Sources and Ages of Carbon and Organic Matter Supporting Macroinvertebrate Production in Temperate Streams

Home , Little Muncy Creek, Towanda Creek

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Amber Renee Bellamy

Graduate Program in Evolution, Ecology and Organismal Biology

The Ohio State University

2017

Dissertation Committee:

Dr. James E. Bauer, Advisor

Dr. Yu-Ping Chin

Dr. Jonathan J. Cole

Dr. Peter S. Curtis

Dr. Andrea G. Grottoli

Dr. S. Mažeika P. Sullivan

Copyrighted by

Amber Renee Bellamy

2017

Abstract

Streams and rivers are tightly integrated into the terrestrial landscape and as a result are responsible for transporting and transforming globally significant amounts of terrestrially-derived materials. Much of this terrestrially-derived organic matter (OM) that is transported by streams and rivers is hundreds to thousands of years in age. Over the past couple decades, research has increasingly demonstrated the importance of terrestrial

OM (i.e., allochthonous) to consumer nutrition in many aquatic ecosystems. However, there has been far less research to date on assessing the ages of carbon (C) and OM from different allochthonous and autochthonous (i.e., derived from within aquatic systems, e.g., algal and macrophytic) sources that are utilized by aquatic consumers and incorporated into aquatic food webs. There is also a need to better understand how human-mediated processes such as land use and climate change may lead to shifts in the inputs and nutritional roles of allochthonous and aged materials in streams and rivers.

Stable isotopes (δ13C, δ15N, and δ2H) have been employed frequently in aquatic food web and ecosystem studies to quantitatively assess contributions of different nutritional resources to the biomass and secondary production of aquatic consumers.

Natural abundance radiocarbon (14C) has been employed far less frequently in food web and ecosystem studies, however, because of its far greater dynamic range than most stable isotopes, the use of natural 14C as a tracer of C and OM inputs and utilization in ii inland waters allows for potentially greater differentiation between potential nutritional resources than stable isotopes alone. Natural 14C also uniquely allows for determination of the ages of the nutritional resources that contribute to aquatic consumer biomass and secondary production.

In the present study, natural stable and radiocarbon isotopes were measured to assess contributions of allochthonous and autochthonous C and OM of varying age to the biomass of aquatic macroinvertebrates belonging to different functional feeding groups

(FFGs) in a number of temperate streams in sub-watersheds of the Mohawk-Hudson

River, New York (USA); Susquehanna River, Pennsylvania (USA); and the Ohio River,

Ohio (USA). The overarching goals of the study were to i) estimate allochthonous and authochthonous nutritional contributions to aquatic macroinvertebrate consumers in these systems, ii) determine relative contributions of C and OM resources of varying ages to macroinvertebrate nutrition, and iii) assess how nutritional resource utilization varied as a function of stream physical characteristics, watershed land use and lithology, and temporal variability.

Aquatic macroinvertebrates collected from the majority of streams studied were found to be primarily reliant on autochthonous OM, with greater contributions to the biomass of some FFGs (scrapers and filtering collectors; up to 99%) compared to others.

However, allochthonous OM in the forms of terrestrial vegetation and soil/sediment OM also made significant contributions to the biomass of multiple FFGs (up to 85% and 46%, respectively), primarily shredders, chironomids, and collector-gatherers, across all systems studied. Macroinvertebrates collected from streams having a large proportion of

iii agricultural land use in their watersheds were more reliant on soil/sediment-derived OM compared to macroinvertebrates collected from streams with low watershed agriculture.

In addition, greater agricultural activity also led to elevated δ15N of algal biomass and macroinvertebrate consumer biomass but had minimal impacts on macroinvertebrate δ13C and δ2H values.

The ∆14C values and ages of macroinvertebrates varied widely across sites and sampling times (-499‰ to 23‰; 5,550 years B.P. equivalent age to modern-aged, respectively). The apparent 14C “ages” observed for many of the macroinvertebrates measured were attributable to some combination of their consumption and utilization of i) truly aged soil/sediment OM and ii) apparently “aged” algae that fixed aged dissolved inorganic carbon (DIC).

Finally, a comprehensive review and synthesis of studies that have used natural abundance 14C as a tracer in inland aquatic food webs was undertaken. With the exception of very large lakes (e.g., Lake Superior), in essentially every system studied one or more forms of aged C and/or OM contributed to the biomass of aquatic consumers, suggesting that geologically aged forms of C and OM are present in, and support to varying extents, modern-day aquatic ecosystems. In future studies, improved knowledge of the ages of C, OM, and consumers in inland water and other aquatic food webs could have fundamental implications for our understanding of terrestrial-aquatic linkages and the role(s) that the ages of nutritional resources may play in aquatic food web and ecosystem structure and function.

Acknowledgments

I would like to thank my major advisor, Dr. James Bauer, for providing guidance throughout the development of my project, supporting my project financially, encouraging me to interact with other scientists, and providing substantial feedback on this document. Thanks go to Dr. Jonathan Cole for his help and advice throughout this process, and his timely responses to my questions as they came up during various phases of my project. I am especially grateful to Dr. Andrea Grottoli who helped me extensively with finding and using the best approaches for analyzing my data. Additional thanks go to my other committee members Dr. Mazeika Sullivan, Dr. Peter Curtis, and Dr. Yu-Ping

Chin who each encouraged me to think about my project from a variety of different perspectives. Without the assistance of Amy Barrett in the field and laboratory this project would not have been possible, and I am extremely grateful for her help and friendship throughout this process. I also thank my former lab mates Thomas Evans,

Steven Gougherty, Scott Kelsey, and Dr. Katie Hossler for field assistance and their experienced advice and support throughout this process. Thanks also to Yohei Matsui in the OSU Stable Isotope Biogeochemistry lab, who always took great care in analyzing my samples. Finally, I would also like to thank the Hudson River Foundation for awarding me a year-long Mark B. Bain Graduate Fellowship that provided me with the resources to conduct the study of the Mohawk-Hudson watershed. v

Vita

2005...... Dobyns-Bennett High School

2009...... B.S. Environmental Science,

Gardner-Webb University

2010-2011 ...... AmeriCorps VISTA

2011-2012 ...... Graduate Research Assistant, Department of

EEOB, Ohio State University

2012-2016 ...... Graduate Teaching Associate, Department

of EEOB, Ohio State University

Fields of Study

Major Field: Evolution, Ecology and Organismal Biology

Table of Contents Abstract ...... ii

Acknowledgments...... v

Vita ...... vi

Fields of Study ...... vi

Table of Contents ...... vii

List of Tables ...... xi

List of Figures ...... xvii

Introduction ...... 1

Chapter 1: Influence of Land Use and Lithology in Sources and Ages of Nutritional

Resources for Stream Macroinvertebrates: A Multi-Isotopic Approach ...... 7

Abstract ...... 8

Introduction ...... 9

Methods ...... 12

Results ...... 18

Discussion ...... 21

vii

Chapter 2: Contributions of Allochthonous and Aged Carbon and Organic Matter to the

Nutrition of Macroinvertebrates in a Headwater Network of the Susquehanna River ..... 43

Abstract ...... 44

Introduction ...... 45

Methods ...... 48

Results ...... 54

Discussion ...... 59

Chapter 3: Temporal Variability in Autochthonous and Allochthonous Nutritional

Sources to Macroinvertebrates in a Temperate Stream ...... 79

Abstract ...... 80

Introduction ...... 81

Methods ...... 83

Results and Discussion ...... 89

Conclusions ...... 101

Chapter 4: Nutritional Support of Inland Aquatic Food Webs by Aged Carbon and

Organic Matter ...... 112

Abstract ...... 113

Introduction ...... 113

Background ...... 114

viii

Variability in the Ages of Carbon and Organic Matter in Inland Waters ...... 117

Sources and Routes of Input of Aged C and OM to Inland Water Systems ...... 119

Ages of POC, DOC, and DIC Pools in Inland Water Systems ...... 124

14C Values and Apparent Ages of Aquatic Consumer Organisms ...... 126

Estimates of the Assimilation of Aged C and OM by Aquatic Consumers ...... 135

Conclusions and Future Directions ...... 142

Conclusions ...... 157

Literature Cited ...... 163

Appendix A: Supplementary Information for Chapter 1 ...... 222

Section 1 ...... 223

Section 2 ...... 224

Section 3 ...... 234

Section 4 ...... 236

Section 5 ...... 239

Appendix B: Supplementary Information for Chapter 2 ...... 241

Section 1 ...... 242

Section 2 ...... 245

Section 3 ...... 252

Section 4 ...... 256

Section 5 ...... 259

Appendix C: Supplementary Information for Chapter 3 ...... 265

Section 1 ...... 266

Section 2 ...... 275

Section 3 ...... 279

List of Tables

Table 1.1. Site and watershed characteristics of the six study streams in the Mohawk- Hudson River watershed in June 2014...... 34

Table 2.1. Sampling site and sub-watershed characteristics of study streams within the upper and middle Susquehanna River watershed...... 70

Table 3.1. Sampling dates and water chemistry parameters for the Paint Creek study site...... 103

Table 3.2. Means ± S.D. of δ13C, δ15N, δ2H, and ∆14C values for macroinvertebrate functional feeding groups (FFGs) in Paint Creek. Isotopic values presented here are raw data and uncorrected for trophic fractionation and the influence of dietary water...... 104

Table 4.1. Summary of inland water food web studies that have employed natural abundance 14C analyses of consumer organisms...... 144

Table A.2. Stable isotope data used in mixing model to examine the effect of using estimated algal isotope values vs. measured algal isotope values. See Appendix A Section 2 for details...... 229

Table A.4. Isotope data used in mixing model with FFG averages collected from all sites. See Appendix A Section 2 for details...... 231

Table A.5. Isotope data used in mixing model for streams from high agriculture watersheds. See Appendix A Section 2 for details...... 231

Table A.6. Isotope data used in mixing model for streams from low agriculture watersheds. See Appendix A Section 2 for details...... 231

Table A.7. Isotope data used in mixing model for Fly Branch (shale absent from watershed). See Appendix A Section 2 for details...... 231

Table A.8. Isotope data used in mixing model for Nowadega Creek (shale present in watershed). See Appendix A Section 2 for details...... 231

xii

Table A.16. Characterization of streambed substrate for each stream sampled...... 240

Table B.1. Percentage of land use within farms by county for the subwatersheds of this study in the upper and middle Susquehanna River network, Pennsylvania, USA...... 242

Table B.3. Lamb’s Creek ...... 248

Table B.4. North Elk Run ...... 248

Table B.5. Little Muncy Creek ...... 248

Table B.6. Elk Run...... 248

Table B.7. Crooked Creek ...... 249

Table B.8. Towanda Creek ...... 249

Table B.9. Cowanesque River ...... 249

xiii

Table B.10. High agriculture sites ...... 249

Table B.11. Low agriculture sites ...... 249

Table B.12. Global mixing model including stable isotope and radiocarbon data ...... 249

Table B.13. Little Muncy Creek mixing model including δ13C, δ2H, and 14C data ..... 250

Table B.20. Lamb’s Creek using δ13C, δ15N, and δ2H ...... 259

Table B.21. North Elk Run using δ13C, δ15N, and δ2H ...... 259

Table B.22. Little Muncy Creek using δ13C, δ15N, and δ2H ...... 260 xiv

Table B.23. Elk Run using δ13C, δ15N, and δ2H ...... 260

Table B.24. Crooked Creek using δ13C, δ15N, and δ2H ...... 261

Table B.25. Towanda Creek using δ13C, δ15N, and δ2H ...... 261

Table B.26. Cowanesque River using δ13C, δ15N, and δ2H ...... 262

Table B.27. High agriculture sites using δ13C, δ15N, and δ2H ...... 262

Table B.28. Low agriculture sites using δ13C, δ15N, and δ2H ...... 263

Table B.29. Global model using δ13C, δ15N, δ2H, and ∆14C data ...... 263

Table B.30. Little Muncy Creek primary consumers using δ13C, δ2H and ∆14C data .... 264

Table B.31. Little Muncy Creek predators using δ13C, δ2H and ∆14C data ...... 264

Table C.1. Genera and functional feeding groups identified in Paint Creek, OH...... 271

Table C.2. November 2012 potential nutritional resource stable isotope data...... 272

Table C.3. May 2013 potential nutritional resource stable isotope data...... 272

Table C.4. July 2013 potential nutritional resource stable isotope data...... 272

Table C.5. November 2013 potential nutritional resource stable isotope data...... 272

Table C.6. May and July 2013 mixing model sources...... 273

Table C.7. November 2012 and 2013 mixing model sources...... 273

Table C.8. Global mixing model sources including stable isotope and 14C data...... 273

Table C.9. Global mixing model sources using stable isotope data only...... 273

Table C.10. One-way ANOSIM test of macroinvertebrates (δ13C, δ15N, and δ2H) at each sampling time point...... 275

Table C.11. One-way ANOSIM test of macroinvertebrates (δ13C, δ15N, and δ2H) between functional feeding groups (FFGs)...... 275

Table C.12. One-way ANOSIM test of macroinvertebrates (δ13C, δ15N, and δ2H) with the combined factor of functional feeding group and sampling date...... 275 xv

Table C.13. Model run using δ13C, δ15N, and δ2H for May and July 2013...... 279

Table C.14. Model run using δ13C, δ15N, and δ2H for November 2012 and 2013...... 279

Table C.15. Model run using global data set and δ13C, δ15N, and δ2H values...... 280

Table C.16. Model run using δ13C, δ15N, δ2H, and ∆14C values...... 280

xvi

List of Figures

Figure 1.1. Sampling site locations for the present study in the Hudson-Mohawk River watershed. The entire Hudson River drainage basin is shown in green in the inset map. Filled dots represent major regional population centers and filled stars represent sampling sites...... 35

Figure 1.2. δ13C, δ15N, and δ2H values of macroinvertebrates and their potential nutritional sources collected from the six study streams in the Mohawk-Hudson River watershed (mean ± SD). A) δ13C vs. δ15N coded by FFG, B) δ13C vs. δ2H coded by FFG, C) δ13C, δ15N, and δ2H NMDS results with all macroinvertebrate individuals across all six sites coded by FFG; Stress: 0.11, D) δ13C vs. δ15N coded by site, E) δ13C vs. δ2H coded by site, F) δ13C, δ15N, and δ2H NMDS results for all macroinvertebrate individuals across all six sites coded by site; Stress: 0.11. Values in panels A, B, D, and E are corrected for trophic fractionation (Post 2002) and dietary water (Wilkinson et al. 2015). Axes for panels C and F are dimensionless (see text for details). CH-Couch Hollow, FB- Fly Branch, NC-Nowadega Creek, OC- Otsquago Creek, ST-Schoharie Tributary, and WC-Wampecack Creek...... 36

Figure 1.3. A) δ13C vs. δ15N values of macroinvertebrates and their potential nutritional sources from high agriculture and low agriculture watersheds (mean ± SD), B) δ13C vs. δ2H of macroinvertebrate individuals collected from high agriculture and low agriculture watersheds and their potential nutritional sources (mean ± SD), and C) δ13C, δ15N, and δ2H NMDS results for all macroinvertebrate individuals from high agriculture and low agriculture watersheds; Stress: 0.11. Values in panels A and B are corrected for trophic fractionation (Post 2002) and the influence of dietary water (Wilkinson et al. 2015). Axes for panel C are dimensionless (see text for details)...... 37

Figure 1.4. A) ∆14C vs. δ13C of macroinvertebrate individuals collected from Fly Branch (no shale in watershed) and Nowadega Creek (shale in watershed) and their potential nutritional sources (mean ± SD), and B) δ13C, δ15N, δ2H, and ∆14C NMDS results for all Fly Branch and Nowadega Creek macroinvertebrates; Stress: 0.02. Axes for panel B are dimensionless (see text for details)...... 38

xvii

Figure 1.5. Posterior distributions of the proportional contributions of A) algal, B) terrestrial vegetation, and C) soil OM potential nutritional sources to the biomass of macroinvertebrates using isotopic averages of δ13C, δ15N, and δ2H across all individuals belonging to a particular FFG for all sites. Scrapers n=5, filtering collectors n=6, collector-gatherers n=3, chironomids n=3, predators n=5. The averages of all potential nutritional sources across all sites were used in the model and algal isotope values were estimated. 5th, 25th, 50th, 75th, and 95th percentiles are shown...... 39

Figure 1.6. Comparison of median proportional nutritional source contributions to macroinvertebrate diet from mixing models in A) sites with high watershed agriculture and B) sites with low watershed agriculture. *n=1 ...... 40

Figure 1.7. Comparison of median proportional nutritional source contributions to macroinvertebrate diet from mixing models for A) Nowadega Creek site with OM-rich shale in its watershed, and B) Fly Branch site without shale in its watershed. These sites are the only the two from which natural radiocarbon analyses were conducted on macroinvertebrates and their potential nutritional sources...... 41

Figure 1.8. Predicted relationships between A) percent allochthony in macroinvertebrates vs. the percentage of shale bedrock substrate in watershed, B) percent similarity between stream DIC ∆14C values and macroinvertebrate ∆14C values vs. the percentage of aged carbon/organic matter (OM) in the watershed lithology, C) percent nutritional contribution of soil OM to macroinvertebrate biomass vs. percent agriculture in the watershed. The predicted relationships are derived from data from the present study and from the studies of Ishikawa et al. (2014) and Wang et al. (2014)...... 42

Figure 2.1. Map of sampling locations in the upper and middle Susquehanna River watershed. The Susquehanna River basin is shown as light blue background, river main stems are shown as red lines, and sampling streams are shown as dark blue lines. Grey lines represent Susquehanna River subwatershed boundaries...... 71

Figure 2.2. δ13C, δ15N, and δ2H values of macroinvertebrate individuals and their potential dietary sources (means ± SD) collected from all seven Susquehanna River sub- watershed study sites. Values are corrected for trophic fractionation (Post 2002) and the influence of dietary water (Wilkinson et al. 2015). Potential nutritional source isotope means and SDs are pooled across sites for visualization purposes, but not for model runs. A) δ13C vs. δ15N, B) δ13C vs. δ2H, and C) NMDS results for δ13C, δ15N, and δ2H for all macroinvertebrate individuals; Stress: 0.15. Panel C axes are dimensionless...... 72

xviii

Figure 2.3. Isotope-isotope plots of macroinvertebrate individuals collected from sites with high percent agriculture (>50% of land use) and low percent agriculture (<50% of land use) in their watersheds and their potential dietary sources (means ± SD). Values are corrected for trophic fractionation (Post 2002) and the influence of dietary water (Wilkinson et al. 2015). Source isotope means and SDs are pooled across sites for visualization purposes only. A) δ13C vs. δ15N, B) δ13C vs. δ2H and C) NMDS results for δ13C, δ15N, and δ2H NMDS of all macroinvertebrate individuals from high and low agriculture sites; Stress: 0.15. Panel C axes are dimensionless...... 73

Figure 2.4. A) ∆14C vs. δ13C of macroinvertebrate individuals collected from six of the study sites (excluding Lamb’s Creek) and their potential dietary sources (means ± SD). δ13C values are corrected for trophic fractionation (Post 2002). B) ∆14C vs. δ13C of macroinvertebrate individuals collected from Little Muncy Creek only and their potential dietary sources (means ± SD). C) NMDS results using δ13C, δ15N, δ2H, and ∆14C for all macroinvertebrates; Stress: 0.11. Panel C axes are dimensionless. For panel C, refer to the legend for panel A...... 74

Figure 2.5. Mixing model estimates of median nutritional contributions in the seven subwatersheds of the upper Susquehanna River using δ13C, δ15N, and δ2H values of macroinvertebrate functional feeding groups (FFGs) and their potential dietary sources. ND: no data. A) Lamb’s Creek, B) North Elk Run, C) Little Muncy Creek, D) Elk Run, E) Crooked Creek, F) Towanda Creek, G) Cowanesque River...... 75

Figure 2.6. Comparison of median proportional source contributions to macroinvertebrate nutrition from mixing models for A) sites with high agriculture (>50% of land use) in their watersheds and B) sites with low agriculture (<50% of land use) in their watersheds...... 76

Figure 2.7. Posterior distributions of the proportional contributions of different nutritional sources to macroinvertebrate biomass from the global mixing model for six of the study sites (excluding Lamb’s Creek) using δ13C, δ15N, δ2H and ∆14C. 5th, 25th, 50th, 75th, and 95th percentiles are shown...... 77

Figure 2.8. A) Median proportional source contribution estimates to macroinvertebrate diet for Little Muncy Creek primary consumers grouped by year using δ13C, δ2H and ∆14C. B) Median proportional source contribution estimates to macroinvertebrate diet for Little Muncy Creek predators grouped by year using δ13C, δ2H and ∆14C...... 78

Figure 3.1. Locations of Paint Creek (light blue line), study site (black circle) and USGS gaging station (gray circle) within the greater Scioto River watershed (light blue shaded area). The Scioto River is a tributary of the Ohio River...... 105

xix

Figure 3.2. Paint Creek discharge (m3/second) as measured at the USGS gaging station downstream of the study reach (see Fig. 1). Data in figure taken from http://waterdata.usgs.gov/usa/nwis/ uv?03234000. Black arrows indicate sampling dates in the present study...... 106

Figure 3.3. δ13C, δ15N, and δ2H values of individual macroinvertebrates and their potential dietary sources (means ± SDs) collected from Paint Creek, Ohio, USA. δ13C and δ15N values are corrected for trophic fractionation according to Post (2002) and δ2H values are corrected for contributions of dietary water. A) δ13C vs. δ15N labelled by sampling date, B) δ13C vs. δ2H labelled by sampling date, C) δ13C vs. δ15N labelled by functional feeding group (FFG), and D) δ13C vs. δ2H labelled by FFG...... 107

Figure 3.4. NMDS plot for Paint Creek macroinvertebrate δ13C, δ15N, and δ2H values in the present study. Significant differences within functional feeding groups by sampling date as determined by one-way ANOSIM are indicated by filled colored circles, whereas those without significant differences are indicated by open circles. Open circles- Nov. 2012, July 2013, and Nov. 2013 predators; Nov. 2012 and July 2013 scrapers; Nov. 2012, May 2013, and Nov. 2013 Anthopotamus spp. Dark blue filled circles- Nov. 2012 filtering collectors. Blue filled circles- May 2013 filtering collectors. Red filled circles- May 2013 scrapers. Cyan filled circles- July 2013 filtering collectors. Maroon filled circles- Nov. 2013 scrapers. Light blue filled circles- Nov. 2013 filtering collectors. ... 108

Figure 3.5. ∆14C and equivalent 14C ages vs. δ13C values of macroinvertebrates and their potential nutritional resources across multiple sampling dates labelled by functional feeding group (light blue and dark blue filled circles). The δ13C values are corrected for trophic fractionation according to Post (2002)...... 109

Figure 3.6. Posterior distributions of the proportional contributions of A) aquatic algae, B) terrestrial vegetation and C) terrestrial soil OM to the biomass of macroinvertebrates in Paint Creek using the global data set. 5th, 25th, 50th, 75th, and 95th percentiles are shown...... 110

Figure 3.7. Mixing model results showing median percent nutritional source contributions to macroinvertebrates for A) combined May and July 2013 sampling dates, B) combined November 2012 and 2013 sampling dates, and C) the global model across all sampling dates with stable isotope (δ13C, δ15N, and δ2H) and natural radiocarbon (∆14C) values of macroinvertebrates and potential nutritional sources for which ∆14C data was available. See Methods for details...... 111

Figure 4.1. Natural abundance isotopes and typical ranges of values in natural waters. Fossil C and OM are defined by the lower detection limit of AMS (50,000 yrs B.P.). ∆14C values >~-50‰ indicate inputs of “bomb” 14C from thermonuclear weapons testing that peaked in the 1960s, or from direct modern day inputs from thermonuclear reactors. When both natural and anthropogenic sources of 14C to the Earth system are considered, the dynamic range of Δ14C is increased to nearly 2,000‰ over exclusively natural 14C sources...... 152

Figure 4.2. Potential sources and ages of allochthonous carbon (C) and organic matter (OM) found in aquatic systems and relative sizes of each reservoir. Units in parenthesis are 1018 g C. Adapted from Bauer and Bianchi (2011) and Hedges (1992)...... 153

Figure 4.3. Routes through which carbon (C) and organic matter (OM) of varying ages enter inland water ecosystems. Images in figure adapted from http://ian.umces.edu/. Sources: Bellamy et al. (submitted a, b, c), Chanton et al. (1995), Clark and Fitz (1997), Currie et al. (1997), Ishikawa et al. (2015), Pessenda et al. (1996), Petsch (2001), and Pohlman et al. (2009), Wozniak et al. (2011, 2012 a, b)...... 154

Figure 4.4. Potential pathways of aged particulate organic matter (POM) and dissolved organic matter (DOM) utilization by aquatic consumers...... 155

Figure 4.5. Literature ∆14C values and equivalent 14C ages for metazoan consumers of various inland waters. See Table 1 for details of systems and organisms analyzed...... 156

Figure A.1. A) Fraction of macroinvertebrate autochthony for FFGs across all model methods (1-4) for three-source model vs. two-source model, and B) comparison of autochthony for FFGs across all model methods (1-4) using three-source and two-source models. G- global model (Method 1), L Ag- low agriculture (Method 2), H Ag- high agriculture (Method 2), NC- Nowadega Creek (Method 3), FB- Fly Branch (Method 3). See Appendix A Section 2 for details...... 232

Figure A.2. Comparison of median proportional source contributions to macroinvertebrate nutrition determined from mixing models using A) estimated algal isotope value (δ13C, δ15N, δ2H) estimates and B) measured algal isotope values. These mixing model outcomes represent data from only the four sites from which algal samples were collected. See Appendix A Section 2 for details...... 233

xxi

Figure B.1. A) Fraction of autochthony (i.e., algal nutrition) for each functional feeding group for mixing models using two nutritional resources (algae, terrestrial vegetation, and soil/sediment) vs. three nutritional resources (algae and terrestrial vegetation). B) Comparison of 3- vs. 2-source mixing model outcomes for all functional feeding groups for all model scenarios. Abbreviations: CrCr-Crooked Creek, CR-Cowanesque River, ER-Elk Run, LC-Lamb’s Creek, LMC-Little Muncy Creek, NER-North Elk Run, TC- Towanda Creek, Low Ag- <50% agricultural land use in watershed, High Ag- >50% agricultural land use in watershed, and G- global model using δ13C, δ15N, δ2H, and ∆14C data...... 251

Figure C.1. A) Relationship between fraction autochthony (i.e., algal contribution) of each functional feeding group (FFG) for mixing model using two nutritional resources (algae and terrestrial vegetation) vs. three nutritional resources (algae, terrestrial vegetation, and soil OM). B) Comparison of 3 and 2 source mixing model outcomes for all FFGs for all model scenarios. MJ- May and July 2013 combined model; Nov- November 2012 and 2013 combined model; G- global model across all sampling times; 14C- model including stable isotope and ∆14C data across all sampling times...... 274

xxii

Introduction

The relative contributions of allochthonous (i.e., terrestrial) and autochthonous

(i.e., aquatic) nutritional resources to secondary production is an important emerging topic in our understanding of the functioning of aquatic food webs and ecosystems

(Minshall 1967, Fisher and Likens 1973, Junger and Planas 1994, Cole and Caraco 2001,

Roach 2013). In many aquatic systems, heterotrophic respiration exceeds gross primary production (GPP), supporting the contention that significant terrestrial organic matter

(OM) subsidies to heterotrophs, ranging from bacteria to apex consumers, support this net heterotrophy (Cole et al. 1994, 2007, Cole and Caraco 2001, Duarte and Prairie 2005).

The growing body of evidence for this contrasts with traditional models which suggest that secondary production is supported almost exclusively by autochthonous primary production in aquatic systems (Thorp and Delong 2002, Torres-Ruiz et al. 2007, Müller-

Navarra 2008). Rivers and streams transport and alter various forms of OM from terrestrial sources and upstream water bodies. Improved qualitative and quantitative understanding of OM sources and subsidies to consumers and food webs in aquatic systems is therefore needed to better understand their structure and function and to more fully constrain their associated carbon and energy flows.

In streams and rivers, the characteristics, reactivity, and ages of allochthonous carbon (C) and OM can vary significantly between systems due to differences in soil and vegetation types, as well as on specific geomorphologic and hydrologic features of the landscape (Zeug and Winemiller 2008, Hossler and Bauer 2013a, Lu et al. 2014). While it is well-established that aged and ancient forms of C and OM contribute to lotic waters and sediments (Meybeck 1993, Battin et al. 2009, Drenzek et al. 2009, Hossler and Bauer

2013a), we have far less understanding of i) whether these aged forms of C and OM are exploited by consumers, and ii) the quantitative contributions of aged allochthonous nutritional resources subsidizing stream and river food webs.

Stream and river C and OM have been found to range from modern- to geologically-aged (Raymond et al. 2004, Marwick et al. 2015). Moderately and highly aged C and OM is in fact more abundant in Earth reservoirs than contemporary OM by several orders of magnitude (Hedges 1992). Thus, mobilization of these old and ancient global reservoirs may contribute significantly to stream and river C and OM pools and their age distributions in these systems (Caraco et al. 2010, Hossler and Bauer 2013b).

The sources and ages of C and OM supporting lower aquatic food webs are important because subsidies of these materials have the ability to influence the magnitude of secondary production, food web stability, community structure (Polis et al. 1997), and consumer reproductive success (Brett et al. 2009). Because of its greater essential fatty acid content and lower C:N ratios (Torres-Ruiz et al. 2007, Brett et al. 2009, Guo et al.

2016), autochthonous OM has historically been assumed to contribute more to aquatic consumer diet than allochthonous OM. At the same time, allochthonous OM has traditionally been considered to be more recalcitrant and less bioavailable to consumers

(Torres-Ruiz et al. 2007, Brett et al. 2009). However, even relatively small contributions of allochthonous OM to the nutrition of organisms and food webs may shift the balance of aquatic metabolism and secondary production in streams and rivers from net autotrophy to net heterotrophy (Caraco and Cole 2004, Kritzberg et al. 2004, Tank et al.

2010). In addition, most previous studies have not considered microbial repackaging of

3 allochthonous OM and the possible contributions of this repackaged material to consumers (Tranvik 1992, Hall and Meyer 1998, France 2011). Overall, little research has been conducted to date to quantify aged C and OM nutritional resources to lower aquatic food webs – in spite of the dominance of aged allochthonous forms of C and OM in most streams and rivers. Streams and rivers are open systems that are tightly linked with the surrounding landscape (Cole et al. 2007, Aufdenkampe et al. 2011), and as a result their food webs may be strongly dependent on materials of varying ages transferred from terrestrial environments.

Macroinvertebrates play a number of important roles in the processing of autochthonous and allochthonous C and OM in aquatic systems (Cummins 1974, 2016,

Moulton et al. 2010). Macroinvertebrates are also responsible for the transfer of C and

OM to higher trophic levels, both in aquatic and terrestrial food webs (Sabo and Power

2002, reviewed in Baxter et al. 2005). Most stream macroinvertebrates are omnivorous and non-selective feeders, but limited previous research suggests that some collectors and shredders may select higher quality food sources (e.g. phytoplankton and leaves that are conditioned to a greater degree by microbes and fungi; Davies 1975, Arsuffi and

Suberkropp 1989, Guo et al. 2016). Morphological and behavioral adaptations may further control the forms and quality of nutritional resources ingested by an organism, and these adaptations may further impact the relative amounts of dietary autochthonous versus allochthonous OM (Merritt and Cummins 1996). Macroinvertebrates from different functional feeding groups have also been observed to compensate for low food quality by increasing the quantity of food consumed (House 1965, Schindler 1971, Cruz-

Rivera and Hay 2000). Low- or non-selectivity by macroinvertebrates may diminish their ability to select younger and more nutritious OM sources. Ingestion and assimilation of aged materials may be more common amongst filterers, collector-gatherers, and generalist consumers because they are able to exploit OM from a larger variety of sources

(Cummins and Klug 1979, Merritt and Cummins 1996, Mihuc 1997).

In order to assess the contributions of nutritional resources and their ages to macroinvertebrate biomass, the research conducted in this dissertation employed a multiple isotope approach using stable isotopes (13C, 15N, 2H) and natural abundance radiocarbon (14C) and Bayesian isotopic mixing models. The overarching goal of this dissertation was to assess the contributions of allochthonous and autochthonous C and

OM to macroinvertebrate diet and nutrition in temperate lotic systems, and to determine how the utilization of allochthonous OM varies as a function of: stream characteristics, land use and lithology in watersheds, and time. Additionally, I aimed to determine if any of these factors affected the assimilation of aged C and OM by macroinvertebrate consumers.

In Chapters 1-3 of this dissertation, potential nutritional resources and macroinvertebrate consumers from multiple temperate streams were collected and analyzed for stable isotopes and radiocarbon, in order to quantify the contributions of allochthonous and autochthonous C and OM of varying ages to organism biomass. In

Chapter 1, the roles of land use and lithology on the isotopic compositions of, and nutritional resources utilized by, stream and river macroinvertebrates were examined in the Mohawk-Hudson, New York watershed. Chapter 2 examined how nutritional

5 resource utilization by macroinvertebrates varied throughout a geographically extensive stream-watershed network in the middle and upper Susquehanna River watershed in northern central Pennsylvania. This chapter also evaluated the effects of land use, including a transient event in the form of natural gas seepage in a stream thought to be possibly due to extensive hydraulic fracturing activities in the region. In Chapter 3, variation in seasonal allochthonous and autochthonous nutritional resource utilization by macroinvertebrates over the course of a year were evaluated at a single site in Paint

Creek, Ohio. Finally, Chapter 4 consists of a comprehensive review and analysis of the entirety of known published studies that have used natural 14C as a food web source-age tracer in all inland water ecosystems (streams, rivers, lakes, and wetlands). The published

∆14C values and equivalent ages of aquatic metazoans were used to i) estimate the relative contributions of fossil, moderately and modern aged C and OM to the biomass of aquatic consumers across inland water systems and ii) provide an overview of whether the aged C and OM common to these waters is nutritionally available to consumers.

Chapter 1: Influence of Land Use and Lithology on Sources and Ages of Nutritional Resources for Stream Macroinvertebrates: A Multi-Isotopic Approach

Amber R. Bellamy

James E. Bauer

Andrea G. Grottoli

Submitted to Aquatic Sciences

Abstract

Terrestrially-derived organic matter (OM) – much of it of significant age - is prevalent in many streams and rivers, yet little is known about the potential for aged OM to contribute to aquatic consumers. The impacts of watershed characteristics (e.g., land use and lithology) on the sources and ages of OM utilized by aquatic consumers are also poorly understood. To assess these factors, aquatic macroinvertebrates belonging to multiple functional feeding groups (FFGs) were collected from six headwater streams having unique watershed lithology and land use in the Hudson-Mohawk River system and analyzed for natural δ13C, δ15N, δ2H, and ∆14C. A Bayesian isotopic-mixing model using δ13C, δ15N and δ2H revealed that autochthonous primary production was of greatest importance (62-92%) to the biomass of all FFGs across all sites. In addition, macroinvertebrates collected from streams with watersheds having >40% or < 20% agricultural land use were estimated to assimilate 4-31% and 0-13% soil-derived OM, respectively. ∆14C values for macroinvertebrates from a shale-rich and a shale-poor site were significantly different (mean ∆14C = -75‰ and -34‰ and equivalent 14C ages = 630 and 280 years B.P., respectively). While inclusion of ∆14C data in the mixing models confirmed the importance of autochthonous primary production, it also demonstrated that lithology indirectly influenced nutritional resource utilization by influencing stream substrate type, and potentially the retention of allochthonous OM. Findings from this study show that autochthonous OM was the primary source of nutrition to all types of macroinvertebrates and FFGs in our study systems, but the degree of allochthonous contributions was dependent on FFG, land use type, and lithology.

Introduction

Nutritional resource utilization by aquatic macroinvertebrates can vary as a function of morphological and behavioral adaptations, as well as the characteristics of the organisms’ habitats and the relative availability of different nutritional resources

(Cummins and Klug 1979, Vannote et al. 1980, Rosi-Marshall and Wallace 2002,

Wallace et al. 2015). Potential nutritional sources comprised of living and non-living organic matter (OM) can be generally classified as autochthonous (produced within the aquatic system, e.g., algae, macrophytes) or allochthonous (produced outside of the aquatic system, e.g., living and senesced leaves, woody debris) in origin (Vannote et al.

1980, McCutchan and Lewis 2002, Ishikawa et al. 2014). The morphological and ecological characteristics of macroinvertebrates have also been used in previous studies to predict the forms of OM contributing to macroinvertebrate nutrition (Vannote et al.

1980, Roach 2013, Cummins 2016).

The morphology of macroinvertebrate feeding appendages and behaviors are often used to classify these organisms by functional feeding group (FFG; Cummins and

Klug 1979, Cummins 2016). An organism’s FFG may provide insight into its utilization of different relative amounts and the quality of autochthonous and allochthonous forms of nutritional materials. Additionally, a number of conceptual models including the River

Continuum Concept (RCC), the Flood Pulse Concept (FPC) for river-floodplain systems, and the Riverine Productivity Model (RPM; Vannote et al. 1980, Junk et al. 1989, Thorp and Delong 1994, 2002) have attempted to explain the controls on nutritional sources contributing to lower trophic levels of stream and river food webs. Both the RCC and the

FPC models highlight the importance of allochthonous OM to the production of aquatic consumers, whereas the RPM stresses the importance of local autochthonous production.

Because resource availability in streams and rivers may fluctuate significantly as a function of hydrology (e.g., discharge) and human disturbances in the watershed (Power et al. 1996, Ledger et al. 2013, Roach 2013), it is unlikely that a “one-size-fits-all” model is appropriate to assess nutritional resource utilization by aquatic consumers.

Natural abundance stable isotopes (δ13C, δ15N, and δ2H) are increasingly used to provide quantitative estimates of autochthonous and allochthonous contributions to aquatic consumer nutrition (Bunn and Boon 1993, McCutchan and Lewis 2002,

Middelburg 2014). However, in some aquatic systems, there may be significant overlap in the δ13C and δ15N values of different sources of nutritional OM, making it difficult to accurately assess resource contributions to organism biomass (Moore and Semmens

2008, Phillips et al. 2014). δ2H is useful for establishing the contributions of terrestrial vs. aquatic OM to consumer biomass because of the large isotopic separation (~100‰) between the two (Doucett et al. 2007, Wilkinson et al. 2015).

In contrast to stable isotopes, natural abundance radiocarbon (14C) is a relatively new and underutilized tracer of carbon sources and nutrition in aquatic food web studies.

Natural 14C has a much greater dynamic range (> 1,000‰) compared to δ13C and δ15N that have maximum ranges of tens of ‰, and δ2H that has a maximum range of ~100‰.

In addition, unlike δ13C and δ15N, ∆14C values do not need to be corrected to account for trophic fractionation (Stuiver and Polach 1977). Natural abundance 14C also uniquely

10 allows for determination of the ages of potential nutritional resources utilized by consumer organisms.

Examination of the 14C ages of carbon (C) and OM pools in stream and rivers has revealed that, in many cases, highly aged carbon, such as soils, weathered sedimentary rock, groundwater, and both natural and anthropogenic petroleum-derived hydrocarbons, can contribute to the particulate and dissolved OM (POM and DOM, respectively) pools

(Raymond and Bauer 2001, Longworth et al. 2007, Hossler and Bauer 2013a, 2013b).

The presence of these aged materials in aquatic systems can be controlled by hydrological and geomorphological factors, as well as human activities (Longworth et al.

2007, Hossler and Bauer 2013b, Butman et al. 2015). Stream and river DOM typically has ∆14C values of slightly < 0‰ to ~ 0‰ (no more than century-aged to modern-aged), while POM generally has ∆14C values in the -1,000 to -100‰ range (fossil to millennial aged; Hossler and Bauer 2013a, Marwick et al. 2015). Aged aquatic OM therefore contains carbon that was fixed and stored in soils and/or sedimentary rocks for between decades to millions of years or more (Tourtelot 1979, Hedges 1992, Trumbore 1997) before mobilization to aquatic systems. To date, there is little known about the potential utilization of aged forms of C and OM by aquatic consumers, and whether or not such materials are of adequate nutritional quality for consumers.

