroinvertebrate abundance (Fig. 5.8a), despite the large rainfall event 15 days PF. However, macroinvertebrate abundance exhibited high variability among the years sampled at each site (CV among years range from 0.50 at B2 to 0.71 at D1) (Appendix 5.1). In contrast, taxa richness (number of taxa/m2) was less variable among years (CV among years ranges from 0.13 to 0.27, Appendix 5.2), but also did not change post-fire relative to pre-fire data and controls (Fig.

245

5.8b). Shannon-Weiner diversity (H’) differed among years (CV among years ranges from 0.11 to 0.34, Appendix 5.3). In general, diversity was higher from 1995 to 2000, and then decreasing in 2001 and 2002, and rebounding in 2003, in all sites (Fig. 5.8c). Despite these changes among years, the pattern of change was similar among the control streams and the burned stream, D1. NMS ordination of all sites showed that there was no difference in community composition between before/after sampling periods or control/impact sites, based on a 3axis solution that represented 91.3% of the variation (Fig. 5.9). Axis 1 represented 36.5% of the variation, axis 2 represented 46.6%, and axis 3 represented 8.2%. Year of sampling was positively correlated with axis 1 (Pearson’s r = 0.64) and axis 3 (r = 0.33), and negatively correlated with axis 2 (r = -0.62). Site was negative correlated with axis 1 (r = -0.438), and with axis 2 (r = -0.43). A dummy variable of time/treatment, representing before/after and control/fire was not correlated with any axis (r < 0.1). The data points from post-fire sampling in D1 fall within the cloud of points from pre-fire and control samples (Fig. 5.9). A comparison of the prescribed burned D1and the adjacent control M1 using NMS showed that there was no difference in community composition after the fire based on a 2-axis solution representing 84.0% of the variance (Fig. 5.10a). Axis 1 represents 54.5% of the variation and axis 2 represents 29.6%. Year of sampling is negatively correlated with axis 1 (r = -0.80) and axis 2 (r = -0.68). A dummy variable representing before/after*control/fire was negatively correlated with axis 2 (r = -0.34), but not axis 1 (r < 0.30). A dummy variable representing before/after sampling only was negatively

246

correlated with both axis 1 (r = -0.34) and axis 2 (r = -0.43), and may represent interoperator error (RBdR collected 1995-98, LAB collected 2000-03). Community composition differed the most between the burned D1 and the adjacent control M1 in 1998 after the ENSO year (Sørensen distance, d = 0.347) and 19 days postfire (Nov 11, 2002, d = 0.32) (Fig. 5.10b). Despite these two events, the distance between D1 and M1 is relatively constant, ranging from 0.21 to 0.28 (0.24 ± 0.03, mean ± SD). Importantly, by one-year post-fire, D1 and M1 communities are the most similar in the entire study period (d = 0.21, Fig. 5.10a-b).

Discussion In this study, I have demonstrated that a moderate-intensity prescribed surface fire that was actively ignited in the riparian zone had minimal effects on Dark Canyon Creek, which drains a small mixed-conifer forest in the central Sierra Nevada during the first year post-fire. This study has many limitations and these results may not be applicable to other regions, prescribed burns, or even other watersheds within the mixed-conifer region of the Sierra Nevada. For example, although multiple control sites were used in this study, it was not possible to replicate the prescribed fire treatment because of logistical constraints. For example, after two watersheds were chosen for treatment, the planning process for the prescribed fires was initiated in early 2001 for implementation in October or November 2001. Permits were filed with the U.S. Forest Service in El Dorado County, and local and state Air Resources Boards. Although the permits were granted for conservative implementation of the burns (e.g., < 5 mph winds, night-time ignition), an early rain event increased fuel moistures and prevented ignition and spread of the fire.

247

Approximately 30 volunteers plus U.S. Forest Service and California Department of Forestry staff (totaling approximately 30 people) were on hand in both 2001 and 2002. After the failed implementation in 2001, the prescribed burn was conducted in late October 2002, however, as a result of time and weather constraints, the second burn could not be started prior to the fall rains. Nevertheless, evidence from both abiotic and biotic components suggest that there were few effects of the prescribed fire on Dark Canyon Creek, and the few parameters that were affected quickly recovered during the study period (e.g., periphyton). In contrast to the effects on the stream, there were marked effects of the fire on the riparian zone (e.g., scorch, fuel consumption). Here, I consider each line of evidence separately, in the context of direct and indirect effects of the fire, immediate and delayed impacts, and patterns described in the literature. Finally, I discuss the factors which may have contributed to the few effects that this prescribed fire had in the short to mid-term.

Prescribed fire in the riparian zone: fuels and vegetation The prescribed fire in the riparian zone was patchy in terms of intensity, consumption, and severity. The riparian zone of these streams is dominated by conifers, in particular, incense cedar, and characteristic riparian trees (e.g., white alder, Alnus rhombifolia) and understory vegetation (Mexican elderberry, Sambucus mexicana; mountain dogwood, Cornus nuttallii) that are restricted to areas immediately adjacent to the stream. The fire was most severe in those areas with large accumulations of conifer litter and debris, and the fire usually self-extinguished when it came in contact with moist soil and characteristic riparian vegetation. High soil and fuel moisture, and high relative

248

humidity can reduce fire intensity and retard fire spread in riparian zones (e.g., Dwire and Kauffman 2003). Surface vegetation was directly affected by the prescribed fire, reducing the amount of cover, taxa richness, and diversity. For example, ground cover in the riparian area was greatly reduced by the prescribed fire (63 to 15%). In an upland-only prescribed fire (prevented from entering the riparian zone), Lamb et al. (2003) found no effect of fire on riparian vegetation abundance or distribution. Similarly, Elliott et al. (1999) found that an upland prescribed fire in the Appalachian mountains had no effect on low slope (near riparian) overstory or understory vegetation cover or species composition. Little tree mortality could be attributed to the fire. Even though 49.4% of all tagged trees and snags were scorched by the prescribed fire, only 4.4% of all tagged trees died between the 2001 and 2003 surveys. The trees that were killed by the prescribed fire were generally small (22.2 ± 12.1 cm DBH, range = 11.4 to 40.4 cm) incense cedar (75% of newly dead trees), which were near areas of high litter accumulation. In comparison, Elliott et al. (1999) found no mortality in near-riparian overstory trees following a prescribed fire in the Appalachian mountains. In general, the riparian zone acted as a buffer between the moderate severity burned portions of the upland zone and the stream, as has been hypothesized by several authors (Lamb et al. 2003, Dwire and Kauffman 2003, Timoney et al. 1997). There are two main factors which may have contributed to this buffering capacity. First, the riparian zone experienced very low to moderate fire intensities, with many areas not burning at all, compared to the upland areas. Additionally, the relatively flat slopes of the majority of the riparian zone may have contributed to this buffering capacity (riparian zone slopes