In the present study, we evaluated the relative contributions of allochthonous and autochthonous sources of nutrition to stream macroinvertebrate consumers belonging to different FFGs in six geographically proximate temperate subwatersheds. We also examined whether watershed land use and lithology influenced the ages of nutritional

11 resources utilized by macroinvertebrate consumers. We hypothesized that allochthonous

OM would contribute more to the nutrition of shredder and collector-gatherer FFGs and chironomids, and that a greater proportion of autochthonous OM would contribute to filtering collector and scraper FFGs. We further predicted that stream macroinvertebrates from watersheds containing significant amounts of fossil shale OM would be depleted in

14C due to their assimilation of this material.

Methods

Site Description. Individual sites from six streams in subwatersheds of varying

2 size (21-149 km ; Table 1.1) were sampled in the Mohawk-Hudson River watershed, New

York, USA (Figure 1.1). These same sites were previously sampled by Longworth et al.

(2007), and were found to vary in the amounts of OM-rich shale and OM-poor shale in the lithology of their watersheds (Table 1.1). OM-rich and OM-poor shale refers to the content of total organic carbon (TOC) in the shale, with OM-rich shales containing 2-4%

TOC and OM-poor shales containing <0.5% TOC (Longworth et al. 2007). In our study system, the OM-rich shales are a part of the Utica shale formation, whereas OM-poor shales consisted of Grey and Frankfort shale, and Gneiss (Longworth et al. 2007). The six subwatersheds also varied over an order of magnitude in the relative amounts of agricultural activity (pasture and row crop) in each (6-65%; Table 1.1).

Field Sampling. Macroinvertebrates belonging to different FFGs (filtering collectors, scrapers, collector-gatherers, shredders, predators, and chironomids) and their potential nutritional sources were collected from each stream in June 2014. Organisms were collected by hand-picking them from rocks and logs and using a kick net where

12 sediments and aquatic vegetation dominated. Predators and primary consumers were separated from each other into stream water from the respective sites filtered through a 47 mm quartz fiber QMA filter, and allowed to void their guts for 24 hours at ambient temperature (Brooke et al. 1996). After gut voidance, organisms were immediately placed in baked (525° C) aluminum foil pouches and frozen on dry ice until processing in the lab.

The dominant potential nutritional sources to macroinvertebrates, including both aquatic and terrestrial OM, were also collected from the individual sites. Sediment samples were collected from the stream using a 60 ml syringe corer and immediately placed on dry ice and frozen. Where cleanly eroded stream banks were present within a sampling reach, terrestrial soil samples were collected by inserting syringe corers horizontally into the eroded stream bank at shallow (surface) and deep (~20 cm) depths.

When there was no eroded stream bank, terrestrial soil samples were collected by excavating a hole of ~20 cm depth on level ground within 10 m of the stream. Soil samples were collected at ~1 cm and ~20 cm depths using a baked spatula, and were wrapped individually in pieces of baked aluminum foil, placed in a Ziploc bag, and frozen on dry ice. When available, shale shards were also collected and frozen until analysis.

Terrestrial vegetation from riparian trees and aquatic vegetation samples were collected by hand using clean disposable nitrile gloves and preserved as for soil and sediment samples. Stream biofilm samples were collected from 2-3 cobbles of similar size (15-20 cm in diameter) by scraping the surface of a ~25 cm2 area with a toothbrush

13 or a baked razor blade and then rinsing the scraped material from the cobble surface using ultra-pure Labconco water. Rinsed material was collected in acid-cleaned (10%

HCl) polycarbonate bottles and frozen on dry ice. Bulk stream suspended POM was collected by filtering water from each sampling site through a baked 47 mm QMA filter

(0.8 µm nominal pore size).

Dissolved organic and inorganic C (DOC and DIC, respectively) were collected form each stream. DIC samples were processed by filtering water from each site through a 47 mm quartz fiber QMA filter into baked, crimp-sealed 125 ml serum bottles containing 200 μL of a saturated HgCl2 solution and sparged with ultra-high purity N2 gas. The DIC bottles were stored in the dark at ambient room temperature until processing. Filtered stream water (~250 ml) for both DOC and nutrients (N and P) was collected in acid-cleaned polycarbonate bottles and frozen at -20° C until processing. A sample of stream water from each site was also collected in a 20 ml baked scintillation

2 vial for  H analysis of H2O. Basic water chemistry data, including temperature, pH, dissolved oxygen (DO), and conductivity, were collected at each site using a Model

AP110 Accumet Portable pH/ORP meter, a YSI ProODO handheld DO meter, and a YSI

Model 30/10 FT multiparameter probe, respectively. Canopy cover was estimated at each site by using a convex spherical densiometer and taking the mean of nine measurements at different points within the reach.

Sample Processing and Analysis. Macroinvertebrates were sorted and identified to genus whenever possible and assigned to the appropriate FFG according to Merritt et al.

(2008; Table A.1). Chironomids were only identified to family (Chironomidae) and were

14 treated as such in the mixing models. They were not assigned to a specific FFG because chironomid FFG and trophic position can vary across subfamily and genus (Reuss et al.

2013). Organisms, terrestrial vegetation, and aquatic vegetation were dried and homogenized to a fine powder in preparation for stable isotope and radiocarbon analysis.

In some cases, multiple individuals of small macroinvertebrates were pooled to increase sample sizes for isotopic (especially 14C) analysis (A1.1). Aquatic vegetation samples, biofilm, terrestrial soil, aquatic sediment, shale samples, and POM filters were acid- fumed with fresh concentrated HCl in a clean glass desiccator prior to homogenization to ensure that carbonates were removed. A subsample of each sample type, with the exception of POM filters and biofilms, was set aside prior to acid fuming for δ2H analysis. POM and biofilm samples could not be analyzed for δ2H because of interference from the quartz filters.

Homogenized samples and filter portions were packed in tin capsules for δ13C and

δ15N analysis. δ13C and δ15N analyses were conducted at the University of California at

Davis Stable Isotope Facility, using a PDZ Europa ANCA-GSL elemental analyzer interfaced to a PDZ Europa 20-20 isotope ratio mass spectrometer. Stable isotope values for δ13C and δ15N are reported relative to the V-PDB and air international standards, respectively. For δ2H analysis, homogenized samples of macroinvertebrate, aquatic and terrestrial vegetation, soil, and sediment were packed in silver capsules. Water samples were placed in baked scintillation vials for bulk δ2H analysis. In order to determine δ2H of the nonexchangeable H fraction for solid samples, a bench-top equilibration method was used to allow for the exchange of H in the samples with H in the local water vapor

(Wassenaar and Hobson 2003, Doucett et al. 2007). Solid samples were analyzed for δ2H on a Thermo-Finnigan TC/EA and DeltaPLUS-XL and stream water samples were analyzed using a Los Gatos Research DLT-100 Liquid Water Isotope Analyzer at the

Colorado Plateau Stable Isotope Laboratory at Northern Arizona University.

Selected homogenized and acid-fumed samples and filters from two of the streams (Fly Branch and Nowadega Creek) were placed in pre-baked quartz tubes with

14 Cu and CuO and combusted to CO2 at 750º C for 4 hours for  C analysis. Within 24 hours of combustion, CO2 from each sealed tube combustion was purified and quantified on a vacuum extraction line and sealed in pre-baked 6 mm Pyrex tubes. The purified CO2 was reduced to graphite and analyzed for Δ14C at the National Ocean Sciences

Accelerator Mass Spectrometry (NOSAMS) Laboratory at Woods Hole Oceanographic

Institution. DIC samples were acidified and sparged using ultra-high purity He gas and the CO2 was collected and purified cryogenically on a vacuum extraction line, and stored and analyzed for 14C as above. Dissolved inorganic N and P nutrient concentrations were analyzed using a Lachat QuickChem 8500, and DOC concentrations were analyzed using a Shimadzu TOC Analyzer 5000.

Statistical Analyses. Multivariate statistical analyses were completed using the

PRIMER software package (Clarke and Gorley 2006; v. 6, PRIMER-E Ltd) in order to assess relationships between organism isotope data from the different sampling sites and

FFGs. Macroinvertebrate δ13C, δ15N, δ2H, and ∆14C values were normalized and evaluated using non-metric multidimensional scaling (NMDS). One-way analysis of similarity (ANOSIM) was completed for macroinvertebrate isotopic data in order to

16 determine whether significant isotopic differences existed between a) individual sites, b)

FFGs, c) degree of agricultural land use in each watershed (high: >40% agricultural land use; low: <20% agricultural land use) and d) the presence or absence of OM-rich shale in watershed lithology. Agricultural land use in the study watersheds fell into two categories: <20% and >40% (Table 1.1). The ANOSIM pairwise test statistic, R, is considered a better indicator of separation between groups than the p value because it is not influenced by sample size, which was relatively small in this study (Clarke and

Gorley 2006). In cases where the overall p values for the ANOSIM model were significant (p<0.05), all pairwise comparisons within the model were examined including those with a p>0.05, but the global R was greater than 0.8.

Isotopic Mixing Models. Proportional contributions of potential nutritional resources to macroinvertebrate biomass were estimated using MixSIAR (Stock and

Semmens 2013), an isotopic mixing model that employs a Bayesian approach (Moore and Semmens 2008) via a graphical user interface (GUI) and R Statistical Software (R

Core Team, 2014). Use of a Bayesian approach allows for the incorporation of uncertainty in source contribution estimates, as there are multiple sources of variation that could impact source contribution estimates. Some of the sources of variability include, but are not limited to, use of multiple nutritional sources by consumers, isotopic fractionation, and spatial and temporal variability of isotopic signatures (Finlay et al.

2002, Moore and Semmens 2008). Details of the isotopic signatures of the potential nutritional sources used in the models, trophic fractionation factors used, the methods

17 associated with the Bayesian model, and specific information about how the models were run are detailed in Appendix A, Section 2.

Results

δ13C, δ15N, δ2H, and ∆14C Values of Macroinvertebrate Consumers

Stable and radiocarbon isotopic data for all macroinvertebrates and FFGs measured in the present study are provided in Appendix A, Section 3. Scraper FFG macroinvertebrates were generally lower in δ13C and δ2H than other FFGs across all sites, with the exception of Fly Branch (Figure 1.2A, B, Table A.9). In contrast, chironomids, collector-gatherers, and shredders were elevated in δ13C and δ2H compared to other FFGs at all sites (Figure 1.2A, B, Table A.9). Filtering collectors from Otsquago Creek were more 13C- and 2H-depleted relative to filtering collectors for the other sites. Predators were the most elevated of all FFGs in their δ15N values (Figure 1.2A, Table A.9). NMDS analysis did not reveal any obvious clustering by FFG (Figure 1.2C). However, one-way

ANOSIM analysis revealed significant differences in composite isotopic compositions

(δ13C, δ15N, and δ2H) between all FFGs (Table A.10A). Significant differences were found between almost all pairwise comparisons of FFG isotopic composition.

When organized by site, stable isotope data showed that Nowadega Creek macroinvertebrates were highest in δ15N, whereas those collected from the Schoharie

Tributary were lowest in δ15N (Figure 1.2D). Couch Hollow macroinvertebrates were generally lower in δ13C compared to those collected from the other five sites (Figure

1.2D, E). In contrast to δ13C and δ15N, macroinvertebrate δ2H values did not show any site-specific patterns (Figure 1.2E), however, NMDS of δ2H values revealed some 18 clustering of macroinvertebrates by site (Figure 1.2F). One-way ANOSIM also indicated significant differences in macroinvertebrate isotopic compositions between sites, and all pairwise site comparisons were significant (Table A.10B).

When macroinvertebrate stable isotopic data were pooled by the amount (high vs. low) of agricultural activity in the watershed, macroinvertebrates collected from streams with high agriculture were more 15N-enriched than those macroinvertebrates collected from streams with low agriculture (Figure 1.3A). There was no apparent 13C- or 2H- depletion or enrichment in macroinvertebrate biomass with respect to the amount of agricultural activity in the watersheds (Figure 1.3A, B). NMDS showed that macroinvertebrate individuals clustered according to high vs. low agriculture in their respective watersheds (Figure 1.3C), and the differences between macroinvertebrate stable isotopic compositions from the high and low agriculture watersheds were significant (R=0.251, p<0.001).

Macroinvertebrate ∆14C values at Fly Branch ranged from -59‰ to 23‰

(equivalent 14C age = 430 years B.P. to modern-aged, respectively) (Figure 1.4A, Table

A.9), with the most 14C-enriched macroinvertebrate being a shredder. Macroinvertebrate

∆14C values at Nowadega Creek ranged from -114‰ to -1‰ (910 years B.P. to modern- aged, respectively) (Figure 1.4A). Inclusion of ∆14C data in NMDS for the two sites for which it was available showed further separation between sites both with and without

OM-rich shale in their watersheds (Nowadega Creek and Fly Branch; R=0.572, p<0.001)

(Figure 1.4B).

Estimation of Source Contributions to Macroinvertebrate Nutrition

Mixing model estimates of nutritional source contributions (5%, 50%, and 95% posterior probabilities) for the different FFGs are provided in Section 4 of Appendix A.

Model results using mean δ13C, δ15N, and δ2H values of macroinvertebrate FFGs indicated that algae were the primary nutritional resource for all FFGs in the streams studied, contributing 63-92% of macroinvertebrate biomass (Figure 1.5; Table A.11).

However, soil and terrestrial vegetation (i.e., allochthonous OM) also comprised up to

21% and 31% of consumer biomass, respectively (Figure 1.5, Table A.11).

When organisms were grouped as a function of percent agriculture in the watershed, mixing model results again indicated that algae were the largest contributor to macroinvertebrate biomass, particularly in the sites with <20% agriculture (Figure 1.6).

However, soil OM contributed up to 31% of macroinvertebrate biomass in organisms collected from the >40% agriculture streams but never more than 13% in organisms from the <20% agricultural streams (Figure1. 6, Tables A.12 and A.13).

Inclusion of 14C data in the mixing model for the two sites for which it was measured indicated a slight decrease (18%, on average) in the assimilation of algae by macroinvertebrates compared to models using only stable isotopes, especially for Fly

Branch, which did not contain OM-rich shale in its watershed (Figure 1.7A, B, Table

A.14, A.15). Algae were still the primary nutritional resource for all FFGs (39-93%) at both sites, with the exception of collector-gathers at Fly Branch, which primarily assimilated terrestrial vegetation (55%) (Figure 1.7B). When natural 14C was used in the mixing models, there was also an increase in the importance of terrestrial vegetation to

FFG biomass, with a subsequent decrease in the importance of soil-derived OM at both

Nowadega Creek and Fly Branch, with terrestrial vegetation contributing more (28-55%) than soil OM (0-21%) to macroinvertebrate biomass at Fly Branch (Figure 1.7A). Soil

OM was also still an important nutritional resource for predators (21% at Fly Branch and

13% at Nowadega Creek), but contributed less than 5% to all other FFGs at these two sites (Table A.14, A.15).

Discussion

The origins and characteristics of nutritional resources, in conjunction with consumer morphological adaptations, may be primary drivers of the absolute and relative amounts of OM that can be utilized, respired, and assimilated by aquatic ecosystems

(Cummins and Klug 1979, Cummins 2016). Macroinvertebrates play a number of important roles in stream ecosystems, including the processing and decomposition of both aquatic and terrestrial OM and support of higher trophic level aquatic and riparian consumers (Baxter et al. 2005, Prather et al. 2013, Carlson et al. 2016, Kautza and

Sullivan 2016). Macroinvertebrate communities are also indicators of overall stream

“health” (Cairns and Pratt 1993, Bunn et al. 1999). The primary goals of the present study were to assess the sources and ages of C and OM contributing to the biomass of macroinvertebrate FFGs, and how contributions of C and OM varied as a function of land use and lithology.

Contributions of Autochthonous vs. Allochthonous Primary Production to Stream

Consumers

Considering all mixing model scenarios, with the exception of collector-gatherers from Fly Branch (Figure 1.7B), stream algae were the dominant contributor to macroinvertebrate biomass (39-99%). This finding was contrary to what was initially predicted because the streams we sampled were small and in small subwatersheds of the

2 th Hudson-Mohawk basin (all streams had watershed areas < 150 km and were < 4 order)

(Table 1.1), and the RCC predicts that allochthonous OM fuels secondary production in headwater and low order streams (Vannote et al. 1980). Findings from the present study support recent research that has reevaluated traditional paradigms about the importance of autochthonous OM to macroinvertebrate nutrition in streams and rivers (Torres-Ruiz et al. 2007, Guo et al. 2016, Hayden et al. 2016, Rosi-Marshall et al. 2016). Autochthonous aquatic primary production is often of higher quality (e.g., lower C:N and higher essential fatty acid [EFA] content) than allochthonous OM (Torres-Ruiz et al. 2007, Müller-

Navarra 2008, Guo et al. 2016). High-quality nutritional resources are important for adequate growth and reproduction in macroinvertebrates (Torres-Ruiz et al. 2007), and macroinvertebrates may be selective in what they consume and assimilate (Goedkoop et al. 2007, Guo et al. 2016).

Studies have shown that filtering collectors and scrapers can selectively consume and assimilate algal material (Finlay 2001, McNeely et al. 2006, Rasmussen 2010). In our global model using consumer and nutritional resource stable isotope signatures across all sites, scraper and filtering collector biomass was predominantly (≥85%) comprised of

22 algae (Figure 1.3, Table A.11). Inclusion of 14C data into the mixing models also suggested that algae contributed the most to filtering collector biomass (Figure 1.7,

Tables A.14 and A.15). Algae were also the most important nutritional resource to predatory macroinvertebrates in all mixing model scenarios, suggesting that autochthonous nutritional resources may be more readily transferred to higher trophic levels than allochthonous resources (Figures 1.5, 1.6, and 1.7). Previous studies have also shown that fatty acid abundances in aquatic consumers increase with increasing trophic level, which further suggests that some macroinvertebrate predators may have the ability to select higher quality prey (Lau et al. 2014, Guo et al. 2016).

In spite of the general dominance of algal material to most of the aquatic consumers in the present study, terrestrial vegetation still contributed measurably, and at times significantly, to macroinvertebrate biomass (Figures 1.5, 1.6, and 1.7) and therefore cannot be ignored in macroinvertebrate nutritional and energy budgets. The Bayesian mixing model output indicated that terrestrial vegetation made up 55% of collector- gatherer biomass in Fly Branch. Collector-gatherers in Fly Branch consisted of gammarid amphipods and Ephemera mayflies, which most likely consume fine particulate OM composed of fragmented terrestrial vegetation and phytoplankton (Cummins and Klug

1979, Cummins 2016). Based on feeding experiments and determination of leaf litter breakdown rates, as well as gut content analysis, terrestrial vegetation is known to be a significant component of the diets of both gammarids and ephemerid mayflies (Hamilton and Clifford 1983, Arsuffi and Suberkropp 1989, Piscart et al. 2009). Shredding macroinvertebrates are also generally thought to depend on terrestrial vegetation,

23 however, other stable isotope studies have shown algae to contribute significantly to their biomass (Leberfinger et al. 2011, Neres-Lima et al. 2016). Unfortunately, there were not enough shredder individuals collected at our study sites to accurately assess nutritional resource utilization by this FFG.

Macroinvertebrates in the present study measured for natural 14C content at two of the sites (Fly Branch and Nowadega Creek) were found to have equivalent 14C “ages” ranging from 970 years B.P. to modern (∆14C = -114‰ to 23‰) (Figure 1.4A, Table

A.9). Mixing models that included both stable isotopes and 14C revealed that algae were still the dominant nutritional resource to filtering collectors, collector-gatherers, and predators (42-92%), with the exception of collector-gatherers from Fly Branch (Figure

1.7A, B). Other studies that have employed natural 14C in stream and river food web studies have observed equivalent organism ages ranging from ~2,000 years B.P. to modern aged (∆14C = -240 to 68‰; Caraco et al. 2010, Ishikawa et al. 2014, 2016).

While some of these studies contend and/or demonstrate that aged 14C-depleted consumer biomass is dependent on assimilation of significantly aged OM from both allochthonous and autochthonous sources, observations of 14C depletion in aquatic consumers may also be due to the co-occurring effect of consumer utilization of “aged” algae that fix 14C- depleted DIC and CO2(aq) (Bellamy and Bauer, submitted, Ishikawa et al. 2014, 2016).

Based on our mixing model results and the similarity in 14C values of the DIC, algae, and macroinvertebrates, this is likely the case for at least some of our study sites (Figure

1.4A). There are a number of sources of CO2 that may contribute to stream DIC, including atmospheric CO2 exchange, weathering of carbonate rocks, and respired soil

24 and sediment OM, with the most aged sources being carbonate rocks and respired soil

CO2 (Broecker and Walton 1959, Butman and Raymond 2011, Keaveney and Reimer

2012, Ishikawa et al. 2016). In either case, the incorporation by aquatic primary and secondary producers of aged (in many cases, highly aged – see review by Bellamy and

Bauer, submitted) inorganic C and/or OM alters our understanding of terrestrial-aquatic linkages and the sources and ages of C and OM supporting aquatic food webs.

Influence of Agriculture on Nutritional Resource Utilization by Macroinvertebrates

Algal material comprised nearly the entirety of stream collector-gatherer and scraper biomass (≥96%) (Figure 1.6B, Table A.13) in low agriculture watersheds according to our mixing models. While filtering collectors, scrapers, and predators from high agriculture streams were still primarily reliant on algae (≥76%) (Figure 1.6A, Table

A.12), there were also significant contributions from allochthonous OM, both as soil- derived OM and terrestrial vegetation (4-31%) (Figure 1.6A, Table A.12).

In streams in high agriculture watersheds, mixing model outcomes indicated that macroinvertebrates were more reliant on soil-derived OM than macroinvertebrates from low agriculture streams (Figure 1.6). The proportional contribution of soil OM was greatest for high agriculture collector-gatherers (up to 31%), compared to <1% in low agriculture streams (Tables A.12 and A.13). Agricultural activity is known to increase exports of so-called “slow” (modern or near-modern in age) and “passive” (103-104 years

B.P. in age) turnover soil OM and associated mineral particles to streams and rivers

(Matthaei et al. 2010, Hossler and Bauer 2012, 2013a, 2013b, Burdon et al. 2013). These

25 inputs will concomitantly limit aquatic primary production due to increased turbidity

(Allan 2004, Roach 2013).

Collector-gatherers (Ephemerella mayflies) in the Schoharie tributary, a stream with low watershed agriculture, were much more reliant (96%; Figure 1.6B, Table A.13) on algae than in a high agriculture stream (61% in Nowadega Creek; Figure 1.6A), which could be attributed to the higher quality habitat in streams less impacted by agriculture

(Lenat 1984, Lenat and Crawford 1994). Increased sedimentation in our streams with greater agricultural impacts may have shifted nutritional resource utilization by macroinvertebrates inhabiting the sediment-water interface such as Ephemerella mayflies from autochthonous to allochthonous OM such as terrestrial vegetation and soils (Yule et al. 2010). The decreased importance of autochthonous OM by this group could potentially be explained by i) sedimentation inhibiting light availability and aquatic primary production, ii) increased scouring and removal of aquatic autotrophs (Horner et al. 1990, Henley et al. 2000, Madsen et al. 2001, Hall et al. 2015), and/or iii) increased availability and consumer utilization of soil OM itself (Wang et al. 2014).

Soil OM may also be incidentally ingested by aquatic consumers as they feed on algae or terrestrial vegetation. However, because soil OM is generally considered to be of lower nutritional quality due to its low C:N ratio and molecular composition (Kleber

2010, Kleber and Johnson 2010), it may be less likely to be directly selected by consumers. Recent reevaluations of the reactivity and bioavailability of soil OM in aquatic systems (Schmidt et al. 2011, Marín-Spiotta et al. 2014) suggest that heterotrophic bacteria and fungi may facilitate utilization of soil OM by

26 macroinvertebrates (Hall and Meyer 1998, Williams et al. 2010, Wang et al. 2014, Weber et al. accepted). Microbial processing and “repackaging” of soil OM may increase its bioavailability to macroinvertebrate consumers, as soil microbes have a lower C:N (~4-9) than soil OM itself (8-25; Finlay and Kendall 2007). Our findings suggest that soil may increasingly support macroinvertebrate biomass where inputs of soil OM from the watershed are greater (e.g., higher agricultural land use). This further implies that human alteration of watersheds and catchments may play a direct role in the sources and characteristics of C and OM supporting aquatic food webs (Lu et al. 2014, de Castro et al.

2016, Docile et al. 2016).

Influence of Lithology on Nutritional Resource Utilization

The use of 14C data at Nowadega Creek (containing significant amounts of shale in its watershed) and Fly Branch (containing little to no shale in its watershed) also revealed significant isotopic separation of macroinvertebrates collected from these two sites (Figure 1.4). Mixing model results suggest that terrestrial vegetation and soil OM were of greater nutritional importance to Fly Branch than to Nowadega Creek macroinvertebrates (Figure 1.7A). We initially predicted that the presence of OM-rich shale in a watershed would influence nutritional resource utilization by potentially providing a more highly aged source of OM (Petsch 2001, Schillawski and Petsch 2008).

This prediction was based on previous work in the main stem of the Hudson River that found significant 14C-depletion in zooplankton (mean 14C = -240‰, equivalent age of

2,000 years B.P.), suggesting that they assimilated 57% moderately aged soil OM (∆14C=

-350%, equivalent age of 3,460 years B.P.) or 21% fossil aged OM (Caraco et al. 2010).

While macroinvertebrates collected from Nowadega Creek were more 14C-depleted than those from Fly Branch (Figure 1.4A), on the basis of their 14C values it is unlikely that they were assimilating shale-derived OM to any significant extent. This is further supported by the fact that algae collected from Nowadega Creek were also more 14C- depleted than those collected from Fly Branch (4C = -114 ± 23‰ and -73 ± 3‰, respectively; equivalent of ~900 years B.P. and ~550 years B.P., respectively).

The differences in nutritional resource utilization by macroinvertebrates at

Nowadega Creek and Fly Branch are most likely a function of the physical characteristics of each stream rather than the presence or lack of shale in the stream watershed. For example, the reach sampled at Nowadega Creek had far less canopy cover (3%) than the reach sampled at Fly Branch (66%; Table 1.1), and this rather than the relative dominance of fossil shale OM may explain the increased utilization (21% greater) of terrestrial vegetation by collector-gatherers and filtering collectors in Fly Branch (Figure

1.7; Table A.14, A.15) (England and Rosemond 2004, Doi et al. 2007, Collins et al.

2015).

Internal characteristics such as streambed substrate can also influence nutritional resource utilization by aquatic consumers (Sullivan 2013, Smits et al. 2015). In the present study, the size and type of substrate in our different stream systems varied (Table

A.16). Streambeds in shale streams were dominated by large unbroken rock outcrop surfaces and small boulders, whereas substrate in non-shale streams were mixed in composition (Table A.16). Large unbroken rock streambed surfaces in shale streams may limit the retention of terrestrial OM and detritus in these systems, consequently reducing

28 the inputs and availability of allochthonous OM to consumers and increasing the assimilation of aquatic primary production (Walters et al. 2007, Smits et al. 2015,

Wallace et al. 2015). Therefore, the presence of shale in our watershed may have indirectly influenced nutritional resource utilization by macroinvertebrates in our study.

Factors Affecting the Availability of Soil and Shale-derived OM to Macroinvertebrates

A previous study in the same six subwatersheds of the Hudson-Mohawk system by Longworth et al. (2007) suggested that both land use and lithology were important but independent factors influencing stream suspended POM composition and 14C values.

We measured 14C from only Nowadega Creek and Fly Branch samples (Figure 1.4;

Tables A.7, A.8, A.9). Both streams were in high agriculture watersheds, preventing us from evaluating the relative influences of lithology and agriculture separately on nutritional resource availability and utilization by macroinvertebrates. However,

Longworth et al. (2007) used linear mixing models to assess source contributions to the

POM pool at Nowadega Creek and found that both shallow and deep soils had greater median proportional contributions (52% and 42%, respectively) to the POM pool than

14C-free (i.e., fossil aged) shale OM (6%, respectively; Longworth et al. 2007).

For Nowadega Creek, we employed a mass balance approach using the C:N and

∆14C values of algae/phytoplankton, soil, shale, and terrestrial vegetation (Table A5.2) to estimate the contribution of each to POM. Details are provided in Section 5 of Appendix

A. Our estimates suggest that modern terrestrial vegetation contributed undetectably (0%) to the POM pool, but that POM contained roughly equal contributions (~50:50) from algae and soil OM from 20 cm depth (Table A5.2). However, if it is reasonably assumed

29 that soil OM in Nowadega Creek consisted of a mixture of surface and 20-cm-depth material, then the contribution of soil OM to the POM pool alone is not 14C-depleted enough to explain the 14C-depletion observed in the POM (Table A5.2). If we use fossil shale as the aged end member instead of soil OM, mass balance again showed that terrestrial vegetation did not contribute (0%), algae comprised an even greater proportion

(81%), and shale-derived OM contributed 19% to suspended POM (Table A5.2).

Contributions of 20-cm-depth soil-derived OM, shale, or both to the POM pool are required to explain the ∆14C values of POM from Nowadega Creek, as algae alone was not 14C-depleted enough to explain the 14C-depletion in the POM.

Different mechanisms of soil- and shale-derived OM transport to aquatic systems may explain the lower-than-expected contribution of shale OM to suspended POM in

Nowadega Creek, and the contributions of soil and shale OM to macroinvertebrate biomass. Tilling associated with agricultural activities increases the erodibility and losses of soils, facilitating inputs of aged soil OM into stream ecosystems (Lal 2003, Longworth et al. 2007, Boix-Fayos et al. 2009). In contrast, shale-derived OM may be less readily mobilized to stream POM, even when shale outcrops are present due to it being more physically stable than soil-derived OM. Increased flow may be necessary to mechanically weather shale, and in some lotic systems, mechanical weathering may be primarily responsible for the input of shale-derived OM (Leithold et al. 2006, Hilton et al. 2011,

Graz et al. 2012).

Chemical weathering of shales is thought depend on the specific form(s) of kerogen in them (Durand 1980, Zhu et al. 2015). There is also experimental evidence that

30 solubilization of OM releases fossil DOM to aqueous systems (Schillawski and Petsch

2008). However, only two (from Otsquago Creek) out of a total of twenty-four ∆14C values of DOC reported from across all streams sampled by Longworth et al. (2007) were even modestly 14C-depleted (∆14C = -40 and -22‰; 330 and 180 years B.P. respectively;

Longworth et al. 2007). Therefore, the input of fossil shale-derived DOM to these and similar systems through chemical weathering is unlikely to be significant. Because of the physical and chemical constraints imposed on the weathering of OM-rich sedimentary rocks, we suggest that any aged allochthonous OM that may subsidize macroinvertebrate nutrition was probably derived from aged “passive” soil OM, with agricultural activity in the surrounding watershed facilitating its movement into adjacent streams.

Use of 14C as a Tracer in Stream Food Webs

Previous aquatic food web studies employing natural abundance 14C found that highly aged forms of OM, such as weathered sedimentary rock, soils, and peat, can significantly contribute (up to 57%) to the biomass of aquatic consumers (Schell 1983,

Caraco et al. 2010, Wang et al. 2014). Fossil shale-derived OM potentially contributed

~20% to stream suspended POM (Table A5.2); this estimate is larger than that of

Longworth et al. (2007), yet we have no direct evidence that a source of fossil shale- derived OM contributed to macroinvertebrate nutrition. However, one or more sources of non-fossil, but still aged, C (weathered carbonates and/or respired soil CO2) contributed up to 87% to macroinvertebrate biomass via fixation of aged DIC by aquatic primary producers and their subsequent consumption and assimilation. We also found that

31 agricultural activity and stream geomorphology (e.g., substrate size/type) may have been more important controls on the availability of C and OM to macroinvertebrate nutrition.

The synthesis of living, 14C-depleted autochthonous OM in aquatic systems has been observed in other lentic and lotic systems (Ishikawa et al. 2013, 2014, Keaveney et al. 2015). Findings from the present study indicate that aged forms of carbon (both organic and inorganic) can be active components of contemporary stream food webs.

Natural abundance 14C can be an important independent tracer in aquatic food web studies, and, as in the case of our study and others, reveals that modern aquatic food webs can be supported by carbon and OM that was formed thousands to millions of years ago

(Bellamy and Bauer, submitted).

Based on the findings of the present study and previous studies that have used 14C in aquatic food webs, we predict the following about the sources of nutrition supporting stream macroinvertebrate biomass: i) when streambed and immediate watershed substrates are dominated by shale bedrock, the importance of allochthony by macroinvertebrates is reduced (Figure 1.8A), ii) a positive relationship may exist between the presence of aged rock (sedimentary or carbonate) and the similarity between ∆14C values of macroinvertebrates and stream DIC, which is best explained by macroinvertebrates assimilating aged autochthonous production that has fixed aged DIC

(Figure 1.8B), and iii) a causative relationship may exist between the amount of agriculture in a watershed and the contributions of aged soil-derived OM to aquatic consumer biomass (Figure 1.8C; Ishikawa et al. 2014, 2016, Wang et al. 2014). Further

32 studies should strive to test the validity of these relationships between aquatic consumer nutritional resources and watershed lithology and land use.

Future studies that employ natural 14C as a tracer in aquatic food webs face challenges in assessing and quantifying organism 14C-depletion and apparent age as a result of actual utilization of one or more forms of aged OM (soil or sedimentary rock) and/or to utilization of apparently aged algae. In this regard, use of compound-specific isotope analysis (CSIA) could provide better differentiation between aged allochthonous and aged autochthonous OM (Ishikawa et al. 2015, Kruger et al. 2016). Finally, greater consideration should be given to the role of internal characteristics of streams (e.g, substrate type) in controlling the sources and ages of C and OM available to consumers, in addition to watershed-scale factors.

Table 1.1. Site and watershed characteristics of the six study streams in the Mohawk-Hudson River watershed in June 2014.

Site Stream Stream Watershed pH Temp. DO OM-Rich OM-Poor Agriculture Canopy Orderb Width Area (km2)d (ºC) (mg/L) Shale Shale by area Cover (m)c (%)d (%)d (%)d (%) Nowadega 3 25.3 82 8.22 20.6 9.7 45 44 49 3 Creek (NC)a Fly Branch 1 5.8 31 8.14 17.4 8.7 0 7 46 66 (FB) Couch Hollow 1 5.7 24 6.77 17.4 8.8 0 0 6 92 Branch (CH) Schoharie 2 7.9 21 8.31 15.9 9.6 0 95 19 6 Tributary (ST) Wampecack 4 8.5 50 8.33 18.9 9.8 0 8 61 0 Creek (WC) 34 Otsquago 4 17.7 149 8.20 21.0 9.5 55 37 65 7 Creek (OC) DO - dissolved oxygen aAbbreviations of sites are given in parentheses after each site name bStrahler stream order determined using the U.S. Geological Survey National Hydrography Dataset cWidth of stream at samping location measured directly dFrom Longworth et al. 2007

-100 8 Chironomids A B C Collector-gatherers -120 Filtering collectors 6 Scrapers -140 Predators Ter. veg. -160 4 Soil Algae- high ag. -180 Algae- low ag. 2 -200

H (‰)

N (‰) 2

0  -220

 -240 -2 -260 -4 -280 -40 -38 -36 -34 -32 -30 -28 -26 -24 -36 -34 -32 -30 -28 -26 -24 -22 -100 CH D E F 36 8 FB -120 NC

6 OC -140 ST WC -160 4 Ter. veg. Soil -180 Algae- high ag. 2 Algae- low ag. N (‰) -200

H (‰)

  0 -220 -240 -2 -260 -4 -280 -40 -38 -36 -34 -32 -30 -28 -26 -24 -36 -34 -32 -30 -28 -26 -24 -22 13C (‰) 13C (‰)

Figure 1.2. δ13C, δ15N, and δ2H values of macroinvertebrates and their potential nutritional sources collected from the six study streams in the Mohawk-Hudson River watershed (mean ± SD). A) δ13C vs. δ15N coded by FFG, B) δ13C vs. δ2H coded by FFG, C) δ13C, δ15N, and δ2H NMDS results with all macroinvertebrate individuals across all six sites coded by FFG; Stress: 0.11, D) δ13C vs. δ15N coded by site, E) δ13C vs. δ2H coded by site, F) δ13C, δ15N, and δ2H NMDS results for all macroinvertebrate individuals across all six sites coded by site; Stress: 0.11. Values in panels A, B, D, and E are corrected for trophic fractionation (Post 2002) and dietary water (Wilkinson et al. 2015). Axes for panels36 C and F are dimensionless (see text for details). CH-Couch Hollow, FB-Fly Branch, NC-Nowadega Creek, OC- Otsquago Creek, ST-Schoharie Tributary, and WC-Wampecack Creek. 10 -80 A -100 B C -120 5 -140 -160 0 -180

N (‰) H (‰) -200

2 15 >20% agriculture



 <20% agriculture -220 -5 Ter. veg. Soil -240 Algae- high ag. -260 Algae- low ag. -280 -36 -34 -32 -30 -28 -26 -24 -22 -36 -34 -32 -30 -28 -26 -24 -22 13C (‰) 13C (‰)

A B 0 0

1,000 Years BP

-200

C (‰)



-600 Fly Branch (no shale) 7,000 Nowadega Creek (shale)

38 Algae

Terrestrial vegetation -800 Soil 13,000

-36 -34 -32 -30 -28 -26 -24 -22 13C (‰)

100 A 80

Percent Contribution Percent 20

scraper predator chironomid

filtering collector 100 collector-gatherer

B 80

Percent Contribution Percent

scraper predator Algae Ter Veg filtering collector collector-gatherer* Soil

100 A

Percent contribution 20

predator

collector-gatherer filtering collector

100 B

Percent contribution 20

predator Algae Ter Veg collector-gatherer filtering collector Soil

Chapter 2: Contributions of Allochthonous and Aged Carbon and Organic Matter to the Nutrition of Macroinvertebrates in a Headwater Network of the Susquehanna River

Amber R. Bellamy

James E. Bauer

Andrea G. Grottoli

Prepared for submission to Freshwater Biology

Abstract

Nutritional resources to macroinvertebrates across seven streams in a headwater network of the upper and west branches of the Susquehanna and Chemung River basins,

Pennsylvania, were assessed using stable isotopes (δ13C, δ15N, and δ2H) and radiocarbon

(∆14C) in order to quantify the sources and ages of carbon (C) and organic matter (OM) contributing to macroinvertebrate biomass. Study streams varied in order and the amount of agricultural land use in their associated watersheds. Autochthonous OM contributed the most to macroinvertebrate biomass of multiple functional feeding groups (44-89%) across most sites, with the exception of some predators and shredders in lower order streams that were more reliant on allochthonous OM. Macroinvertebrates from streams with greater (>50% of land use) agricultural land use in their watershed were more 15N- enriched than organisms from streams with lower (<50%) agricultural land use.

Allochthonous OM, as soil and sediment rather than terrestrial vegetation, also contributed more (15-48% soil and sediment OM and 2-10% terrestrial vegetation) to macroinvertebrate biomass in high-agriculture streams. Contributions of aged C and OM to macroinvertebrate biomass varied with land use and site. The ∆14C values of macroinvertebrates from Little Muncy Creek collected in 2011, when a methane seep was present, were the most 14C-depleted organisms observed at our study sites (-188 ±

197‰). Local land use effects appeared to influence nutritional resource utilization to a greater extent than watershed-scale factors in some of the study streams, possibly a result of short-term hydrological and other episodic events.

Introduction

The relative contributions of allochthonous vs. autochthonous sources of carbon

(C) and organic matter (OM) and their impacts on the structure and function of aquatic ecosystems have been the subject of an increasing number of studies (Nakano and

Murakami 2001, Atlas et al. 2013, Richardson and Sato 2015, Sitters et al. 2015).

Primary sources of allochthonous OM in stream and river systems include terrestrial vegetation, woody debris, soils, terrestrial invertebrates, and weathered sedimentary rocks (Harmon et al. 1986, Wallace et al. 1997, Nakano and Murakami 2001, Schillawski and Petsch 2008), among others. Autochthonous OM includes such sources as phytoplankton, benthic algae, macrophytes, and cyanobacteria (Graham and Wilcox

2000, Hall et al. 2001, Allan and Castillo 2007).

Allochthonous C and OM can play a quantitatively important role in stream ecosystems, supporting both autotrophic and consumer (e.g., bacteria, invertebrates, and fish) production (Allan et al. 2003, Agren et al. 2008, Berggren and del Giorgio 2015,

Wallace et al. 2015). Much of the research conducted to date in lotic systems has emphasized the importance of leaf litter and terrestrially-derived dissolved OM (DOM;

(Wallace et al. 1997, 2015, Tank et al. 2010, France 2011). However, little research has been conducted so far on the potential for moderately to highly-aged (hundreds to thousands of years old) OM derived from soils and weathered sedimentary rocks to contribute either directly (i.e., via direct consumption) or indirectly (e.g., via microbially

“repackaging”) to macroinvertebrate nutrition, or on the controls of aged OM contributions on aquatic consumer nutrition.

Utilization of allochthonous OM by stream and river consumers and food webs is dependent upon the availability and nutritional quality of the potential dietary sources.

Hydrologic and geomorphic properties of streams (e.g., flow, substrate type, habitat heterogeneity, etc.) can influence the storage vs. mobilization of allochthonous and autochthonous OM to and within lotic ecosystems (Walters et al. 2007, Sullivan 2013,

Smits et al. 2015, Wallace et al. 2015). Human activities such as riparian-zone alterations

(Bunn et al. 1999, Giling et al. 2012) and land-use change (Lal 2003, Sickman et al.

2010, Butman et al. 2014) may also influence the amounts, bioavailability, and quality of dietary resources in numerous ways.

Natural abundance stable isotopes (e.g., 13C, 15N, and 2H) are integrative tracers that provide more quantitative estimates of nutritional contributions to aquatic food webs than most other approaches (Peterson and Fry 1987, Finlay and Kendall 2007, Fry 2007).

δ13C and δ15N have relatively small dynamic ranges (tens of per mil [‰] at most) in most natural systems, and thus their differentiation of multiple OM sources can be challenging in some systems (Phillips and Gregg, 2003). In contrast, stable isotope ratios of H (δ2H or

δD), which have a much larger dynamic range (often 100‰ or more), may be used in conjunction with δ13C and δ15N to improve quantitative estimates of consumer nutritional sources (Doucett et al. 2007, Finlay et al. 2010, Wilkinson et al. 2015). Natural 14C, the radioactive isotope of C, has an even greater dynamic range than stable isotopes (i.e.,

Δ14C range ≥ ~1,000‰) and uniquely allows for determination of the ages (modern to ≥

50,000 years B.P.) of different C and OM nutritional sources utilized by organisms

(Bellamy and Bauer, submitted). The simultaneous use of natural stable isotopic and 14C ratios therefore allows for a more robust assessment of C and OM sources and ages in aquatic systems, and contributions to secondary production.

Numerous studies have explored the sources and ages of C and OM transported by the main stem of the Susquehanna River (Raymond and Bauer 2001, Raymond et al.

2004, Hossler and Bauer 2013a, 2013b), the major source of freshwater to the

Chesapeake Bay estuary – the largest estuary in the United States and third largest in the world. The overarching goal of the present study was to assess the contributions of allochthonous and autochthonous OM to macroinvertebrate consumers in part of the

Susquehanna River headwater-stream network and determine how these contributions vary as a function of stream size, land use, and related factors in the surrounding watershed. On the basis on the conceptual model proposed by the River Continuum

Concept (RCC), we hypothesized that contributions of allochthonous OM would be more important to macroinvertebrate nutrition in smaller, low order streams, whereas autochthonous OM would be more important in larger, higher order streams (Vannote et al. 1980). We also predicted that agricultural activity in sub-watersheds would enhance i) allochthonous support of macroinvertebrate nutrition, and ii) contributions of aged terrestrial (e.g., soils) OM to macroinvertebrate biomass (Lal 2003, Allan 2004, Burdon et al. 2013).

Methods

Site Description. Seven low- to mid-order streams were sampled (Figure 2.1) within the Upper and West Branch watersheds of the Susquehanna River Basin and the

Chemung River, a tributary of the Susquehanna River. Streams varied in order, watershed size, canopy cover, and their proximity to agricultural activities (Table 2.1). Percent agricultural land use (crop plus pasture, Table B.1) was determined as detailed in

Appendix B, Section 1.

Sampling was conducted in August 2011 and 2012, May 2013, and June 2014.

Little Muncy Creek was not sampled in 2012, and Towanda Creek was not sampled in

2011. At the time of sampling Little Muncy Creek in August of 2011 we observed elevated concentrations of methane in the water column (220.6 ± 5.2 μM), but methane was not detectable in subsequent years (May 2013 and June 2014). We sampled only one reach in each system using the rationale that repeated sampling of the same reaches for 3-

4 years would provide information about site-specific differences in nutritional resource utilization by macroinvertebrates vs. sampling multiple reaches less frequently. However, we recognize that small spatial-scale and local effects may influence our findings and limit our ability to draw conclusions about effects at the watershed scale.

Field Sampling. Macroinvertebrates belonging to different functional feeding groups (FFGs) and their potential nutritional resources were collected from each site.

Individual macroinvertebrates were collected by hand-picking them from rocks and logs and using a kick net where sediments and aquatic vegetation were more abundant.

Predators and primary consumers were separated and placed in stream water from their

48 respective sites that was filtered through a baked (525º C) 47 mm quartz fiber QMA filter, and allowed to void their guts for 24 hours (Brooke et al. 1996). After gut voidance, organisms were immediately placed in baked aluminum foil pouches and

Ziploc bags and frozen on dry ice until processing back in the lab.

Dominant potential aquatic and terrestrial nutritional sources of varying ages

(aquatic and terrestrial vegetation, soils, and sediments) available to macroinvertebrates were also collected, immediately placed in baked (525º C) aluminum foil pouches and

Ziploc bags and frozen on dry ice until processing back in the lab. Sediment samples were collected from each stream using a 60 ml syringe with the top cut off and immediately frozen on dry ice. Terrestrial soil samples were collected by digging a 20 cm depth hole on level ground within 10 m of the stream. Soil samples were collected from the hole at the soil surface and at 20 cm depths using a baked spatula. At sites where there was a cleanly eroded embankment adjacent to the stream, terrestrial soil samples were collected at 1 and 20 cm depths using a baked spatula.

Terrestrial vegetation (from riparian trees and shrubs) and aquatic vegetation samples were collected by hand using clean disposable nitrile gloves. Aquatic vegetation was not collected in 2011 because it was not available in adequate quantities for isotopic analyses. Biofilm samples were collected from 2-3 cobbles of similar size (15-20 cm in diameter) that were removed from each stream, the surface scraped with a toothbrush or a baked razor blade, and rinsed with ultra-pure water. Rinsed and scraped material was stored in acid-cleaned (10% HCl) polycarbonate bottles and frozen on dry ice. Bulk stream particulate OM (POM) was collected by filtering water from each sampling site

49 through a baked 47 mm QMA filter (0.8 µm nominal pore size) and wrapped in baked aluminum foil and frozen in a Ziploc bag on dry ice.

Dissolved inorganic C (DIC) samples were collected by filtering stream water from each site through a baked 47 mm QMA filter and then injecting the filtered water into baked, 125 ml serum bottles that had been pre-sparged with N2 gas, poisoned with

200 μL of HgCl2, and stored in the dark until processing. Filtered stream water (~250 ml) for both DOC and nutrients (N and P) was collected in acid-cleaned polycarbonate bottles and frozen at -20ºC until processing. A sample of stream water from each site was also

2 collected in 2013 and 2014 in a 20 ml baked scintillation vial for δH -H2O analysis.

Water chemistry data, including temperature, pH, dissolved oxygen (DO), and conductivity, were collected at each site using a Model AP110 Accumet Portable pH/ORP meter, a YSI ProODO handheld DO meter, and a Model 30/10 FT YSI multiparameter probe, respectively. Canopy cover was estimated at each site by using a convex spherical densiometer and taking the mean of nine measurements at different points within the reach. Dissolved stream water methane samples were collected by carefully drawing water into a clean 60 ml syringe and injecting the full volume into a

125 ml gas-tight serum bottle that was pre-sparged with N2 gas and poisoned with 1 ml of

2M NaOH to halt microbial activity (Hoehler et al. 2000).

Laboratory Methods. Upon return to the lab, macroinvertebrates were sorted and, whenever possible, identified to genus. Based on genus, organisms were assigned to the appropriate FFG according to Merritt, Cummins, and Berg 2008 (Table B.2).

Chironomids could only be identified to family and were not placed in one or more FFGs

50 because chironomid FFG and trophic position is known to vary across subfamily and genus (Reuss et al. 2013). Macroinvertebrates, terrestrial vegetation, and aquatic vegetation were dried and ground to a fine powder using solvent-cleaned mortars and pestles in preparation for stable isotope and radiocarbon analyses. Smaller macroinvertebrates were pooled to ensure adequate sample size for isotopic analysis

(Table B.2). Algae, biofilm, terrestrial soil, and aquatic sediment samples were acid- fumed with fresh concentrated HCl in a clean glass desiccator prior to homogenization to ensure that carbonates were removed. POM-containing filters were also acid-fumed. A subsample of each sample was set aside prior to acid-fuming for δ2H analysis.

Ground, homogenized samples and divided filter portions were packed in tin capsules for δ13C and δ15N analyses. For samples collected in 2011 and 2012, δ13C and

δ15N were measured at the Ohio State University Stable Isotope Laboratory using a

Costech elemental analyzer, connected to a Delta IV stable isotope ratio mass spectrometer (IRMS) via a Conflo III interface. δ13C and δ15N values were determined for samples collected in 2013 and 2014 by the University of California at Davis Stable

Isotope Facility, using a PDZ Europa ANCA-GSL elemental analyzer interfaced with a

PDZ Europa 20-20 IRMS. Stable isotope values for δ13C and δ15N were reported relative to international standards, V-PDB and air, respectively. Selected DOC samples were analyzed for δ13C based on a modified method of Osburn and St-Jean (2007) using an

Oceanographics International model 1030C DOC analyzer, connected to a modified

Graden CO2 trap and a Thermo-Finnigan Delta Plus XL isotope ratio mass spectrometer

(IRMS).

For δ2H analysis, non-acidified homogenized samples of macroinvertebrate, aquatic and terrestrial vegetation, soil, and sediments were packed in silver capsules.

Stream water samples were placed in baked scintillation vials for bulk δ2H analysis. δ2H values for all samples were measured at the Colorado Plateau Stable Isotope Laboratory at Northern Arizona University. In order to determine the δ2H of the non-exchangeable H component for solid samples, a bench-top equilibration method was used to estimate the exchange of H between the samples and local water vapor (Wassenaar and Hobson 2003,

Doucett et al. 2007). Solid samples were analyzed for δ2H on a Thermo-Finnigan TC/EA and DeltaPLUS-XL and stream water samples were analyzed for δ2H using a Los Gatos

Research DLT-100 Liquid Water Isotope Analyzer.

For ∆14C analysis, homogenized subsamples of selected macroinvertebrate and potential nutritional sources were placed in quartz tubes with Cu and CuO and combusted to CO2 at 750º C for 4 hours. Within 24 hours of combustion, the CO2 from each sealed tube was purified and quantified on a vacuum extraction line and sealed in a pre-baked 6 mm Pyrex tube. DIC samples were acidified and the CO2 was collected and purified as

14 above. The purified CO2 was reduced to graphite, and analyzed for Δ C at the NSF

Arizona Accelerator Mass Spectrometry (AMS) Facility at the University of Arizona, and at the National Ocean Sciences Accelerator Mass Spectrometry (NOSAMS) laboratory at

Woods Hole Oceanographic Institution. 14C data were collected from all sites except

Lamb’s Creek.

Stream water dissolved methane concentrations were determined immediately upon return to the lab by gas chromatography using a Shimadzu GC-8A gas

52 chromatograph with a flame ionization detector and calculated using equations from

Kampbell and Vandegrift (2008). Nutrient concentrations were analyzed using a Lachat

QuickChem 8500, and DOC concentrations were analyzed using a Shimadzu TOC 5000

Analyzer.

Statistical Analyses. In order to determine if there were significant differences in the isotopic composition of macroinvertebrates across sampling years, sites, and FFGs, a multivariate statistical approach was employed using the PRIMER software package

(Clarke and Gorley 2006; v. 6, PRIMER-E Ltd). Macroinvertebrate δ13C, δ15N, δ2H, and

∆14C values were normalized and visually evaluated using non-metric multidimensional scaling (NMDS). One-way analyses of similarity (ANOSIM) were conducted for macroinvertebrate isotopic data in order to determine whether significant isotopic differences existed between a) sampling years, b) sites, c) FFGs, and d) the degree of agricultural land use in each watershed (>50% and <50%). The ANOSIM pairwise test statistic R is a better indicator of separation between groups than the p-value because it is not influenced by sample size (Clarke and Gorley 2006). All pairwise comparisons within the model were examined in cases where overall p-values for the ANOSIM model were significant (p<0.05). Pairwise comparisons were considered significant even if the pairwise comparisons had p>0.05, but the R>0.8.

Isotopic Mixing Models. In order to estimate relative contributions of different potential nutritional sources to macroinvertebrate biomass, we used the Bayesian isotopic mixing model MixSIAR (Stock and Semmens 2013), which employs a graphical use interface (GUI) with R Statistical software (R Core Team, 2014) and allows for the

53 incorporation of uncertainty. Sources of uncertainty associated with isotope values include isotopic fractionation, isotopic variability in potential nutritional sources, and utilization of multiple nutritional sources (Finlay et al. 2002, Moore and Semmens 2008,

Phillips et al. 2014). Details of the nutritional sources used in mixing models, trophic fractionation and dietary water corrections, and other details about mixing model procedures are provided in Section 2 of Appendix B.

Results

Stable Isotopic Composition (δ13C, δ15N, and δ2H) of Macroinvertebrate Consumers

Macroinvertebrates collected from Lamb’s Creek were consistently lower in δ13C and had the lowest δ15N values compared to macroinvertebrates collected from the other sites (Figure 2.2A, Table B.14). In contrast, macroinvertebrates from Little Muncy Creek

(2013 and 2014) and the Cowanesque River had greater δ13C values compared to the other five sites (Figure 2.2A, Table B.14). For δ2H, organisms from Lamb’s Creek, North

Elk Run, and Little Muncy Creek were enriched compared to those macroinvertebrates collected from the other four sites (Figure 2.2B, Table B.14). NMDS and ANOSIM revealed that macroinvertebrates from Crooked Creek and Elk Run clustered closely together and no significant differences in macroinvertebrate isotopic composition existed between these two sites (Figure 2.2C; Table B.15).

Predators were consistently the most 15N-enriched across all sampling times and years (Table B.14). Filtering collectors, scrapers, and in some cases collector-gatherers, had the lowest δ13C and δ2H values compared to the other FFGs across all years (Table

B.14). Despite the small amount of isotopic variability across FFGs within a single

54 isotope, differences among FFGs emerged when all stable isotopes were pooled. One- way ANOSIM revealed differences in isotopic composition among FFGs with few exceptions (R= 0.221, p=0.001; Table B.16).

One-way ANOSIM of macroinvertebrate isotopic compositions by year were significant (p= 0.002), however the global R statistic (0.064) was low, suggesting significant overlap in isotopic compositions of macroinvertebrates across years (Table

B.17). As a result, macroinvertebrate and potential nutritional resource isotope data were combined across years for mixing model runs.

Macroinvertebrate isotopic composition differed significantly between watersheds with high vs. low agricultural land use (defined as >50% and <50% of watershed area, respectively; Figure 2.3A). Macroinvertebrates collected from streams with high agriculture were higher in δ15N than those collected from streams with low agriculture

(Figure 2.3A). However, there was no apparent separation between the high and low agriculture land use groups for δ13C vs. δ2H (Figure 2.3B). NMDS and ANOSIM further confirmed that macroinvertebrates clustered according to the amount of watershed agriculture (Figure 2.3C; R= 0.631, p=0.001).

Radiocarbon (14C) Isotopic Compositions of Macroinvertebrate Consumers

The most highly 14C-depleted organisms were from Little Muncy Creek and the

Cowanesque River in 2011 and from Crooked Creek in 2012 (Figure 2.4A, Table B.14).

In comparison, macroinvertebrates from North Elk Run and Little Muncy Creek in 2013 and 2014 were 14C-enriched (Figure 2.4A, Table B.14). The most highly 14C-enriched macroinvertebrates (1 ± 6‰, modern in age) were predators from North Elk Run in 2014

(Figure 2.4A, Table B.14). However, none of the FFGs significantly differed overall from each other (ANOSIM R = 0.079, p = 0.085). NMDS and ANOSIM revealed that macroinvertebrates from the Cowanesque River did not differ from those at Elk Run,

Towanda Creek, and Crooked Creek (Figure 2.4C; Table B.18). One-way ANOSIM by year across all sites also showed significant differences in isotopic composition of macroinvertebrates (Table B.19). Unfortunately, the small size of the ∆14C dataset prevented us from running the mixing model separately for each year sampled.

Examination of ∆14C values across all years for Little Muncy Creek alone revealed much larger variability and lower values of macroinvertebrate ∆14C in 2011 (the sampling time that methane bubbles were observed) than in subsequent years, and macroinvertebrates from 2011 differed from those in 2013 and 2014 (Figure 2.4B).

Nutritional Source Contribution Estimates to Macroinvertebrates Using Stable Isotopes

Nutritional source contribution estimates and associated posterior probabilites

(5%, 50%, 95%) using Bayesian isotopic mixing models for different FFGs using 13C,

15N, and 2H are provided in Section 5 of Appendix B. Mixing model outputs showed that autochthonous production (algae) was of primary importance to most FFGs but was also highly variable (7-89%; Figure 2.5). Macroinvertebrates not primarily reliant on autochthonous OM included i) shredders at Lamb’s Creek (Figure 2.5A), North Elk Run

(Figure 2.5B), and Little Muncy Creek (Figure 2.5C) and ii) predators at North Elk Run and Little Muncy Creek (Figure 2.5B, C). For Elk Run, autochthonous production contributed most to collector-gatherer biomass (74%), in contrast to the other sites where scraper biomass was dominated by autochthonous production (62-81%, Figure 2.5).

Compared to the other sites, autochthonous production contributed the most to filtering collector biomass in Crooked Creek, with a maximum contribution of 89% (Figure 2.5E).

Collector-gatherer and scraper biomass was still primarily derived from autochthonous production (82% and 81%, respectively) in Crooked Creek as well. Model outputs for the two sites with the largest watershed areas (Towanda Creek and Cowanesque River; Table

2.1) suggest similarities in resource utilization between FFGs in these two systems, with little variability in autochthonous and allochthonous sources (Figure 2.5F, G).

Allochthonous OM in the form of terrestrial vegetation contributed most to organism biomass in only three cases: shredders from Lamb’s Creek and North Elk Run and predators from North Elk Run (Figure 2.5A, B). Additionally, terrestrial vegetation contributed >20% to macroinvertebrate biomass of some FFGs in Little Muncy Creek,

Elk Run, Crooked Creek, and Towanda Creek (Figure 2.5). Allochthonous OM in the form of soil and sediment-derived OM was the dominant contributor to the biomass of shredders from Little Muncy Creek (Figure 2.5C), but was also a large contributor (20-

30%) to the biomass of all FFGs from the two largest watershed sites (Towanda Creek and Cowanesque River; Figure 2.5F, G).

Autochothonous OM was the dominant contributor to the biomass of scrapers, filtering collectors, collector-gatherers, and chironomids across both high and low agricultural sites (Figure 2.6). These same organisms in high agriculture streams had greater contributions from allochthonous production in the forms of soil and sediment

OM (15-48%) than from terrestrial vegetation (2-10%; Figure 2.6A). In low agriculture streams, contributions of terrestrial vegetation also increased to all FFGs compared to

57 high agriculture streams, except for filtering collectors and chironomids, for which soil and sediment OM still contributed more than terrestrial vegetation (Figure 2.6B). Overall, soil and sediment OM contributed ~10% more on average to the biomass of macroinvertebrates in the high vs. the low agricultural watershed streams.

Mixing Model Outcomes Using Stable Isotopes and ∆14C

Nutritional source contribution estimates and associated posterior probabilities

(5%, 50%, 95%) using a Bayesian isotopic mixing model for different FFGs using 13C,

15N, 2H, and ∆14C are also provided in Section 5 of Appendix B. Mixing models that included both stable isotope and ∆14C data for macroinvertebrates and potential nutritional resources pooled across sites produced similar findings as the stable isotope- only models and confirmed the primary importance of autochthonous OM (Figure 2.7).

Terrestrial vegetation and soil/sediment OM also contributed to varying degrees to all

FFGs (18-43% and 5-16%, respectively; Figure 2.7). In contrast to the use of stable isotopes alone, inclusion of ∆14C showed that terrestrial vegetation became the most important nutritional resource for chironomids (44% contribution; Figure 2.7). Filtering collectors also assimilated more (~19%) terrestrial vegetation in the model that included

∆14C (Figure 2.7). Also in contrast to the stable isotope-only model, autochthonous production was shown to be of primary importance to shredders when ∆14C was included

(Figure 2.7).

In August 2011 when active methane seepage was observed in Little Muncy

14 Creek, the combined stable isotope and ∆ C model showed that CH4 contributed only

~4% to macroinvertebrate primary consumer biomass (Figure 2.8A). In subsequent years

58 without seepage, contributions of methane C to macroinvertebrate biomass were nominal

(0.5%; Figure 2.8A, B). Soil and sediment OM were the largest contributors to Little

Muncy primary consumer and predator biomass in 2011 (40% and 82%, respectively;

Figure 2.8A, B). In contrast, algae was the largest contributor to all Little Muncy macroinvertebrates in 2013 and 2014 (67-98%; Figure 2.8A, B).

Discussion

Previous work in the main stem of the Susquehanna River indicates that much of the transported C and OM is terrestrially-derived and of varying age (Hossler and Bauer

2012, 2013a, 2013b). The tributaries and sub-watersheds of the Susquehanna form a complex network through which C, OM, nutrients, and even organisms are exchanged and transported (Gomi et al. 2002, Lowe et al. 2006, Campbell Grant et al. 2007). The structure of drainage networks may influence the availability of nutritional resources at different points within the network (Power and Dietrich 2002, Campbell Grant et al.

2007). In-stream processing may also have a significant impact on flows of C, energy, and OM and on consumer biomass, and this is predicted to be more pronounced and variable, as well as more tightly linked to watershed processes, in the upper reaches of stream networks (Power and Dietrich 2002, Campbell Grant et al. 2007). Findings from this study in the upper and middle Susquehanna and Chemung River watershed network suggest that differences in size, land use, and associated stream characteristics along a network gradient may influence nutritional resource utilization by macroinvertebrates.

Stream Size and Physical Characteristics as Controls on Nutritional Resource Utilization

As stream order increases along a network, physical characteristics of lotic systems change (e.g., increased channel width, reduced canopy cover, reduced watershed slope, etc.) in a predictable manner, and may influence nutritional resource availability and utilization (Gomi et al. 2002, Benda et al. 2004, Smits et al. 2015). Similar physical attributes in some of our streams may explain observed similarities in nutritional resource utilization in these systems. Consumers in North Elk Run and Lamb’s Creek were most reliant on terrestrial vegetation compared to the other sites (maximum contributions of

85% vs. 45%) (Figure 2.5A, B). Lamb’s Creek and North Elk Run were also the smallest streams (2nd order) and were heavily shaded (Table 2.1). Macroinvertebrate biomass in

Elk Run, a third order stream, was also largely derived from allochthonous OM, both in the form of terrestrial vegetation and soil and sediment OM (5-37% and 14-44%, respectively) (Figure 2.5D). Algal contributions were larger (13% on average) to all

FFGs in Elk Run vs. North Elk Run and Lamb’s Creek FFGs. This may be due to increased availability of autochthonous OM due to the somewhat larger size of, and lower canopy cover at Elk Run (Hall et al. 2001, Doi et al. 2007, Collins et al. 2016).

The greater dependence of macroinvertebrate consumers on terrestrial vegetation in these same three systems supports predictions of the RCC (Vannote et al. 1980).

However, in contrast to the contention of the RCC that allochthonous OM primarily supports secondary production in low order streams, algae was of greatest importance to most FFGs, with the exception of shredders in North Elk Run and Lamb’s Creek and predators in North Elk Run (Figure 2.5A, B). Our findings of algal contributions dominating macroinvertebrate biomass in low order streams are similar to those of a

60 recent study in the Salmon River Basin (Idaho, USA), where autochthonous OM contributed 30-45% to the diet of macroinvertebrates in a 2nd-order stream (Rosi-

Marshall et al. 2016). While algae may be less abundant in lower order streams because of light limitation, algal quality may have a compensatory effect on nutrition (Mayer and

Likens 1987, McCutchan and Lewis 2002, Guo et al. 2016, Rosi-Marshall et al. 2016).

Crooked Creek, a 4th order stream, drains a larger area than Lamb’s Creek, North

Elk Run, and Elk Run, and macroinvertebrates across multiple FFGs in Crooked Creek were more reliant on algae (up to 89%) compared to the other six streams (up to 79%), with the exception of chironomids (Figure 2.5E). There was an increase in autochthonous

OM contributions to macroinvertebrate biomass from 3rd to 4th order streams (Figure

2.5D, E), but there was a slight reduction in autochthony (10% on average) in our 5th order streams (Little Muncy, Cowanesque, and Towanda) (Figure 2.5C, F, G) compared to Crooked Creek. There were also similarities in nutritional resource utilization by macroinvertebrates in Little Muncy Creek and Lamb’s Creek, but allochthonous OM contributions to most FFGs in Lamb’s Creek were derived primarily from terrestrial vegetation. In contrast, soil and sediment OM contributed more than terrestrial vegetation to macroinvertebrate biomass, with the exception of predators, in Little Muncy, Towanda and Cowanesque (Figure 2.5A, C, E and F). Crooked Creek (a 4th order stream) was unique in that autochthonous production contributed more to macroinvertebrate biomass compared to two of the 5th order streams (Towanda and Little Muncy).

Differences in nutritional resource contributions to macroinvertebrate primary consumer biomass in Towanda Creek and the Cowanesque River, the two largest systems

(Table 2.1), were nominal (Figure 2.5F, G), despite difference in watershed size (~400 km2). This suggests that in our study systems, stream order and stream size may be more important in controlling nutritional resource availability than watershed size (Collins et al. 2015). 13C-enrichment of some of the organisms from these two sites relative to potential nutritional resources measured (Figure 2.2A, B) may also indicate one or more nutritional sources that were not identified and measured.

Previous studies have also observed relationships between stream size and nutritional resource utilization by aquatic consumers (Vannote et al. 1980, Collins et al.

2015). However, our findings at Crooked, Little Muncy, and Towanda Creeks and the

Cowanesque River differed from what might be predicted by the RCC. Nutritional contributions to macroinvertebrates in Little Muncy were not similar to the other 5th order streams sampled (Figure 2.5C, F, G). Additionally, autochthonous OM contribution to macroinvertebrates in Crooked Creek compared to our higher order streams was not consistent with the RCC (Figure 2.5E). However, Crooked Creek was not heavily shaded, therefore light was unlikely limiting autochthonous production.

Stream order may by itself not be a good predictor for assessing nutritional resource contributions to consumers in stream ecosystems (Gomi et al. 2002, Rosi-

Marshall et al. 2016). One possible explanation for this is that small spatial scale (i.e., local) effects may have greater-than-expected impacts on consumer nutrition compared to watershed-scale effects (Strayer et al. 2003, Lowe et al. 2006), and the relative importance of local vs. watershed scale effects may depend on the stream order of the system (Buck et al. 2004). In order to assess local vs. watershed scale effects on

62 nutritional resource contributions to lotic consumers, multiple reaches in a given stream must be measured and compared. Sampling single reaches in each of our systems limited our ability to assess whether shifts in nutritional resource utilization along a stream size gradient were reflective of reach- or stream-scale patterns. Hydrological and geomorphological differences between these systems may also help explain the patterns in nutritional resource utilization by macroinvertebrates by influencing the availability of different forms of OM (Roach and Winemiller 2015, Smits et al. 2015).

Influence of Agriculture on Isotopic Composition and Nutritional Resource Utilization

Macroinvertebrates and algae were higher in δ15N in streams with higher agriculture in their watersheds (North Elk Run, Elk Run, Crooked Creek, Towanda

Creek, and Cowanesque River; Figure 2.3A). Inputs of 15N-enriched N from agricultural fertilizers can be significant in lotic systems (Heathwaite and Johnes 1996, Arheimer and

Liden 2000, Mallin et al. 2015), and may be transferred to aquatic consumers (Harrington et al. 1998, Anderson and Cabana 2005, Vander Zanden et al. 2005, Diebel and Vander

Zanden 2009).

While croplands dominated agricultural land use in our study watersheds, there were significant contributions by pasturelands (6.5-13.6%; Table B.1). Previous research suggests that elevated δ15N values observed in aquatic consumers may result from denitrification of inorganic fertilizers (Diebel and Vander Zanden 2009) rather than, or in addition to, inputs from 15N-enriched animal manure (Karr et al. 2001, Anderson and

Cabana 2005, Kendall et al. 2007) and human septic waste (Aravena et al. 1993,

Harrington et al. 1998, Fenech et al. 2012). Regardless of the mechanism(s), inputs of

63 inorganic and organic fertilizer N are likely potentially causes of elevated consumer δ15N in our study streams. Because algae collected from high-agriculture streams were more

15N-enriched compared to low agriculture streams (Figures 2.2A, 2.3A), we suggest macroinvertebrates were utilizing 15N-enriched algae, rather than directly assimilating animal manure as found in other studies (del Rosario et al. 2002, Mesa et al. 2016).

Mixing models also suggested that macroinvertebrates collected from high agriculture streams generally had greater contributions of soil and sediment OM to their biomass (Figure 2.6A, B). Row crop agriculture can be a significant source of soil- derived OM and sedimentation in lotic systems (Lal 2003, Allan 2004, Zaimes et al.

2004, Wilson and Xenopoulos 2009). Grazing cattle can also increase erosional soil inputs via the removal of vegetation (Belsky et al. 1999, Brand et al. 2015). Soil and sediment OM and bacteria contribute to consumable detrital pools (Cummins and Klug

1979, Carlough 1994), and can be used by aquatic consumers (Evans and Bauer 2016a,

2016b). However, increased inputs of soils to streams may also lead to higher turbidity and reduce stream primary production, as well as negatively affect pollution-sensitive macroinvertebrate taxa (Ephemeropterans, Plecopterans, and Trichopterans; Matthaei et al. 2010, Wagenhoff et al. 2012, Murphy et al. 2015). Soil and sediment OM were also generally more 14C-depleted than other potential nutritional sources in our streams, which may explain some part of the 14C-depletion in macroinvertebrate consumers that utilized this aged source of OM (Figure 2.4A).

Soil and sediment OM contributions to macroinvertebrate biomass in Little

Muncy Creek, a low agriculture stream, were 10% greater on average compared to

64 contributions in Lamb’s Creek, the other low agriculture stream (Figure 2.5A, C). There was a large section of cornfield (~0.7 km in length) in close proximity (within ~100 m) to our study reach at Little Muncy. Local effects of this agricultural activity may have attenuated watershed-scale effects that we might otherwise have observed (Buck et al.

2004, Zhou et al. 2012), and help explain the increased assimilation of soil and sediment

OM by macroinvertebrates in this system. However, canpy cover was also more extensive at Lamb’s Creek (80%) compared to Little Muncy Creek (28%), which may have resulted in the increased avaibility of fresh terrestrial vegetation to macroinvertebrate consumers in Lamb’s Creek.

Little Muncy Creek: A Case-Study of Anthropogenically Induced Fossil OM Inputs

Much of the temporal variability in potential nutritional resource inputs to lotic systems can be attributed to factors such as seasonal and storm-driven hydrology and temporal changes in aquatic and terrestrial primary production (Junk et al. 1989, Dalzell et al. 2007, Sobczak and Raymond 2015). Human activities such as resource extraction may also play a role in specific systems or regions. Hydraulic fracturing activity in our study region was extensive during the period of this study. According to data reported from the Pennsylvania Department of Environmental Protection Office of Oil and Gas, between 2010 and 2014, 2,730 unconventional gas wells were drilled in our study area’s counties, with more than half of these being drilled in 2010 and 2011 (PA DEP 2016).

High methane concentrations have been observed in groundwater and drinking water wells in this region, and this has been proposed to be associated with the increase of hydraulic fracturing wells and activity in Pennsylvania (Osborn et al. 2011, Jackson et

65 al. 2013, Llewellyn et al. 2015, Hammond 2016). Ground and surface water connectivity may also lead to increased concentrations of methane in surface waters (Vengosh et al.

2014, Heilweil et al. 2015). The presence of methane in ground and surface water is not, however, always diagnostic of hydraulic fracturing activity (Molofsky et al. 2013, Siegel et al. 2015). Thermogenic and biogenic methane can both occur naturally in aquatic systems, and distinguishing between thermogenic and biogenic methane can be accomplished through the use of stable isotopes and natural radiocarbon (Whiticar et al.

1986, Osborn et al. 2011, Trimmer et al. 2012).

Little Muncy Creek had active methane seepage in the study reach when we sampled it in 2011, but we have no knowledge of the duration and magnitude of seepage before and after our sampling. While the isotopic signatures of the methane were not measured directly, the low δ13C (-20.0 ± 0.2‰) and ∆14C (-253 ± 4‰) values of DIC collected from Little Muncy Creek in August 2011 compared to the relatively more elevated δ13C and ∆14C values of DIC observed in May 2013 and June 2014 (-10.7 ±

0.6‰ and -18 ± 8‰, respectively), suggest that thermogenic and/or biogenic methane in

August 2011 was effluxing from the stream bed and being oxidized by methane-oxidizing bacteria. Both primary consumer and predator macroinvertebrates collected in August

2011 were the most 13C and 14C-depleted in this study (Figure 2.4B), suggesting that this episodic input of highly aged methane was incorporated into at least lower trophic levels of the stream food web. This aged material may have entered the food web through two possible mechanisms: 1) algae fixed the highly aged DIC from methane oxidation in the stream, and the “aged” algae were then consumed by macroinvertebrates (Bellamy and

Bauer, submitted), and/or 2) methanotrophic bacteria utilized the methane, and the bacteria were consumed by macroinvertebrates either directly or via the microbial loop

(Trimmer et al. 2009, DelVecchia et al. 2016). While there is evidence of methane contributing to macroinvertebrate biomass in other lotic systems (Kohzu et al. 2004,

Takagi et al. 2006, Trimmer et al. 2009, DelVecchia et al. 2016, Grey 2016), we cannot definitively identify the exact mechanism for the assimilation of the aged material by macroinvertebrates in Little Muncy Creek in 2011.

Mixing model outcomes further suggest contributions in Little Muncy of fossil- aged methane-derived C to macroinvertebrate nutrition at all sampling times (4.1%,

0.5%, and 0.5% in 2011, 2013, and 2014, respectively; Figure 2.8). However, there was large overlap in the ∆14C values of the potential nutritional sources (methane and soil/sediment OM) used in the model, which may have overestimated the proportional contribution of soil and sediment OM to macroinvertebrate biomass. Further, if the methane contributing to macroinvertebrate biomass was younger than fossil-aged, contributions of methane derived C would be greater and our estimates may be conservative. Regardless, the evidence suggests that a transient episodic input of aged methane C contributed to macroinvertebrate biomass in Little Muncy Creek in August

2011. Whether these types of inputs are natural or anthropogenic in origin, their periodicity and magnitude has implications for both short and long term shifts in the sources of OM contributing to consumer nutrition and food web energetics in such systems, as has been observed in freshwater hyporheic (DelVecchia et al. 2016) and marine systems (Brooks et al. 1987, Bauer et al. 1990, Levin and Mendoza 2007).

Conclusions and Implications

In the present study of a headwater stream network, our findings suggest that in some cases the location of a stream in the network, as well as its size, may lead to predictable shifts in the importance of allochthonous vs. autochthonous OM to macroinvertebrate nutrition. Canopy cover also appeared to be a good predictor of increased allochthonous OM utilization by macroinvertebrates, regardless of stream order and size. Canopy cover can directly and indirectly influence nutritional resource utilization by serving as a source of OM and limiting the growth of autochthonous OM via light limitation and influences on water temperature (Doi et al. 2007, Webb et al.

2008, Taipale et al. 2013, Collins et al. 2015, Hofmeister et al. 2015). However, land use in watershed networks and short temporal and spatial differences in inputs of C and OM to streams can limit our use of single stream characteristics and existing conceptual models for explaining the controls on nutritional support of aquatic consumers (Vannote et al. 1980, Collins et al. 2015, Rosi-Marshall et al. 2016). Consideration must also be given to whether or not local (e.g., canopy cover and land use within a reach) or watershed scale effects have a greater impact on contributions of allochthonous OM to aquatic food webs.

The findings of the present study and previous studies show that incorporation of both moderately and highly aged C and OM into aquatic food webs suggests that modern aquatic food webs and the C and OM supporting them may be temporally and/or spatially disconnected (Caraco et al. 2010, Ishikawa et al. 2014, 2016, Fellman et al. 2015). Aged sources of C and OM represent nutritional resources that have been traditionally

68 considered unavailable to consumers (Kleber et al. 2011, Marín-Spiotta et al. 2014). The inclusion of aged C and OM into aquatic food webs could influence ecosystem community structure, and potentially affect other ecosystem processes and biogeochemical cycles. Further assessments are also needed to determine whether aged

OM acts to displace other sources of nutrition to aquatic food webs, or if it represents a nutritional supplement. These issues warrant further examination in order to assess how general the utilization of aged C and OM in aquatic systems may be, and what naturally occurring and human-induced controls may lead to this phenomenon.

Table 2.1. Sampling site and sub-watershed characteristics of study streams within the upper and middle Susquehanna River watershed.

Sampling Site Dates Stream Watershed pHb Temperature DO Ag. Cover Canopy Sampled Ordera Area (km2) (ºC)b (mg/L)b (%) Cover (%)c Crooked Creek 2011-2014 4 110 8.1 ± 0.3 20.5 ± 2.6 9.3 ± 1.1 71 10 Cowanesque 2011-2014 5 556 8.2 ± 0.4 24.6 ± 2.3 8.7 ± 0.2 52 0 River Lamb’s Creek 2011-2014 2 6 7.3 ± 0.3 16.3 ± 1.5 8.1 ± 0.2 17 80 North Elk Run 2011-2014 2 17 7.4 ± 0.2 21.0 ± 1.0 6.7 ± 2.6 85 66 Elk Run 2011-2014 3 65 8.5 ± 0.5 24.0 ± 2.6 8.9 ± 0.7 76 43 Little Muncy 2011, 2013, 5 113 7.5 ± 0.6 18.6 ± 1.3 8.9 ± 0.4 41 28

70 Creek 2014

Towanda Creek 2012-2014 5 169 8.5 ± 0.4 23.7 ± 1.9 10.0 ± 1.0 61 31 S.D. - standard deviation. aFrom the Susquehanna River Basin Coalition Water Resource Portal; http://gis.srbc.net/wrp/. bMeans ± S.D. for all years. cMeasured in June 2014.

15 Crooked Creek A Cowanesque River Little Muncy Creek Towanda Creek Elk Run North Elk Run 10 Lamb's Creek Terrestrial vegetation Sediment Soil POM

(‰) Biofilm N 5 Algae- high ag.