249

range from 6.2 to 30.6%, mean = 19.9 ± 6.9%) by keeping erosion minimal, and increasing infiltration of precipitation (e.g., Townsend and Douglas 2000). In contrast, steep slopes (51 to 85%), and narrow riparian zones (usually < 2 m wide, L.Bêche, unpublished data) may have contributed to the negative effects of prescribed fires on streams in Sequoia National Park (southern Sierra Nevada) as described by Chan (1998). Several case studies have documented effects when prescribed fires were allowed to extend into the riparian zone, particularly in the national parks of the western U.S. (e.g., Chan 1998, Huntzinger 2003). However, I have found no studies on the effects of riparian prescribed fires on riparian vegetation. In contrast, however, there are published studies on the effects of upland prescribed fires on riparian vegetation (Lamb et al. 2003, Elliott et al. 1999) and on the effects of wildfires on riparian vegetation (Ellis 2001, Russell and McBride 2001).

Large woody debris Even though several snags fell over as a result of the fire, either in the riparian zone or over the stream channel (forming a bridge), there was no overall increase in the amount or movement of LWD in Dark Canyon Creek relative to the control streams. However, I did find that in one riparian plot, LWD increased one-year post-fire (from 13 to 18 pieces of LWD 20-50cm diameter). The discrepancy in these results is because the former result is based on surveys conducted 300 m downstream of the latter result, and the latter result is based on much less thorough surveys (LWD was only counted, it was not tagged, measured, or mapped). In comparison to my results, Chan (1998) found great increases in LWD in two 1st order streams following prescribed fire in Sequoia National

250

Park. There have been few studies on the effects of prescribed fire on instream LWD; however, it has been documented that prescribed fire often increases the number of standing snags and surface woody debris (e.g., Waltz et al. 2003), and thus may contribute to increased instream LWD in the long-term, as compared to unburned watersheds (e.g., Bragg 2000). In simulations of LWD recruitment, Bragg (2000) found that moderately-severe fire increased the amount of instream LWD over the long-term (300 + yr) in old-growth forested headwater streams. In contrast to prescribed burns, wildfires often have dramatic and immediate effects on large woody debris (e.g., Young 1994, Minshall et al. 1997). For example, severe wildfires can consume instream large woody debris, decrease LWD during uncharacteristically harsh post-fire floods, and can ultimately result in dramatic increases in instream LWD as a result of fire-felled trees (e.g., Minshall et al. 1997). LWD movement also increased in burned streams as a result of higher magnitude floods and decreased bank stability, which reduced the stability of LWD (Young 1994, Minshall et al. 1997). The effects of wildfire on LWD are often more pronounced in smaller streams (1st-3rd order) than in larger streams (4th + order) because LWD is a major structural element, forming pools and trapping fine sediment (Keller and Swanson 1979, but see Berg et al. 1998).

Sediment composition and channel morphology Benthic sediments in riffle areas became finer after post-fire snowmelt but returned to pre-fire values four months after the snowmelt (Table 5.8). Pools, however, did not experience a change in the amount of fine sediment. The streams at BFRS are sandy-

251

bottom streams with a high percentage of fine sediment in both riffles and pools (except Gaddis and Deep Canyon Creeks, which drain watersheds of different underlying geology). Thus, post-fire changes in the amount of fine sediment composition in riffles are a relatively small percentage increase (29%). In comparison, in the steep canyons of Sequoia National Park (mixed-conifer forest), fine sediment increased one year following fire (Chan 1998), suggesting that hillside erosion increased as a result of the fire. If surface soils, overstory trees, and ground vegetation are sufficiently altered as a result of a fire, the effects on hillside erosion and instream sediment composition, could be similar to, if not more exacerbated than, logging (e.g., Benda et al. 2003). For example, after severe wildfires, benthic sediment became finer and more embedded in burned streams as compared to unburned reference streams (Minshall et al. 2001c). Whereas severe wildfires often denude the entire watershed, leaving few living trees, this prescribed fire only removed surface vegetation from 70% of the total area burned, which only represents 20% of the total watershed area. As shown by Wondzell and King (2003), hillside erosion risk is much reduced as the amount of vegetation, litter, and debris cover increases, thereby stabilizing and protecting forest soils. Thus, the erosional capacity of the watershed is potentially buffered by the unburned areas in the watershed. Biswell and Schultz (1957) found that there was no increase in surface runoff or erosion following a prescribed fire in the coast ranges of California (ponderosa pine forest), and attributed this to the maintenance of partially decomposed duff and litter, which buffered the action of rain on fresh ash. In fact, they found significant increases in erosion only in areas where bare soil was exposed (as a result of fire lanes, or other activities), leading to the conclusion that the low intensity prescribed fire which left some litter and duff intact

252

had no effect on erosion rates. Likewise, the historically low to moderate severity fire regime that is typical of the mixed-conifer region of the Sierra Nevada mountains (e.g., Kilgore and Taylor 1979, Stephens and Collins 2004), may interact with the soil’s relatively low sensitivity to erosion (as a result of low to moderate slopes) to produce minimal effects of fire on geomorphic processes (Swanson 1981). Other studies have shown that fire severity may have a large effect on erosion rates, sediment yields, and time to recovery (e.g., Swanson 1981, DeBano et al. 1996, Wondzell and King 2003). For example, areas of low severity wildfire in eastern Oregon resulted in small increases in sediment yield which recovered within 3 years post-fire; whereas moderate to severe severity burn areas took much longer to recover (7 and 14 years, respectively) (DeBano et al. 1996). Considering that DeBano et al. (1996) results were based on wildfires where the entire watershed was affected, it is not surprising that the low to moderate intensity prescribed fire in this study resulted in minimal hillside erosion, geomorphic change, and sediment composition change. The channel surveys I conducted do not provide conclusive evidence of erosion or aggradation as a result of the prescribed fire because I did not survey pre-fire in 2002 to document change between 2001 and 2002. However, the pre-fire survey can be used as a baseline to make conservative estimates of the effects of the fire on channel morphology, especially since there were no floods larger than 0.06 m3/s in 2002. Because the majority of hillside erosion usually occurs during the first year post-fire (Agee 1993, DeBano et al. 1998, DeBano et al. 1996, Robichaud and Brown 1999), these results should be representative of the effect of the prescribed fire on channel morphology. Post-fire surveys indicated that there were areas where erosion (e.g., Fig. 5.4) and deposition (e.g.,

253

Fig. 5.5c) occurred, even if these changes are subtle in comparison to post-wildfire channel changes reported elsewhere (Roby 1989, Moody and Martin 2001, Wondzell and King 2003). Erosion appears to have occurred in areas where there were large accumulations of litter and branches in 2001, suggesting that the differences in elevation between the 2001 and 2003 longitudinal surveys (from 60-80 m, Fig. 5.4) could be partially explained by shifts in litter deposits. Deposition occurred in areas where dense streamside vegetation was reduced (e.g., Fig. 5.5c), however, deposition did not occur in many areas along the stream channel. Even though channel erosion and deposition was documented after the prescribed fire, the resultant change in channel morphology is negligible. The estimated deposition and erosion within the channel reach is very small when compared to changes occurring after wildfire (e.g., Moody and Martin 2001).

Water chemistry Ash deposition from the prescribed fire appeared to have a minimal impact on stream water chemistry. Only nitrate, TKN, and total P increased immediately post-fire (1 d and 1 week post-fire, respectively) relative to control streams, and at the end of the Nov 11 flood both had returned to pre-fire levels. Although the post-fire increase in nitrate, TKN, and total P was not seen in the control streams, the measured post-fire concentration was within the pre-fire range of concentrations (Fig. 5.6c,e,h). In general, the direct effects of fire (prescribed or wildfire) on streams (ash deposition and fireinduced temperature increases) are usually negligible (e.g., Minshall 2003, Minshall et al. 2001a), however, in the case of some severe wildfires, ash deposition and diffusion of smoke into the water can dramatically increase phosphorous and nitrogen (e.g, Spencer

254

and Hauer 1991). Similarly, Minshall et al. (1997) found high ammonia levels in burned streams immediately after the fire, which probably was the cause of fish mortality. Other chemical components either increased after the first major post-fire flood (19 d post-fire, SO4-), after the spring snowmelt (Ca2+, Mg2+), or did not exhibit any change relative to controls (K+, ammonium, sol P). The increase in SO4- after the first post-fire flood was not long-lasting, and returned to post-fire concentrations within two months. Similarly, Stephens et al. (2004) found that after a high consumption prescribed fire with moderate intensity, SO4- , Ca2+, and Mg2+ increased for up to 4 months post-fire in intermittent streams in the Lake Tahoe Basin (northern Sierra Nevada, CA, USA). In contrast, SO4- , nitrate, K+, Ca2+, increased post prescribed-fire in Sequoia National Park (southern Sierra Nevada, CA, USA) (Williams and Melack 1997), and slowly decreased for three years post-fire. Previous studies have found that hillside slopes, fuel consumption, and the pattern of post-fire precipitation will influence changes in stream water chemistry following prescribed fire. For example, Townsend and Douglas (2000) found that particularly in areas with gentle slopes, low to moderate intensity prescribed fires do not cause an increase in sediment export (and thus, many chemical components). Dark Canyon Creek has low to moderate hillside slopes (average 20%), thus, accelerated erosion and sediment and ash export would not be expected. In contrast, steep slopes may have contributed to the dramatic and long-lasting (3-5 yrs) increases in anions and cations following prescribed fire reported by Williams and Melack (1997). Payne (1999) also found that most major chemical constituents increased following prescribed fire in the Lake Tahoe basin, CA, USA. In contrast, both this study and Stephens et al. (2004)

255

report minimal effects of prescribed fire on water chemistry in areas of gentle hillside slopes (< 20%). Low-severity fires which result in partial fuel consumption of the duff and litter layer may be effective in maintaining high infiltration and preventing surface runoff (Robichaud 2000). Biswell and Schultz (1957) found that following a prescribed fire in the coast ranges, CA, USA, there was no indication of fire induced surface runoff or erosion. Similarly, Richter and Ralston (1982) found that there were no changes in chemical constituents of stream water following multiple prescribed fires in a southeastern, USA watershed. Post-fire precipitation patterns may influence the amount of ash that is transported to streams. Although there are no published accounts specifically examining this issue, anecdotal evidence suggests that large rainfall events may increase the amount of solutes in streams as compared to smaller rainfall or snowfall events. For example, Spencer et al. (2003) noted that the Red Bench fire occurred during a prolonged drought and produced little opportunity for water-borne transport of nutrients to the stream; thus any increases in nutrient concentrations that they observed must have been a result of ash and smoke. Unlike low to moderate-intensity prescribed fires, large wildfires often result in dramatic increases in stream solutes, which may last for years post-fire (e.g., Spencer et al. 2003, Hauer and Spencer 1998, Chorover et al. 1994, Tiedemann et al. 1979). In most of these studies, post-fire increases in nutrient concentrations which occur after rainfall or snowmelt have been attributed to overland flow or subsurface transport, except in cases where there were high amounts of ash deposition and/or smoke diffusion into the streams (e.g., Earl and Blinn 2003, Spencer and Hauer 1991, Roby and Azuma 1995). For

256

example, Hauer and Spencer (1998) found that nutrient concentrations increased dramatically immediately after the wildfire as a result of ash and smoke deposition, falling back to baseline conditions within 3 weeks, and then periodically increased over the next several years, particularly during spring run-off.

Periphyton Despite the obvious response of periphyton biomass to flooding, the prescribed fire resulted in a statistically significant decrease in periphyton biomass in Dark Canyon Creek relative to control streams starting 7 weeks post-fire and recovering by one-year post-fire (Fig. 5.7). Post-fire, periphyton biomass was lowest post-fire during the April sampling period, which coincides with the snowmelt, and flows that reached 0.07 m3/s (baseflow prior to the snowmelt was 0.01 m3/s). Similarly, a dip in pre-fire periphyton biomass occurred in March 2002, which was probably a result of several early snowmelts during that winter, each resulting in flows of approximately 0.06 m3/s. In fact, the March 6, 2002 day of sampling coincided with the 2002 peak flow event (0.063 m3/s). The decrease in periphyton could be a result of increases in fine sediment in riffle areas that was found after the spring snowmelt, but which recovered to pre-fire values within one-year post-fire (Table 5.8). Although there have been no published studies investigating the effects of prescribed fire on periphyton biomass, wildfire has resulted in decreases in periphyton biomass, which was attributed to increases in sediment (Robinson et al. 1994, Minshall et al. 1995). In contrast, Earl and Blinn (2003) found no change in periphyton biomass following wildfire; however, they only focused on the immediate impacts of fire as a result of smoke and ash. However, all three studies found

257

that wildfire resulted in changes in diatom taxonomic composition towards an increase in adnate (flat-lying) species. In this study, I did not examine changes in diatom taxonomic composition. However, composition changes were generally short-lived (Earl and Blinn 2003, Robinson et al. 1994) most likely because diatoms have rapid turnover and are generally resilient to short-term changes in the physical or chemical environment (e.g., Duncan and Blinn 1989).