15 Algae- low ag.



-45 -40 -35 -30 -25 -20 13C (‰) -100 B

-150

H (‰)

 -200

-250

-300 -45 -40 -35 -30 -25 -20 13  C (‰)

Figure 2.2. δ13C, δ15N, and δ2H values of macroinvertebrate individuals and their potential dietary sources (means ± SD) collected from all seven Susquehanna River sub-watershed study sites. Values are corrected for trophic fractionation (Post 2002) and the influence of dietary water (Wilkinson et al. 2015). Potential nutritional source isotope means and SDs are pooled across sites for visualization purposes, but not for model runs. A) δ13C vs. δ15N, B) δ13C vs. δ2H, and C) NMDS results for δ13C, δ15N, and δ2H for all macroinvertebrate individuals; Stress: 0.15. Panel C axes are dimensionless. 72

High ag.- macros. A 15 Low ag.- macros. High ag.- algae Low ag.- algae Terrestrial vegetation 10 Sediment Soil POM Biofilm

(‰) 5 N

 0

-5 -45 -40 -35 -30 -25 -20 13  C (‰) -100 B

-150

(‰) -200



-250

-300 -45 -40 -35 -30 -25 -20 13C (‰)  C

200 Crooked Creek A Cowanesque River Little Muncy Creek 0 0 Towanda Creek Elk Run North Elk Run -200 algae terrestrial vegetation Years BP soil sediment -400 biofilm 4,000 C (‰) POM 14  -600

13,000 -800

-1000 -50 -45 -40 -35 -30 -25 -20 13C (‰) 200 2011 B 2013 2014 0 Algae 0 Terrestrial vegetation Soil/ Sediment

-200 Methane Years BP 2,500

-400 C (‰) 5,000

 -600 10,000 -800

-1000 50,000 -50 -45 -40 -35 -30 -25 -20 13C (‰) C

100 100 A) Lamb's Creek B) North Elk Run 80 80

60 60

40 40

20 20

Percent Contribution

ND ND 0 0 ND 100 100 C) Little Muncy Creek D) Elk Run 80 80

60 60

40 40

20 20

Percent Contribution Percent Contribution ND 0 0 100 100 E) Crooked Creek F) Towanda Creek 80 80

60 60

40 40

20 20

Percent Contribution

Percent Contribution ND ND 0 0 100 G) Cowanesque River

80 scrapers shredderspredators chironomids 60 filteringcollector-gatherers collectors

40 Algae 20 Terrestrial vegetation

Percent Contribution Soil/Sediment 0

scrapers shredderspredators chironomids

filteringcollector-gatherers collectors

100 A Algae Terrestrial vegetation 80 Soil/Sediment

20 Percent ContributionPercent

scrapers shredders predators chironomids filtering collectors 100 collector-gatherers

B 80

Percent ContributionPercent 20

scrapers shredders predators chironomids

filtering collectorscollector-gatherers

100 A Algae Ter. veg. 80 Soil/ Sediment CH4

20 contributionPercent

0 2011 2013 2014 100 B 80

Percent contributionPercent

0 2011 2014 Year

Chapter 3: Temporal Variability in Autochthonous and Allochthonous Nutritional Sources to Macroinvertebrates in a Temperate Stream

Amber R. Bellamy

James E. Bauer

Andrea G. Grottoli

Submitted to Marine and Freshwater Research

Abstract

Temporal variability is a key driver of many aspects of aquatic ecosystem structure and function. The amounts and characteristics of organic matter (OM) available for consumption and assimilation by aquatic consumers can vary temporally on scales from less than a day to annual or more through changes in insolation, discharge, and seasonal to inter-annual inputs of nutritional resources. In the present study we sought to determine how nutritional resource inputs and utilization by aquatic macroinvertebrates varied intra-annually at one site in a temperate sub-tributary of the Scioto River (which flows into the Ohio River) using natural abundance stable isotopes (δ13C, δ15N, and δ2H) and radiocarbon (14C). The δ15N values of algae and macroinvertebrates varied with season, whereas δ13C and δ2H values did not. Bayesian isotopic mixing models revealed that autochthonous aquatic nutritional resources contributed more to macroinvertebrate primary consumer biomass in spring and summer (56-75%) than in autumn (41-57%).

For predators, fresh allochthonous terrestrial vegetation was the dominant nutritional source regardless of season (43-45%). Degraded terrestrial vegetation, as soil-derived

OM, also contributed significantly to consumer biomass, primarily to scrapers, predators, and potamanthid mayflies (24-46%). In addition, filtering collectors and potamanthid mayflies were depleted in 14C throughout the year (mean ∆14C range -58‰ to -41‰; equivalent 14C ages of 480 and 340 yrs. B.P., respectively) compared to modern, contemporary biomass, suggesting that one or more forms of aged carbon and/or OM contributed to their biomass. Better understanding of nutritional resource variation on

80 different temporal scales is important for evaluating the changes that nutritional inputs and utilization may undergo and their implications for aquatic community and ecosystem structure and function, and biogeochemical cycling.

Introduction

The availability of internal aquatic (autochthonous) and external terrestrial

(allochthonous) materials in streams and rivers can be dynamic and highly variable in space and time (Bianchi et al. 2007, Dodds et al. 2013, Ran et al. 2013). This variability may have significant impacts on the sources and timing of nutritional resources available to stream and river consumers (Pringle et al. 1988, Ceola et al. 2013, Junker and Cross

2014). For example, short- and long-term changes in factors such as discharge-related turbidity, nutrient inputs, and tree canopy cover in riparian areas may impact the amounts and ratios of autochthonous vs. allochthonous carbon (C) and organic matter (OM) inputs to streams and rivers (Horner and Welch 1981, Dekar et al. 2009, Roach 2013, Roach and

Winemiller 2015). Changes in the absolute and relative inputs of autochthonous and allochthonous C and OM by these and other mechanisms may thus regulate time- dependent differences in consumer production in streams and rivers (Vannote et al. 1980,

Roach 2013, Guo et al. 2016a, Rosi-Marshall et al. 2016).

Nutrient inputs and light availability also vary temporally in steam and river systems and can influence their rates and extents of primary production (Francoeur et al.

1999, Rosemond et al. 2000, Guo et al. 2016b). Nutrient concentrations fluctuate seasonally in these systems as a function of in-stream processing (e.g., uptake by primary producers and remineralization by consumers), as well as inputs from the surrounding

81 landscape (Hoellein et al. 2007, Roberts and Mulholland 2007, Kaushal et al. 2014).

Light limitation by riparian vegetation on autochthonous production also varies with season in temperate climates, with streams typically experiencing the greatest riparian canopy cover in summer (Rosemond et al. 2000, Hill et al. 2001, Hill and Dimick 2002).

However, the influence of riparian vegetation on autochthonous primary production is likely to be the greatest in small, forested streams compared to larger streams and rivers having more open canopies (Vannote et al. 1980). Temporal controls on the presence of autochthonous and allochthonous OM available for consumption and assimilation by consumers in streams and rivers are numerous, and interactions between factors such as hydrology, light, nutrient availability, and temperature can co-limit or act as stressors on autochthonous primary production (Dodds et al. 1996, Piggott et al. 2015, Collins et al.

2016, Warren et al. 2016). Changes in availability of autochthonous vs. allochthonous C and OM could influence consumer community and functional feeding group composition

(Thompson and Townsend 2005, Ledger et al. 2013, Wallace et al. 2015).

Temporal variability in numerous physical, chemical, and biological factors may also alter isotopic signatures of stream and river consumers and their nutritional resources. For example, changes in runoff, discharge, and flow can lead to variations in aquatic-atmospheric CO2 exchange and in the sources of inorganic nutrients (Finlay

2004, Finlay and Kendall 2007, Hadwen et al. 2010, Hladyz et al. 2012, Peipoch et al.

2012), which can affect the δ13C, δ15N, δ2H, and 14C values of aquatic primary producers. Time-dependent changes in isotopic signatures of aquatic and terrestrial

82 primary producer OM may therefore influence assessments of nutritional contributions supporting stream consumers (Hadwen et al. 2010).

In order to assess the temporal scales of the factors influencing nutritional resource inputs and utilization by stream macroinvertebrates, a field study was conducted in a temperate and open-canopy stream using natural abundance stable isotopes (δ13C,

δ15N, and δ2H) and radiocarbon (∆14C) measurements of both macroinvertebrates and their potential nutritional resources. Our use of this system was intended to reduce the often confounding and synergistic effects of riparian canopy cover and leaf inputs, and to include a potential, isotopically unique source of nutritional OM (i.e., shale). We hypothesized that i) autochthonous OM would contribute more to multiple functional feeding groups (FFGs) during periods of high aquatic primary production (i.e., spring and summer), and ii) allochthonous OM contributions would increase in autumn with decreasing aquatic primary production and increased inputs of terrestrial C and OM in

Paint Creek, a tributary of the Scioto River, Ohio.

Methods

Site Description. Paint Creek is a 5th order tributary of the Scioto River (itself a tributary of the Ohio River) in south-central Ohio, USA with a drainage area of 2,960 km2 (Figure 3.1; ODNR Division of Wildlife, 2001). Land use is dominated by agriculture (cultivated crops and pasture/hay, 60% and 10%, respectively), with forest and urban/suburban development accounting for 20% and 4%, respectively (OEPA,

Division of Surface Water, 2012). Along an approximately 370 m reach of Paint Creek, the lower 15 m of the Ohio Shale formation is exposed along the stream bank

(Thornberry-Ehrlich 2010). Our sampling site was approximately 50 m in length adjacent to this exposed shale. Discharge data was collected from a U.S. Geological Survey gaging station ~7 km downstream of the sampling reach (Figure 3.2).

Field Sampling. Macroinvertebrates of different FFGs and their potential nutritional resources were collected from Paint Creek in November 2012, and May, July, and November 2013. In November 2013, an additional site ~4 km upstream of the shale outcrop was also sampled in order to compare the isotopic composition of macroinvertebrates collected from a reach influenced and a reach not influenced by the shale outcrop. Careful attention was given to where in a given reach organisms were collected from (e.g., from riffles or runs), since small scale spatial heterogeneity can influence the inputs and sources of OM consumed and the stable isotope values of aquatic primary producers (Pringle et al. 1988, Finlay 2001, Finlay et al. 2002, McNeely et al.

2007). Macroinvertebrates were collected by hand-picking them from rocks and logs, and a kick net was used where sediments and aquatic vegetation were more extensive.

Primary consumers and predators were placed in separate, acid-cleaned, Rubbermaid storage boxes filled with filtered stream water, and allowed to void their guts for 24 hours

(Brooke et al. 1996). After gut voidance, organisms were immediately placed in baked

(525° C) aluminum foil pouches and frozen on dry ice.

Dominant potential aquatic and terrestrial (including shale OM) nutritional resources were collected at all sampling times. Stream sediment samples were collected using 12 cm-long 60 cc clean plastic syringe “corers” that had the syringe tops cut off. In

November 2012, terrestrial soils were collected by digging a hole of 20 cm depth on level

84 ground within 10 m of the stream and sampling material at 1 cm and 20 cm depths using a baked spatula. Terrestrial vegetation (riparian shrubs and deciduous trees) and aquatic vegetation samples were collected by hand using clean disposable nitrile gloves. Aquatic vegetation was not collected in July or November 2013 because it was not present in the stream in quantities large enough for isotopic analyses. Suspended POM was collected by filtering stream water from each sampling site through a baked 47 mm QMA filter (0.8

µm nominal pore size). All solid-phase materials were individually wrapped in pieces of baked aluminum foil, placed in a Ziploc bag, and immediately frozen in the field on dry ice. Stream biofilm samples were collected from 2-3 cobbles of ~15-20 cm diameter by scraping the surface with a toothbrush or a baked razor blade and rinsing the cobble surfaces with DI water into acid-cleaned (10% HCl) polycarbonate bottles that were immediately frozen on dry ice.

Stream dissolved inorganic and organic C (DIC and DOC, respectively) and nutrient samples were also collected. For DIC, stream water was filtered through baked

47 mm QMA filters and injected by syringe into baked crimp-sealed 125 ml serum bottles poisoned with 200 μL HgCl2 that had been sparged free of air with ultra-high purity N2 gas. DIC bottles were stored in the dark at ambient temperature until processing. For DOC and nutrients (N and P), stream water was filtered (~250 ml) into acid-cleaned polycarbonate bottles and frozen at -20º C until processing. Water chemistry data, including temperature, pH, dissolved oxygen (DO), and conductivity were collected using a Model AP110 Accumet Portable pH/ORP meter, a YSI ProODO handheld DO meter, and a Model 30/10 FT YSI multiparameter probe, respectively (Table 3.1).

Laboratory Methods. Macroinvertebrates were sorted and identified to genus whenever possible and assigned to a FFG according to Merritt, Cummins, and Berg

(2008) (Table C.1). Due to their different isotopic composition and known differences in feeding behaviors (see Results and Discussion), for purposes of statistical analysis and mixing model runs (see below), Anthopotamus spp. mayflies were treated as their own group, rather than grouping them with the other filtering collectors (isonychiid mayflies and hydropsychiid caddisflies).

In preparation for stable isotope (δ13C, δ15N, δ2H) and natural radiocarbon (∆14C) analysis, macroinvertebrates, terrestrial vegetation, and algae were dried in a drying oven for 24 hours at 60º C and homogenized to fine powders using a solvent-cleaned mortar and pestle. Smaller macroinvertebrates were pooled to ensure adequate sample size for isotopic analyses (Table 3.2). Biofilm, terrestrial soil, and aquatic sediment samples, and

QMA filters containing suspended POM, were dried as for the macroinvertebrate and vegetation samples and then acid-fumed using fresh concentrated HCl in a clean glass desiccator prior to homogenization to ensure removal of inorganic carbonates. A fraction

2 of each sample was set aside prior to acid fumigation for δ H analysis. Following removal of carbonates, biofilms, soils, and sediments were again dried and homogenized as for the macroinvertebrate and vegetation samples.

Homogenized samples and filter portions were packed in tin capsules for δ13C and

δ15N analysis at the Ohio State University Stable Isotope Laboratory using a Costech elemental analyzer, connected to a Delta IV stable isotope mass spectrometer (IRMS) via a Conflo III interface, or at the University of California at Davis Stable Isotope Facility,

86 using a PDZ Europa ANCA-GSL elemental analyzer interfaced with a PDZ Europa 20-

20 isotope ratio mass spectrometer. Measured δ13C and δ15N values are reported relative to the V-PDB and air international standards, respectively. For δ2H analyses, homogenized, non-acid-fumed subsamples of macroinvertebrates, aquatic and terrestrial vegetation, soil, and sediment were packed in silver capsules. δ2H values were measured by the Colorado Plateau Stable Isotope Laboratory at Northern Arizona University. In order to determine the δ2H values of the nonexchangeable H fraction of solid samples, a benchtop equilibration method was used that allowed for the exchange of sample H with

H in local water vapor (Wassenaar and Hobson 2003, Doucett et al. 2007). Samples were then analyzed on a Thermo-Finnigan elemental analyzer coupled to a Delta Plus XL

IRMS.

For 14C analyses, selected acid-fumed macroinvertebrate and potential nutritional sources from the November 2012, May 2013, and November 2013 samplings were placed in quartz tubes with Cu and CuO and combusted to CO2 at 750ºC for 4 hours.

Within 24 hours, the CO2 from each sealed tube combustion was purified and quantified on a vacuum extraction line and sealed in pre-baked 6 mm Pyrex tubes. The CO2 was reduced to graphite and analyzed for Δ14C at the NSF Arizona Accelerator Mass

Spectrometry (AMS) Facility at the University of Arizona or the National Ocean

Sciences Accelerator Mass Spectrometry (NOSAMS) laboratory in Woods Hole, MA

(USA).

DIC samples were acidified in sealed serum bottles and the evolved CO2 gas was collected cryogenically on a vacuum extraction line and stored and analyzed for ∆14C as

87 above. Selected DOC samples were analyzed for δ13C following a modification of the method described by Osburn and St-Jean (2007) using an Oceanographics International model 1030C DOC analyzer, connected to a modified Graden CO2 trap and a Thermo-

Finnigan Delta Plus XL IRMS at the U.S. Geological Survey, Woods Hole, MA (USA).

Statistical Analyses. To evaluate temporal variation in macroinvertebrate and FFG isotope compositions, statistical analyses were conducted using the PRIMER software package (Clarke and Gorley 2006; v. 6, PRIMER-E Ltd). Macroinvertebrate δ13C, δ15N,

δ2H, and ∆14C values were normalized and visually evaluated using non-metric multidimensional scaling (NMDS). One-way analyses of similarity (ANOSIM) were conducted for macroinvertebrate isotopic data in order to determine whether significant isotopic differences existed between a) sampling dates, b) FFGs, c) and the combined factors of sampling date and FFG. The ANOSIM pairwise test statistic R is a better indicator of separation between groups than the p-value because it is not influenced by sample size (which was small in this case; (Clarke and Gorley 2006). In cases where the overall p-values for the ANOSIM model were significant (p<0.05), significant pairwise comparisons (i.e., p ≤ 0.05) were examined as well as pairwise comparisons that were not statistically significant but with R-values greater than 0.8.

Isotopic Mixing Models. Contributions of potential sources of macroinvertebrate nutrition were estimated using MixSIAR (Stock and Semmens 2013), an isotopic mixing model that uses a Bayesian approach (Moore and Semmens 2008) via a graphical user interface (GUI) and R Statistical Software (R Core Team, 2014). Bayesian isotopic mixing models are better than simple isotopic mass balance models at incorporating

88 uncertainty in source contribution estimates (Moore and Semmens 2008). Sources of variation that may impact nutritional source contribution estimates include, but are not limited to, possible utilization of multiple nutritional sources by consumers, isotopic fractionation, and spatial and temporal variability of isotopic signatures of both sources and consumers (Finlay et al. 2002, Moore and Semmens 2008). Details of how isotopic data were prepared and handled in mixing models, and how mixing models were run, are provided in Appendix C, Section 1.

Results and Discussion

Stable Isotopic Composition of Potential Nutritional Resources

Biofilm (when available), POM, and algae were the most 15N-enriched potential nutritional resources compared to terrestrial vegetation, soil, and sediment at all sampling times (Tables C.2-5). Algae were more 15N-depleted in May 2013 (5.0 ± 1.5‰) compared to both measured and estimated algal δ15N in November 2012, 2013, and July

2013 (range of δ15N means 9.0‰ to 15.1‰) (Tables C.2-5). Similar levels of 15N- depletion were observed for POM in May 2013 compared to other sampling dates (Table

C.3). Stream sediments were the most 15N-depleted of all potential sources of nutrition

(Tables C.2-5). Similar to δ15N trends, algae and POM were also consistently the most

13C-depleted at all sampling times, but only algae were consistently the most 2H-depleted

(Tables C.2-5). Algae, terrestrial vegetation, and soil OM alone did not appear to constrain all macroinvertebrate δ13C and δ15N values (Figure 3.3A). However, δ2H values for algae, terrestrial vegetation, and soil adequately constrained most macroinvertebrates

(Figure 3.3B), and appeared to better explain their nutritional resources in Paint Creek than δ13C and δ15N.

The observed temporal variability in δ15N of algae and POM is likely attributable to agricultural activity in the Paint Creek watershed (Tables C.2-5), as human activities may cause extreme seasonal shifts in δ15N values in aquatic systems (Diebel and Vander

Zanden 2009, Nishikawa et al. 2009, Peipoch et al. 2012, Pastor et al. 2014). The 15N- depleted algae and POM may at least in part be explained by increased inputs of 15N- depleted synthetic fertilizer use in spring (May) compared to later in the year (July and

November) (Tables C.2-5), and the subsequent uptake of this fertilizer nitrogen (N) by primary producers (Peterson and Fry 1987, Vitosh et al. 1995, Randall et al. 2008, Pastor et al. 2014). We also cannot eliminate the possibility that increases in available inorganic

N (Table 3.1) led to greater discrimination against 15N by primary producers, leading to a more 15N-depleted signature in them and their macroinvertebrate consumers (Evans 2001,

Pastor et al. 2014). The 15N-enriched algae observed in July and November (Table C.2) may result from the addition of organic (animal manure) fertilizer to crops later in the growing season, denitrification of inorganic fertilizers, or from human septic waste

(Aravena et al. 1993, Harrington et al. 1998, Vander Zanden et al. 2005, Lefebvre et al.

2007, Diebel and Vander Zanden 2009).

In contrast to δ15N, algal δ13C values showed much lower temporal variability

(δ13C range of means: -32.3 to -30.6‰), in spite of significant changes in precipitation and discharge over the study (Figure 3.2). We predicted greater variability in δ13C of algae because flow in lotic systems influences CO2 availability and fractionation of

13C/12C by algae (Finlay 2001, 2004). However, most of our algal samples were collected from riffle areas in Paint Creek, where flow is consistently higher and turbulent mixing between the atmosphere and water column is greater (Finlay et al. 1999), which may explain the lower variability in our algal δ13C values.

Natural 14C Composition of Potential Nutritional Resources

Soil, sediment, and shale OM in the Paint Creek basin were the most 14C-depleted of the potential nutritional resources measured (mean ∆14C = -316‰, -768‰, and

-999‰, respectively; equivalent 14C ages of 3,050, 11,736, and 50,000 years B.P., respectively) (Figure 3.5). On the basis of terrestrial plants being known to fix modern

14 atmospheric CO2 (mean ∆ C=39 ± 2‰ in 2011) (Randerson et al. 2002, Levin et al.

2013), terrestrial vegetation was the most 14C-enriched potential nutritional resource

(Table C.8). ∆14C values of algae, POM, and biofilm were slightly below modern in age, and the range of variability of ∆14C for these groups was only ~30‰ (∆14C range of means -82 to -52‰; equivalent 14C ages of 690 and 430 years B.P., respectively) (Table

C.8). Across all sampling times, the slight 14C-depletion observed in algal primary production (mean ∆14C = -52 ± 7‰, 430 years B.P. equivalent age) (Table C.8) and biofilm (mean ∆14C = -63 ± 8‰, 525 years B.P. equivalent age) (Table C.8) is best explained by the ∆14C values of dissolved inorganic carbon (DIC; mean ∆14C = -54 ±

7‰, 450 years B.P. equivalent age) (data not shown).

It is important to note that our DIC findings represent specific points in time at the

Paint Creek sampling site, whereas algal ∆14C and 13C values are presumably more

14 13 integrative of the  C and  C values of DIC and CO2(aq) fixed over longer periods of

91 time. Concentrations and 14C and 13C values of DIC in lotic systems can vary significantly over hours to days (Finlay 2003, Raymond et al. 2004, Parker et al. 2010), and it is likely that we did not capture their full temporal variability with our measurements. In Paint Creek, relatively small contributions from either respired aged soil OM or dissolution of fossil aged carbonates (Division of Geological Survey, ODNR) could account for some or all of the 14C-depletion observed in the DIC pool (Raymond et al. 2004, Marwick et al. 2015), algae, and biofilm OM.

Stable Isotopic Signatures of Macroinvertebrates

Macroinvertebrates collected in November 2013 were generally the most 15N- enriched of all animals collected over this study (range of means 14.2‰ to 17.3‰)

(Table 3.2, Figure 3.3A). However, some overlap in organism δ15N values between sampling dates was observed because of their generally high variability (δ15N range of means 10.4‰ to 17.3‰) (Table 3.2, Figure 3.3A). Macroinvertebrates collected in May

2013 were the most 15N-depleted, with the exception of several July 2013 organisms. The variability in δ15N values of macroinvertebrates collected in May 2013 was lower (range of means 10.4‰ to 11.3‰) than that observed for any other sampling date (Table 3.2,

Figure 3.3A).

The δ13C values of macroinvertebrates were also highly variable (δ13C range of means -31.9‰ to -27.0‰) and had no visually apparent relationship to sampling date

(Table 3.2, Figure 3.3A). Similarly, macroinvertebrate δ2H values spanned a range of nearly 80‰ (δ2H range of means -192.2‰ to -113.8‰), also with no visually apparent temporal pattern (Table 3.2, Figure 3.3B). Variability in the stable isotopic composition

92 of macroinvertebrates can be due to the assimilation of different nutritional resources or changes in the isotopic signatures of nutritional resources over time (Finlay 2004, Finlay and Kendall 2007, Middelburg 2014). Nevertheless, one-way ANOSIM by sampling date confirmed that macroinvertebrate isotopic compositions varied significantly by sampling date, and all pairwise comparisons were significant with the exception of November 2012 vs. July 2013 (Table C.10). This suggests that the types and/or stable isotopic composition of nutritional resources available for macroinvertebrate assimilation were significantly different in November 2012 compared to July 2013.

Stable Isotopic Relationships Between Functional Feeding Groups

Examination of macroinvertebrate stable isotopic compositions by FFG revealed that predators, as well as some filtering collectors and scrapers, were the most 15N- enriched (Table 3.2). Filtering collectors were also consistently the most 13C-depleted

FFG on all sampling dates, whereas some individual scraper organisms were the most

13C-enriched (Table 3.2, Figure 3.3C). Anthopotamus spp. mayflies, which are also considered filtering collectors, were more 13C-enriched compared to the other filtering collectors (mean δ13C range -28.5‰ to -27.4‰ vs. -31.9‰ to -30.8‰, respectively)

(Table 3.2, Figure 3.3C). The difference in δ13C values of Anthopotamus spp. mayflies compared to other filtering collectors (isonychiid mayflies and hydropsychiid caddisflies) may be due to Anthopotamus spp. being sediment burrowers that filter materials from the sediment-water interface, whereas hydropsychids and isonychiids primarily filter materials from the water column (McCafferty and Bae 1992, Merritt, Cummins, and Berg

2008). Filtering collectors were generally the most 2H-depleted FFG (Table 3.2, Figure

3.2D). Examination of composite stable isotopic composition by FFG suggests that different FFGs are relying on different nutritional resources, leading to variability in the stable isotopic compositions between different FFGs (Cremona et al. 2010, Kobayashi et al. 2011, Cummins 2016). However, differences in stable isotopic composition by FFG could also be at least in part explained by differences in isotopic fractionation during catabolic processes, lipid synthesis, or deamination (Bunn et al. 2013, de Carvalho et al.

2015).

Stable isotope biplots suggested clustering of macroinvertebrate individuals by

FFG (Figure 3.3C, D). Filtering collectors and Anthopotamus spp. mayflies showed the overall lowest stable-isotopic variability (Figure 3.3C, D). One-way ANOSIM by FFG also showed significant differences in the stable isotopic composition of macroinvertebrates belonging to different FFGs (R=0.38, p=0.001), but pairwise comparisons revealed no significant differences between predators and Anthopotamus spp. mayflies, between predators and scrapers, or between scrapers and Anthopotamus mayflies (Table C.11). Similarities in the stable isotopic composition between these FFGs suggests that predators, scrapers, and Anthopotamus spp. assimilated similar nutritional resources.

When macroinvertebrates were organized by both sampling date and FFG, one- way ANOSIM revealed several significant differences in the isotopic compositions of specific FFGs between sampling dates (Figure 3.4, Table C.12). In particular, macroinvertebrates belonging to the filtering collector FFG significantly differed between

November 2012 and 2013, May and July 2013, and July and November 2013 (Figure 3.4,

Table C.12). Macroinvertebrates belonging to the scraper FFG also differed between May and November 2013 (Figure 3.4, Table C.12). As filtering collector isotopic composition was consistently different across different sampling times, it suggests a shift in nutritional resource utilization over time by this FFG, or a change in the isotopic composition of their nutritional resources with time.

The δ15N values of all FFGs followed a seasonal pattern similar to that of algal

δ15N, with the lowest values being observed in May (Table 3.2). Seasonal variability in

δ15N was generally greatest for filtering collectors (Figures 3.3A, C), and the δ15N values alone are likely the primary factor contributing to the significant clustering of filtering collectors by season in the NMDS analysis (Figure 3.4). This is expected because filtering collectors assimilated the most algal biomass in all mixing model scenarios compared to other FFGs (see Bayesian Isotope Mixing Model Outcomes section).

Similarities in the δ15N values of basal nutritional resources (at the base of the food chain, e.g., terrestrial and aquatic primary producers) and macroinvertebrates, as well as the influences of agricultural activity on δ15N values of nutritional resources and consumers, are well-documented (Anderson and Cabana 2005, Vander Zanden et al. 2005, Diebel and Vander Zanden 2009, Peipoch et al. 2012). Agricultural activity in the Paint Creek watershed may have influenced sources of N available to algae and their δ15N values, and macroinvertebrate consumption of this algal OM may explain the observed δ15N values of macroinvertebrates in this study.

Apparent 14C Ages of Aquatic Macroinvertebrates

∆14C values for both filtering collectors and Anthopotamus spp. mayflies spanned a relatively small range (-58‰ to -41‰; equivalent 14C ages of 480 and 340 years B.P., respectively) (Table 3.2, Figure 3.5). Filtering collectors from May and November 2013 were the most 14C-depleted and 14C-enriched FFG, respectively (Table 3.2). The same pattern was observed for Anthopotamus spp. mayflies (Table 3.2), though sample size for

Anthopotamus spp. was small (n=2). One-way ANOSIM by FFG including both stable isotope and natural abundance 14C data further showed a significant difference in the stable and radiocarbon isotopic composition of Anthopotamus spp. mayflies and filtering collectors (isonychiid mayflies and hydrpsychid caddisflies), even though both are categorized in the filtering collector FFG (p=0.015, R=0.589). Similar to observations using macroinvertebrate stable isotopic compositions alone, differences in 14C signatures and apparent ages of Anthopotamus spp. and other filtering collectors suggest reliance on different forms of nutrition between these two types of consumers.

Examination of macroinvertebrate ∆14C values (Table 3.2) suggests that one or more forms of aged carbon or OM contributed to macroinvertebrate biomass in Paint

Creek. Macroinvertebrate ∆14C-13C isotopic space could be established by direct consumption and assimilation of a combination of modern (e.g., terrestrial vegetation) and aged (e.g., soil, sediment and suspended POM; Figure 3.5) nutritional sources, as has been documented in a few other studies (Caraco et al. 2010, Wang et al. 2014, Evans and

Bauer 2016a, 2016b). Macroinvertebrate ∆14C values were, however, also well- constrained by DIC and algal ∆14C values (Figure 3.5), suggesting that 14C-depletion in

96 macroinvertebrates could also be accounted for by utilization of 14C-depleted algae that appear “aged” because of their fixation of aged DIC (Ishikawa et al. 2014, 2016). We chose to conduct this study at Paint Creek because of the proximity of the sampling reach to a major outcrop of fossil-aged shale, and because previous studies have shown that fossil OM from exposed shale is biologically labile and assimilated by heterotrophic bacteria (Petsch 2001, Schillawski and Petsch 2008). As a result, we also cannot rule out the potential for a very small contribution (e.g., < ~ 7%) of fossil (∆14C = -1,000‰;

Figure 3.5) shale-derived OM to explain the relatively small 14C-depletion that we observed in Paint Creek macroinvertebrate biomass. Respiration of shale-derived OM by bacteria could have also contributed to the age of the DIC pool (Schillawski and Petsch

2008, McCallister and del Giorgio 2012). Whatever the specific reason(s) for the observed 14C-depletion of macroinvertebrates in Paint Creek, these consumers depend in part on inputs of aged C and OM, as has been found for the majority of all consumers in inland waters (Bellamy and Bauer, submitted), thus linking geologically aged C and OM reservoirs with contemporary aquatic food web nutrition and energetics.

Bayesian Isotope Mixing Model Outcomes Using Stable Isotopes

Nutritional source contribution estimates and associated distributions (5%, 50%, and 95% posterior probabilities) for each mixing model run are provided in Appendix C,

Section 3. Mixing model outcomes using the global data set across all sampling dates revealed that algae were the primary nutritional resource to all FFGs (48-67%; Figure

3.6A). However, terrestrial vegetation also made significant contributions to macroinvertebrate biomass (16-41%), with the exception of scrapers (Figure 3.6B). In

97 addition, soil OM contributed significantly, and more than terrestrial vegetation, to the biomass of certain groups, i.e., both scrapers and Anthopotamus spp. mayflies (37% and

34%, respectively) (Figure 3.6C).

Mixing model results for the combined May and July 2013 datasets also revealed that algae were the primary contributor to macroinvertebrate biomass for all FFGs (56-

75%) (Figure 3.7A) with the exception of predators. Predator biomass was instead dominated by contributions from terrestrial vegetation (43%) (Figure 3.7A). Soil-derived

OM also comprised relatively large contributions to the nutrition of scrapers,

Anthopotamus spp. mayflies, and predators (24-31%) (Figure 3.7A).

Results of the mixing model for the combined November 2012 and 2013 data sets suggested 16% less algal assimilation by all FFGs compared to the May and July 2013 model outcomes, with the exception of predators (Figure 3.7B). Terrestrial vegetation was also estimated to contribute on average 8% more to the biomass of all FFGs in

November 2012 and 2013, and estimated soil-derived OM contributed 15% and 16% more to scraper and Anthopotamus spp. mayfly biomass, respectively (Figure 3.7B), compared to May and July 2013. In general, mixing model outcomes suggest that allochthonous OM was collectively more important for consumer nutrition in November

2012 and 2013 than it was in May and July 2013.

The increased utilization of autochthonous primary production by all FFGs except predators in May and July 2013 compared to November 2012 and 2013 may be a result of differences in the relative availability of different nutritional resources across sampling dates and/or seasons. Increased utilization of autochthonous primary production by

98 macroinvertebrates during times of higher discharge was unexpected, as greater inputs of allochthonous OM are expected during high flow periods (Junk et al. 1989, Atkinson et al. 2009, Cook et al. 2015). However, the absolute and relative amounts of a given nutritional resource are not necessarily related to its utilization by and bioavailability to consumers (Marcarelli et al. 2011, Pingram et al. 2012, Guo et al. 2016a). For example, aquatic organisms often preferentially consume and assimilate nutritional resources with a high essential fatty acid concentration (e.g., autochthonous production), as these forms of nutrition better support somatic growth and reproduction (Torres-Ruiz et al. 2007, Guo et al. 2016a).

Evidence also exists that even during periods of high discharge there is little change in contributions of allochthonous vs. autocthonous OM supporting aquatic food webs (Delong et al. 2001, Hladyz et al. 2012). Furthermore, the location of our sampling reach downstream of a dam may have potentially provided additional sources of autochthonous nutrition to macroinvertebrates that were stored and released from upstream reservoirs over different periods of time (Doi et al. 2008, Wellard Kelly et al.

2013). Mixing models suggested that terrestrial vegetation comprised a larger proportion of consumer biomass in November compared to May and July, suggesting that the increased availability of terrestrial OM via leaf senescence did contribute significantly to macroinvertebrate biomass in the present study (Figure 3.5A, B) which has been observed in other lotic systems (Huryn et al. 2001, Hladyz et al. 2012, Roach 2013).

Macroinvertebrate tissues integrate the contributions of nutritional resources over some extended period of time, whereas our collection of potential nutritional resources

99 was temporally discrete. Such factors must be considered when attempting to reconcile what nutritional resources best explain the macroinvertebrate stable and radiocarbon isotopic signatures in this and any study. For example, mixing model outputs suggested that predatory macroinvertebrates appeared more reliant on allochthonous OM in May and July, as well as in November, in contrast to what was observed for primary consumers (Figure 3.7 A, B). Tissue turnover for aquatic primary consumers can be relatively rapid (days to weeks; Fry and Arnold 1982, McIntyre and Flecker 2006). In contrast, tissue turnover in predatory macroinvertebrates likely takes longer due to their generally larger body sizes compared to most primary consumers, and predators may integrate the stable and radiocarbon isotopic values of nutritional resources over longer periods of time than primary consumers (Atkinson et al. 2014, Zanden et al. 2015).

Collection of potential nutritional resources over longer and more continuous periods of time is thus important for more accurate assessment of temporal changes in nutritional resource utilization, especially for primary consumers. Nonetheless, longer tissue turnover time may help explain why predatory macroinvertebrates did not appear to use the same basal nutritional resources as primary consumers in the present study.

Impact of the Inclusion of Natural 14C on Model Outcomes

Inclusion of ∆14C data in the mixing model runs for Anthopotamus spp. and filtering collectors supported the stable isotope-only findings of algae as the primary contributor to the nutrition of these two groups (51-69%) (Figure 3.7C). Soil OM contributions using the ∆14C data were slightly more important to the nutrition of

Anthopotamus spp. than for other filtering collectors (14 and 7% contributions,

100 respectively) (Figure 3.7C) compared to using stable isotopes alone. Also relative to the stable isotope-only findings, there was an 18% increase in the importance of terrestrial vegetation to Anthopotamus spp. when ∆14C was included (Figure 3.7C).

Inclusion of ∆14C data in the model also suggests that modern allochthonous OM, in the form of recently living terrestrial vegetation, was of greater importance to

Anthopotamus spp. mayflies than was moderately aged soil OM (Figure 3.7C). Fossil shale-derived OM, while abundant in the immediate watershed surrounding our sampling site, contributed negligible amounts to macroinvertebrate biomass in Paint Creek (Figure

3.7C). This further confirms our contention that the amounts of a potential nutritional resource do not necessarily predict its availability to and utilization by different consumer groups. However, mixing models still suggested that relatively small amounts (7-13%) of moderately aged soil-derived OM may contribute to macroinvertebrate biomass in Paint

Creek. This finding also potentially explains at least some of the 14C-depletion in macroinvertebrate tissues that we observed.

Conclusions

Findings from the present study indicate that OM from aquatic primary production represents the greatest contributor of various potential nutritional sources to macroinvertebrate primary (non-predatory) consumers in a temperate watershed, even in the presence of large amounts of fossil OM shale. These autochthonous contributions were greatest in spring and summer, and decreased in the fall. In contrast, predators were primarily reliant on allochthonous OM regardless of season.

101

Our study somewhat surprisingly did not reveal any direct impacts of river discharge on nutritional resource utilization. ∆14C data also revealed that one or more forms of aged C and/or OM were small but measureable contributors to aquatic consumer biomass. Inclusion of ∆14C data in aquatic food web studies not only provides information about the age of C and OM assimilated, but it can also provide better mixing model nutritional source resolution and potentially alters the interpretation of mixing model outcomes that use stable isotopes alone. The present study also found that the stable isotopic composition — in particular δ15N values — of potential nutritional resources can also change over time in agriculturally influenced watersheds.

Our findings collectively support the use of natural abundance isotopes as useful tracers of the nutritional resources supporting aquatic secondary production. However, different consumer groups and tissues reflect the isotopic signatures of nutritional resources integrated over different periods of time, while sampling of potential nutritional resources typically occurs only at discrete points in time. Therefore, the relationships between these temporal linkages and discontinuities must be taken into account when designing and conducting nutritional resource studies, especially in lotic systems where short-term changes in nutritional resource inputs and abundances may occur.

102

Table 3.1. Sampling dates and water chemistry parameters for the Paint Creek study site.

- - Date Temperature (ºC) pH Dissolved O2 (mg/L) Conductivity (μS) NO3 +NO2 (μg N/L)

11/07/2012 8.8 8.1 ± 0.2 11.3 ± 0.01 ND 709 ± 3

05/06/2013 14.9 8.0 ± 0.1 10.0 ± 0.01 273.3 ± 0.1 2,230 ± 7

07/29/2013 22.5 7.9 ± 0.1 8.1 ± 0.02 280.4 ± 8.1 1,860 ± 0

11/15/2013 12.3 8.1 ± 0.1 9.7 ± 0.5 201.1 ± 27.9 507 ± 4 ND - not determined

103

FFG (n) δ13C (‰) δ15N (‰) δ2H (‰) ∆14C (‰) November 2012 Filtering collectors (5)* -31.1 ± 0.6 12.4 ± 0.8 -176.7 ± 12.7 -48 ± 6 (4)* Anthopotamus spp. (2) -27.4 ± 0.3 12.3 ± 0.4 -154.9 ± 1.2 -51 (1) Scrapers (2) -27.0 ± 0.2 12.0 ± 1.0 -113.8** ND Predators (5) -29.1 ± 1.2 14.2 ± 0.4 -136.7 ± 18.7 ND May 2013 Filtering collectors (7) -31.8 ± 0.3 10.4 ± 0.2 -192.2 ± 5.1 -58 ± 10 (4) Anthopotamus spp. (2) -28.5 ± 0.1 11.3 ± 0.1 -161.0 ± 2.1 -56 (1) Scrapers (2) -29.5 ± 0.3 10.6 ± 0.6 -191.9 ± 0.9 ND July 2013 Filtering collectors (3) -30.8 ± 0.4 10.5 ± 1.4 -158.0 ± 17.9 ND Scrapers (3) -27.2 ± 2.2 13.7 ± 1.1 -142.7 ± 15.3 ND Predators (2) -28.7 ± 0.2 12.5 ± 0.8 -139.1 ± 5.5 ND November 2013 Shale Filtering collectors (8) -31.4 ± 1.1 15.9 ± 0.8 -183.1 ± 11.8 -41 ± 7 (3) Anthopotamus spp. (1) -28.3 14.2 -159.5 ND Scrapers (4) -27.2 ± 1.9 15.0 ± 1.3 -158.4 ± 12.0 ND Predators (4) -28.0 ± 1.9 15.7 ± 1.1 -179.9 ± 24.4 ND November 2013 No Shale Filtering collectors (3) -31.9 ± 0.4 17.3 ± 0.2 -182.0 ± 9.5 ND Predators (3) -30.4 ± 0.8 14.6 ± 0.4 -165.4 ± 40.8 ND *value in parentheses is number of replicate samples, n ND - not determined

104

105

106

16 -80 A November 2012 B -100 14 May 2013 July 2013 November 2013 -120 12 Algae -140 Ter. veg. 10 Soil -160 Sediment 8 POM -180

N (‰) Biofilm H (‰)

  -200 6 -220 4 -240 2 -260

0 -280 -34 -32 -30 -28 -26 -24 -22 -20 -34 -32 -30 -28 -26 -24 -22 -20 13 13

107  C (‰)  C (‰) 16 -80 Filtering collectors C -100 D 14 Anthopotamus spp. Predators Scrapers -120 12 Algae -140 Ter. veg. 10 Soil -160 Sediment 8 POM -180

N (‰)

H (‰)

Biofilm 2

15 

 -200 6 -220 4 -240

2 -260

0 -280 -34 -32 -30 -28 -26 -24 -22 -20 -34 -32 -30 -28 -26 -24 -22 -20 13 13C (‰)  C (‰) Figure 3.3. δ13C, δ15N, and δ2H values of individual macroinvertebrates and their potential dietary sources (means ± SDs) collected from Paint Creek, Ohio, USA. δ13C and δ15N values are corrected for trophic fractionation according to Post (2002) and δ2H values are corrected for contributions of dietary water. A) δ13C vs. δ15N labelled by sampling date, B) δ13C 2 13 15 13 2 vs. δ H labelled by sampling date, C) δ C vs. δ N labelled by 107functional feeding group (FFG), and D) δ C vs. δ H labelled by FFG.

108

Anthopotamus spp. Filtering collectors Algae 0 0 Terr. vegetation Equivalent Soil OM

Sediment OM

Suspended POM

Biofilm

Fossil shale OM 1790 14 -200 C Age (Years B.P.)

C (‰)

 -400 5,000

-600 10,000 -800

-1000 50,000

-34 -32 -30 -28 -26 -24 -22 -20

13C (‰) Figure 3.5. ∆14C and equivalent 14C ages vs. δ13C values of macroinvertebrates and their potential nutritional resources across multiple sampling dates labelled by functional feeding group (light blue and dark blue filled circles). The δ13C values are corrected for trophic fractionation according to Post (2002).

109

110

100 A Algae Terr. veg 80 Terr. soil OM

Percent Contribution Percent 20

0 100 B

Percent Contribution Percent 20

Scrapers Predators Anthpotamus

Filtering collectors 100 C

Percent Contribution Percent 20

Anthopotamus Filtering collectors Figure 3.7. Mixing model results showing median percent nutritional source contributions to macroinvertebrates for A) combined May and July 2013 sampling dates, B) combined November 2012 and 2013 sampling dates, and C) the global model across all sampling dates with stable isotope (δ13C, δ15N, and δ2H) and natural radiocarbon (∆14C) values of macroinvertebrates and potential nutritional sources for which ∆14C data was available. See Methods for details. 111

Chapter 4: Nutritional Support of Inland Aquatic Food Webs by Aged Carbon and Organic Matter

Amber R. Bellamy

James E. Bauer

Submitted to Limnology and Oceanography Letters

112

Abstract

Aged (typically hundreds to thousands of years old) forms of non-living carbon

(C) and organic matter (OM) predominate in many inland water ecosystems. Advances in the methodologies used to measure natural abundance radiocarbon (14C) have led to increased use of natural 14C as both a source and age tracer in aquatic ecosystem and food web studies. Here we reviewed i) ∆14C values and ages of C and OM typically found in different inland water systems, ii) the mechanisms through which these materials enter inland water ecosystems, and iii) all known 14C data on aquatic consumers across a range of inland water ecosystem types. Using ∆14C values of aquatic consumers and their potential nutritional resources, we estimated contributions of aged C and OM to aquatic consumer biomass. We conclude that in nearly every case, one or more forms of aged C or OM contribute to aquatic consumer nutrition in inland water ecosystems.

Introduction

A growing body of evidence shows that aged and even ancient forms of carbon

Hossler and Bauer 2013; Wu et al. 2013; Spencer et al. 2014a) from both external (e.g., watersheds, atmospheric deposition, etc.) and internal (e.g., sediments) sources. The sizes of Earth’s geologically aged C and OM reservoirs (e.g., sedimentary rocks and carbonates) are orders of magnitude larger than modern to moderate-aged reservoirs (e.g., terrestrial vegetation and soils and aquatic sediments; Hedges 1992). These aged C and

OM sources are mobilized and potentially available to inland water ecosystems (for

113 purposes of this review and synthesis, inland waters are hereafter referred to as "aquatic") through various natural and anthropogenically altered routes (Butman et al. 2015;

Marwick et al. 2015). However, the extent to which aged forms of C and OM contribute to aquatic consumer nutrition and food webs, and the factors controlling these contributions, are poorly understood and have not been accounted for in the vast majority of C and energetic budgets (Thorp & Delong 2002; Caraco et al. 2010; Rosi-Marshall et al. 2016). We propose that greater consideration should be given to the potential for incorporation of aged forms of C and OM to aquatic food webs, as this could have profound implications for our conceptual and quantitative models of C and energy flow in, as well as the community structure of, aquatic ecosystems.

The goals of this review and synthesis are to i) conduct an exhaustive compilation and synthesis of literature data on the potential sources and inputs of C and OM to inland water ecosystems using natural abundance 14C as a tracer, ii) assess the ∆14C values and ages of particulate and dissolved organic C (POC and DOC, respectively) and dissolved inorganic C (DIC) pools in streams, rivers, and lakes available for possible utilization by aquatic consumers, iii) assess the ∆14C values and apparent ages of aquatic consumer organisms, and iv) use these data to estimate aged C contributions to aquatic metazoan consumer biomass and nutrition.

Background

Stable Isotopes as Tracers of Aquatic C and OM Nutritional Sources

Natural abundance stable isotopes (e.g., 13C, 15N, and 2H) have been used extensively for evaluating OM utilization by aquatic food webs and have been shown to

114 be far more quantitative than classical gut content analysis (Peterson and Fry 1987;

Junger and Planas 1994; Leberfinger et al. 2011). However, in most aquatic systems 13C and 15N have relatively small dynamic ranges (tens of ‰ at most; Figure 4.1).

Therefore, differentiation of multiple dietary and nutritional sources with overlapping

13C and 15N signatures can be challenging and is often non-definitive (Phillips and

Gregg 2003). Stable isotope ratios of H (δ2H or δD) have a much larger dynamic range than 13C and 15N in natural systems (up to ~100‰ or more; Figure 4.1) and when used in conjunction with 13C and 15N can provide greater (i.e., isotopically three- dimensional) differentiation of dietary and nutritional contributions (Doucett et al. 2007;

Deines et al. 2009; Cole et al. 2011). δ2H has been increasingly used to distinguish between allochthonous and autochthonous sources of OM in aquatic systems because terrestrial vegetation can be enriched in 2H by ~100‰ or more over aquatic vegetation

(Doucett et al. 2007; Finlay et al. 2010). The main limitation of using stable isotopes in aquatic dietary and nutritional studies is that they cannot easily discriminate between forms of C and OM that are derived from rapidly cycled contemporary (i.e., modern- aged) sources and those that may originate from the mobilization of aged (i.e., century to fossil aged) sources that are far more abundant in different Earth reservoirs.

Natural 14C as a Source and Age Tracer of C and OM in Aquatic Systems

14 Natural abundance radiocarbon ( C), the radioactive isotope of carbon (t1/2=5,568 years; Stuiver and Polach 1977), has at least a one to two order-of-magnitude greater dynamic range than 13C, 15N, and 2H. Natural (“pre-bomb”) Δ14C values range over

~1,000‰ (i.e., from -1,000‰ to ~0‰; McNichol and Aluwihare 2007; Taylor 2016) and 115 are even greater (up to ~1,900‰ range) when considering anthropogenic 14C inputs such as thermonuclear weapons testing and nuclear reactors (McNichol and Aluwihare 2007;

Taylor 2016; Figure 4.1). In addition to being a highly sensitive source tracer, the radioactive nature of 14C uniquely allows for determination of the ages of C and OM in both non-living and living aquatic C-containing components in which Δ14C is measured.

Natural abundance 14C was historically employed much less frequently than stable isotopes in studies of both aquatic and terrestrial food webs. This was due in large part to the challenges in obtaining adequate sample C quantities, lengthy sample processing times, and analytical cost (McNichol and Aluwihare 2007). However, the advent of accelerator mass spectrometry (AMS) has resulted in increased use of natural 14C measurements in ecosystem studies. Prior to the development and widespread use of

AMS, 14C analyses were conducted using -decay methods such as gas proportional and liquid scintillation counting (McNichol and Aluwihare 2007; Taylor 2016). Decay counting requires grams of C and up to days of counting time per sample for an accurate assessment of 14C content, whereas AMS methods require only tens to hundreds of micrograms of C and minutes or less analysis times (McNichol and Aluwihare 2007;

Taylor 2016).

Most inland water food web studies utilizing natural 14C have been conducted in subtropical, temperate, and subarctic North American and European systems, but a growing number are being conducted in Asian systems (Table 4.1). To our knowledge, no studies have yet been carried out in desert (e.g., hypersaline) and tropical systems. While the use of natural 14C for determining the ages of C and OM utilized in aquatic food webs

116 is in principal straightforward, caution must be applied when attributing the apparent ages of consumer organisms exclusively to their utilization of non-living aged OM because living aquatic autotrophs may in fact also have apparent age due to their fixation of often aged DIC (as CO2(aq)) in inland waters (Ishikawa et al. 2013; Ishikawa et al. 2014).

Variability in the Ages of Carbon and Organic Matter in Inland Waters

Modern-Aged Forms of Carbon and Organic Matter in Inland Waters

Non-living C and OM in aquatic systems has been found to range from modern

(∆14C>~0‰) to fossil (∆14C<-1,000‰) in age (Table 4.2; Hedges et al. 1997; Raymond et al. 2004; Butman et al. 2012) and may be either autochthonous or allochthonous in origin. Modern living or recently living aquatic and terrestrial vegetation is derived from

14 modern-day CO2 recently fixed from the atmosphere ( C>0‰; Table 4.2; Garnett and

Billett 2007; Gaudinski et al. 2009; Carbone et al. 2013). However, modern terrestrial vegetation is a small global reservoir of organic C compared to moderately aged (i.e., that having ∆14C<0 but >-1,000‰) and fossil aged C reservoirs (i.e., those having ∆14 C=

~-1,000‰; Table 4.2, Figure 4.2). Potential sources of young autochthonous OM in freshwater systems include macrophytes, benthic algae, phytoplankton, cyanobacteria

(Allan and Castillo 2007) and young OM-utilizing heterotrophic bacteria (Table 4.2;

Stevenson et al. 1996; Graham and Wilcox 2000; Hall et al. 2000; Cashman et al. 2013), and biogenic methane (Table 4.2; Chanton et al. 1995; Grey 2016).

Moderately to Fossil Aged Forms of Carbon and Organic Matter in Inland Waters

Non-living aged autochtonous and allochthonous OM sources to aquatic systems include aquatic sediments, terrestrial soils, thermogenic and biogenic methane, and

117 sedimentary rocks that have been stored in watersheds, streams, and rivers for decades to millions of years since their deposition (Table 4.2; Tourtelot 1979; Hedges 1992; Copard et al. 2007; Battin et al. 2009). Soil OM from leaf litter, roots, and woody debris derived from recent primary production are typically modern- or near-modern (i.e., decadal or less) aged (Table 4.2; Gaudinski et al. 2000; Trumbore 2009). Century to millennial aged soil OM typically arises from more highly degraded terrestrial vegetation (Table 4.2,

Figure 4.2; Trumbore, 2009; Rumpel and Kögel-Knabner 2011; Schmidt et al. 2011) and depends on soil depth, horizon, etc. (Trumbore, 2000; Jenkinson et al. 2008; Schrumpf et al. 2013). The resuspension of aged fine soil and sedimentary POM can lead to commonly observed moderately aged suspended POM in the water column as well, especially in turbulent flowing waters (Bianchi and Bauer 2011 and Hossler and Bauer

2013). Biogenic methane, depending on the age of its OM source (e.g., peat) supporting methanogenesis, can also be moderately to highly aged (Chanton et al. 1995; DelVecchia et al. 2016).

Aquatic primary production in aquatic systems is generally assumed to be a source of modern- or near-modern aged OM (Abbott and Stafford Jr., 1996; Zigah et al.

14 2011; Kruger et al. 2016). However, C is often depleted in DIC and CO2(aq) in inland waters relative to atmospheric CO2 (known as the freshwater reservoir offset; Keaveney and Reimer 2012; Philippsen and Heinemeier 2013; Fernandes et al. 2015) and can lead to 14C-depletion and apparent “age” in living autotrophic biomass (Gaudinski et al. 2000;

14 Trumbore 2000). C-depleted DIC and CO2(aq) in aquatic systems can originate from fossil carbonaceous rock weathering and respiration of non-living, aged water column

118 soil and sediment OM (Table 4.2; Broecker and Walton 1959; Hope et al. 2004; Butman and Raymond 2011; Keaveney and Reimer 2012; Ishikawa et al. 2014). This 14C-depleted living or recently living aquatic OM may thus serve as a source of “aged” nutrition to aquatic consumers (Broecker and Walton 1959; Ishikawa et al. 2013, 2014).

Ancient or fossil-aged (∆14C<-1,000‰) OM reservoirs derive from materials stored on geological timescales in sedimentary rocks (e.g., shales, kerogens, coal, petroleum, and thermogenic methane; Table 4.2, Figure 4.2). Fossil aged C and OM are far more abundant in global reservoirs than contemporary or moderately aged C and OM by several orders of magnitude (Figure 4.2; Hedges 1992). Thus, the mobilization of even small amounts these abundant and highly aged materials has the potential to contribute significantly to the amounts and 14C ages of C and OM in modern-day aquatic systems

(Caraco et al. 2010; Hossler and Bauer 2013; Marwick et al. 2015).

Sources and Routes of Input of Aged C and OM to Inland Water Systems

Aquatic systems and terrestrial landscapes are closely linked, and their connectivity facilitates hydrologic inputs of different sources of C and OM of varying age

(Figure 4.2; Likens and Bormann 1974; Hynes 1975; Williamson et al. 2008). While inland waters cover a small fraction of the earth’s surface (~3%; Raymond et al. 2013), C and OM yields from the surrounding landscape are large and can support ecosystem processes such as metabolism and secondary production, leading to inland waters being largely net heterotrophic (Allan 2004; Cole et al. 2007; McCallister and del Giorgio

2012; Wilkinson et al. 2013). Factors such as geomorphology, lithology, climate (e.g.,

119 temperature, precipitation), and land use can influence the forms and corresponding ages of C and OM mobilized to inland waters (Marwick et al. 2015).

Aged materials were historically not thought to be significant components of aquatic bulk C and OM pools, as they are known to be today (Hedges et al. 1986;

Raymond and Bauer 2001a). Human activities have dramatically increased the mobilization of these materials from their storage reservoirs (Figure 4.3; Post et al. 1990;

Falkowski et al. 2000; Regnier et al. 2013; Butman et al. 2015). Consequently, increasing inputs of allochthonous aged materials to aquatic systems may be impacting aquatic food webs to an even greater extent as agriculture, urbanization, and other land use changes have expanded (Aufdenkampe et al. 2011; Butman et al. 2015; Fellman et al. 2015).

Because of the great ages of global geological stores of C and OM (Table 4.2) and their very large reservoir sizes (Figure 4.2), the mobilization of even small amounts may constitute significant impacts on aquatic C and OM and consumer nutrition. Historical paradigms purporting the biological recalcitrance of allochthonous OM based on molecular structure and age are also changing (McCallister and del Giorgio 2012;

Fellman et al. 2014; Marín-Spiotta et al. 2014).

Extensive aging of C and OM results from long-term storage in forms (e.g., sorbed to mineral particles) and environments (e.g., suboxic and anoxic) that prevent their degradation and utilization by heterotrophic bacteria and higher organisms (Salmon et al. 2000; Lützow et al. 2006; Kleber 2010). However, in many cases when aged OM is released from its points of storage and protection, it becomes increasingly susceptible to degradation and utilization (Petsch et al. 2001; Schillawski and Petsch 2008; Kleber and

120

Johnson 2010). Interactions between terrestrial and aquatic environments via the hydrologic cycle therefore likely i) increase the rates and extents of processing of aged C and OM (Blair and Aller 2012; Marín-Spiotta et al. 2014), and ii) modify the routes through which C and OM of differing sources and ages enter streams, rivers, and lakes

(Figure 4.3).

Streams and Rivers

The C and OM in streams and small rivers is typically dominated by allochthonous materials (Figure 4.2) for two primary reasons. First, stream waters arise from surface and subsurface runoff in watersheds, and these waters are in intimate contact with both living and non-living terrestrial biomass, as well as with soils and their associated terrestrial OM (Hope et al. 1997; Aitkenhead et al. 1999; Boix-Fayos et al.

2009). Second, shading from riparian vegetation and canopy cover often strongly limits autochthonous production in streams and small rivers (Vannote et al. 1980; Finlay 2001).

Both natural and anthropogenic controls play important roles in the sources and ages of C and OM entering streams and rivers (Figure 4.2; Hossler and Bauer 2013,

Marwick et al. 2015). Natural controls include hydrogeomorphology, lithology, and climate, all of which can have dramatic temporal and spatial variability (Hossler and

Bauer 2012, 2013). For example, small mountainous rivers (SMRs) – typically underlain by OM-rich kerogen-containing sedimentary rocks - are responsible for the transport and delivery of large amounts of moderately to fossil aged C and OM to streams, rivers, and ocean margins (Kao and Liu 1996; Leithold and Blair 2001; Hilton et al. 2008). Low- gradient watersheds containing fossil carbonates and shales may also introduce aged C

121 and OM to aquatic systems (Figure 4.3; Leithold et al. 2006; Longworth et al. 2007;

Marwick et al. 2015).

Human activities in watersheds are increasingly important controls on the sources and ages of C and OM in streams and rivers, and land use change can alter the ages of materials transported to aquatic systems (Figure 4.3; Butman et al. 2015). Agricultural activity and/or removal of riparian vegetation can lead to higher inputs of moderately aged soil C to streams and rivers (Lal 2003; Syvitski et al. 2005; Longworth et al. 2007;

Restrepo et al. 2015). In urbanized regions, petroleum hydrocarbons in runoff from impervious surfaces, wastewater treatment plants, and aerosol deposition can all transport aged C and OM to streams and rivers (Figure 4.3; Mitra et al. 2002; Paul & Meyer 2008;

Griffith et al. 2009; Wozniak et al. 2012a, b). Increasing stream water temperatures resulting from global climate warming have also led to observed increases in aged C and

OM inputs to streams and rivers (Kaushal et al. 2010, van Vliet et al. 2013), especially in high latitude regions associated with thawing permafrost soils and glaciers (Vonk et al.

2013; Hood et al. 2015).

Lakes

Depending upon their size, lakes are also integrated to a greater or lesser extent with the terrestrial landscape, and are influenced by many of the same mechanisms that affect C and OM inputs to streams and rivers (Caraco and Cole 2004; Cole et al. 2007).

However, in contrast to streams and rivers, lake basins typically serve to a greater extent as storage reservoirs for imported C and OM from the surrounding watershed (Figure 4.3;

Søballe and Kimmel 1987; Cole et al. 2007; Tranvik et al. 2009). Streams are a primary

122 mechanism for the delivery of C and OM of various origins and ages from the surrounding watershed to lakes (Figure 4.3; Kaushal and Binford 1999; Spencer et al.

2014b; Zigah et al. 2014). In Lake Superior, peak stream flow was found to be associated with increased delivery of modern-aged dissolved organic carbon (DOC) derived from upper soil layers, whereas at low flows streams delivered more moderately aged DOC from deep soils and groundwaters to the lake (Zigah et al. 2011).

The surface area and volume of lakes influence the sources and ages of C and OM that predominate in them (Wetzel 1990; Zigah et al. 2012 a, b). Higher watershed:lake surface area ratios generally lead to proportionately greater inputs and amounts of allochthonous and aged C and OM in small vs. large lakes, where modern-aged autochthonous production tends to be quantitatively more important (Zigah et al. 2012;

Cole 2013; Wilkinson et al. 2013).

Anthropogenic activities such as fossil fuel combustion also lead to increased inputs of aged C and OM to lakes and their watersheds via direct deposition of atmospheric particulates and runoff of watershed-deposited particulates (Figure 4.3; Guo et al. 2010; Wozniak et al. 2012 a, b; Barakat et al. 2013). Alpine lakes are especially subject to aged C and OM inputs from combustion-derived aerosol deposition because these aerosols have a tendency to accumulate in cold, high latitude regions (Fernández et al. 2003; Mladenov et al. 2009; Spencer et al. 2014a; Hood et al. 2015). Glacial meltwaters are increasingly recognized sources of aged C and OM from both natural and anthropogenic aerosol deposition (Xu et al., 2009; Spencer et al. 2014a; Spencer et al.

123

2014b). The ages of C and OM in and transported by aquatic systems can thus vary widely, and will depend on the sources and ages of C and OM mobilized to them.

Ages of POC, DOC, and DIC Pools in Inland Water Systems

Streams and Rivers

The majority of studies have found POC to be more highly aged compared to

DOC and DIC in streams and rivers (Table 4.3), and is thought to be a result of moderately to highly aged soil OM and fossil aged bedrock contributing to the POC pool

(Leithold and Blair 2001; Raymond & Bauer 2001a; Blair et al. 2003; Blair et al. 2010;

Marwick et al. 2015). The modern to at most century ages of most river DOC (Hossler and Bauer 2013, Marwick et al. 2015) are attributed to its origination from leachates of fresh surface soil litter and root exudates (Aitkenhead-Peterson et al. 2003; Bianchi 2007;

Schlesinger and Bernhardt 2013). When more highly aged riverine DOC has been found, human activities such as agricultural, wastewater treatment plant, and petroleum inputs are often involved (Wang et al. 2012; Butman et al. 2015; Marwick et al. 2015). Glacial melt water also contributes moderately aged DOC to subarctic streams and rivers (Hood et al. 2009, 2015). With few exceptions (e.g., some agricultural and high relief watersheds), stream and river DIC is generally modern or near modern in age (Table 4.3), indicating that it is predominantly derived from respiration of recently produced forms of organic matter (e.g., terrestrial and aquatic plants; Marwick et al. 2015). 14C and 13C values of DIC are often more challenging to interpret than those of POC and DOC due to effects of respiration, carbonate system equilibrium, and both turbulent and diffusive atmospheric exchange (Finlay 2003; Raymond et al. 2004; Geldern et al. 2015).

124

Lakes

Fewer studies have used natural 14C in lakes than in streams and rivers, but this number has been gradually increasing. The 14C values and ages of DIC, DOC, and POC in the lakes studied to date are variable, but less so compared to streams and rivers

(Tables 4.3 and 4.4). Similar to streams and rivers, POC in lakes is generally more highly and variably aged than DOC or DIC (Zigah et al. 2011, 2012a, b; McCallister and del

Giorgio 2012; Fernandes et al. 2013; Keaveney et al. 2015). This aged lake POC is thought to be derived from often significantly aged deeper soil layers and/or resuspended lake sediments (Meyers and Ishiwatari 1993; Zigah et al. 2011). Also similar to streams and rivers, the DOC and DIC in lakes is often modern- to century-aged (Tables 4.3 and

4.4) and thought to be a product of some combination of recent lake primary production,

POC dissolution to DOC, and respiration of aged DOC and POC to DIC (Zigah et al.

2011; McCallister and del Giorgio 2012). Runoff from the surrounding watershed and inputs from groundwater can also be important sources of aged DOC and DIC to lakes

(Brady et al. 2009; Melymuk et al. 2014). Greater amounts of aged C and OM tend to occur in smaller lakes (i.e., higher watershed:lake area), those occurring in drainage basins with lithologies containing aged forms of organic and inorganic carbon, and those at higher elevations and latitudes receiving glacial runoff (Broecker and Walton 1959;

Keaveney and Reimer 2012; Spencer et al. 2014a; Keaveney et al. 2015).

125

14C Values and Apparent Ages of Aquatic Consumer Organisms

Utilization of Aged C and OM by Aquatic Consumers

When aged C and OM is mobilized from its points of preservation and storage to aquatic systems, it may be assimilated by aquatic bacterial and metazoan consumers as particulate and/or dissolved OM (POM and DOM, respectively; note that POC and DOC refer to only the carbon-containing components of POM and DOM). POM and DOM can contribute to consumer nutrition both directly and indirectly (Figure 4.4). Metazoan consumers may directly ingest aged POM (which also includes aquatic primary producers that have fixed aged aquatic DIC), as well as flocculated aged DOM (Figure 4.4; Wallace et al. 1997; Kerner et al. 2003; Kürten et al. 2013; Scharnweber et al. 2014). Some soft- bodied organisms also have the ability to take up DOM directly (Figure 4.4; Roditi et al.

2000; Baines et al. 2005). In addition to being consumers of aquatic OM, bacteria may also “repackage” aged DOM and POM into aged bacterial biomass that can then be consumed by higher consumers directly or by grazing on protozoan bacterivores (Figure

4.4; Azam et al. 1983; Cherrier et al. 1999; McCallister et al. 2004; Berggren et al. 2010,

2014).

Bacterial Utilization of Aged C and OM in Temperate and Subtropical Systems

It is well-documented that heterotrophic bacteria play an important role in the degradation and utilization of fossil aged petroleum hydrocarbons in many aquatic systems (Atlas 1981; Heitkamp and Johnson 1984; Ahad and Pakdel 2013). Bacteria, through their enzymatic and hydrolytic activities, are also known to mediate the mobilization of C and OM to inland waters (Attermeyer et al. 2013; Collins et al. 2016).

126

Natural 14C measurements show that modern- to moderate-aged (i.e., non-fossil) OM contributes to heterotrophic bacterial biomass (∆14C range -153 to 214‰; 1,330 years to

14 modern-aged, respectively; McCallister et al. 2004) and/or respired CO2(aq) (∆ C range

-172 to 94‰; 1,470 years to modern-aged, respectively; McCallister and del Giorgio

2012) in lake and riverine systems (Hood et al. 2009; McCallister and del Giorgio 2012;

14 Fellman et al. 2014). ∆ C values of respired CO2 from aquatic bacterial incubation studies independently support the contention that aged OM can serve as a source of bacterial energy as well as biomass (Karlsson et al. 2007; McCallister and del Giorgio

2012; Guillemette et al. 2016).

Fossil DOC leached from shales in watersheds may also enter streams and rivers and be utilized by heterotrophic bacteria (Schillawski and Petsch 2008). Petsch et al.

(2001) determined that 74-94% of the lipid C in bacteria growing on watershed shales was derived from fossil OM. Rapid bacterial utilization of fossil DOC solubilized from shale OM (80% loss over 2-week incubation; Schillawski and Petsch 2008) further illustrates that even highly aged forms of OM can be bioavailable once mobilized to aquatic systems.

Metazoan Utilization of Aged C and OM in Temperate and Subtropical Systems

Invertebrates in inland waters play an important role in the processing of autochthonous and allochthonous C and OM through their feeding and metabolic activities (Cummins 1974; Anderson and Sedell 1979; Wallace et al. 2015). They are also an important source of nutrition for higher consumers (e.g., vertebrates). Natural 14C measurements have been conducted on both herbivorous aquatic consumers as well as

127 secondary consumers and predators (Table 4.1), and studies suggest that aged C and OM can be transferred to progressively higher trophic levels (Schell 1983; Wang et al. 2014;

Hagvar et al. 2016).

Streams and Rivers

Natural 14C has been employed as a food web tracer in multiple stream and river systems (Table 4.1). 14C-depletion in living consumer biomass may result from direct consumption and assimilation of i) non-living aged OM (Schell 1983; Caraco et al. 2010;

Hagvar and Ohlson 2013; Wang et al. 2014) and/or ii) living 14C-depleted OM derived

14 from fixation of C-depleted DIC and CO2(aq) by aquatic primary producers (Table 4.2,

Figure 4.5; Fernandes et al. 2012, 2015; Philippsen and Heinemeier 2013; Ishikawa et al.

2014).

A recent study by Caraco et al. (2010) showed that zooplankton (cladocerans and copepods) from the Hudson River, NY had a mean Δ14C of -240‰ (equivalent age of

~2,200 years B.P.) due to utilization of millennial-aged terrestrial soil OM (Figure 4.5).

In this study, zooplankton ∆14C were nearly 200‰ (~1,790 years) lower than the average

∆14C of phytoplankton and about 300‰ (~2,870 years) lower than modern-aged terrestrial plant biomass. Similar conclusions of aged OM assimilation by macroinvertebrate and freshwater mussel consumers were reached by studies in the

Hudson-Mohawk watershed and the Muskingum River (Bellamy et al., submitted a,

Weber et al., in press; Figure 4.5). In the Hudson-Mohawk, utilization of 14C-depleted

“aged” algae by macroinvertebrates could also not be ruled out (Bellamy et al., submitted a). However, in the Hudson-Mohawk tributaries, aged soil OM contributions to

128 macroinvertebrate nutrition were greater in streams with high agriculture vs. low agriculture watersheds (Bellamy et al., submitted a).

In the Kanno River, Japan, deforestation was found to lead to inputs and autotrophic fixation of 14C-depeted DIC from fossil carbonate weathering, resulting in

14C-depeted aquatic primary producer biomass (Ishikawa et al. 2016). Similar to Hudson-

Mohawk tributaries, findings in the Kanno River watershed also showed that modern- aged sources of OM contributed more to stream macroinvertebrate biomass in watersheds having less disturbance (Ishikawa et al. 2016). These same authors further suggested that during forest reestablishment, root biomass of Cryptomeria japonica (Japanese cedar) increased bacterial respiration of soil OM to CO2, and consequently led to greater carbonic acid weathering of soil fossil carbonates and inputs of 14C-depleted DIC to headwater streams (Ishikawa et al. 2016). Thus, various forms of watershed disturbance may facilitate the movement of aged C and OM into aquatic systems, and impact the 14C content and apparent ages of consumers (Caraco et al. 2010, Ishikawa et al. 2016; Figure

4.5). Naturally-occurring forms of highly aged C have also been observed to contribute to aquatic consumer nutrition. Recent work in the Nyack floodplain on the Middle Fork of the Flathead River, Montana, USA revealed that stoneflies in the hyporheic zone were highly aged (up to 6,900 years B.P.; Table 4.1, Figure 4.5), suggesting that they assimilated aged biogenic, and possibly thermogenic, methane by grazing on methanotrophic bacteria (DelVecchia et al. 2016).

129

Lakes and Wetlands

Natural 14C has been used in food web studies of lakes to a lesser extent than those of streams and rivers (Table 4.1, Figure 4.5). In the lakes studied, fixation of aged

DIC and CO2(aq) by lake primary producers is often the dominant pathway through which lake consumers assimilate aged C (Keaveney and Reimer 2012; Fernandes et al.

2013; Keaveney et al. 2015 a). In Eastern Townships lakes (Quebec, Canada), zooplankton (cladocerans and copepods) were used as a proxy for autochthonous primary production by assuming that zooplankton would select algal over detrital components of the POM pool (McCallister and Giorgio 2008). Zooplankton ∆14C values ranged from -

2‰ to 40‰ (modern in age) and overlapped with the ∆14C values of the DIC, confirming the validity of the assumption in this case. As a result, aged terrestrial OM was deduced not to contribute significantly to zooplankton biomass in these lakes (Figure 4.5;

McCallister and Giorgio 2008).

14C-depletion (∆14C=-130‰ to 33‰; 1,120 years to modern-aged) in vertebrates and molluscs in German lakes (Lakes Rosenfeld, Schwerin, Ostorf, and Schloss

Wilhelosthal; Figure 4.5) was also attributed to their consumption and utilization of 14C- depleted algae (∆14C=-44‰; 270 years equivalent age) and macrophytes (∆14C=-33‰;

360 years equivalent age) fixing the 14C-depleted DIC pool, as in Lake Schwerin DIC was 240 years in equivalent age (∆14C=-29‰; Fernandes et al. 2013). However, mussel tissue collected from Lake Ostorf was more 14C-depleted (∆14C=-130‰; 1,120 years equivalent age) than fish and mussels collected from Lake Schwerin. It was assumed by

14 the authors that C-depleted, respired CO2 from aged peat C in the Lake Ostorf

130 watershed contributed to the lake DIC pool and phytoplankton (neither measured directly) that served as a nutritional source to mussels there (Fernandes et al., 2013).

Extensive work in Lough Erne, United Kingdom has similarly shown 14C- depleted algae (mean ∆14C=-59±9‰; 490 years equivalent age), suggesting a significant freshwater reservoir offset, presumably from inputs of weathered fossil carbonates and/or aged soil CO2 to the lake DIC pool (Keaveney et al. 2015 a, b). In contrast, some consumers in Lough Erne were enriched in 14C (maximum invertebrate and fish ∆14C values were 0‰ and 10‰, respectively, and modern-aged; Figure 4.5), suggesting that modern terrestrial OM contributed to their biomass (Keaveney et al. 2015 a). POC was more 14C-depleted (-122±61‰; 1,050 years in age) than DIC or algae, suggesting that there was a moderately aged, but relatively non-utilized, source of OM in the lake. While copepods in Lough Erne apparently selected against significant amounts of this aged OM since their ∆14C values (-50‰ to -4‰; 410 years to modern-aged; Figure 4.5) closely matched algae and DOC (depending on season), cladocerans derived some of their biomass from moderately aged detrital carbon (∆14C=-138‰ to -67‰; equivalent ages of

1,190 to 560 years B.P.; Keaveney et al. 2015 a).

Natural 14C measurements of lake zooplankton and fish suggest that their utilization of aged C and OM is limited. Due to their greater watershed:lake surface area ratios, small lakes are more likely to contain greater amounts aged allochthonous materials, whereas larger lakes (smaller watershed:lake surface areas) contain greater amounts of 14C-enriched modern C and OM (Zigah et al. 2012 a, b; Wilkinson et al.

14 2013) due to autotrophic utilization of modern, C-enriched CO2 from atmosphere-lake

131 exchange. Studies in Lake Superior (the largest lake yet studied) showed that 14C- enriched (young), recently produced aquatic OM was the preferred source of nutrition to zooplankton (cladocerans and copepods; Figure 4.5). Zooplankton ∆14C values were slightly elevated (∆14C=44-65‰; modern-aged; Figure 4.5) compared to POC and DOC

(∆14C=10±29‰ and 38±21‰, respectively; both modern-aged), and contributions of aged OM to the Lake Superior pelagic food web were minimal (Zigah et al. 2012 a).

More recent work in Lake Superior has shown that aged (940 years equivalent age;

∆14C=-117‰; Li et al. 2013) sedimentary OM was assimilated by the benthic amphipod

Diaporeia, which was 14C-depleted relative to non-benthic organisms (e.g., zooplankton and fish; Kruger et al. 2016). However, recently produced 14C-enriched (modern-aged)

OM contributed more than aged sources even to benthic consumer biomass.

In contrast to most lake consumers, the 14C contents of wetland consumers suggest that they rely to varying extents on aged OM. Wang et al. (2014) examined the impacts of agriculture on nutritional resource utilization by fish and macroinvertebrates in the Florida Everglades (Wang et al. 2014). Largemouth bass and sunfish collected from agriculturally impacted regions of the Florida Everglades were 14C-depleted (∆14C range: -52‰ to -22‰; 430 to 180 years equivalent age; Figure 4.5). The authors assumed that OM derived from peat deposits (estimated at ~2,000 years old) exposed by nearby agricultural activity was exported into Everglades wetlands. This aged OM was assimilated by aquatic invertebrates, and in turn consumed by fish (Wang et al. 2014).

Wetlands differ from most lakes in a number of important ways. Their generally shallower nature, smaller volumes, and proximity to often abundant aged sources of OM

132 may help explain the relatively greater reliance of wetland consumers on aged sources of

C and OM.

Bacterial Utilization of Aged C and OM in Subarctic Systems

A commonly used method of assessing the age of OM utilized by aquatic bacteria is to measure changes in the amounts and ∆14C values of DOC before and after a period

(generally days to weeks) of dark incubation (Raymond & Bauer 2001b; Hood et al.

2009; Fellman et al. 2014). Isotopic mass balance between the amounts and ∆14C values of DOC at the start and end of the incubation is then used to infer the ∆14C of the utilized

DOC, and by association, of the biomass of bacteria utilizing it. Laboratory incubation experiments using this approach with DOC from thawing, millennial-aged permafrost

(from Siberian Yedoma deposits) revealed that this highly aged (2010: 24,100 years equivalent age; 2011: 21,700 years equivalent age), 14C-depleted (2010: ∆14C =

-946±25‰; 2011: ∆14C = -933‰) form of OM was utilized by native stream heterotrophic bacteria from the Kolyama River, Siberia (Vonk et al. 2013).

Metazoan Utilization of Aged C and OM in Subarctic Systems

Growing evidence suggests that aged C and OM from permafrost and melting glaciers can also contribute to higher aquatic consumers in downstream rivers and lakes in subarctic ecosystems (Hood et al. 2009; Vonk et al. 2013; Fellman et al. 2014; Spencer et al. 2014a; Fellman et al. 2015). One of the earliest applications of natural 14C in aquatic food web studies was conducted by Schell (1983) using -decay counting in an

Alaskan river-estuarine food web (Table 4.1, Figure 4.5). Anadromous and riverine fish and oldsquaw ducks collected from the Colville River, Alaska, were significantly more

133

14C-depleted (∆14C=-153‰ to -55‰, 1,330 to 460 years equivalent age) than marine fish from the Beaufort Sea (∆14C=17‰ to 204‰, modern-aged) due to the reliance of aquatic invertebrates on millennial aged (~8,700 years B.P.) terrestrial peat and their subsequent consumption by the fish and ducks (Schell 1983).

Recent studies in Norwegian subarctic systems found that tissues from predatory terrestrial invertebrates (spiders, harvestmen, and beetles) were moderately aged (∆14C=

-130‰ to -41‰; 1,100 to 340 years equivalent age), and gut content analysis revealed that aged chironomids (∆14C=-121‰; 1,040 years equivalent age; Figure 4.5) were a primary nutritional source (Hagvar and Ohlson 2013). Hagvar and Ohlson (2013) further postulated that chironomid larvae relied on aged, glacier-derived OM and transferred it to terrestrial invertebrate predators (Hagvar and Ohlson 2013). In the same study, a predaceous diving beetle larva and adult collected from a nearby dam had equivalent 14C ages of 1,200 (∆14C=-139‰) and 1,130 years (∆14C=-131‰), respectively (Figure 4.5).

These values were similar to those for terrestrial predators, which were also likely consuming aged chironomid larvae (Hagvar and Ohlson 2013). Subsequent studies in small lakes in the same region of Norway confirmed these earlier studies’ original hypothesis that the aged glacial C and OM were passing through chironomid larvae before being transferred to terrestrial and aquatic predatory invertebrates (Hagvar et al.

2016).

Studies by Fellman et al. (2015) in the heavily-glaciated Herbert River, AK also provide evidence of the direct assimilation of glacial meltwater-derived C and OM by aquatic organisms. In a non-glacial stream, ∆14C values of macroinvertebrates and fish

134 ranged from -14‰ to 52‰ (115 years to modern-aged), while in the in the glacially impacted upper and lower Herbert River, macroinvertebrates and fish had ∆14C values as low as -171‰ (equivalent age of 1,500 years; Fellman et al. 2015; Figure 4.5). Inputs of aged DOC from glaciers and thawing permafrost to streams, rivers, lakes, and other inland waters are predicted to increase with increasing global temperatures (Hood et al.

2015; Spencer et al. 2015), and this aged OM could increasingly contribute to the secondary production of both bacteria and higher consumers (Fellman et al. 2015; Hagvar et al. 2016).