Macroinvertebrates Prior to the first flood, there were no direct effects of the prescribed fire on aquatic macroinvertebrates in Dark Canyon Creek (Figs. 5.8-5.10), based on sampling 2 and 11 days post-fire. Generally, the direct effects of fire on aquatic macroinvertebrates are usually minimal or non-existent, as the heating and ash deposition that my be associated with some fires rarely exceeds the lethal limit for macroinvertebrates (Rinne 1996, Minshall et al. 1997, Minshall 2003). However, some severe fires have resulted in temperature or ash-induced mortality (e.g., Roby and Azuma 1995). In addition, the prescribed fire and post-fire flood had no effect on aquatic macroinvertebrate communities as measured by abundance, taxa richness, and diversity relative to unburned control streams (Fig. 5.8). For example, community measures were within the measured range for both Dark Canyon Creek and the control streams. The lack of response in the macroinvertebrate communities to the prescribed fire is probably because the fire resulted in no changes in hydrology and minimal, but not long-lasting, changes in sediment composition and channel morphology (Table 5.8, Figs. 5.4, 5.5). Similarly, low- to moderate-intensity prescribed fire has been shown in other studies to

258

have no or a subtle effect on aquatic macroinvertebrate communities (Britton 1991b, Chan 1998). For example, Britton (1991b) found no change in abundance, dominance, diversity, or functional feeding groups in response to a prescribed fire in a South African watershed dominated by fynbos vegetation (similar to southern California chaparral). The prescribed fire resulted in very little riparian vegetation mortality and thus did not appear to have affected sediment inputs (Britton 1991b). In contrast, Chan (1998) found that prescribed fires in Sequoia National Park resulted in decreases of aquatic macroinvertebrate diversity one-year post-fire, which he attributed to increases in instream fine sediment. The decrease in diversity was a result of a decrease in evenness, as taxa richness remained the same post-fire (Chan 1998). Likewise, there was no change in abundance in response to the prescribed fire. However, sampling was not continued after the first year post-fire, so the time to recovery could not be determined (Chan 1998). Multivariate ordination of all of the study sites together also indicated that the prescribed fire had no effect on aquatic macroinvertebrate communities relative to controls. However, when only the adjacent stream communities (Dark Canyon: prescribed fire treatment, and Mutton: control, Fig. 5.1a) were ordinated together, it appears that the 19d post-fire flood (Fig. 5.2) may have impacted stream communities in Dark Canyon Creek differently than in Mutton Creek (Fig. 5.10a-b). The distance between the prescribed burned community and the adjacent control was greatest in October 1998, followed by the sample period 19d post-fire (Fig. 5.10b). The 1997-98 winter was the most dramatic El Niño oscillation year on record, producing a peak flood of 1.85 m3/s after a rain-on-snow event where 25 cm of rain fell on 60 cm of snow in a 36 hour period (January 2, 1998, Fig. 5.2). However, the effects of the El Nino were seen in

259

the Oct 1998 sampling period (del Rosario 2000, Fig. 5.10). The differences between the communities following a single flood or an abnormally wet year suggests that the two streams respond differently to large floods, irrespective of whether they follow a fire. Because the effects of fire on macroinvertebrates are mediated through changes in the physical stream habitat, the recovery of communities is dependent on the recovery of the riparian zone and stream channel, in particular (Minshall et al. 2001b,c). Following severe canopy fires, macroinvertebrate communities may return to pre-fire conditions in as little as one to two years (if measuring richness or abundance), to five to 10 years (e.g., predominance of disturbance adapted species) (Minshall 2003, Vieira 2003, Vieira et al. 2004). Additionally, burned streams may exhibit higher variability than adjacent unburned streams for an undetermined period of time (Minshall et al. 2001b,c). However, the effects of wildfire are not always negative for aquatic habitats and communities (Gresswell 1999). For example, the 1999 16,000 ha “Sixteen complex” wildfire in Napa and Yolo counties, CA has shown that fire resulted in an overall increase in aquatic macroinvertebrate diversity (Greaves 2002, Resh, McElravy and Bêche, unpublished data). In general, stream recovery following even large and severe wildfires is rapid on an ecological time scale, and even in the midst of recovery, streams maintain the ability to support aquatic macroinvertebrates and other biota (Gresswell 1999, Minshall 2003).

Conclusion In this study, the prescribed fire treatment was not replicated because of logistical constraints, which is a common problem in fire effects monitoring (Schindler 1998, van

260

Mantgem et al. 2001); thus, a more rigorous statistical analysis is not possible. However, the trends observed in examining the multiple abiotic and biotic parameters measured in this study suggest that the prescribed fire had no or short-lasting (≤ 1 year) impacts on Dark Canyon Creek and its riparian zone (Table 5.11). Based on multiple abiotic and biotic components that were measured, I postulate that the relatively intact riparian zone (14% of the watershed area was burned, and the burn was patchy in the riparian zone) acted as a filter between the upland area and the stream (e.g., Naiman and Décamps 1997), thereby buffering the effects of the prescribed fire from the stream (also reported in Britton 1991b). Similarly, Minshall (2003) noted that the recovery of streams post-fire is largely dependent on the recovery of the riparian zone, which is usually more rapid than upland recovery. Thus, the extent of the upland area that is burned may be of equal, or even greater, importance than the amount of riparian burn. In conclusion, the exclusion of fire from riparian zones may be contributing to the accumulation of fuels and increasing the risk of an uncharacteristically large and severe wildfire (e.g., Ellis 2001, Skinner 2003), particularly in areas characterized by frequent, low to moderate severity fires (e.g., mixed-conifer zone of the Sierra Nevada, California, USA). Nevertheless, prescribed fire is generally not allowed to burn into riparian zones (e.g., Arno 1996), with a few exceptions (e.g., Chan 1998, Huntzinger 2003). In this study, a small prescribed fire that was allowed to burn and actively ignited in the riparian zone reduced fuel loads in the riparian zone up to 80%. The observations from this study suggest that it is possible for a prescribed fire that does not affect the entire area of a moderately sloped watershed (e.g., create burn “patchiness”) to enter the riparian zone

261

with few adverse effects. However, several factors may have contributed to the minimal effects of this prescribed fire on this mixed-conifer watershed. For example, watershed area burned, fire intensity (and, thus severity), slope, amount of vegetation consumed in upland and riparian areas, post-fire precipitation and weather are all factors that contribute to fire effects on streams. In this study, a small area was burned with a low to moderate severity fire on relatively low slopes, and post-fire precipitation was low. Furthermore, because this is a case-study in only one watershed, the results of this research may have limited applicability to other regions, or even to other watersheds within the mixed-confer region of the Sierra Nevada of California (e.g., Chan 1998). Ultimately, fire history, watershed condition (i.e., prior impairment may exacerbate the effects of fire; Minshall 2003), and management goals must be carefully considered prior to the implementation of prescribed fire as a management strategy.