Estimates of the Assimilation of Aged C and OM by Aquatic Consumers

The use of mixing models in food web studies employing stable isotopes has provided valuable insight to the nutritional sources utilized by consumers (Fry 2007;

Moore and Semmens 2008; Layman et al. 2012). Consumers are often not reliant on a single nutritional source, and using the isotopic values of organisms and their potential sources along with known isotopic fractionation factors, mixing models provide estimates of the relative contributions of the different potential nutritional sources to the consumer

(Phillips and Gregg 2003; Solomon et al. 2011; Phillips et al. 2014). Early linear mixing models relied on a system of algebraic equations (Phillips and Greg 2001), whereas more recent models employ a Bayesian approach and allow for improved incorporation of uncertainty (Phillips & Gregg 2003; Moore & Semmens 2008). Use of multiple (2-4) isotopes in a mixing model can additionally improve estimates of source contributions to consumer biomass by further reducing estimate uncertainty, especially when multiple isotopically unique potential nutritional sources are being considered (Finlay et al. 2010;

135

Phillips et al. 2014). In addition, when natural 14C is used in conjunction with stable isotopes in mixing models, estimates of the ages of potential nutritional resources to consumer biomass are also obtained.

In order to estimate the maximum possible C contributions of varying ages to aquatic consumers in different aquatic systems where natural 14C has been used, we used the most 14C-depleted values for consumer organisms from each study system shown in

Figure 4.5. For illustrative purposes, we employed a simple, linear two endmember mixing model approach with one endmember (nutritional source) ranging from fossil to millennial-aged and the other endmember being modern-aged (i.e., zero age). Since we used only two dominant potential nutritional sources, we could use the following mass balance equations to estimate the proportional contribution of each to the consumer

(Phillips and Gregg 2001):

∆mix = pA*∆A + pB*∆B (1)

pA + pB = 1 (2) where mix represents the mixture (consumer), and A and B represent each nutritional source or endmember, and p represents the proportion of each nutritional source.

14 and ∆mix are the  C values of A, B, and the mixture, respectively.

In our calculations we used three different aged endmembers (A) in order to cover a reasonable range of aged C and OM sources to aquatic systems and consumers, consisting of: 1) a fossil aged endmember (either carbonate or shale-derived OM; 14C=

-1,000‰, age > 50,000 years B.P.), 2) a moderately aged “passive” soil C and OM endmember (∆14C=-540‰, age=6,240 years B.P.) as defined by Hossler and Bauer

136

(2012), and 3) a 50:50 mixture of the “passive” soil endmember (∆14C = -540‰) and a

“slow turnover” soil C and OM endmember (∆14C = 157‰, modern in age; Hossler and

Bauer 2012), yielding an endmember of ∆14C=-192‰ (age=1,710 years B.P.). In their synthesis of the ∆14C and 13C values of POC, DOC, and DIC in eight northeast U.S. rivers, Hossler and Bauer (2012) determined that river OM consisted primarily of varying combinations of the “passive” and “slow” turnover endmembers. The results of the calculations using the three different aged endmembers are shown in Table 4.5.

Streams and Rivers

Using the fossil endmember (∆14C=-1,000‰), our model estimates indicate that fossil C could potentially contribute from 9% to 58% to consumer biomass in all streams and rivers included in our study (Table 4.5). When “passive” and “slow/passive” soil aged endmembers were used, the estimates of aged nutritional contributions to consumers increased correspondingly. Using the “passive” aged soil C and OM endmember resulted in aged nutritional contributions ranging from 17% to >100%, and the combined

“slow/passive” aged soil endmember yielded contributions of 48% to >100% (Table 4.5).

Estimates >100% indicate that contributions from the “passive” and “slow/passive” aged materials are not adequate to account for the entirety of aged consumer biomass.

Therefore, contributions from more highly aged (e.g., fossil or “passive” age ranges) sources of C and OM are necessary to attain the observed ∆14C values for these consumers (Table 4.5).

Some of the studies reviewed here (Caraco et al. 2010; Fellman et al. 2015;

Ishikawa et al. 2014, 2016; del Vecchia et al. 2016) provided estimates of the

137 contributions of aged C and OM to consumer biomass, but the others did not. Our modeled estimates for the contribution of fossil aged C and OM to zooplankton in the

Hudson River (35%) falls within the range of estimates for zooplankton presented by

Caraco et al. (2010), thus supporting the validity of our simple linear 14C mixing approach. However, our estimated 65% to >100% contributions to Hudson zooplankton using “passive” aged and “slow/passive” aged soil C (Table 4.5), respectively, are higher than Caraco et al.’s (2010) estimates (21% and 57% aged C, assuming fossil and moderately aged C endmembers, respectively). The reason for the disparity in our model estimates and Caraco et al.’s (2010) are due to differences in our chosen moderately aged endmember ∆14C values (soil OM; ∆14C = -540‰ and -192‰; 6,240 and 1,710 years equivalent ages) compared to theirs (assumed to be equivalent to their most highly aged zooplankton sample [∆14C=-350‰; 3,460 years equivalent age]). Regardless of the exact age of the terrestrial endmember used, moderately (i.e., millennial) aged C contributed to zooplankton biomass in the Hudson River.

Fellman et al. (2015) used nutritional source modeling to estimate that millennial- aged, glacial meltwater-derived C and OM contributed 25-36% to consumer biomass in the Herbert River, AK. This is greater than our estimates using fossil and “passive” aged soil endmember contributions (9% and 17%, respectively; Table 4.5), but lower than our estimate using the “slow/passive” aged soil endmember (48%; Table 4.5). Ishikawa et al.

(2014, 2016) estimated mean contributions of periphyton (which was the aged endmember in their studies) to consumer biomass to range from 23% to 68% to stream consumers in the Lake Biwa watershed, and over 70% in Kanno River consumers. Mean

138 contributions of periphyton to consumer biomass were dependent on the specific functional feeding group (defined as the organism’s mechanism of acquiring food) and site (Ishikawa et al. 2014, 2016). These workers’ aged C estimates were also larger than our fossil aged C contribution estimate (18-20%; Table 4.5), but our estimate using the

“passive” aged soil endmember fell within the range of their estimates (33-36%; Table

4.5). Using the “slow/passive” soil C endmember produced an overestimate of aged C contribution (93% to >100%; Table 4.5) compared to Ishikawa et al.’s (2014, 2016) estimates. DelVecchia et al. (2016) estimated that aged (∆14C=-580‰ to -48‰; 6,910 to

335 years equivalent age) methane could have contributed up to 100% to hyporheic zone stonefly biomass in several cases in the Nyack floodplain (Middle Fork of the Flathead

River), whereas we estimated that fossil-derived C contributed 58% and both soil C and

OM pools contributed >100%.

Differences between our fossil endmember estimates and those of the studies of

Fellman et al. (2015), Ishikawa et al. (2014, 2016), and DelVecchia et al. (2016) may be explained by the fact that the glacially derived C, periphyton C, and methane endmembers that each used, respectively, were not as 14C-depleted as our fossil C endmember. Use of soil C and OM endmembers in our model led to estimates closer to those calculated by Fellman et al. (2015) and Ishikawa et al. (2014, 2016). However, for

DelVecchia et al. (2016), our soil OM was not old enough to explain their observed stonefly ages. Our estimates based on C and OM of typical soil ages are potentially more realistic, as Hossler and Bauer (2012) noted that most rivers contain admixtures of

“passive” and “slow” turnover aged soil C and OM, which is consistent with our

139 predicted contributions of the sources and ages of materials to consumer biomass in other studies.

Lakes and Wetlands

In contrast to rivers and streams, our model indicates that fossil, “passive” aged, and “slow/passive” aged soil C and OM did not contribute to consumer biomass in all temperate and subtropical lake and wetland systems (Figure 4.5), with contributions ranging from 0-17%, 0-31%, and 0-86%, respectively (Table 4.5). These findings suggest that temperate and subtropical lake food webs are less likely to be impacted by inputs of aged C and OM than streams and rivers. Similar to studies conducted in streams and rivers (Table 4.1), estimates of aged C and OM contributions to consumer biomass were not always conducted in lake studies. We estimated no contributions from fossil,

“passive” or “slow/passive” soil aged C and OM to zooplankton in Quebec Eastern

Township lakes (McCallister and Giorgio 2008; Table 4.5), which was in agreement with these authors’ suggestion that zooplankton depended on modern aquatic primary production. In Lake Superior, contributions of aged sedimentary C (∆14C=-117‰ to

-20‰; 1,000 to 160 years equivalent age) to zooplankton biomass were estimated to be small (~3%), with modern-aged forms of aquatic primary production dominating (91%) zooplankton biomass using a combined ∆14C-δ13C isotopic mixing model (Zigah et al.

2012 a). Similar to the Quebec lakes, our estimates of the contributions of fossil,

“passive” and “slow/passive” aged soil C and OM to Lake Superior zooplankton were all

0% (Table 4.5), consistent with the dominance of modern-aged planktonic biomass compared to aged C and OM from any source to large lakes.

140

Wang et al.’s (2014) estimates of 31±8% contributions of aged C to subtropical

Everglades fish are higher than our estimates using the fossil C and “passive” soil aged C endmembers (7% and 13%, respectively), but similar to our estimate (38%) using the

“slow/passive” endmember (Table 4.5). This may be due to Wang et al.’s (2014) assumption that the ∆14C of POC in agricultural runoff was the same as that of peat soil

OM (-212‰; 1,910 years equivalent age), which was more 14C-enriched than our fossil and “passive” soil aged C endmembers, but close to the “slow/passive” soil C endmember used in our estimates (Wang et al. 2014).

In contrast to temperate and subtropical lakes, in small subarctic Norway lakes, fossil, “passive” and “slow/passive” aged soil C and OM contributed significantly to invertebrate consumer biomass in our model estimates (33%, 62%, and >100%, respectively; Table 4.5). While Hagvar et al. (2016) do not provide estimates for the proportions of aged OM contributing to their invertebrate biomass, they provide the age of silt (3,220±20 to 5,160±40 years B.P., ∆14C=-330‰ to -474‰, respectively) from areas of glacial retreat. In comparison to using the ages of silt they provided for the assumed aged endmembers, our models using fossil and “passive” aged soil C (33% and

62%, respectively; Table 4.5) likely underestimated contributions of aged C and OM to their consumers. However, using the lower 14C of glacier-derived C (silt) in the model that Hagvar et al. (2016) provided, glacially derived OM could have contributed up to

100% to consumer biomass.

141

Conclusions and Future Directions

The ages of the nutritional resources utilized by aquatic consumers have generally not been evaluated in aquatic food web studies. However, in every inland water study that has so far measured natural abundance 14C in consumers, consumer biomass is supported by one or more forms of aged C or OM, with the exception of the largest lakes

(e.g., Lake Superior; Table 4.5). Minimal contributions of aged C and OM to the Lake

Superior food web can be explained by the lake’s relatively small ratio of watershed:lake surface area, and the greater abundance and importance of modern-aged aquatic primary production to consumer nutrition in such systems (Zigah et al. 2012 a; Kruger et al.

2016).

The use of natural abundance 14C in aquatic food web studies not only provides information on the utilization of aged C and OM by different living components, but also provides a unique source tracer with a far greater dynamic range than stable isotopes. If, as shown here for numerous studies (Tables 4.3 and 4.4), aged C and OM are present to varying degrees in many aquatic systems (especially in turbulent, flowing waters). From our analysis, these aged C and OM sources contribute in the majority of cases to consumer biomass (Table 4.5, Figure 4.5), indicating that modern aquatic food webs depend not just on living or recently produced sources of nutrition, but that they are also supported by geological reservoirs of C and OM. The two primary mechanisms by which aged C can enter aquatic food webs are 1) consumption of aged OM (derived, e.g., from soils, sedimentary rocks, petroleum hydrocarbons, etc.) or as bacterially “repackaged” biomass and 2) photosynthetic fixation of aged DIC and CO2(aq) by aquatic primary

142 producers and the subsequent consumption of this “aged” living or recently living aquatic

OM. While these mechanisms are unique from each other, both still represent contributions of moderately to geologically-aged C to present-day aquatic food webs.

Future studies using natural 14C to assess contributions of aged C and OM to aquatic consumer biomass must consider both the aged OM and aged DIC pathways as entry points of aged C to aquatic food webs. Use of methods such as compound specific radiocarbon analysis (CSRA) may help differentiate between these two mechanisms

(Ishikawa et al. 2015, Kruger et al. 2016).

A number of important research questions remain to be addressed on the importance of the ages of C and OM contributing to aquatic food webs. These include:

1) does utilization of aged C and OM augment or supplant modern-aged sources of nutrition?; 2) do contributions of aged C and OM to aquatic consumer nutrition alter food web community structure, ecosystem function, and biogeochemical cycling relative to aquatic systems that do not contain aged C and OM or utilize aged nutritional resources?; and 3) what watershed- and larger-scale factors (e.g., changes in land use, hydrology, climate [temperature, rainfall, etc.]) may influence the mobilization of aged C and OM to aquatic systems and how might these inputs alter food web nutrition and dynamics?

143

Table 4.1. Summary of inland water food web studies that have employed natural abundance 14C analyses of consumer organisms. Location Date Organisms studied Sourcea

North America Great Basin Lakes, USA and Canada 1950's Fish, brine shrimp 1 Colville River/ Beaufort Sea, AK, USA 1980 Fish, invertebrates, birds 2 Hudson River, NY, USA 2000-2001 Bacteria 3 Eastern Townships Lakes, Quebec, 2004 Bacteriab, cladocerans, copepods 4 Canada Hudson River, NY, USA 2004-2005 Cladocerans and copepods 5 Everglades, FL, USA 2006-2007 Fish 6 Lake Superior, USA and Canada 2009 Cladocerans and copepods 7 Pigeon River, MI, USA 2011 Ammocoetes (lamprey larva) 8 Clear Fork River, OH, USA 2011 Ammocoetes (lamprey larva) 8 Lake Superior, USA and Canada 2012 Cladocerans, copepods, mysids, 9 amphipods, fish Herbert River, AK, USA 2012 Macroinvertebrates and fish 10 Paint Creek, OH, USA 2012-2013 Macroinvertebrates 11 Muskingum River, OH, USA 2013 Freshwater mussels 12 Susquehanna River watershed, PA, USA 2011-2014 Macroinvertebrates 11 Mohawk-Hudson, NY, USA 2014 Macroinvertebrates 11 Nyack Floodplain, MT, USA 2013-2015 Stoneflies 13 Europe Trave River, Germany 2007-2010 Fish, crayfish 14 Weibe River, Lake Rosenfieldc, 2009/2011 Mussels 15 Germany Midtdalsbreen Glacier, Norway 2010/2014 Fish, chironomids, aquatic 16 beetles Lough Erne, Ireland 2011 Fish, cladocerans, copepods, 17 mysids Lakes Schwerinc and Ostorfc, Germany 2011 Fish, eels, zebra mussels 18 Rivers Rebbe and Schloss Wilhelosthalc, 2013 Fish 19 Germany Asia Lake Biwa watershed, Japan 2006-2008 Macroinvertebrates and fish 20 Kanno River watershed, Japan 2011-2012 Macroinvertebrates 20 aSources: 1- Broecker and Walton (1959), 2- Schell (1983), 3- McCallister et al. (2004), 4- McCallister and del Giorgio (2008) and (2012), 5- Caraco et al. (2010), 6-Wang et al. (2014), 7-Zigah et al. (2012), 8- Evans 2012, 9-Kruger et al. (2015), 10- Fellman et al. (2015), 11- Bellamy et al. (submitted, a, b, c), 12- Weber et al., (in press), 13- DelVecchia et al. (2016) 14- Phillipsen and Heinemeier (2013), 15- Fernandes et al. (2012), 16- Hagvar and Olson (2013) and Hagvar et al. (2016), 17-Keaveney et al. (2015), 18- Fernandes et al. (2013), 19- Fernandes et al. (2015), 20- Ishikawa et al. (2010), (2014), and (2016). b 14 ∆ C of bacterial respired CO2 was measured as a proxy for bacterial biomass. cDue to small sample sizes from each of these lakes, data were grouped for Figure 4.5 as “German Lakes.” 144

Table 4.2. Representative literature values of ∆14C values, equivalent 14C ages, and δ13C values of potential carbon (C) and organic matter (OM) sources to various inland water ecosystems. Source of C or OM Location Δ14C (‰) 14C Age (Yrs B.P.) δ13C (‰) Source Algae/cyanobacteria Lough Erne, Northern Ireland, UK -94 to -33 733 – 209 -39.1 to -26.6 Keaveney et al. (2015) Periphyton Seri and Inukami Rivers, Japan -361 to 21 3,540 – Modern -28 to 13.2 Ishikawa et al. (2014), (2015) Cladophora spp. Seri River, Japan -199 1,720 -23.0 Ishikawa et al. (2015) Macrophytes Furen Lagoon Japan -40 to -10 270 – 20 -17 to – 5 Watanabe and Kuwae (2015) Terrestrial vegetation Seri River watershed, Japan 27 Modern -30.9 Ishikawa et al. (2015) (Quercus glauca) Temperate soils Harvard Forest, MA, USA -171 to 146 1,460 - Modern -27.0 to -24.8 Gaudinski et al. (2000) Woodland and grassland Rothamsted Experimental Farm, UK -508 to 44 5,650 - Modern -26.3 to -25.2 Jenkinson et al. (2008) Soils Arctic Tundra soils Edge Island, Norway -364 to 48 3,700 - Modern ND Cherkinsky (1996) Arctic Tundra soils Russia -468 to 152 5,190 - Modern ND Cherkinsky (1996)

145 Taiga soils Russia -424 to 264 4,760 - Modern ND Cherkinsky (1996)

Rainforest soil Brazil -453 to 34 4,800 - Modern -26.8 to - 25.0 Pessenda et al. (1996) Atlantic forest soils Brazil -783 9,340 - Modern -25.8 to -15 Pessenda et al. (1996) Temperate soils PA, OH, NY, USA -789 to 16 12,450 - Modern -29.6 to -13.8 Bellamy et al. (submitted a, b, c) Monterey shale (weathered) Santa Clara River Basin, CA, USA -885 17,320 -22.3 Komada et al. (2004) Monterey shale Santa Clara River Basin, CA, USA -995 42,510 -21.7 Komada et al. (2004) (unweathered) Biogenic methane Esthwaite Water, UK; Mekkojärvi, Finland; 135.4 to 160.6 Modern <-45 Deines and Grey (2006); Glacial Lake Agassiz region, MN Taipale et al. (2007); Chanton et al. (1995) Thermogenic methane Pacific, Atlantic, and Gulf of Mexico ocean -1000 to -978 50,000 ≤ -45 to -40 Pohlman et al. (2009); margins Grey (2016) Polycyclic aromatic Washington, DC, USA ~-1,000 50,000 ≤ -26.0 to -24.3 Currie et al. (1997) hydrocarbons (PAHs) continued

145

Table 4.2, continued

Stream sediments PA, OH, NY, USA -929 to 16 21,100 - Modern -31.0 to -9.6 Bellamy et al. (submitted a, b, c) Aerosols (Particulates) NY, VA, USA -539 to -45 6,220 - 370 -25.9 to -21.4 Wozniak et al. (2011) Aerosols (WSOC*) NY, VA, USA -193 to 106 1,720 - mModern -27.6 to -21.1 Wozniak et al. (2012 a, b) Carbonate rock ND -1,000 50,000≤ ~0 Clark and Fritz (1997)

Atmospheric CO2 Mauna Loa, HI, USA; South Pole ~30 Modern -8.5 to -8.2 Ciais (2013), Turnbull et al. (2016) *Water soluble organic carbon ND – no data

146

Table 4.3. Representative literature values (means ± standard deviations, SD) of ∆14C, equivalent 14C ages and δ13C values of particulate organic carbon (POC), dissolved organic carbon (DOC) and dissolved inorganic carbon (DIC) measured in various streams and rivers. Replicate numbers of samples analyzed, n, are given in parentheses. Location Date ∆14C (‰) Age (yrs B.P.) δ13C (‰) Source* POC Amazon River Basin (mountain rivers) 1991-2003 -202 ± 198 (8) 1,813 ± 1800 -25.7 ± 1.7 (8) 1 Amazon River Basin (lowland rivers) 1991-2003 90 ± 55 (10) Modern -29.8 ± 1.8 (10) 1 Amazon River Basin (mixed mountain and lowland 1991-2003 -135 ± 141 (10) 1,165 ± 1429 -28.2 ± 0.9 (10) 1 rivers) Amazon River Basin (carbonate free rivers) 1991-2003 129 ± 10 (6) Modern -29.2 ± 2.1 (6) 1 Santa Clara River, CA, USA 1997-1998 -333 ± 251 (6) 3253 ± 2322 -23.8 ± 4.6 (7) 6 Susquehanna River, PA, USA 2000 -61 (1) 506 -25.6 (1) 2 Hudson River, NY, USA 1998/2000 -281 ± 181 (4) 2,650 ± 1604 -27.8 ± 1.5 (2) 2

147 Delaware River, NY, USA 2000 6 (1) Modern -25.6 (1) 2 Parker River, MA, USA 1998/2000 -89 ± 94 (5) 749 ± 875 -31.8 ± 1.5 (5) 2 Mackenzie River, Canada 2004 -578 ± 67 (2) 6,930 ± 1389 -27.8 ± 1.5 (2) 7 Sag River, AK 2004 -466 ± 1 (2) 5,040 ± 15 -26.9 ± 0.1 (2) 7 Yukon River, AK 2004 -545 ± 19 (2) 6,326 ± 342 -27.4 ± 1.0 (2) 7 Congo River, Africa 2008 -62 ± 13 (5) 514 ± 112 -28.2 ± 0.5 (5) 8 Lower Chesapeake Bay Watershed streams 2008-2009 88 ± 7 (4) Modern -28.9 ± 0.5 (15) 3 (Forested), VA, USA Lower Chesapeake Bay Watershed streams 2008-2009 75 ± 1 (2) Modern -29.5 ± 1.5 (10) 3 (Pasture), VA, USA Lower Chesapeake Bay Watershed streams (Crop), 2008-2009 68 (1) Modern -29.9 ± 0.3 (6) 3 VA, USA Lower Chesapeake Bay Watershed streams (Urban), 2008-2009 -90 ± 13 (2) 758 ± 115 -30.1 ± 1.5 (8) 3 VA, USA Yellow River, China 2009 -531 ± 88 (4) 6,082 ± 1669 -24.2 ± 0.8 (4) 9 Continued 147

Table 4.3, continued Changjiang, China 2009 -114 ± 13 (3) 972 ± 119 -24.4 ± 1.2 (3) 9 Seri River, Japan 2009-2010 -109 ± 52 (4) 927 ± 483 -25 ± 3.5 (4) 5 Fudoji River, Japan 2009-2010 -7 ± 38 (4) Modern -24.9 ± 0.2 (4) 5 DOC Susquehanna River, PA, USA 2000 -236 (1) 2,162 -27.1 (1) 2 Hudson River, NY, USA 1998/2000 -89 ± 77 (4) 748 ± 710 -26.7 ± 0.8 (4) 2 Delaware River, NY, USA 2000 -114 (1) 972 -27.1 (1) 2 Parker River, MA, USA 1998/2000 82 ± 48 (5) Modern -29.0 ± 0.8 (5) 2 Mackenzie River, Canada 2004 -112 ± 81 (2) 954 ± 767 -26.4 ± 0.2 (2) 7 Sag River, AK 2004 -353 ± 156 (2) 3,498 ± 2216 -27.5 ± 0.4 (2) 7 Yukon River, AK 2004 -154 ± 12 (2) 1,343 ± 115 -27.6 ± 0.6 (2) 7 Congo River, Africa 2008 73 ± 16 (5) Modern -28.9 ± 0.6 (5) 8 148 Lower Chesapeake Bay Watershed streams 2008-2009 100 ± 11 (5) Modern -28.4 ± 0.3 (6) 3

(Forested), VA, USA Lower Chesapeake Bay Watershed streams 2008-2009 76 ± 10 (4) Modern -29.0 ± 3.0 (4) 3 (Pasture), VA, USA Lower Chesapeake Bay Watershed streams (Crop), 2008-2009 89 ± 11 (2) Modern -29.9 ± 1.8 (2) 3 VA, USA Lower Chesapeake Bay Watershed streams (Urban), 2008-2009 -222 ± 24 (3) 2,017 ± 251 -27.4 ± 1.0 (3) 3 VA, USA Rote Mulde, WeiBe Mulde, and Saubach, Germany 2008-2010 7 ± 27 (12) Modern -26.6 ± 0.5 (12) 4 Yellow River, China 2009 -96 ± 37 (4) 811 ± 335 -29.2 ± 2.4 (4) 9 Changjiang, China 2009 -108 ± 57 (4) 918 ± 531 -30.3 ± 1.4 (4) 9 Seri River, Japan 2009-2010 -248 ± 111 (4) 2,290 ± 1282 -24.2 ± 2.9 (4) 5 Fudoji River, Japan 2009-2010 645 ± 47 (4) Modern -28.4 ± 0.1 (4) 5

Continued

148

Table 4.3, continued DIC Amazon River Basin (mountain rivers) 1991-2003 -240 ± 233 (14) 2,205 ± 2,940 -4.9 ± 2.7 (14) 1 Amazon River Basin (lowland rivers) 1991-2003 89 ± 44 (43) Modern -17 ± 5.9 (43) 1 Amazon River Basin (mixed mountain and lowland 1991-2003 -14 ± 99 (11) 113 ± 850 -14.2 ± 2.9 (11) 1 Rivers) Amazon River Basin (carbonate free rivers) 1991-2003 98 ± 20 (38) Modern -17.1 ± 6.2 (38) 1 Susquehanna River, PA, USA 2000 - - -8.5 (1) 2 Hudson River, NY, USA 1998/2000 37 ± 33 (4) Modern -9.5 ± 1.9 (4) 2 Delaware River, NY, USA 2000 43 (1) Modern -7 (1) 2 Parker River, MA, USA 1998/2000 78 ± 18 (4) Modern -13.8 ± 1.5 (4) 2 149 Lower Chesapeake Bay Watershed streams 2008-2009 69 ± 14 (4) Modern -22.4 ± 2.8 (13) 3

(Forested), VA, USA Lower Chesapeake Bay Watershed streams 2008-2009 70 ± 0 (2) Modern -11.4 ± 6.6 (11) 3 (Pasture), VA, USA Lower Chesapeake Bay Watershed streams (Crop), 2008-2009 81 (1) Modern -15.6 ± 1.4 (4) 3 VA, USA Lower Chesapeake Bay Watershed streams (Urban), 2008-2009 -271± 38 (2) 2,540 ± 430 -12.4 ± 1.1 (8) 3 VA, USA Rote Mulde, WeiBe Mulde, and Saubach, Germany 2008-2010 15 ± 17 (11) Modern ND 4 Seri River, Japan 2009-2010 -217 ± 31 (4) 1,970 ± 320 -7.2 ± 0.2 (4) 5 Fudoji River, Japan 2009-2010 -9 ± 14 (4) 73 ± 114 -7.9 ± 1.4 (4) 5 *Sources: 1- Mayorga et al. (2005), 2- Raymond et al. (2004), 3- Lu et al. (2014), 4- Tittel et al. (2013), 5- Ishikawa et al. (2015), 6- Masiello and Druffel (2001), 7- Guo et al. (2006), 8-Spencer et al. (2012), 9- Wang et al. (2012). ND - no data

149

Table 4.4. Representative literature values of ∆14C values, equivalent 14C ages, and δ13C values of particulate organic carbon (POC), dissolved organic carbon (DOC) and dissolved inorganic carbon (DIC) pools measured in various lakes. Location Date ∆14C range (‰) Age (yrs B.P.) δ13C (‰) Source* POC Lake Superior, USA/ON Sep 2007 - Aug 2009 -354 to -68 3,892 to 566 -34.9 to -26.7 1 Eastern Townships, Quebec Sep 2004 73 to 179 Modern ND 2 Lough Erne, Northern Ireland, UK Feb - May 2011 -191 to -82 1,703 to 687 -32.7 to -28.5 3 Lake Schwerin, Germany Sep 2011 -85** 714 -25.2 5 DOC Lake Superior, USA/ON Sep 2007 - Aug 2009 -77 to 104 644 to Modern -28.3 to -19.1 1 Eastern Townships, Quebec Sep 2004 14 to 110 Modern ND 2 Lough Erne, Northern Ireland, UK Mar - Aug 2011 -23 to -9 187 to 73 -13.6 to -21.3 3

150 Namtso Lake, Tibetan Plateau - -70 583 -22.9 6 Yamdrok Lake, Tibetan Plateau - -87 731 -22.4 6 Lake Schwerin, Germany Sep 2011 -100** 846 -24.9 5 DIC Lake Superior, USA/ON Sep 2007 - Aug 2009 32 to 77 Modern -0.1 to 1.5 1 Eastern Townships, Quebec Sep 2004 -91 to 52 766 to Modern ND 2 Lough Erne, Northern Ireland, UK Feb - Aug 2011 -75 to -21 626 to 170 ND 3 Pavilion Lake, Marble Canyon, BC Aug 2006 -45 370 -4.4 to -1.9 4 Lake Schwerin, Germany Sep 2011 -42** 345 -3.4 5 *Sources: 1- Zigah et al. (2011), (2012a), (2012b), 2- McCallister and Del Giorgio (2012), 3- Keaveney et al. (2015a), 4-Brady et al. (2009), 5- Fernandes et al. (2013), 6-Spencer et al. (2014). **Data were originally presented as pMC (percent Modern carbon) in the referenced source and converted to ∆14C values and equivalent 14C ages. ND – no data

150

Table 4.5. Estimated maximum contributions of fossil aged C and OM, “passive” aged soil C and OM, and 50:50 mixture of “passive” and “slow turnover” aged soil C and OM endmembers to metazoan consumer biomass for lakes, wetlands, streams, and rivers in which organism 14C was measured. See text for details. System Type Fossil “Passive” “Passive”/ aged (%)a aged soil (%)b “slow turnover” aged soil (%)c Streams and Rivers Lake Biwa watershed, Japan 18 33 93 Kanno River, Japan 20 36 >100d Susquehanna River, PA 50 92 >100d Pigeon River, MI 11 21 58 Clear Fork River, OH 11 21 58 Muskingum River, OH 15 29 80 Colville River, AK 15 28 80 Hudson River, NY 35 65 >100d Mohawk-Hudson, NY 11 21 60 Weibe and Rebbe Rivers 13 24 69 Trave River 19 35 97 Herbert River, AK 9 17 48 Nyack Floodplain, MT 58 >100d >100d Lakes and Wetlands Everglades, FL 7 13 38 Norway Glacier 33 62 >100d Great Basin Lakes 17 31 86 Eastern Townships lakes, 0 0 0 Quebec Lake Superior 0 0 0 Lough Erne 14 26 72 German Lakes 13 24 68 aFossil C and OM endmember ∆14C= -1,000‰ (equivalent 14C age > 50,000 yrs B.P.). bEstimated ∆14C values for the passive soil C and OM pool were -538‰ (equivalent 14C age = 6,200 yrs B.P.) for Inceptisols and -541 (equivalent 14C age = 6.260 yrs B.P.) for Ultisols according to Hossler and Bauer (2012). We used the average of these two values for a passive aged soil OC endmember ∆14C value of -540‰ (equivalent 14C age = 6,240 yrs B.P.). See text for calculation details. c50:50 contributions from “passive” soil C and OM (mean ∆14C = -540‰; equivalent 14C age = 6,240 yrs B.P.) and “slow turnover” soil C and OM (∆14C = 157‰; modern-aged) pools (Hossler and Bauer 2012), giving a passive/slow turnover aged soil OC endmember ∆14C value of -192‰ (equivalent 14C age = 1,710 yrs B.P.). See text for calculation details. dEstimates > 100% indicate that inputs of C and OM having ∆14C < -192‰ are required.

151

152

153

153 154

154

Figure 4.4. Potential pathways of aged particulate organic matter (POM) and dissolved organic matter (DOM) utilization by aquatic consumers.

155

156

Figure 4.5. Literature ∆14C values and equivalent 14C ages for metazoan consumers of various inland waters. See Table 1 for details of systems and organisms analyzed.

156

Conclusions

157

Use of stable isotopes (δ13C, δ15N, and δ2H), radiocarbon (∆14C), and Bayesian mixing models in multiple temperate stream ecosystems revealed that autochthonous C and OM were generally the primary nutritional resources to most macroinvertebrate functional feeding groups (FFGs). However, allochthonous resources were also important and comprised significant proportions of macroinvertebrate nutrition and biomass formation. In some cases, the apparent 14C ages of consumers were attributable to utilization of aged OM, while in other cases, nutritional resources and consumers were

14 “aged” due to fixation of C-depleted DIC and CO2(aq) by aquatic photoautotrophs. The differentiation of these two distinct yet interconnected mechanisms of imparting 14C- depletion and apparent age to aquatic consumers represents a major future research challenge.

In streams of the Mohawk-Hudson watershed, autochthonous OM was of primary importance to the nutrition all FFGs (Chapter 1). Soil-derived OM also contributed more to the biomass of macroinvertebrates that were collected from streams with higher amounts of agriculture in their watersheds. Macroinvertebrates collected from a stream with OM-rich fossil aged shale in its watershed (Nowadega Creek) were more 14C- depleted than those macroinvertebrates collected from a stream with no shale in its watershed (Fly Branch). While it was initially predicted that shale OM would contribute to stream OM pools and macroinvertebrate nutrition, the greater apparent ages of macroinvertebrates from Nowadega Creek compared to those from Fly Branch were most likely explained by increased utilization of 14C-depleted algae by macroinvertebrates in

158

Nowadega Creek. However, the possibility that the apparent ages of macroinvertebrates could be attributed to the utilization of aged soil OM could also not be ruled out.

In streams of the Susquehanna River watershed network (Chapter 2), autochthonous OM also contributed the most to macroinvertebrate biomass to most FFGs in all of the streams studied. Macroinvertebrates collected from streams in the

Susquehanna having greater agricultural land use in their watersheds were more reliant on soil and sediment-derived OM, similar to findings for the Mohawk-Hudson system

(Chapter 1). The δ15N values of macroinvertebrates collected from high agriculture streams were also found to be greater than those collected from low agriculture streams, and likely resulted from the application of fertilizer N to agricultural lands. In addition, macroinvertebrates from Little Muncy Creek in the Susquehanna watershed were significantly depleted in 14C in 2011 when active natural gas seepage was observed in this system.

Overall, physical characteristics (e.g., stream order and canopy cover) of each of the streams studied in the headwaters of Susquehanna River may be important controls on the nutritional resources utilized by macroinvertebrates in these streams. Land use at small spatial scales, rather than at the watershed scale, may also have a potentially greater effect on the inputs of aged allochthonous OM (such as soil) to streams, and consequently the utilization of this aged OM by macroinvertebrate consumers.

At the Paint Creek time-series sampling site (Chapter 3), nutritional resource utilization by macroinvertebrates was found to vary seasonally. Macroinvertebrate primary (i.e., herbivorous) consumer biomass was supported to a greater extent by

159 autochthonous OM in the spring and summer than in the fall when allochthonous OM contributed more to primary consumer biomass. In contrast, macroinvertebrate predators were primarily reliant on terrestrial vegetation regardless of season. A seasonal shift in the δ15N values of algae and macroinvertebrates was also observed and attributed to agricultural activity in the watershed. Mixing model outcomes indicated that aged soil

OM contributed to consumer biomass in Paint Creek, however, similar to the Mohawk-

Hudson and Susquehanna watersheds, 14C-depleted OM in the form of DIC assimilated by algae also contributed to macroinvertebrate biomass was also likely utilized by macroinvertebrates there.

The natural radiocarbon (∆14C) signatures and apparent ages of particulate and dissolved organic C (POC and DOC, respectively) and dissolved inorganic C (DIC) pools that have been found to be prevalent in streams, rivers, lakes, and wetlands are reviewed in Chapter 4. The ∆14C values and equivalent ages of aquatic metazoans from all known published studies are also synthesized. Using this information, the contributions of aged

C and OM to consumer biomass across inland aquatic systems were assessed. It was found that in all inland water systems studied to date, with the exception of the largest lakes, one or more forms of aged C or OM contribute to aquatic consumer biomass. This general finding suggests that aged C and OM reservoirs, often of great age, are actively incorporated to modern-day aquatic ecosystems and their carbon budgets and energetic processes.

Overall, findings from this dissertation research highlight several important points. First, autochthonous forms of aquatic OM were the dominant nutritional resource

160 to multiple stream macroinvertebrate FFGs at multiple study sites. Macroinvertebrates belonging to the filtering collector and scraper FFGs were consistently most reliant on autochthonous OM across many of the streams studied. Collector-gatherers and chironomids were more likely to rely to a greater extent on soil OM, sediment OM, and terrestrial vegetation. Predator nutritional resource utilization varied across all of the study systems. Stream size and related characteristics (e.g., canopy cover and substrate type), and human activity (e.g., agriculture) within the watershed may have also influenced the utilization of autochthonous vs. allochthonous nutritional resources by macroinvertebrates in the study systems. The assimilation of aged C by macroinvertebrates was found to be ubiquitous across all of the streams for which 14C data was collected in the present study. A similar conclusion was reached in the review and synthesis of the existing literature. However, further exploration of the mechanism(s) through which aged C and OM are made available to aquatic consumers are necessary.

Determining if aged C and OM contributes to aquatic food webs through the consumption and assimilation by aquatic consumers of apparently “aged” algae or through the consumption and assimilation of aged soil or sediment OM is important for interpreting the routes through which geologically aged C enters aquatic food webs.

Finally, the present study also suggests that human activities, in the forms of land use alteration and possibly hydraulic fracturing activities may facilitate the mobilization and movement of aged C and OM into stream, river, and other inland water systems, and consequently the utilization of these materials by aquatic consumers.

161

A number of future research questions and challenges regarding the utilization of

C and OM from different sources (i.e., allochthonous vs. autochthonous) and of different ages in aquatic food webs arise from these study findings. These include: i) Do aged forms of C and OM act to augment or supplant modern forms of C and OM in aquatic food webs?, ii) Does the utilization of allochthonous vs. authochthonous and aged vs. modern C and OM impact community structure and ecosystem function?, and iii) How do factors such as land use, geomorphology, hydrology and changing climate impact the relative roles of allochthonous and autochthonous and different ages of C and OM to aquatic consumers across different stream inland water systems? Additional approaches that could help to answer some of these questions and provide a better understanding of the types of aged C and OM that are assimilated by aquatic consumers would include the use of compound-specific isotopic analysis (CSIA).

162

Literature Cited

Abbott, M. B., & Stafford Jr, T. W. 1996. Radiocarbon geochemistry of modern and

ancient Arctic lake systems, Baffin Island, Canada. Quaternary Research 45: 300–

311.

Agren, A., M. Berggren, H. Laudon, and M. Jansson. 2008. Terrestrial export of highly

bioavailable carbon from small boreal catchments in spring floods. Freshwater

Biology 53: 964–972.

Ahad, J. M., and H. Pakdel. 2013. Direct evaluation of in situ biodegradation in

Athabasca oil sands tailings ponds using natural abundance radiocarbon.

Environmental Science & Technology 47: 10,214–10,222.

Aitkenhead, J. A., Hope, D., and M. F. Billett. 1999. The relationship between dissolved

organic carbon in stream water and soil organic carbon pools at different spatial

scales. Hydrological Processes 13: 1,289–1,302.

Aitkenhead-Peterson, J. A., W.H. McDowell, J. C. Neff, E. Stuart, and L. Robert. 2003.

Sources, production, and regulation of allochthonous dissolved organic matter

inputs to surface waters. In S. E. G. Findlay and R. L. Sinsabaugh (Eds.), Aquatic

Ecosystems: Interactivity of Dissolved Organic Matter (pp. 26–70). San Diego,

CA: Academic Press.

163

Allan, J. D. 2004. Landscapes and riverscapes: the influence of land use on stream

ecosystems. Annual Review of Ecology, Evolution, and Systematics:257–284.

Allan, J. D., and M. M. Castillo. 2007. Stream ecology: structure and function of running

waters. Springer Netherlands: Springer Science+ Business Media BV.

Allan, J. D., M. S. Wipfli, J. P. Caouette, A. Prussian, and J. Rodgers. 2003. Influence of

streamside vegetation on inputs of terrestrial invertebrates to salmonid food webs.