References

Agee, J.K. 1993. Fire ecology of the Pacific Northwest forests. Island Press, Washington, D.C. Arno, M.K. 1996. Reestablishing fire-adapted communities to riparian forests in the ponderosa pine zone. INT-GTR-341. USDA Forest Service, Ogden, UT. Benda, L., Miller, D., Bigelow, P., and Andras, K. 2003. Effects of post-wildfire erosion on channel environments, Boise River, Idaho. Forest Ecology and Management 178: 105-119.

262

Berg, N., Carlson, A., and Azuma, D. 1998. Function and dynamics of woody debris in stream reaches in the central Sierra Nevada, California. Canadian Journal of Fisheries and Aquatic Sciences 55: 1807-1820. Biswell, H.H. and Schultz, A.M. 1957. Surface runoff and erosion as related to prescribed burning. Journal of Forestry 55: 372-374. Bragg, D.C. 2000. Simulating catastrophic and individualistic large woody debris recruitment for a small riparian system. Ecology 81: 1383-1394. Britton, D.L. 1990. Fire and the dynamics of allocthonous detritus in a South African mountain stream. Freshwater Biology 24: 360Britton, D.L. 1991a. Fire and the chemistry of a South African mountain stream. Hydrobiologia 218: 117-192. Britton, D.L. 1991b. The benthic macroinvertebrate fauna of a South African mountain stream and its response to fire. South African Journal of Aquatic Science 17: 51-64. Brown, J.K. 1974. Handbook for inventorying downed wood material. Intermountain Forest and Range Experimental Station General Technical Report INT-16. USDA Forest Service, Ogden, UT. Byram, G.M. 1959. Combustion of forest fuels. Pages 113-126 in Davis, K.P. (Ed.), Forest fire: control and use. McGraw-Hill, New York. Chan, I.A. 1998. Aquatic macroinvertebrates of small streams in the Mineral King region of Sequoia National Park: Baseline communities and response to prescribed fire. M.S., University of California, Davis; Davis, California, USA.

263

Chorover, J., Vitousek, P.M., Everson, D.A., Esperanza, A.M., and Turner, D. 1994. Solution chemistry profiles of mixed-conifer forests before and after fire. Biogeochemistry 26: 115-144. Clesceri, L.S., Greenberg, A.E., and Eaton, A.D. (Eds.). 1998. Standard Methods for the Examination of Water and Wastewater. 20th ed. American Public Health Association, Washington, D.C. DeBano, L.F., Ffolliott, P.F., and Baker, M.B. 1996. Fire severity effects on water resources. Pages 77-84 in DeBano, L.F., Baker, Jr., M.B., Gottfried, G.J., SolisGarza, G. Edminster, C.B., Neary, D.G., Allen, L.S., Hamre, R.H. (Eds.). Effects of fire on Madrean province ecosystems: A symposium proceedings. U.S.D.A. Forest Service, Rocky Mountain Forest and Range Experiment Station, Fort Collins, CO. DeBano, L.F. and Neary, D.G. 1996. Effects of fire on riparian systems. Pages 69-76 in DeBano, L.F., Baker, Jr., M.B., Gottfried, G.J., Solis-Garza, G. Edminster, C.B., Neary, D.G., Allen, L.S., Hamre, R.H. (Eds.). Fire effects on Madrean Province ecosystems: A symposium proceedings. U.S.D.A. Forest Service, Rocky Mountain Forest and Range Experiment Station, Fort Collins, CO. DeBano, L.F., Neary, D.G., and Ffolliott, P.F. 1998. Fire' s Effects on Ecosystems. John Wiley and Sons, New York. del Rosario, R.B. 2000. Aquatic insect responses to hydrologic intermittency and organic enrichment in headwater streams. Ph.D., University of California, Berkeley; Berkeley, California, USA.

264

Duncan, S.W. and Blinn, D.W. 1989. Importance of physical variables on the seasonal dynamics of epilithic algae in a highly shaded canyon stream. Journal of Phycology 25: 455-461. Dwire, K.A. and Kauffman, J.B. 2003. Fire and riparian ecosystems in landscapes of the western USA. Forest Ecology and Management 178: 61-74. Earl, S.R. and Blinn, D.W. 2003. Effects of wildfire ash on water chemistry and biota in South-Western U.S.A. streams. Freshwater Biology 48: 1015-1030. Elliott, K.J., Hendrick, R.L., Major, A.E., Vose, J.M., and Swank, W.T. 1999. Vegetation dynamics after a prescribed fire in the southern Appalachians. Forest Ecology and Management 114: 199-213. Ellis, L.M. 2001. Short-term response of woody plants to fire in a Rio Grande riparian forest, Central New Mexico, USA. Biological Conservation 97: 159-170. Erman, N.A. 1996. Status of aquatic invertebrates. vol. II, chap. 35. University of California Centers for Water and Wildland Resources, Davis, CA. Everett, R., Schellhass, R., Ohlson, P., Spurbeck, D., and Keenum, D. 2003. Continuity in fire disturbance between riparian and adjacent sideslope Douglas-fir forests. Forest Ecology and Management 175: 31-47. Greaves, M.E. 2002. Aquatic insect diversity and community structure in temporary streams of serpentine outcrops. M.S., University of California, Davis; Davis, California, USA. Gresswell, R.E. 1999. Fire and aquatic ecosystems in forested biomes of North America. Transactions of the American Fisheries Society 128: 221-

265

Hauer, F.R. and Spencer, C.N. 1998. Phosphorus and nitrogen dynamics in streams associated with wildfire: a study of immediate and longterm effects. International Journal of Wildland Fire 8: 183-198. Hilton, S. and Lisle, T.E. 1993. Measuring the fraction of pool volume filled with fine sediment. Research Note PSW-RN-414-WEB. Pacific Southwest Research Station, Forest Service, U.S. Department of Agriculture, Albany, CA. Huntzinger, M. 2003. Effects of fire management practices on butterfly diversity in the forested western United States. Biological Conservation 113: 1-12. Kattelmann, R. and Embury, M. 1996. Riparian areas and wetlands. Vol. 3, Sierra Nevada Ecosystem Project, Wildland Resources Center Report No. 38. University of California, Davis. Keifer, M., Stephenson, N.L., and Manley, J. 2000. Prescribed fire as the minimum tool for wilderness forest and fire regime restoration: a case study from the Sierra Nevada, CA. Pages 266-269 in Proceedings of the wilderness science in a time of change conference, May 23-27, 1999, Missoula, MT. U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station, Ogden, UT. Keller, E.A and Swanson, F.J. 1979. Effects of large organic material on channel form and fluvial processes. Earth Surface Processes. 4:361-380. Earth Surface Processes and Landforms 4: 361-380. Kilgore, B.M. and Taylor, D. 1979. Fire history of a sequoia mixed conifer forest. Ecology 60: 129-142. Kondolf, G.M. and Li, S. 1992. The pebble count technique for quantifying surface bed material size in instream flow studies. Regulated Rivers 3: 87-

266

Kruskal, J.B. 1964. Nonmetric multidimensional scaling: a numerical method. Psychometrika 29: 115-129. Lamb, E.G., Mallik, A.U., and Mackereth, R.W. 2003. The early impact of adjacent clearcutting and forest fire on riparian zone vegetation in northeastern Ontario. Forest Ecology and Management 177: 529-538. Lamberti, G.A. and Resh, V.H. 1985. Comparability of introduced tiles and natural substrata for sampling lotic bacteria, algae and macroinvertebrates. Freshwater Biology 15: 21-30. Legendre, P. and Legendre, L. 1998. Numerical Ecology. 2nd ed. Elsevier, Amsterdam. Lienkaemper, G.W. and Swanson, F.J. 1987. Dynamics of large woody debris in streams in old-growth douglas-fir forests. Canadian Journal of Forest Research 17: 150-156. Martin, R.E. and Dell, J.D. 1978. Planning for prescribed burning in the inland northwest. GTR-PNW-175. USDA Forest Service, Pacific Southwest Forest and Range Experiment Station, Portland, OR, USA. McCune, B. and Mefford, M.J. 1999. PC-ORD: Multivariate analysis of ecological data version 4.27. MjM Software, Gleneden Beach, Oregon, USA. McCune, B., Grace, J.B., and Urban, D.L. 2002. Analysis of Ecological Communities. MjM Software Design, Gleneden Beach, Oregon. Miller, C. and Urban, D.L. 2000. Modeling the effects of fire management alternatives on Sierra Nevada mixed-conifer forests. Ecological Applications 10: 85-94. Minshall, G.W., Brock, J.T., and Varley, J.D. 1989. Wildfires and Yellowstone's stream ecosystems. BioScience 39: 707-715.

267

Minshall, G.W., Robinson, C.T., Royer, T.V., and Rushforth, S.R. 1995. Benthic community structure in two adjacent streams in Yellowstone National Park five years after the 1988 wildfires. The Great Basin Naturalist 55: 193-200. Minshall, G.W., Robinson, C.T., and Lawrence, D.E. 1997. Postfire responses of lotic ecosystems in Yellowstone National Park, U.S.A. Canadian Journal of Fisheries and Aquatic Sciences 54: 2509-2525. Minshall, G.W., Brock, J.T., Andrews, D.A., and Robinson, C.T. 2001a. Water quality, substratum and biotic responses of five central Idaho (USA) streams during the first year following the Mortar Creek fire. International Journal of Wildland Fire 10: 185199. Minshall, G.W., Robinson, C.T., Lawrence, D.E., Andrews, D.A., and Brock, J.T. 2001b. Benthic macroinvertebrate assemblages in five central Idaho (USA) streams over a 10-year period following disturbance by wildfire. International Journal of Wildland Fire 10: 201-213. Minshall, G.W., Royer, T.V., and Robinson, C.T. 2001c. Response of the Cache Creek macroinvertebrates during the first 10 years following disturbance by the 1988 Yellowstone wildfires. Canadian Journal of Fisheries and Aquatic Sciences 58: 10771088. Minshall, G.W. 2003. Responses of stream benthic macroinvertebrates to fire. Forest Ecology and Management 178: 155-161. Moody, J.A. and Martin, D.A. 2001. Initial hydrologic and geomorphic response following a wildfire in the Colorado Front Range. Earth Surface Processes and Landforms 26: 1049-1070.

268

Naiman, R.J. and Décamps, H. 1997. The ecology of interfaces: riparian zones. Annual Review of Ecology and Systematics 28: 621-658. Olson, D.L. 2000. Fire in riparian zones: a comparison of historical fire occurrence in riparian and upslope forests in the Blue Mountains and Southern Cascades of Oregon. M.S., University of Washington, Seattle; Seattle, Washington, USA. Payne, D.Y. 1999. Prescribed fire effects on water quality in the Lake Tahoe Basin, California. M.S., San Jose State University; San Jose, California, USA. Platts, W.S., Armour, C., Booth, B.D., Bryant, M., Buford, J.L., Cuplin, P., Jensen, S., Lienkaemper, G.W., Minshall, G.W., Monsen, S.B., Nelson, R.L., Sedell, J.R., and Tuhy, J.S. 1987. Methods for evaluating riparian habitats with applications to management. INT-GTR 221. USDA Forest Service, Ogden, UT. Resh, V.H., Brown, A.V., Covich, A.P., Gurtz, M.E., Li, H.W., Minshall, G.W., Reice, S.R., Sheldon, A.L., Wallace, J.B., and Wissmar, R.C. 1988. The role of disturbance in stream ecology. Journal of the North American Benthological Society 7: 433-455. Richter, D.D. and Ralston, C.W. 1982. Prescribed fire: effects on water quality and forest nutrient cycling. Science 215: 661-663. Rinne, J.N. 1996. Short-term effects of wildfire on fishes and aquatic macroinvertebrates in the southwestern United States. North American Journal of Fisheries Management 16: 653-658. Robichaud, P.R. and Brown, R.E. 1999. What happened after the smoke cleared: onsite erosion rates after a wildfire in Eastern Oregon. Pages 419-426 in Wildland Hydrology Conference. Bozeman, MT,

269

Robichaud, P.R. 2000. Fire effects on infiltration rates after prescribed fire in Northern Rocky Mountain forests, USA. Journal of Hydrology 231-232: 220-229. Robinson, C.T., Rushforth, S.R., and Minshall, G.W. 1994. Diatom assemblages of streams influenced by wildfire. Journal of Phycology 30: 209-216. Roby, K.B. 1989. Watershed response and recovery from the Will fire: ten years of observation. General Technical Report PSW-109. USDA Forest Service. Roby, K.B. and Azuma, D.L. 1995. Changes in a reach of a northern California stream following wildfire. Environmental Management 19: 591-600. Romme, W.H. 1982. Fire and landscape diversity in subalpine forests of Yellowstone National Park. Ecological Monographs 52: 199-221. Russell, W.H. and McBride, J.R. 2001. The relative importance of fire and watercourse proximity in determining stand composition in mixed conifer riparian forests. Forest Ecology and Management 150: 259-265. Schindler, D.W. 1998. Replication versus realism: the need for ecosystem-scale experiments. Ecosystems 1: 323-334. Skinner, C.N. 2003. A tree-ring based fire history of riparian reserves in the Klamath mountains. Pages 116-119 in California Riparian Systems: Processes and Floodplain Management, Ecology, and Restoration. Pickleweed Press, Mill Valley, CA, March 12, 2001. Spencer, C.N. and Hauer, F.R. 1991. Phosphorous and nitrogen dynamics in streams during a wildfire. Journal of the North American Benthological Society 10: 24-30.