Canadian Journal of Fisheries and Aquatic Sciences 60: 309–320.

Anderson, C., and G. Cabana. 2005. δ15N in riverine food webs: effects of N inputs from

agricultural watersheds. Canadian Journal of Fisheries and Aquatic Sciences 62:

333–340.

Anderson, N. H., J. R. Sedell. 1979. Detritus processing by macroinvertebrates in stream

ecosystems. Annual Review of Entomology 24: 351–377.

Aravena, R., M. L. Evans, and J. A. Cherry. 1993. Stable isotopes of oxygen and nitrogen

in source identification of nitrate from septic systems. Ground water 31: 180–186.

Arheimer, B., and R. Liden. 2000. Nitrogen and phosphorus concentrations from

agricultural catchments—influence of spatial and temporal variables. Journal of

Hydrology 227: 140–159.

Arsuffi, T. L., and K. Suberkropp. 1989. Selective feeding by shredders on leaf-

colonizing stream fungi: comparison of macroinvertebrate taxa. Oecologia 79:

30–37.

164

Atkinson, C. L., S. W. Golladay, S. P. Opsahl, and A. P. Covich. 2009. Stream discharge

and floodplain connections affect seston quality and stable isotopic signatures in a

coastal plain stream. Journal of the North American Benthological Society 28:

360–370.

Atkinson, C. L., J. F. Kelly, and C. C. Vaughn. 2014. Tracing consumer-derived nitrogen

in riverine food webs. Ecosystems 17: 485–496.

Atlas, R. M. 1981. Microbial degradation of petroleum hydrocarbons: an environmental

perspective. Microbiological Reviews 45: 180-209.

Atlas, W. I., W. J. Palen, D. M. Courcelles, R. G. Munshaw, and Z. L. Monteith. 2013.

Dependence of stream predators on terrestrial prey fluxes: food web responses to

subsidized predation. Ecosphere 4: 1-16.

Attermeyer, K., K. Premke, T. Hornick, S. Hilt, and H.P. Grossart. 2013. Ecosystem-

level studies of terrestrial carbon reveal contrasting bacterial metabolism in

different aquatic habitats. Ecology: 2,754–2,766.

Aufdenkampe, A. K., E. Mayorga, P. A. Raymond, J. M. Melack, S. C. Doney, S. R.

Alin, R. E. Aalto, and K. Yoo. 2011. Riverine coupling of biogeochemical cycles

between land, oceans, and atmosphere. Frontiers in Ecology and the Environment

9: 53–60.

Azam, F., T. Fenchel, J. Field, J. Gray, L. Meyer-Reil., and F. Thingstad. 1983. The

ecological role of water-column microbes in the sea. Marine Ecology Progress

Series 10: 257–263.

165

Baines, S. B., N. S. Fisher, and J. J. Cole. 2005. Uptake of dissolved organic matter

(DOM) and its importance to metabolic requirements of the zebra mussel,

Dreissena polymorpha. Limnology and Oceanography 50: 36–47.

Balesdent, J., C. Girardin, and A. Mariotti. 1993. Site-Related δ13C of Tree Leaves and

Soil Organic Matter in a Temperate Forest. Ecology 74: 1,713–1,721.

Barakat, A.O., A. Mostafa, N. B. El-Sayed, T. L. Wade, and S. T. Sweet. 2013.

Polycyclic aromatic hydrocarbons (PAHs) in surface sediments of Lake Manzala,

Egypt. Soil and Sediment Contamination: An International Journal 22: 315–331.

Battin, T. J., S. Luyssaert, L. A. Kaplan, A. K. Aufdenkampe, A. Richter, and L. J.

Tranvik. 2009. The boundless carbon cycle. Nature Geoscience 2: 598–600.

Bauer, J. E., R. B. Spies, J. S. Vogel, D. E. Nelson, and J. R. Southon. 1990. Radiocarbon

evidence of fossil-carbon cycling in sediments of a nearshore hydrocarbon seep.

Nature 348: 230–232.

Bauer, J. E., and T. S. Bianchi. 2011. Dissolved organic carbon cycling and

transformation. In J. Middelburg and R. Laane (Eds.), Treatise on Estuarine and

Coastal Science, Volume 5: Biogeochemistry (pp.7–67). Waltham, MA:

Academic Press.

Baxter, C. V., K. D. Fausch, and W. Carl Saunders. 2005. Tangled webs: reciprocal flows

of invertebrate prey link streams and riparian zones. Freshwater Biology 50: 201–

220.

166

Bellamy, A. R., J. E. Bauer, and A. G. Grottoli. (a). Influence of land use and lithology

on sources and ages of nutritional subsidies to stream macroinvertebrates: a multi-

isotopic approach. Submitted to Food Webs.

Bellamy, A. R., J. E. Bauer, and A. G. Grottoli. (b). Contributions of allochthonous

organic matter to the diet and nutrition of macroinvertebrates in the upper

Susquehanna River watershed. Submitted to Freshwater Biology.

Bellamy, A. R., J. E. Bauer, and A. G. Grottoli. (c). Temporal variability in the utilization

of autochthonous and allochthonous organic matter by macroinvertebrates in

Paint Creek, OH. Submitted to Aquatic Sciences.

Belsky, A. J., A. Matzke, and S. Uselman. 1999. Survey of livestock influences on stream

and riparian ecosystems in the western United States. Journal of Soil and Water

Conservation 54: 419–431.

Benda, L. E. E., N. L. Poff, D. Miller, T. Dunne, G. Reeves, G. Pess, and M. Pollock.

2004. The network dynamics hypothesis: how channel networks structure riverine

habitats. BioScience 54: 413–427.

Berggren, M., Ström, L., Laudon, H., and others. 2010. Lake secondary production fueled

by rapid transfer of low molecular weight organic carbon from terrestrial sources

to aquatic consumers. Ecology Letters 13: 870–880.

Berggren, M., S. E. Ziegler, N. F. St-Gelais, B. E. Beisner, and P. A. del Giorgio. 2014.

Contrasting patterns of allochthony among three major groups of crustacean

zooplankton in boreal and temperate lakes. Ecology 95: 1,947–1,959.

167

Berggren, M., and P. A. del Giorgio. 2015. Distinct patterns of microbial metabolism

associated to riverine dissolved organic carbon of different source and quality.

Journal of Geophysical Research: Biogeosciences 120: 989-999.

Bianchi, T. S., L. A. Wysocki, M. Stewart, T. R. Filley, and B. A. McKee. 2007.

Temporal variability in terrestrially-derived sources of particulate organic carbon

in the lower Mississippi River and its upper tributaries. Geochimica et

Cosmochimica Acta 71: 4,425–4,437.

Bianchi, T. S. 2007. Biogeochemistry of estuaries. New York, NY: Oxford University

Press. 720 pp.

Bianchi, T. S., and J. E. Bauer. 2011. Particulate organic carbon cycling and

transformation. In J. Middelburg and R. Laane (Eds.), Treatise on Estuarine and

Coastal Science, Volume 5: Biogeochemistry (pp. 69–117). Waltham, MA:

Academic Press.

Blair, N. E., E. L. Leithold, S. T. Ford, K. A. Peeler, J. C. Holmes, and D. W. Perkey.

2003. The persistence of memory: The fate of ancient sedimentary organic carbon

in a modern sedimentary system. Geochimica et Cosmochimica Acta 67: 63–73.

Blair, N. E., E. L. Leithold, H. Brackley, N. Trustrum, M. Page, and L. Childress. 2010.

Terrestrial sources and export of particulate organic carbon in the Waipaoa

sedimentary system: Problems, progress and processes. Marine Geology 270:

108–118.

Blair, N. E., and R. C. Aller. 2012. The fate of terrestrial organic carbon in the marine

environment. Annual Review of Marine Science 4: 401–423.

168

Boix-Fayos, C., J. de Vente, J. Albaladejo, and M. Martínez-Mena. 2009. Soil carbon

erosion and stock as affected by land use changes at the catchment scale in

Mediterranean ecosystems. Agriculture, Ecosystems & Environment 133: 75–85.

Bowling, D. R., D. E. Pataki, and J. T. Randerson. 2008. Carbon isotopes in terrestrial

ecosystem pools and CO2 fluxes. New Phytologist 178: 24–40.

Brady, A. L., G. Slater, B. Laval, and D. S. Lim. 2009. Constraining carbon sources and

growth rates of freshwater microbialites in Pavilion Lake using 14C analysis.

Geobiology 7: 544–555.

Brand, G., B. Vondracek, and N. R. Jordan. 2015. Influence of grazing and land use on

stream-channel characteristics among small dairy farms in the Eastern United

States. Renewable Agriculture and Food Systems 30: 524–536.

Brett, M. T., M. J. Kainz, S. J. Taipale, and H. Seshan. 2009. Phytoplankton, not

allochthonous carbon, sustains herbivorous zooplankton production. Proceedings

of the National Academy of Sciences 106: 21,197–21,201.

Broecker, W. S., and A. Walton. 1959. The geochemistry of 14C in fresh-water systems.

Geochimica et Cosmochimica Acta 16: 15–38.

Brooks, J. M., M. Kennicutt, C. Fisher, S. Macko, K. Cole, J. Childress, R. Bidigare, and

R. Vetter. 1987. Deep-sea hydrocarbon seep communities: evidence for energy

and nutritional carbon sources. Science 238: 1,138–1,142.

Buck, O., D. K. Niyogi, and C. R. Townsend. 2004. Scale-dependence of land use effects

on water quality of streams in agricultural catchments. Environmental Pollution

130:287–299.

169

Bunn, S. E., and P. I. Boon. 1993. What sources of organic carbon drive food webs in

billabongs? A study based on stable isotope analysis. Oecologia 96: 85–94.

Bunn, S. E., P. M. Davies, and T. D. Mosisch. 1999. Ecosystem measures of river health

and their response to riparian and catchment degradation. Freshwater Biology 41:

333–345.

Bunn, S. E., C. Leigh, and T. D. Jardine. 2013. Diet-tissue fractionation of δ15N by

consumers from streams and rivers. Limnology and Oceanography 58: 765–773.

Burdon, F. J., A. R. McIntosh, and J. S. Harding. 2013. Habitat loss drives threshold

response of benthic invertebrate communities to deposited sediment in

agricultural streams. Ecological Applications 23: 1,036–1,047.

Butman, D. E., H. F. Wilson, R. T. Barnes, M. A. Xenopoulos, and P. A. Raymond.

2015. Increased mobilization of aged carbon to rivers by human disturbance.

Nature Geoscience 8: 112–116.

Butman, D., P. A. Raymond, K. Butler, and G. Aiken. 2012. Relationships between ∆14C

and the molecular quality of dissolved organic carbon in rivers draining to the

coast from the conterminous United States. Global Biogeochemical Cycles 26:

doi:10.1029/2012GB004361.

Butman, D., and P. A. Raymond. 2011. Significant efflux of carbon dioxide from streams

and rivers in the United States. Nature Geoscience 4: 839–842.

170

Cairns, J., and J. R. Pratt. 1993. A history of biological monitoring using benthic

macroinvertebrates. Pages 10-27 in D. Rosenberg and V. Resh, Editors.

Freshwater biomonitoring and benthic macroinvertebrates. Chapman and Hall,

New, York, New York, USA.

Campbell Grant, E. H., W. H. Lowe, and W. F. Fagan. 2007. Living in the branches:

population dynamics and ecological processes in dendritic networks. Ecology

Letters 10: 165–175.

Caraco, N., J. E. Bauer, J. J. Cole, S. Petsch, and P. Raymond. 2010. Millennial-aged

organic carbon subsidies to a modern river food web. Ecology 91: 2,385–2,393.

Caraco, N., and J. Cole. 2004. When terrestrial organic matter is sent down the river: the

importance of allochthonous carbon inputs to the metabolism of lakes and rivers.

Pages 301-316 in G. Polis, M. Power, and G. Huxel, Editors. Food webs at the

landscape level. University of Chicago Press, Chicago, Illinois, USA.

Carbone, M. S., Czimczik, C. I., Keenan, T. F., and others. 2013. Age, allocation and

availability of nonstructural carbon in mature red maple trees. New Phytologist

200: 1,145–1,155.

Carlough, L. A. 1994. Origins, structure, and trophic significance of amorphous seston in

a blackwater river. Freshwater Biology 31: 227–237.

Carlson, P. E., B. G. McKie, L. Sandin, and R. K. Johnson. 2016. Strong land-use effects

on the dispersal patterns of adult stream insects: implications for transfers of

aquatic subsidies to terrestrial consumers. Freshwater Biology 61: 848-861.

171 de Carvalho, A. P. C., B. Gücker, M. Brauns, and I. G. Boëchat. 2015. High variability in

carbon and nitrogen isotopic discrimination of tropical freshwater invertebrates.

Aquatic Sciences 77: 307–314.

Cashman, M. J., Wehr, J. D., and Truhn, K. 2013. Elevated light and nutrients alter the

nutritional quality of stream periphyton. Freshwater Biology 58: 1,447-1,457. de Castro, D. M. P., D. R. de Carvalho, P. dos Santos Pompeu, M. Z. Moreira, G. B.

Nardoto, and M. Callisto. 2016. Land Use Influences Niche Size and the

Assimilation of Resources by Benthic Macroinvertebrates in Tropical Headwater

Streams. PLOS ONE 11:e0150527. doi:10.1371/journal.pone.0150527.

Ceola, S., I. Hödl, M. Adlboller, G. Singer, E. Bertuzzo, L. Mari, G. Botter, J. Waringer,

T. J. Battin, and A. Rinaldo. 2013. Hydrologic variability affects invertebrate

grazing on phototrophic biofilms in stream microcosms. PLOS ONE 8:e60629.

doi:10.1371/journal.pone.0060629

Chanton, J. P., Bauer, J. E., Glaser, P. A., and others. 1995. Radiocarbon evidence for the

substrates supporting methane formation within northern Minnesota peatlands.

Geochimica et Cosmochimica Acta 59: 3,663–3,668.

Cherkinsky, A. E. 1986. 14C dating and soil organic matter dynamics in Arctic and

subarctic ecosystems. Radiocarbon 38: 241–245.

Cherrier, J., J. E. Bauer, E. R. M. Druffel, R. B. Coffin, and J. P. Chanton. 1999.

Radiocarbon in marine bacteria: Evidence for the ages of assimilated carbon.

Limnology and Oceanography 44: 730–736.

172

Ciais, P., C. Sabine, G. Bala, and others (2013). Carbon and other biogeochemical cycles.

In T.F. Stocker, D. Qin, G.K. Platner, M. Tignor, S.K. Allen, J. Boschung, A.

Nauels, Y. Xia, V. Bex, P. M. Midgley (Eds.), Climate Change 2013: The

Physical Science Basis. Contribution of Working Group I to the Fifth Assessment

Report of the Intergovernmental Panel on Climate Change (pp. 465-570).

Cambridge, UK: Cambridge University Press.

Clarke, K. R., and R. N. Gorley. 2006. User manual/tutorial. Primer-E Ltd., Plymouth,

192pp.

Clark, I. D., and P. Fritz. 1997. Environmental isotopes in hydrogeology. Boca Raton,

FL: CRC Press. 352 pp.

Cole, J. J. (2013). Freshwater ecosystems and the carbon cycle. Oldendorf/Luhe:

International Ecology Institute. 146 pp.

Cole, J. J., and N. F. Caraco. 2001. Carbon in catchments: connecting terrestrial carbon

losses with aquatic metabolism. Marine and Freshwater Research 52: 101–110.

Cole, J. J., N. F. Caraco, G. W. Kling, and T. K. Kratz. 1994. Carbon dioxide

supersaturation in the surface waters of lakes. Science 265:1,568–1,570.

Cole, J. J., S. R. Carpenter, J. Kitchell, M. L. Pace, C. T. Solomon, and B. Weidel. 2011.

Strong evidence for terrestrial support of zooplankton in small lakes based on

stable isotopes of carbon, nitrogen, and hydrogen. Proceedings of the National

Academy of Sciences 108: 1,975–1,980.

173

Cole, J. J., Y. T. Prairie, N. F. Caraco, W. H. McDowell, L. J. Tranvik, R. G. Striegl, C.

M. Duarte, P. Kortelainen, J. A. Downing, and J. J. Middelburg. 2007. Plumbing

the global carbon cycle: integrating inland waters into the terrestrial carbon

budget. Ecosystems 10: 172–185.

Collins, S. M., T. J. Kohler, S. A. Thomas, W. W. Fetzer, and A. S. Flecker. 2015. The

importance of terrestrial subsidies in stream food webs varies along a stream size

gradient. Oikos 125: 674-685.

Collins, S. M., J. P. Sparks, S. A. Thomas, S. A. Wheatley, and A. S. Flecker. 2016.

Increased Light Availability Reduces the Importance of Bacterial Carbon in

Headwater Stream Food Webs. Ecosystems 19: 396–410.

Cook, R. A., B. Gawne, R. Petrie, D. S. Baldwin, G. N. Rees, D. L. Nielsen, and N. S.

Ning. 2015. River metabolism and carbon dynamics in response to flooding in a

lowland river. Marine and Freshwater Research 66: 919–927.

Copard, Y., P. Amiotte-Suchet, and C. Di-Giovanni. 2007. Storage and release of fossil

organic carbon related to weathering of sedimentary rocks. Earth and Planetary

Science Letters 258: 345–357.

Coplen, T. B., and C. Kendall (2000) Stable hydrogen and oxygen isotope ratios for

selected sites of the US Geological Survey’s NASQAN and benchmark surface-

water networks. DTIC Document.

Cremona, F., D. Planas, and M. Lucotte. 2010. Influence of functional feeding groups

and spatiotemporal variables on the δ15N signature of littoral macroinvertebrates.

Hydrobiologia 647: 51–61.

174

Cruz-Rivera, E., and M. E. Hay. 2000. Can quantity replace quality? Food choice,

compensatory feeding, and fitness of marine mesograzers. Ecology 81: 201–219.

Cummins, K. W. 1974. Structure and function of stream ecosystems. BioScience 24:631–

641.

Cummins, K. W. 2016. Combining taxonomy and function in the study of stream

macroinvertebrates. Journal of Limnology 75: 235-241.

Cummins, K. W., and M. J. Klug. 1979. Feeding ecology of stream invertebrates. Annual

Review of Ecology and Systematics 10: 147–172.

Currie, L. A., T. I. Eglinton, B. A. Benner, and A. Pearson. 1997. Radiocarbon “dating”

of individual chemical compounds in atmospheric aerosol: First results comparing

direct isotopic and multivariate statistical apportionment of specific polycyclic

aromatic hydrocarbons. Nuclear Instruments and Methods in Physics Research

Section B: Beam Interactions with Materials and Atoms 123: 475–486.

Dalzell, B. J., T. R. Filley, and J. M. Harbor. 2007. The role of hydrology in annual

organic carbon loads and terrestrial organic matter export from a midwestern

agricultural watershed. Geochimica et Cosmochimica Acta 71: 1,448–1,462.

Davies, I. J. 1975. Selective feeding in some arctic Chironomidae. Verhandlungen der

Internationalen Vereinigung für theoretische und angewandte Limnologie

19:3,149–3,154.

Deines, P., and J. Grey. 2006. Site-specific methane production and subsequent midge

mediation within Esthwaite Water, UK. Archiv Für Hydrobiologie 167: 317–334.

175

Deines, P., M. J. Wooller, and J. Grey. 2009. Unravelling complexities in benthic food

webs using a dual stable isotope (hydrogen and carbon) approach. Freshwater

Biology 54: 2,243–2,251.

Dekar, M. P., D. D. Magoulick, and G. R. Huxel. 2009. Shifts in the trophic base of

intermittent stream food webs. Hydrobiologia 635: 263–277.

Delong, M. D., J. H. Thorp, K. S. Greenwood, and M. C. Miller. 2001. Responses of

consumers and food resources to a high magnitude, unpredicted flood in the upper

Mississippi River basin. Regulated Rivers: Research & Management 17: 217–

234.

DelVecchia, A. G., J. A. Stanford, and X. Xu. 2016. Ancient and methane-derived carbon

subsidizes contemporary food webs. Nature Communications

7:doi:10.1038/ncomms13163.

Diebel, M. W., and M. Vander Zanden. 2009. Nitrogen stable isotopes in streams: effects

of agricultural sources and transformations. Ecological Applications 19: 1,127–

1,134.

Diefendorf, A. F., K. E. Mueller, S. L. Wing, P. L. Koch, and K. H. Freeman. 2010.

Global patterns in leaf 13C discrimination and implications for studies of past and

future climate. Proceedings of the National Academy of Sciences 107: 5,738–

5,743.

Docile, T., D. C. Rosa, R. Figueiró, and J. Nessimian. 2016. Urbanisation alters the flow

of energy through stream food webs. Insect Conservation and Diversity 9: 416-

246.

176

Dodds, W. K., R. E. Hutson, A. C. Eichem, M. A. Evans, D. A. Gudder, K. M. Fritz, and

L. Gray. 1996. The relationship of floods, drying, flow and light to primary

production and producer biomass in a prairie stream. Hydrobiologia 333:151–159.

Dodds, W. K., A. M. Veach, C. M. Ruffing, D. M. Larson, J. L. Fischer, and K. H.

Costigan. 2013. Abiotic controls and temporal variability of river metabolism:

multiyear analyses of Mississippi and Chattahoochee River data. Freshwater

Science 32: 1,073–1,087.

Doi, H., K.-H. Chang, T. Ando, H. Imai, S. Nakano, A. Kajimoto, and I. Katano. 2008.

Drifting plankton from a reservoir subsidize downstream food webs and alter

community structure. Oecologia 156: 363–371.

Doi, H., Y. Takemon, T. Ohta, Y. Ishida, and E. Kikuchi. 2007. Effects of reach-scale

canopy cover on trophic pathways of caddisfly larvae in a Japanese mountain

stream. Marine and Freshwater Research 58: 811–817.

Doucett, R. R., J. C. Marks, D. W. Blinn, M. Caron, and B. A. Hungate. 2007. Measuring

terrestrial subsidies to aquatic food webs using stable isotopes of hydrogen.

Ecology 88: 1587–1592.

Drenzek, N. J., K. A. Hughen, D. B. Montluccon, J. R. Southon, G. M. dos Santos, E. R.

Druffel, L. Giosan, and T. I. Eglinton. 2009. A new look at old carbon in active

margin sediments. Geology 37: 239–242.

Duarte, C. M., and Y. T. Prairie. 2005. Prevalence of Heterotrophy and Atmospheric CO2

Emissions from Aquatic Ecosystems. Ecosystems 8: 862–870.

177

Durand, B. 1980. Kerogen: Insoluble Organic Matter from Sedimentary Rocks. Editions

Technip, Paris, France. 519 pp.

Ehleringer, J. R., N. Buchmann, and L. B. Flanagan. 2000. Carbon isotope ratios in

belowground carbon cycle processes. Ecological Applications 10: 412–422.

England, L. E., and A. D. Rosemond. 2004. Small reductions in forest cover weaken

terrestrial-aquatic linkages in headwater streams. Freshwater biology 49:721–734.

Evans, R. D. 2001. Physiological mechanisms influencing plant nitrogen isotope

composition. Trends in Plant Science 6: 121–126.

Evans, T. M. 2012. Assessing Food and Nutritional Resources of Native and Invasive

Lamprey Larvae Using Natural Abundance Isotopes. M.S. Thesis, The Ohio State

University.

Evans, T. M., and J. E. Bauer. 2016a. Identification of the nutritional resources of larval

sea lamprey in two Great Lakes tributaries using stable isotopes. Journal of Great

Lakes Research 1: 99–107.

Evans, T. M., and J. E. Bauer. 2016b. Using stable isotopes and C: N ratios to examine

the life-history strategies and nutritional sources of larval lampreys. Journal of

fish biology 88: 638–654.

Falkowski, P., Scholes, R. J., Boyle, E., and others. 2000. The global carbon cycle: a test

of our knowledge of earth as a system. Science 290: 291–296.

Fellman, J. B., E. Hood, P. A. Raymond, J. Hudson, M. Bozeman, and M. Arimitsu.

2015. Evidence for the assimilation of ancient glacier organic carbon in a

proglacial stream food web. Limnology and Oceanography 60: 1,118-1,128.

178

Fellman, J. B., R. G. Spencer, P.A. Raymond, N. E. Pettit, G. Skrzypek, P. J. Hernes, and

P. F. Grierson. 2014. Dissolved organic carbon biolability decreases along with its

modernization in fluvial networks in an ancient landscape. Ecology, 95: 2,622–

2,632.

Fenech, C., L. Rock, K. Nolan, J. Tobin, and A. Morrissey. 2012. The potential for a suite

of isotope and chemical markers to differentiate sources of nitrate contamination:

a review. Water Research 46: 2023–2041.

Fernandes, R., Bergemann, S., Hartz, S., and others. 2012. Mussels with meat: bivalve

tissue-shell radiocarbon age differences and archaeological implications.

Radiocarbon 54: 953–965.

Fernandes, R., A. Dreves, M. J. Nadeau, and P. Grootes. 2013. A freshwater lake saga:

carbon routing within the aquatic food web of Lake Schwerin. Radiocarbon 55:

1,102–1,113.

Fernandes, R., C. Rinne, M. J. Nadeau, and P. Grootes. 2015. Towards the use of

radiocarbon as a dietary proxy: Establishing a first wide-ranging radiocarbon

reservoir effects baseline for Germany. Environmental Archaeology 21: 285-294.

Fernández, P., G. Carrera, J. O. Grimalt, and others. 2003. Factors governing the

atmospheric deposition of polycyclic aromatic hydrocarbons to remote areas.

Environmental Science & Technology 37: 3,261–3,267.

Finlay, J. C. 2001. Stable-carbon-isotope ratios of river biota: implications for energy

flow in lotic food webs. Ecology 82: 1,052–1,064.

179

Finlay, J. C. 2003. Controls of streamwater dissolved inorganic carbon dynamics in a

forested watershed. Biogeochemistry 62: 231–252.

Finlay, J. C. 2004. Patterns and controls of lotic algal stable carbon isotope ratios.

Limnology and Oceanography 49: 850–861.

Finlay, J. C., R. R. Doucett, and C. McNeely. 2010. Tracing energy flow in stream food

webs using stable isotopes of hydrogen. Freshwater Biology 55: 941–951.

Finlay, J. C., and C. Kendall. 2007. Stable isotope tracing of temporal and spatial

variability in organic matter sources to freshwater ecosystems. Pages 283-333 in

R. Michener and K. Lajtha, editors. Stable Isotopes in Ecology and

Environmental Science. Blackwell Publishing, Malden, Massachusetts, USA.

Finlay, J. C., S. Khandwala, and M. E. Power. 2002. Spatial Scales of Carbon Flow in a

River Food Web. Ecology 83: 1,845–1,859.

Finlay, J. C., M. E. Power, and G. Cabana. 1999. Effects of water velocity on algal

carbon isotope ratios: implications for river food web studies. Limnology and

Oceanography 44: 1,198–1,203.

Fisher, S. G., and G. E. Likens. 1973. Energy flow in Bear Brook, New Hampshire: an

integrative approach to stream ecosystem metabolism. Ecological monographs 43:

421–439.

Fogel, M. L., and N. Tuross. 1999. Transformation of plant biochemicals to geological

macromolecules during early diagenesis. Oecologia 120: 336–346.

180

France, R. L. 2011. Leaves as “crackers”, biofilm as “peanut butter”: Exploratory use of

stable isotopes as evidence for microbial pathways in detrital food webs.

Oceanological and Hydrobiological Studies 40: 110–115.

Francis, T. B., D. E. Schindler, G. W. Holtgrieve, E. R. Larson, M. D. Scheuerell, B. X.

Semmens, and E. J. Ward. 2011. Habitat structure determines resource use by

zooplankton in temperate lakes. Ecology Letters 14: 364–372.

Francoeur, S. N., B. J. Biggs, R. A. Smith, and R. L. Lowe. 1999. Nutrient limitation of

algal biomass accrual in streams: seasonal patterns and a comparison of methods.

Journal of the North American Benthological Society 18: 242–260.

Fry, B. 2007. Stable isotope ecology. Springer Science & Business Media, New York,

New York, USA. 308 pp.

Fry, B., and C. Arnold. 1982. Rapid 13C/12C turnover during growth of brown shrimp

(Penaeus aztecus). Oecologia 54: 200–204.

Garnett, M. H., & Billett, M. F. 2007. Do riparian plants fix CO2 lost by evasion from

surface waters? An investigation using carbon isotopes. Radiocarbon 49: 993–

1,002.

Garten, C. T., and G. E. Taylor. 1992. Foliar δ13C within a temperate deciduous forest:

spatial, temporal, and species sources of variation. Oecologia 90: 1–7.

Gaudinski, J. B., S.E. Trumbore, E.A. Davidson, and S. Zheng. 2000. Soil carbon cycling

in a temperate forest: radiocarbon-based estimates of residence times,

sequestration rates and partitioning of fluxes. Biogeochemistry 51: 33–69.

181

Gaudinski, J. B., Torn, M. S., Riley, W. J., and others. 2009. Use of stored carbon

reserves in growth of temperate tree roots and leaf buds: analyses using

radiocarbon measurements and modeling. Global Change Biology 15: 992–1,014.

Geldern, R., P. Schulte, M. Mader, A. Baier, and J. A. Barth. 2015. Spatial and temporal

variations of pCO2, dissolved inorganic carbon and stable isotopes along a

temperate karstic watercourse. Hydrological Processes 29: 3,423–3,440.

Giling, D. P., P. Reich, and R. M. Thompson. 2012. Riparian vegetation removal alters

consumer–resource stoichiometry in an Australian lowland stream. Marine and

Freshwater Research 63: 1–8.

Goedkoop, W., M. Demandt, and G. Ahlgren. 2007. Interactions between food quantity

and quality (long-chain polyunsaturated fatty acid concentrations) effects on

growth and development of Chironomus riparius. Canadian Journal of Fisheries

and Aquatic Sciences 64: 425–436.

Goni, M. A., J. A. Hatten, R. A. Wheatcroft, and J. C. Borgeld. 2013. Particulate organic

matter export by two contrasting small mountainous rivers from the Pacific

Northwest, USA. Journal of Geophysical Research: Biogeosciences 118: 112–

134.

Gomi, T., R. C. Sidle, and J. S. Richardson. 2002. Understanding processes and

downstream linkages of headwater systems. BioScience 52: 905–916.

Graham, L. E., and L. W. Wilcox. 2000. Algae. Upper Saddler River. NJ: Prentice Hall.

616 pp.

182

Graz, Y., C. Di-Giovanni, Y. Copard, N. Mathys, A. Cras, and V. Marc. 2012. Annual

fossil organic carbon delivery due to mechanical and chemical weathering of

marly badlands areas. Earth Surface Processes and Landforms 37: 1,263–1,271.

Grey, J. 2016. The incredible lightness of being methane-fuelled: stable isotopes reveal

alternative energy pathways in aquatic ecosystems and beyond. Frontiers in

Ecology and Evolution 4: doi.org/10.3389/fevo.2016.00008.

Griffith, D. R., R. T. Barnes, and P. A. Raymond. 2009. Inputs of Fossil Carbon from

Wastewater Treatment Plants to U.S. Rivers and Oceans. Environmental Science

& Technology 43: 5,647–5,651.

Grunsky, E. C., L. J. Drew, and D. M. Sutphin. 2009. Process recognition in multi-

element soil and stream-sediment geochemical data. Applied Geochemistry 24:

1,602–1,616.

Guillemette, F., S. L. McCallister, and P. A. del Giorgio. 2016. Selective consumption

and metabolic allocation of terrestrial and algal carbon determine allochthony in

lake bacteria. The ISME Journal 10: 1,373-1,382.

Guo, F., M. J. Kainz, F. Sheldon, and S. E. Bunn. 2016a. The importance of high-quality

algal food sources in stream food webs–current status and future perspectives.

Freshwater Biology 61: 815-831.

Guo, F., M. J. Kainz, D. Valdez, F. Sheldon, and S. E. Bunn. 2016b. The effect of light

and nutrients on algal food quality and their consequent effect on grazer growth in

subtropical streams. Freshwater Science 35: 1,202-1,212.

183

Guo, J., F. Wu, X. Luo, and others. 2010. Anthropogenic input of polycyclic aromatic

hydrocarbons into five lakes in Western China. Environmental Pollution 158:

2,175–2,180.

Guo, L., & Macdonald, R. W. 2006. Source and transport of terrigenous organic matter in

the upper Yukon River: Evidence from isotope (δ13C, ∆14C, and δ15N)

composition of dissolved, colloidal, and particulate phases. Global

Biogeochemical Cycles 20:doi: 10.1029/2005GB002593.

Hadwen, W. L., M. Spears, and M. J. Kennard. 2010. Temporal variability of benthic

algal δ13C signatures influences assessments of carbon flows in stream food webs.

Hydrobiologia 651: 239–251.

Hagvar, S., and M. Ohlson. 2013. Ancient carbon from a melting glacier gives high 14C

age in living pioneer invertebrates. Scientific Reports 3:doi: 10.1038/srep02820.

Hagvar, S., M. Ohlson, and J. E. Brittain. 2016. A melting glacier feeds aquatic and

terrestrial invertebrates with ancient carbon and supports early succession. Arctic,

Antarctic, and Alpine Research 48: 551–562.

Hall, S. J., R. L. Hale, M. A. Baker, D. R. Bowling, and J. R. Ehleringer. 2015. Riparian

plant isotopes reflect anthropogenic nitrogen perturbations: robust patterns across

land use gradients. Ecosphere 6: 1–16.

Hall, R. O., G. E. Likens, and H. M. Malcom. 2001. Trophic basis of invertebrate

production in 2 streams at the Hubbard Brook Experimental Forest. Journal of the

North American Benthological Society 20: 432–447.

184

.Hall, R. O., and J. L. Meyer. 1998. The trophic significance of bacteria in a detritus-

based stream food web. Ecology 79: 1,995–2,012.

Hall, R. O., C. B. Yackulic, T. A. Kennedy, M. D. Yard, E. J. Rosi-Marshall, N.

Voichick, and K. E. Behn. 2015. Turbidity, light, temperature, and hydropeaking

control primary productivity in the Colorado River, Grand Canyon. Limnology

and Oceanography 60: 512–526.

Hall, R. O., J. B. Wallace, and S. L. Eggert. 2000. Organic matter flow in stream food

webs with reduced detrital resource base. Ecology 81: 3,445–3,463.

Hamilton, H., and F. Clifford. 1983. The seasonal food habits of mayfly (Ephemeroptera)

nymphs from three Alberta, Canada, streams, with special reference to absolute

volume and size of particles ingested. Archiv für Hydrobiologie 65: 197–234.

Hammond, P. A. 2016. The relationship between methane migration and shale-gas well

operations near Dimock, Pennsylvania, USA. Hydrogeology Journal 24: 503–519.

Harmon, M. E., J. F. Franklin, F. J. Swanson, P. Sollins, S. V. Gregory, J. D. Lattin, N.

H. Anderson, S. P. Cline, N. G. Aumen, J. R. Sedell, and others. 1986. Ecology of

coarse woody debris in temperate ecosystems. Advances in Ecological Research

15: 133-302.

Harrington, R. R., B. P. Kennedy, C. P. Chamberlain, J. D. Blum, and C. L. Folt. 1998.

15N enrichment in agricultural catchments: field patterns and applications to

tracking Atlantic salmon (Salmo salar). Chemical Geology 147: 281–294.

185

Hayden, B., S. M. McWilliam-Hughes, and R. A. Cunjak. 2016. Evidence for limited

trophic transfer of allochthonous energy in temperate river food webs. Freshwater

Science 35: 544-558.

Heathwaite, A. L., and P. J. Johnes. 1996. Contribution of nitrogen species and

phosphorus fractions to stream water quality in agricultural catchments.

Hydrological processes 10: 971–983.

Hedges, J. I., J. R. Ertel., P.D. Quay, and others. 1986. Organic carbon-14 in the Amazon

River system. Science 231: 1,129–1,131.

Hedges, J. I. 1992. Global biogeochemical cycles: progress and problems. Marine

Chemistry 39:67–93.

Hedges, J. I., R. G. Keil, and R. Benner. 1997. What happens to terrestrial organic matter

in the ocean? Organic Geochemistry 27: 195–212.

Heilweil, V. M., P. L. Grieve, S. A. Hynek, S. L. Brantley, D. K. Solomon, and D. W.

Risser. 2015. Stream measurements locate thermogenic methane fluxes in

groundwater discharge in an area of shale-gas development. Environmental

Science & Technology 49: 4,057–4,065.

Heitkamp, M. A., and B. T. Johnson. 1984. Impact of an oil field effluent on microbial

activities in a Wyoming river. Canadian Journal of Microbiology 30: 786–792.

Henley, W. F., M. A. Patterson, R. J. Neves, and A. D. Lemly. 2000. Effects of

sedimentation and turbidity on lotic food webs: a concise review for natural

resource managers. reviews in Fisheries Science 8: 125–139.

186

Hill, W. R., and S. M. Dimick. 2002. Effects of riparian leaf dynamics on periphyton

photosynthesis and light utilisation efficiency. Freshwater Biology 47: 1,245–

1,256.

Hill, W. R., P. J. Mulholland, and E. R. Marzolf. 2001. Stream ecosystem responses to

forest leaf emergence in spring. Ecology 82: 2,306–2,319.

Hilton, R. G., A. Galy, N. Hovius, M.-J. Horng, and H. Chen. 2011. Efficient transport of

fossil organic carbon to the ocean by steep mountain rivers: An orogenic carbon

sequestration mechanism. Geology 39: 71–74.

Hladyz, S., D. L. Nielsen, P. J. Suter, and E. S. Krull. 2012. Temporal variations in

organic carbon utilization by consumers in a lowland river. River Research and

Applications 28: 513–528.

Hoefs, J. 2008. Stable isotope geochemistry. Springer Science & Business Media, Berlin,

Germany. 201 pp.

Hoehler, T. M., W. S. Borowski, M. J. Alperin, N. M. Rodriguez, and C. K. Paull. 2000.

Model, stable isotope, and radiotracer characterization of anaerobic methane

oxidation in gas hydrate-bearing sediments of the Blake Ridge. Pages 79–85 in

Proceedings of the Ocean Drilling Program, Scientific Results. Ocean Drilling

Program College Station, TX, USA.

Hoellein, T. J., J. L. Tank, E. J. Rosi-Marshall, S. A. Entrekin, and G. A. Lamberti. 2007.

Controls on spatial and temporal variation of nutrient uptake in three Michigan

headwater streams. Limnology and Oceanography 52: 1,964–1,977.

187

Hofmeister, K. L., C. M. Cianfrani, and W. C. Hession. 2015. Complexities in the stream

temperature regime of a small mixed-use watershed, Blacksburg, VA. Ecological

Engineering 78: 101–111.

Hood, E., J. Fellman, R. G. Spencer, P. J. Hernes, R. Edwards, D. D’Amore, and D.

Scott. 2009. Glaciers as a source of ancient and labile organic matter to the marine

environment. Nature 462: 1,044–1,047.

Hood, E., T. J. Battin, J. Fellman, S. O’Neel, and R. G. Spencer. 2015. Storage and

release of organic carbon from glaciers and ice sheets. Nature Geoscience 8: 91–

96.

Hope, D., M. F. Billett, and M. S. Cresser. 1997. Exports of organic carbon in two river

systems in NE Scotland. Journal of Hydrology 193: 61–82.

Hope, D., S. M. Palmer, M. F. Billett, and J. J. Dawson. 2004. Variations in dissolved

CO2 and CH4 in a first-order stream and catchment: an investigation of soil–

stream linkages. Hydrological Processes 18: 3,255–3,275.

Horner, R. R., and E. B. Welch. 1981. Stream periphyton development in relation to

current velocity and nutrients. Canadian Journal of Fisheries and Aquatic

Sciences 38: 449–457.

Horner, R. R., E. B. Welch, M. R. Seeley, and J. M. Jacoby. 1990. Responses of

periphyton to changes in current velocity, suspended sediment and phosphorus

concentration. Freshwater biology 24: 215–232.

188

Hossler, K., and J. E. Bauer. 2012. Estimation of riverine carbon and organic matter

source contributions using time-based isotope mixing models. Journal of

Geophysical Research 117: doi:10.1029/2012JG001988.