270

Spencer, C.N., Gabel, K.O., and Hauer, F.R. 2003. Wildfire effects on stream food webs and nutrient dynamics in Glacier National Park, USA. Forest Ecology and Management 178: 141-153. Steinman, A.D. and Lamberti, G.A. 1996. Biomass and pigments of benthic algae. Pages 1-674 in Hauer, F.R. and Lamberti, G.A. (Eds.), Academic Press, San Diego, CA. Stephens, S.L. 1998. Effects of fuels and silvicultural treatments on potential fire behavior in mixed conifer forests of the Sierra Nevada, CA. Forest Ecology and Management 105, 21-34. Stephens, S.L. 2001. Fire history differences in adjacent Jeffrey pine and upper montane forests in the eastern Sierra Nevada. International Journal of Wildland Fire 10, 161167. Stephens, S.L. and Finney, M.A. 2002. Prescribed fire mortality of Sierra Nevada mixed conifer tree species: effects of crown damage and forest floor combustion. Forest Ecology and Management 162: 261-275. Stephens, S.L., Meixner, T., Poth, M., McGurk, B., and Payne, D.Y. 2004. Prescribed fire, soils, and stream water chemistry in a watershed in the Lake Tahoe Basin, California. International Journal of Wildland Fire 13: 27-35. Stephens, S.L. and Collins, B.M. 2004. Fire regimes of mixed conifer forests in the north central Sierra Nevada at multiple spatial scales. Northwest Science 78: 12-23. Swanson, F.J. 1981. Fire and geomorphic processes. Pages 401-420 in Proceedings of the conference on fire regimes and ecosystem properties. USDA Forest Service, Washington, D.C.

271

Tiedemann, A.R., Conrad, C.E., Dieterich, J.H., Hornbeck, J.W., Megahan, W.F., Viereck, L.A., and Wade, D.D. 1979. Effects of fire on water: a state-of-knowledge review. General Technical Report WO-10. USDA Forest Service, Denver, CO. Timoney, K.P., Peterson, G., and Wein, R. 1997. Vegetation development of boreal riparian plant communities after flooding, fire, and logging, Peace River, Canada. Forest Ecology and Management 93: 101-120. Townsend, S.A. and Douglas, M.M. 2000. The effect of three fire regimes on stream water quality, water yield and export coefficients in a tropical savanna (northern Australia). Journal of Hydrology 229: 118-137. Underwood, A.J. 1994. On beyond BACI: sampling designs that might reliably detect environmental disturbances. Ecological Applications 4: 3-15. van Mantgem, P., Schwartz, M., and Keifer, M. 2001. Monitoring fire effects for managed burns and wildfires: coming to terms with pseudoreplication. Natural Areas Journal 21: 266-273. van Wagtendonk, J.W., Benedict, J.M., and Sydoriak, W.M. 1996. Physical properties of woody fuel particles of Sierra Nevada conifers. International Journal of Wildland Fire 6: 117-123. van Wagtendonk, J.W., Benedict, J.M., and Sydoriak, W.M. 1998. Fuel bed characteristics of Sierra Nevada conifers. Western Journal of Applied Forestry 13: 7384. Vieira, N.K.M. 2003. Impacts of wildfire on benthic insects in burned streams: community and population responses at multiple scales. Ph.D., Colorado State University; Fort Collins, Colorado, USA.

272

Vieira, N.K.M., Clements, W.H., Guevara, L.S., and Jacobs, B.F. 2004. Resistance and resilience of stream insect communities to repeated hydrologic disturbances after a wildfire. Freshwater Biology 49: 1243-1259. Waltz, A.E.M., Fulé, P.Z., Covington, W.W., and Moore, M.M. 2003. Diversity in ponderosa pine forest structure following ecological restoration treatments. Forest Science 49: 885-900. Williams, M.R. and Melack, J.M. 1997. Effects of prescribed burning and drought on the solute chemistry of mixed-conifer forest streams of the Sierra Nevada, California. Biogeochemistry 39: 225-253. Wolman, M.G. 1954. A method of sampling coarse river-bed material. Transactions, American Geophysical Union 35: 951-956. Wondzell, S.M. and King, J.G. 2003. Postfire erosional processes in the Pacific Northwest and Rocky Mountain regions. Forest Ecology and Management 178: 7587. Young, M.K. 1994. Movement and characteristics of stream-borne coarse woody debris in adjacent burned and undisturbed watersheds in Wyoming. Canadian Journal of Forest Research 24: 1933-1938.

273

274

Elevation (m) Stream order Watershed area (ha) Avg. stream width (m) Avg. discharge (m3/s) % of watershed in BFRS Current land-use

deviation.

Prescribed fire Dark Canyon 1452 1 129 1.13 ± 0.25 0.025 ± 0.018 98 Ecological reserve Bacon (1 &2) 1380 1 81 1.22 ± 0.40 0.011 ± 0.007 100 Light cattle grazing, housing, timber harvest

Control sites Deep Canyon Mutton 1280 1310 1 1 139 244 1.93 ± 0.50 1.70 ± 0.18 0.041 ± 0.019 0.021 ± 0.012 51 80 Light timber Light timber harvest harvest, ecological reserve

Gaddis 1243 2 486 2.43 ± 0.34 0.057 ± 0.027 80 Light timber harvest

Table 5.1. Description of prescribed burned and control watersheds in this study. Averages are presented as average ± standard

Table 5.2. Basal area (m2 ha-1) of live riparian tree species (> 11 cm DBH) at each stream based on 400m2 plots.

Alnus rhombifolia Pseudotsuga menziesii Calocedrus decurrens Pinus ponderosa Pinus lambertiani Abies concolor Taxus brevifolia

Dark Canyon (PF) (n =5) 6.3 15.8 26.0 3.9 9.0 17.0 1.1

Bacon (Control) (n = 2) 5.7 19.5 41.5 -----

Gaddis (Control) (n = 2) 0.7 0.3 26.7 25.5 -0.9 --

Mutton (Control) (n = 2) 0.1 32.1 32.4 7.8 11.1 16.0 0.6

275

Table 5.3. Timing and duration of sampling and measurements made pre- and post-fire. . Timing of sampling Hydrology Macroinvertebrates Riparian vegetation Water chemistry Periphyton Sediment Channel morphology Large woody debris

Hourly October July/Aug Monthly Monthly June/Oct. October October

Pre-fire Start Jun 95 Oct 95 July 97 Jun 01 Sept 01 Jun 01 Oct 01 Sept 01

End Oct 02 Oct 02 Aug 01 Oct 02 Oct 02 Oct 02 Oct 01 Sept 02

Post-fire Start Oct 02 Jun 03 July 03 Oct 02 Oct 02 Jun 03 Oct 03 Sept 03

End Dec 03 Oct 03 Aug 03 Oct 03 Oct 03 Oct 03 Oct 03 Sept 03

276

Table 5.4. Average fuel loads (mean ± standard deviation) in the prescribed fire unit pre(2001), 2 weeks post-fire (Nov. 6, 2002), and 1-year post-fire (Oct 2003).

Fuel type (timelag)

1-h 10-h 100-h 1000-h Sound Rotten Duff Litter Total surface Total load

Pre-fire fuel load (kg/m2) n = 10 0.13 ± 0.08 0.63 ± 0.37 0.70 ± 0.51

2 wk PF fuel load (kg/m2) n=5 0.02 ± 0.02 0.20 ± 0.19 0.28 ± 0.32

% consumed (2 wk PF)

1-yr PF fuel load (kg/m2) n = 10

% consumed (1 yr PF)

89 68 60

0.04 ± 0.03 0.11 ± 0.11 0.23 ± 0.27

72 83 67

2.48 ± 3.92 2.13 ± 4.45 4.88 ± 2.79 3.16 ± 1.48 6.07 ± 5.10 14.11 ± 4.93

2.33 ± 3.37 0±0 0.15 ± 0.22 0.39 ± 0.20 2.82 ± 3.27 2.86 ± 3.51

6 100 92 88 34 79

2.71 ± 5.71 2.17 ± 4.50 1.03 ± 1.68 0.63 ± 0.66 5.27 ± 6.70 6.92 ± 6.80

-2 -9 79 80 13 51

277

Table 5.5. Riparian vegetation summary measures before (1997, 2001) and after (2003) the prescribed fire in Dark Canyon Creek (n = 10 plots) and in 3 control streams (n = 10 plots). Values are presented as mean ± standard deviation. % cover vegetation

taxa richness

diversity (H')

Prescribed fire 1997 2001 2003

42.8 ± 24.7 63.3 ± 32.7 15.3 ± 5.2

8.6 ± 2.6 10.1 ± 2.4 6.3 ± 1.4

1.79 ± 0.27 1.75 ± 0.35 1.41 ± 0.30

Control 1997 2001 2003

45.0 ± 24.6 61.2 ± 24.9 67.8 ± 39.6

7.8 ± 2.5 9.9 ± 4.5 8.1 ± 3.3

1.51 ± 0.42 1.69 ± 0.41 1.41 ± 0.30

278

Table 5.6. Large woody debris movement from 2001-2003 in control (C) and prescribedfire (PF) streams. Movement was determined to be detected if the piece moved > 1 m.

Deep Cyn (C) Gaddis 1 (C) Gaddis 2 (C) Mutton (C) Dark Cyn (PF)

# pieces moved 15 11 13 12 8

2001-2002 Avg. Range distance moved (m) (m) 1.4 1.1 – 2.1 2.4 1.1 – 7.1 6.7 1.2 – 38.0 1.9 1.3 – 2.5 1.5 1.1 – 2.7

# pieces moved 3 8 2 7 3

2002-2003 Avg. Range distance moved (m) (m) 2.1 1.1 - 3.6 3.2 1.3 – 8.4 6.5 3.9 – 9.0 2.5 1.2 – 6.7 5.2 1.3 – 11.2

279

Table 5.7. Characteristics of LWD, and changes in LWD volume before (2001, 2002) and after (2003) the prescribed fire. C = control sites, and PF = prescribed fire site. Stream length containing 50 pieces of LWD is also presented.

Bacon 1 (C) Bacon 2 (C) Deep Cyn (C) Gaddis 1 (C) Gaddis 2 (C) Mutton (C) Dark Cyn (PF)

Average Length (m) 2001 diameter Average with 50 volume length (m) LWD (m) (m3) 0.13 1.92 68.8 2.39 0.16 3.53 58.6 9.55 0.13 2.92 86.5 9.01 0.14 3.28 92.0 5.46 0.17 3.60 105.5 7.98 0.19 2.80 59.0 3.63 0.16 3.59 67.3 9.09

2002 volume (m3) 2.42 (2) 9.88 (4) 9.97 (4) 5.48 (1) 7.98 (1) 16.79 (7) 9.12 (1)

2003 volume (m3) 2.42 (0) 9.88 (0) 9.97 (0) 5.50 (2) 8.52 (3) 16.95 (7) 9.25 (5)

280

Table 5.8. Percentage of fine sediment based on pebble counts in each stream before (2001, 2002) and after the fire (2003). PF = prescribed fire, C = control.

Dark Cyn (PF) Mutton (C) Deep Cyn (C) Gaddis 1 (C) Gaddis 2 (C)

Oct 2001 65 35 24 -2

% fine sediment (≤ 4 mm diameter) June 2002 Oct 2002 June 2003 47 52 61 57 41 36 0 15 -1 4 11 0 3 --

Oct 2003 48 47 20 4 11

281

Table 5.9. Percentage of pool volume comprised of fine sediment (V*) before (1997, 2001) and after (2003) the prescribed fire in control (n = 3 streams) and fire (n =1 stream) units. Data are presented as mean ± standard deviation (the number of pools surveyed).

1997 2001 2003

Control 75.98 ± 9.29 81.81 ± 11.39 71.32 ± 14.14

(9) (6) (8)

Prescribed Fire (10) 82.33 ± 8.11 (10) 72.95 ± 12.37 (7) 80.70 ± 8.99

282

Table 5.10. Results of an asymmetrical ANOVA of periphyton biomass (AFDM). A significant CF*BA term, using BA*Site(CF) as a denominator, indicates an effect of the prescribed fire on periphyton biomass. Significant results are indicated in bold.

Source Control/Fire (CF) Before/After (BA) CF*BA Time (BA) Site (CF) BA*Site(CF) CF*Time(BA) Error

SS 0.088 0.061 0.520 5.831 0.074 0.070 2.223 0.629

df 1 1 1 16 2 2 16 31

MS 0.088 0.061 0.520 0.364 0.037 0.035 0.139 0.020

F-ratio 4.318 2.996 25.635 17.958 1.827 1.723 6.847 --

p-value 0.046 0.093