Hossler, K., and J. E. Bauer. 2013a. Amounts, isotopic character, and ages of organic and

inorganic carbon exported from rivers to ocean margins: 2. Assessment of natural

and anthropogenic controls. Global Biogeochemical Cycles 27: 347–362.

Hossler, K., and J. E. Bauer. 2013b. Amounts, isotopic character and ages of organic and

inorganic carbon exported from rivers to ocean margins: 1. Estimates of terrestrial

losses and inputs to the Middle Atlantic Bight. Global Biogeochemical Cycles 27:

331–346.

House, H. L. 1965. Effects of low levels of the nutrient content of a food and of nutrient

imbalance on the feeding and the nutrition of a phytophagous larva, Celerio

euphorbiae (Lepidoptera: Sphingidae). The Canadian Entomologist 97: 62–68.

Huryn, A. D., R. H. Riley, R. G. Young, C. J. Arbuckle, K. Peacock, and G. Lyon. 2001.

Temporal shift in contribution of terrestrial organic matter to consumer

production in a grassland river. Freshwater Biology 46:213–226.

Hynes, H. B. N. (1975). Edgardo Baldi memorial lecture. The stream and its valley.

Verhandlungen Der Internationalen Vereinigung Fur Theoretische Und

Angewandte Limnologie 19: 1–15.

Ishikawa, N. F., F. Hyodo, and I. Tayasu. 2013. Use of carbon-13 and carbon-14 natural

abundances for stream food web studies. Ecological Research 28: 759-769.

189

Ishikawa, N. F., H. Togashi, Y. Kato, M. Yoshimura, Y. Kohmatsu, C. Yoshimizu, N. O.

Ogawa, N. Ohte, N. Tokuchi, N. Ohkouchi, and others. 2016. Terrestrial–aquatic

linkage in stream food webs along a forest chronosequence: multi-isotopic

evidence. Ecology 97: 1,146-1,158.

Ishikawa, N. F., M. Uchida, Y. Shibata, and I. Tayasu. 2014. Carbon storage reservoirs in

watersheds support stream food webs via periphyton production. Ecology 95:

1,264-11,271.

Ishikawa, N. F., M. Yamane, H. Suga, N. O. Ogawa, Y. Yokoyama, and N. Ohkouchi.

2015. Chlorophyll a-specific Δ14C, δ13C and δ15N values in stream periphyton:

implications for aquatic food web studies. Biogeosciences 12: 6,781–6,789.

Jackson, R. B., A. Vengosh, T. H. Darrah, N. R. Warner, A. Down, R. J. Poreda, S. G.

Osborn, K. Zhao, and J. D. Karr. 2013. Increased stray gas abundance in a subset

of drinking water wells near Marcellus shale gas extraction. Proceedings of the

National Academy of Sciences 110: 11,250–11,255.

Jenkinson, D. S., P. R. Poulton, and C. Bryant. 2008. The turnover of organic carbon in

subsoils. Part 1. Natural and bomb radiocarbon in soil profiles from the

Rothamsted long-term field experiments. European Journal of Soil Science 59:

391–399.

Junger, M., and D. Planas. 1994. Quantitative use of stable carbon isotope analysis to

determine the trophic base of invertebrate communities in a boreal forest lotic

system. Canadian Journal of Fisheries and Aquatic Sciences 51: 52–61.

190

Junk, W. J., P. B. Bayley, and R. E. Sparks. 1989. The flood pulse concept in river-

floodplain systems. Canadian Special Publication of Fisheries and Aquatic

Sciences 106: 110–127.

Junker, J. R., and W. F. Cross. 2014. Seasonality in the trophic basis of a temperate

stream invertebrate assemblage: Importance of temperature and food quality.

Limnology and Oceanography 59: 507–518.

Kao, S. J., and K. K. Liu. 1996. Particulate organic carbon export from a subtropical

mountainous river (Lanyang Hsi) in Taiwan. Limnology and Oceanography 41:

1,749–1,757.

Karlsson, J., M. Jansson, and A. Jonsson. 2007. Respiration of allochthonous organic

carbon in unproductive forest lakes determined by the Keeling Plot method.

Limnology and Oceanography 52: 603–608.

Karr, J. D., W. J. Showers, J. W. Gilliam, and A. S. Andres. 2001. Tracing nitrate

transport and environmental impact from intensive swine farming using delta

nitrogen-15. Journal of Environmental Quality 30: 1,163–1,175.

Kaushal, S., and M. W. Binford. 1999. Relationship between C:N ratios of lake

sediments, organic matter sources, and historical deforestation in Lake Pleasant,

Massachusetts, USA. Journal of Paleolimnology 22: 439–442.

Kaushal, S. S., Likens, G. E., Jaworski, N. A., and others 2010. Rising stream and river

temperatures in the United States. Frontiers in Ecology and the Environment 8:

461–466.

191

Kaushal, S. S., P. M. Mayer, P. G. Vidon, R. M. Smith, M. J. Pennino, T. A. Newcomer,

S. Duan, C. Welty, and K. T. Belt. 2014. Land use and climate variability amplify

carbon, nutrient, and contaminant pulses: a review with management implications.

JAWRA Journal of the American Water Resources Association 50: 585–614.

Kautza, A., and S. M. P. Sullivan. 2016. The energetic contributions of aquatic primary

producers to terrestrial food webs in a mid-size river system. Ecology 97: 694–

705.

Keaveney, E. M., and P. J. Reimer. 2012. Understanding the variability in freshwater

radiocarbon reservoir offsets: a cautionary tale. Journal of Archaeological Science

39: 1,306–1,316.

Keaveney, E. M., P. J. Reimer, and R. H. Foy. 2015 (a). Young, old, and weathered

carbon- part 2: using radiocarbon and stable isotopes to identify terrestrial carbon

support of the food web in an alkaline, humic lake. Radiocarbon 57: 425-438.

Keaveney, E. M., P. J. Reimer, and R. J. Foy. 2015 (b). Young, old and weathered

carbon- part 1: using radiocarbon and stable isotopes to identify carbon sources in

an alkaline, humic lake. Radiocarbon 57: 407-423.

Kendall, C., E. M. Elliott, and S. D. Wankel. 2007. Tracing anthropogenic inputs of

nitrogen to ecosystems. Pages 375-449 in R. Michener and K. Lajtha, editors.

Stable Isotopes in Ecology and Environmental Science. Blackwell Publishing,

Malden, Massachusetts, USA. 336 pp.

192

Kerner, M., H. Hohenberg, S. Ertl, M. Reckermann, and A. Spitzy. 2003. Self-

organization of dissolved organic matter to micelle-like microparticles in river

water. Nature 422: 150–154.

Kleber, M. 2010. What is recalcitrant soil organic matter? Environmental Chemistry 7:

320–332.

Kleber, M., and M. G. Johnson. 2010. Advances in understanding the molecular structure

of soil organic matter: implications for interactions in the environment. Advances

in Agronomy 106: 77–142.

Kleber, M., P. S. Nico, A. Plante, T. Filley, M. Kramer, C. Swanston, and P. Sollins.

2011. Old and stable soil organic matter is not necessarily chemically recalcitrant:

implications for modeling concepts and temperature sensitivity. Global Change

Biology 17: 1,097–1,107.

Kobayashi, S., F. Akamatsu, K. Amano, S. Nakanishi, and Y. Oshima. 2011.

Longitudinal changes in δ13C of riffle macroinvertebrates from mountain to

lowland sections of a gravel-bed river. Freshwater Biology 56: 1,434–1,446.

Kohzu, A., C. Kato, T. Iwata, D. Kishi, M. Murakami, S. Nakano, and E. Wada. 2004.

Stream food web fueled by methane-derived carbon. Aquatic Microbial Ecology

36: 189–194.

Komada, T., E. R. M. Druffel, and S. E. Trumbore. 2004. Oceanic export of relict carbon

by small mountainous rivers. Geophysical Research Letters 31:doi:

10.1029/2004GL019512.

193

Kritzberg, E. S., J. J. Cole, M. L. Pace, W. Granéli, and D. L. Bade. 2004. Autochthonous

versus allochthonous carbon sources of bacteria: Results from whole-lake 13C

addition experiments. Limnology and Oceanography 49: 588–596.

Kruger, B. R., J. P. Werne, D. K. Branstrator, T. R. Hrabik, Y. Chikaraishi, N. Ohkouchi,

and E. C. Minor. 2016. Organic matter transfer in Lake Superior’s food web:

Insights from bulk and molecular stable isotope and radiocarbon analyses.

Limnology and Oceanography 61: 149–164.

Kürten, B., S. J. Painting, U. Struck, N. V. Polunin, and J. J. Middelburg. 2013. Tracking

seasonal changes in North Sea zooplankton trophic dynamics using stable

isotopes. Biogeochemistry 113: 167–187.

Lal, R. 2003. Soil erosion and the global carbon budget. Environment International 29:

437–450.

Lau, D. C., I. Sundh, T. Vrede, J. Pickova, and W. Goedkoop. 2014. Autochthonous

resources are the main driver of consumer production in dystrophic boreal lakes.

Ecology 95: 1,506–1,519.

Leberfinger, K., I. Bohman, and J. Herrmann. 2011. The importance of terrestrial

resource subsidies for shredders in open-canopy streams revealed by stable

isotope analysis. Freshwater Biology 56: 470–480.

Ledger, M. E., L. E. Brown, F. K. Edwards, A. M. Milner, and G. Woodward. 2013.

Drought alters the structure and functioning of complex food webs. Nature

Climate Change 3: 223–227.

194

Lefebvre, S., J. C. Clement, G. Pinay, C. Thenail, P. Durand, and P. Marmonier. 2007.

15N-Nitrate signature in low-order streams: effects of land cover and agricultural

practices. Ecological Applications 17: 2.333–2.346.

Leithold, E. L., and N. E. Blair. 2001. Watershed control on the carbon loading of marine

sedimentary particles. Geochimica et Cosmochimica Acta 65: 2,231–2,240.

Leithold, E. L., N. E. Blair, and D. W. Perkey. 2006. Geomorphologic controls on the age

of particulate organic carbon from small mountainous and upland rivers. Global

Biogeochemical Cycles 20:doi:10.1029/2005GB002677.

Lenat, D. R. 1984. Agriculture and stream water quality: a biological evaluation of

erosion control practices. Environmental Management 8: 333–343.

Lenat, D. R., and J. K. Crawford. 1994. Effects of land use on water quality and aquatic

biota of three North Carolina Piedmont streams. Hydrobiologia 294: 185–199.

14 Levin, I., B. Kromer, and S. Hammer. 2013. Atmospheric Δ CO2 trend in Western

European background air from 2000 to 2012. Tellus B 65:

doi.org/10.3402/tellusb.v65i0.20092.

Levin, L. A., and G. F. Mendoza. 2007. Community structure and nutrition of deep

methane‐seep macrobenthos from the North Pacific (Aleutian) Margin and the

Gulf of Mexico (Florida Escarpment). Marine Ecology 28: 131–151.

Li, H., E.C. Minor, E. C., and P. K. Zigah. 2013. Diagenetic changes in Lake Superior

sediments as seen from FTIR and 2D correlation spectroscopy. Organic

Geochemistry 58: 125–136.

195

Likens, G. E., and F. H. Bormann. 1974. Linkages between terrestrial and aquatic

ecosystems. BioScience 24: 447–456.

Llewellyn, G. T., F. Dorman, J. L. Westland, D. Yoxtheimer, P. Grieve, T. Sowers, E.

Humston-Fulmer, and S. L. Brantley. 2015. Evaluating a groundwater supply

contamination incident attributed to Marcellus Shale gas development.

Proceedings of the National Academy of Sciences 112: 6,325–6,330.

Longworth, B. E., S. T. Petsch, P. A. Raymond, and J. E. Bauer. 2007. Linking lithology

and land use to sources of dissolved and particulate organic matter in headwaters

of a temperate, passive-margin river system. Geochimica et Cosmochimica Acta

71: 4,233–4,250.

Lowe, W. H., G. E. Likens, and M. E. Power. 2006. Linking scales in stream ecology.

BioScience 56: 591–597.

Lu, Y. H., E. A. Canuel, J. E. Bauer, and R. M. Chambers. 2014. Effects of watershed

land use on sources and nutritional value of particulate organic matter in

temperate headwater streams. Aquatic Aciences 76: 419–436.

Lützow, M., I. Kögel-Knabner, K. Ekschmitt, E. Matzner, G. Guggenberger, B.

Marschner, and H. Flessa. 2006. Stabilization of organic matter in temperate soils:

mechanisms and their relevance under different soil conditions–a review.

European Journal of Soil Science 57: 426–445.

Madsen, J. D., P. A. Chambers, W. F. James, E. W. Koch, and D. F. Westlake. 2001. The

interaction between water movement, sediment dynamics and submersed

macrophytes. Hydrobiologia 444: 71–84.

196

Mallin, M. A., M. R. McIver, A. R. Robuck, and A. K. Dickens. 2015. Industrial Swine

and Poultry Production Causes Chronic Nutrient and Fecal Microbial Stream

Pollution. Water, Air, & Soil Pollution 226: 1–13.

Marcarelli, A. M., C. V. Baxter, M. M. Mineau, and R. O. Hall Jr. 2011. Quantity and

quality: unifying food web and ecosystem perspectives on the role of resource

subsidies in freshwaters. Ecology 92: 1,215–1,225.

Marín-Spiotta, E., K. E. Gruley, J. Crawford, E. E. Atkinson, J. R. Miesel, S. Greene, C.

Cardona-Correa, and R. G. M. Spencer. 2014. Paradigm shifts in soil organic

matter research affect interpretations of aquatic carbon cycling: transcending

disciplinary and ecosystem boundaries. Biogeochemistry 117: 279–297.

Marwick, T. R., F. Tamooh, C. R. Teodoru, A. V. Borges, F. Darchambeau, and S.

Bouillon. 2015. The age of river-transported carbon: A global perspective. Global

Biogeochemical Cycles 29: 122–137.

Masiello, C. A., and E. R. M. Druffel. 2001. Carbon isotope geochemistry of the Santa

Clara River. Global Biogeochemical Cycles 15: 407–416.

Matthaei, C. D., J. J. Piggott, and C. R. Townsend. 2010. Multiple stressors in

agricultural streams: interactions among sediment addition, nutrient enrichment

and water abstraction. Journal of Applied Ecology 47: 639–649.

Mayer, M. S., and G. E. Likens. 1987. The importance of algae in a shaded headwater

stream as food for an abundant caddisfly (Trichoptera). Journal of the North

American Benthological Society 6:262–269.

197

Mayorga, E., A. K. Aufdenkampe, C.A. Masiello, and others. 2005. Young organic

matter as a source of carbon dioxide outgassing from Amazonian rivers. Nature

436: 538–541.

McCafferty, W. P., and Y. J. Bae. 1992. Filter-feeding habits of the larvae of

Anthopotamus (Ephemeroptera: Potamanthidae). Annales de Limnologie-

International Journal of Limnology 28: 27-34.

McCallister, S. L., J. E. Bauer, J. E. Cherrier and H. W. Ducklow. 2004. Assessing

sources and ages of organic matter supporting river and estuarine bacterial

production: a multiple-isotope (∆14C, δ13C, and δ15N) approach. Limnology and

Oceanography 49: 1,687–1,702.

McCallister, S. L., and P. A. del Giorgio. 2008. Direct measurement of the δ13C signature

of carbon respired by bacteria in lakes: Linkages to potential carbon sources,

ecosystem baseline metabolism, and CO2 fluxes. Limnology and Oceanography,

53: 1,204-1,216.

McCallister, S. L., and P. A. del Giorgio. 2012. Evidence for the respiration of ancient

terrestrial organic C in northern temperate lakes and streams. Proceedings of the

National Academy of Sciences 109: 16,963–16,968.

McConnachie, J. L., and E. L. Petticrew. 2006. Tracing organic matter sources in riverine

suspended sediment: implications for fine sediment transfers. Geomorphology 79:

13–26.

198

McCutchan, J. H., and W. M. Lewis. 2002. Relative importance of carbon sources for

macroinvertebrates in a Rocky Mountain stream. Limnology and Oceanography

47: 742–752.

McIntyre, P. B., and A. S. Flecker. 2006. Rapid turnover of tissue nitrogen of primary

consumers in tropical freshwaters. Oecologia 148: 12–21.

McNeely, C., S. M. Clinton, and J. M. Erbe. 2006. Landscape variation in C sources of

scraping primary consumers in streams. Journal of the North American

Benthological Society 25: 787–799.

McNeely, C., J. C. Finlay, and M. E. Power. 2007. Grazer traits, competition, and carbon

sources to a headwater-stream food web. Ecology 88: 391–401.

McNichol, A. P., and L.I. Aluwihare. 2007. The power of radiocarbon in biogeochemical

studies of the marine carbon cycle: Insights from studies of dissolved and

particulate organic carbon (DOC and POC). Chemical Reviews 107: 443–466.

Melymuk, L., M. Robson, S. A. Csiszar, and others. 2014. From the city to the lake:

loadings of PCBs, PBDEs, PAHs and PCMs from Toronto to Lake Ontario.

Environmental Science & Technology 48: 3,732–3,741.

Merritt, R. W., K. W. Cummins, and M. B. Berg. 2008. An introduction to the aquatic

macroinvertebrates of North America. Fourth edition. Kendall-Hunt Publishing

Company, Dubuque, Iowa, USA.

199

Mesa, L. M., C. Maldini, G. Mayora, M. Saigo, M. Marchese, and F. Giri. 2016.

Decomposition of cattle manure and colonization by macroinvertebrates in

sediment of a wetland of the Middle Paraná River. Journal of Soils and Sediments

16: 2,316-2,325.

Meybeck, M. 1993. Riverine transport of atmospheric carbon: sources, global typology

and budget. Water, Air, and Soil Pollution 70: 443–463.

Meyers, P. A., and R. Ishiwatari. 1993. Lacustrine organic geochemistry—an overview of

indicators of organic matter sources and diagenesis in lake sediments. Organic

Geochemistry 20: 867–900.

Middelburg, J. J. 2014. Stable isotopes dissect aquatic food webs from the top to the

bottom. Biogeosciences 11: 2,357–2,371.

Mihuc, T. B. 1997. The functional trophic role of lotic primary consumers: generalist

versus specialist strategies. Freshwater Biology 37: 455–462.

Minshall, G. W. 1967. Role of allochthonous detritus in the trophic structure of a

woodland springbrook community. Ecology 48: 139–149.

Mitra, S., T. S. Bianchi, B. A. McKee, and M. Sutula. 2002. Black carbon from the

Mississippi River: Quantities, sources, and potential implications for the global

carbon cycle. Environmental Science & Technology 36: 2,296–2,302.

Mladenov, N., J. Lípez-Ramos, D.M. McKnight, and I. Rechea. 2009. Alpine lake optical

properties as sentinels of dust deposition and global change. Limnology and

Oceanography 54: 2,386–2,400.

200

Molofsky, L. J., J. A. Connor, A. S. Wylie, T. Wagner, and S. K. Farhat. 2013.

Evaluation of methane sources in groundwater in northeastern Pennsylvania.

Groundwater 51: 333–349.

Mook, W. G., J. C. Bommerson, and W. H. Staverman. 1974. Carbon isotope

fractionation between dissolved bicarbonate and gaseous carbon dioxide. Earth

and Planetary Science Letters 22: 169–176.

Moore, J. W., and B. X. Semmens. 2008. Incorporating uncertainty and prior information

into stable isotope mixing models. Ecology Letters 11: 470–480.

Moulton, T. P., S. A. Magalhaes-Fraga, E. F. Brito, and F. A. Barbosa. 2010.

Macroconsumers are more important than specialist macroinvertebrate shredders

in leaf processing in urban forest streams of Rio de Janeiro, Brazil. Hydrobiologia

638: 55–66.

Müller-Navarra, D. C. 2008. Food web paradigms: The biochemical view on trophic

interactions. International Review of Hydrobiology 93: 489–505.

Murphy, J. F., J. I. Jones, J. L. Pretty, C. P. Duerdoth, A. Hawczak, A. Arnold, J. H.

Blackburn, P. S. Naden, G. Old, D. A. Sear, and others. 2015. Development of a

biotic index using stream macroinvertebrates to assess stress from deposited fine

sediment. Freshwater Biology 60: 2,019–2,036.

Nakano, S., and M. Murakami. 2001. Reciprocal subsidies: dynamic interdependence

between terrestrial and aquatic food webs. Proceedings of the National Academy

of Sciences 98:166-170.

201

Neres-Lima, V., E. F. Brito, F. A. Krsulović, A. M. Detweiler, A. E. Hershey, and T. P.

Moulton. 2016. High importance of autochthonous basal food source for the food

web of a Brazilian tropical stream regardless of shading. International Review of

Hydrobiology 101: 132–142.

Nishikawa, J., A. Kohzu, N. Boontanon, T. Iwata, T. Tanaka, N. O. Ogawa, R. Ishii, and

E. Wada. 2009. Isotopic composition of nitrogenous compounds with emphasis

on anthropogenic loading in river ecosystems. Isotopes in Environmental and

Health Studies 45: 27–40.

Ohio Environmental Protection Agency (OEPA), Division of Surface Water. 2012. Paint

Creek Watershed TMDL Report.

http://www.epa.state.oh.us/Portals/35/tmdl/PaintCr_factsheet_FINAL.pdf

Osborn, S. G., and J. C. McIntosh. 2010. Chemical and isotopic tracers of the

contribution of microbial gas in Devonian organic-rich shales and reservoir

sandstones, northern Appalachian Basin. Applied Geochemistry 25: 456–471.

Osborn, S. G., A. Vengosh, N. R. Warner, and R. B. Jackson. 2011. Methane

contamination of drinking water accompanying gas-well drilling and hydraulic

fracturing. Proceedings of the National Academy of Sciences 108: 8,172-8,176.

Osburn, C. L., and G. St-Jean. 2007. The use of wet chemical oxidation with high-

amplification isotope ratio mass spectrometry (WCO-IRMS) to measure stable

isotope values of dissolved organic carbon in seawater. Limnology and

Oceanography Methods 5: 296–308.

202

Parker, S. R., S. R. Poulson, M. G. Smith, C. L. Weyer, and K. M. Bates. 2010. Temporal

variability in the concentration and stable carbon isotope composition of dissolved

inorganic and organic carbon in two Montana, USA rivers. Aquatic Geochemistry

16: 61–84.

Pastor, A., J. L. Riera, M. Peipoch, L. Cañas, M. Ribot, E. Gacia, E. Martí, and F.

Sabater. 2014. Temporal variability of nitrogen stable isotopes in primary uptake

compartments in four streams differing in human impacts. Environmental science

& technology 48: 6,612–6,619.

Pataki, D. E., J. T. Randerson, W. Wang, M. Herzenach, and N. E. Grulke. 2010. The

carbon isotope composition of plants and soils as biomarkers of pollution. Pages

407–423 in J. West, G. Bowen, T. Dawsos, and K. Tu, editors. Isoscapes.

Springer Science and Business Media, Cham, Switzerland.

Paul, M. J., and J. L. Meyer. 2008. Streams in the urban landscape. Urban Ecology 32:

333–365.

Peipoch, M., E. Martí, and E. Gacia. 2012. Variability in δ15N natural abundance of basal

resources in fluvial ecosystems: a meta-analysis. Freshwater Science 31: 1,003–

1,015.

Pennsylvania Department of Environmenta Protection, Office of Oil and Gas. Spud

Report. http://www.depreportingservices.state.pa.us/ReportServer/Pages/

ReportViewer.aspx?/Oil_Gas/Spud_External_Data. 29 November 2016.

203

Pessenda, L. C. R., E. P. E. Valencia, P. B. Camargo, and others. 1996. Natural

radiocarbon measurements in Brazilian soils developed on basic rocks.

Radiocarbon 38: 203–208.

Peterson, B. J., and B. Fry. 1987. Stable isotopes in ecosystem studies. Annual Review of

Ecology and Systematics 18:293–320.

Petsch, S. T., T.I. Eglinton, and K.J. Edwards. 2001. 14C-dead living biomass: evidence

for microbial assimilation of ancient organic carbon during shale weathering.

Science 292: 1,127–1,131.

Philippsen, B., and J. Heinemeier. 2013. Freshwater reservoir effect variability in

Northern Germany. Radiocarbon 55: 1,085–1,101.

Phillips, D. L., and J. W. Gregg. 2001. Uncertainty in source partitioning using stable

isotopes. Oecologia 127: 171–179.

Phillips, D. L., and J. W. Gregg. 2003. Source partitioning using stable isotopes: coping

with too many sources. Oecologia 136: 261–269.

Phillips, D. L., R. Inger, S. Bearhop, A. L. Jackson, J. W. Moore, A. C. Parnell, B. X.

Semmens, and E. J. Ward. 2014. Best practices for use of stable isotope mixing

models in food-web studies. Canadian Journal of Zoology 92: 823–835.

Piggott, J. J., D. K. Niyogi, C. R. Townsend, and C. D. Matthaei. 2015. Multiple stressors

and stream ecosystem functioning: climate warming and agricultural stressors

interact to affect processing of organic matter. Journal of Applied Ecology 52:

1,126–1,134.

204

Pingram, M. A., K. J. Collier, D. P. Hamilton, B. O. David, and B. J. Hicks. 2012.

Carbon sources supporting large river food webs: a review of ecological theories

and evidence from stable isotopes. Freshwater Reviews 5: 85-103.

Piscart, C., R. Genoel, S. Doledec, E. Chauvet, and P. Marmonier. 2009. Effects of

intense agricultural practices on heterotrophic processes in streams.

Environmental Pollution 157: 1,011–1,018.

Polis, G. A., W. B. Anderson, and R. D. Holt. 1997. Toward an Intergration of Landscape

and Food Web Ecology: The Dynamics of Spatially Subsidized Food Webs.

Annual Review of Ecology and Systematics 28: 289–316.

Pohlman, J. W., J. E. Bauer, E. A. Canuel, and others. 2009. Methane sources in gas

hydrate-bearing cold seeps: evidence from radiocarbon and stable isotopes.

Marine Chemistry 115: 102–109.

Post, D. M. 2002. Using stable isotopes to estimate trophic position: models, methods,

and assumptions. Ecology 83: 703–718.

Post, W. M., T. H. Peng, W. R. Emanuel, A. W. King, V. H. Dale, and D. L. DeAngelis.

1990. The global carbon cycle. American Scientist 78: 310–326.

Power, M. E., and W. E. Dietrich. 2002. Food webs in river networks. Ecological

Research 17: 451–471.

Power, M. E., W. E. Dietrich, and J. C. Finlay. 1996. Dams and downstream aquatic

biodiversity: potential food web consequences of hydrologic and geomorphic

change. Environmental Management 20: 887–895.

205

Prather, C. M., S. L. Pelini, A. Laws, E. Rivest, M. Woltz, C. P. Bloch, I. Del Toro, C. K.

Ho, J. Kominoski, T. A. Newbold, and others. 2013. Invertebrates, ecosystem

services and climate change. Biological Reviews 88: 327–348.

Pringle, C. M., R. J. Naiman, G. Bretschko, J. R. Karr, M. W. Oswood, J. R. Webster, R.

L. Welcomme, and M. J. Winterbourn. 1988. Patch dynamics in lotic systems: the

stream as a mosaic. Journal of the North American Benthological Society 7: 503–

524.

Ran, L., X. X. Lu, H. Sun, J. Han, R. Li, and J. Zhang. 2013. Spatial and seasonal

variability of organic carbon transport in the Yellow River, China. Journal of

Hydrology 498: 76–88.

Randall, G. W., O. Center, and others. 2008. Nitrogen Application Timing, Forms, and

Additives. Final Report: Gulf Hypoxia and Local Water Quality Concerns

Workshop. Upper Mississippi River Sub-basin Hypoxia Committee. pp. 73-85.

Randerson, J. T., I. G. Enting, E. A. G. Schuur, K. Caldeira, and I. Y. Fung. 2002.

14 Seasonal and latitudinal variability of troposphere Δ CO2: Post bomb

contributions from fossil fuels, oceans, the stratosphere, and the terrestrial

biosphere. Global Biogeochemical Cycles 16: doi:10.1029/2002GB001876.

Rasmussen, J. B. 2010. Estimating terrestrial contribution to stream invertebrates and

periphyton using a gradient-based mixing model for δ13C. Journal of Animal

Ecology 79: 393–402.

Raymond, P. A., and J.E. Bauer. 2001 (a). Riverine export of aged terrestrial organic

matter to the North Atlantic Ocean. Nature 409: 497–500.

206

Raymond, P. A., and J. E. Bauer. 2001 (b). Use of 14C and 13C natural abundances for

evaluating riverine, estuarine, and coastal DOC and POC sources and cycling: a

review and synthesis. Organic Geochemistry 32: 469–485.

Raymond, P. A., J. E. Bauer, N.F. Caraco, J.J. Cole., B. Longworth., and S.T. Petsch.

2004. Controls on the variability of organic matter and dissolved inorganic carbon

ages in northeast US rivers. Marine Chemistry 92: 353–366.

Raymond, P. A., J. Hartmann, R. Lauerwald, and others. 2013. Global carbon dioxide

emissions from inland waters. Nature 503: 355–359.

Reuss, N. S., L. Hamerlik, G. Velle, A. Michelsen, O. Pedersen, and K. P. Brodersen.

2013. Stable isotopes reveal that chironomids occupy several trophic levels within

West Greenland lakes: implications for food web studies. Limnology and

Oceanography 58: 1,023–1,034.

Richardson, J. S., and T. Sato. 2015. Resource subsidy flows across freshwater–terrestrial

boundaries and influence on processes linking adjacent ecosystems.

Ecohydrology 8: 406–415.

Roach, K. A. 2013. Environmental factors affecting incorporation of terrestrial material

into large river food webs. Freshwater Science 32: 283–298.

Roach, K. A., and K. O. Winemiller. 2015. Hydrologic regime and turbidity influence

entrance of terrestrial material into river food webs. Canadian Journal of Fisheries

and Aquatic Sciences 72: 1,099–1,112.

207

Roberts, B. J., and P. J. Mulholland. 2007. In-stream biotic control on nutrient

biogeochemistry in a forested stream, West Fork of Walker Branch. Journal of

Geophysical Research: Biogeosciences 112: doi:10.1029/2007JG000422.

Roditi, H. A., N. S. Fisher, and S. A. Sañudo-Wilhelmy. 2000. Uptake of dissolved

organic carbon and trace elements by zebra mussels. Nature 407:78–80. del Rosario, R. B., E. A. Betts, and V. H. Resh. 2002. Cow manure in headwater streams:

tracing aquatic insect responses to organic enrichment. Journal of the North

American Benthological Society 21: 278–289.

Rosemond, A. D., P. J. Mulholland, and S. H. Brawley. 2000. Seasonally shifting

limitation of stream periphyton: response of algal populations and assemblage

biomass and productivity to variation in light, nutrients, and herbivores. Canadian

Journal of Fisheries and Aquatic Sciences 57: 66–75.

Rosi-Marshall, E. J., K. L. Vallis, C. V. Baxter, and J. M. Davis. 2016. Retesting a

prediction of the River Continuum Concept: autochthonous versus allochthonous

resources in the diets of invertebrates. Freshwater Science 35: 534-543.

Rosi-Marshall, E. J., and J. B. Wallace. 2002. Invertebrate food webs along a stream

resource gradient. Freshwater Biology 47: 129–141.

Rumpel, C., & Kögel-Knabner, I. 2011. Deep soil organic matter—a key but poorly

understood component of terrestrial C cycle. Plant and Soil 338:143–158.

208

Ruppenthal, M., Y. Oelmann, H. F. del Valle, and W. Wilcke. 2015. Stable isotope ratios

of nonexchangeable hydrogen in organic matter of soils and plants along a 2100-

km climosequence in Argentina: New insights into soil organic matter sources

and transformations? Geochimica et Cosmochimica Acta 152: 54–71.

Ruppenthal, M., Y. Oelmann, and W. Wilcke. 2010. Isotope ratios of nonexchangeable

hydrogen in soils from different climate zones. Geoderma 155: 231–241.Sabo, J.

L., and M. E. Power. 2002. River–watershed exchange: effects of riverine

subsidies on riparian lizards and their terrestrial prey. Ecology 83: 1,860–1,869.

Schell, D. M. 1983. Carbon-13 and carbon-14 abundances in Alaskan aquatic organisms:

delayed production from peat in arctic food webs. Science 219:1,068-1,071.

Schillawski, S., and S. Petsch. 2008. Release of biodegradable dissolved organic matter

from ancient sedimentary rocks. Global Biogeochemical Cycles 22:

doi:10.1029/2007GB002980.

Schindler, J. E. 1971. Food quality and zooplankton nutrition. The Journal of Animal

Ecology 40: 589–595.

Schlesinger, W. H., and E. S. Bernhardt. 2013. Biogeochemistry: an analysis of global

change. Waltham, MA: Academic Press. 688 pp.

Schmidt, M. W., M. S. Torn, S. Abiven, T. Dittmar, G. Guggenberger, I. A. Janssens, M.

Kleber, I. Kögel-Knabner, J. Lehmann, D. A. Manning, and others. 2011.

Persistence of soil organic matter as an ecosystem property. Nature 478: 49–56.

209

Schrumpf, M., K. Kaiser, G. Guggenberger, T. Persson, I. Kögel-Knabner, I., and E.D.

Schulze. 2013. Storage and stability of organic carbon in soils as related to depth,

occlusion within aggregates, and attachment to minerals. Biogeosciences 10:

1,675–1,691.

Sickman, J. O., C. L. DiGiorgio, M. L. Davisson, D. M. Lucero, and B. Bergamaschi.

2010. Identifying sources of dissolved organic carbon in agriculturally dominated

rivers using radiocarbon age dating: Sacramento–San Joaquin River Basin,

California. Biogeochemistry 99: 79–96.

Siegel, D. I., N. A. Azzolina, B. J. Smith, A. E. Perry, and R. L. Bothun. 2015. Methane

concentrations in water wells unrelated to proximity to existing oil and gas wells

in northeastern Pennsylvania. Environmental science & technology 49: 4,106–

4,112.

Sitters, J., C. L. Atkinson, N. Guelzow, P. Kelly, and L. L. Sullivan. 2015. Spatial

stoichiometry: cross-ecosystem material flows and their impact on recipient

ecosystems and organisms. Oikos 124: 920-930.

Smits, A. P., D. E. Schindler, and M. T. Brett. 2015. Geomorphology controls the trophic

base of stream food webs in a boreal watershed. Ecology 96:1,775–1,782.

Søballe, D. M., and B. L. Kimmel. 1987. A large-scale comparison of factors influencing

phytoplankton abundance in rivers, lakes, and impoundments. Ecology 68: 1,943–

1,954.

210

Sobczak, W. V., and P. A. Raymond. 2015. Watershed hydrology and dissolved organic

matter export across time scales: minute to millennium. Freshwater Science 34:

392–398.

Solomon, C. T., Carpenter, S. R., Clayton, M. K., and others. 2011. Terrestrial, benthic,

and pelagic resource use in lakes: results from a three-isotope Bayesian mixing

model. Ecology 92: 1,115–1,125.

Solomon, C. T., J. J. Cole, R. R. Doucett, M. L. Pace, N. D. Preston, L. E. Smith, and B.

C. Weidel. 2009. The influence of environmental water on the hydrogen stable

isotope ratio in aquatic consumers. Oecologia 161: 313–324.

Solomon, Hilary. 2002. Paint Creek Watershed Management Plan.

http://water.ohiodnr.gov/portals/soilwater/downloads/wap/PaintCr.pdf

Soto, D. X., L. I. Wassenaar, and K. A. Hobson. 2013. Stable hydrogen and oxygen

isotopes in aquatic food webs are tracers of diet and provenance. Functional

Ecology.

Spencer, R. G., P.J. Hernes, A. K. Aufdenkampe, and others. 2012. An initial

investigation into the organic matter biogeochemistry of the Congo River.

Geochimica et Cosmochimica Acta 84: 614–627.

Spencer, R. G., W. Guo, P. A. Raymond, T. Dittmar, E. Hood, J. Fellman., and A.

Stubbins. 2014 (a). Source and biolability of ancient dissolved organic matter in

glacier and lake ecosystems on the Tibetan Plateau. Geochimica et Cosmochimica

Acta 142:64–74.

211

Spencer, R. G., A. Vermilyea, J. Fellman, P. Raymond, A. Stubbins, D. Scott, and E.

Hood. 2014 (b). Seasonal variability of organic matter composition in an Alaskan

glacier outflow: insights into glacier carbon sources. Environmental Research

Letters 9: doi:10.1088/1748-9326/9/5/055005.

Spencer, R. G., P. J. Mann, T. Dittmar, and others. 2015. Detecting the signature of

permafrost thaw in Arctic rivers. Geophysical Research Letters 42: 2,830–2,835.

Strayer, D. L., R. E. Beighley, L. C. Thompson, S. Brooks, C. Nilsson, G. Pinay, and R.

J. Naiman. 2003. Effects of land cover on stream ecosystems: roles of empirical

models and scaling issues. Ecosystems 6: 407–423.

Stevenson, R. J., M. L. Bothwell, R. L. Lowe, and J. H. Thorp. 1996. Algal ecology:

Freshwater benthic ecosystems. Academic Press, San Diego, CA.

Stuiver, M., and H. A. Polach. 1977. Discussion; reporting of C-14 data. Radiocarbon 19:

355–363.

Sullivan, S. M. P. 2013. Stream foodweb δ13C and geomorphology are tightly coupled in

mountain drainages of northern Idaho. Freshwater Science 32: 606–621.

Syvitski, J. P., C. J. Vörösmarty, A. J. Kettner, and P. Green. 2005. Impact of humans on

the flux of terrestrial sediment to the global coastal ocean. Science 308: 376–380.

Taipale, S., P. Kankaala, and R. I. Jones. 2007. Contributions of different organic carbon

sources to Daphnia in the pelagic foodweb of a small polyhumic lake: results

from mesocosm DI13C-additions. Ecosystems 10: 757–772.

212

Taipale, S., U. Strandberg, E. Peltomaa, A. W. Galloway, A. Ojala, and M. T. Brett.

2013. Fatty acid composition as biomarkers of freshwater microalgae: analysis of

37 strains of microalgae in 22 genera and in seven classes. Aquatic Microbial

Ecology 71: 165-178.

Takagi, A., C. Mizota, J. Okano, S. Nakano, E. Kikuchi, and others. 2006. Contribution

of chemoautotrophic production to freshwater macroinvertebrates in a headwater

stream using multiple stable isotopes. International Review of Hydrobiology 91:

501–508.

Tank, J. L., E. J. Rosi-Marshall, N. A. Griffiths, S. A. Entrekin, and M. L. Stephen. 2010.

A review of allochthonous organic matter dynamics and metabolism in streams.

Journal of the North American Benthological Society 29: 118–146.

Taylor, R. E. 2016. Radiocarbon Dating: Development of a Nobel Method. In E. A. G.

Schurr, E. R. M. Druffel, & S. E. Trumbore (Eds.), Radiocarbon and Climate

Change (pp. 21–44). Cham, Switzerland: Springer International. 315 pp.

Thompson, R. M., and C. R. Townsend. 2005. Energy availability, spatial heterogeneity

and ecosystem size predict food-web structure in streams. Oikos 108: 137–148.

Thornberry-Ehrlich, T. L. 2010. Geologic Resources Inventory Scoping Summary:

Hopewell Culture National Historical Park, Ohio. National Park Service,

Geological Resources Division. https://www.nature.nps.gov/geology/

inventory/publications/s_summaries/HOCU_Scoping_Summary_2010-0128.pdf

213

Thorp, J. H., and M. D. Delong. 1994. The riverine productivity model: an heuristic view

of carbon sources and organic processing in large river ecosystems. Oikos 70:

305–308.

Thorp, J. H., and M. D. Delong. 2002. Dominance of Autochthonous Autotrophic Carbon

in Food Webs of Heterotrophic Rivers. Oikos 96: 543–550.

Tittel, J., O. Büttner, K. Freier, A. Heiser, R. Sudbrack, R., and G. Ollesch. 2013. The age

of terrestrial carbon export and rainfall intensity in a temperate river headwater

system. Biogeochemistry 115: 53–63.

Torres-Ruiz, M., J. D. Wehr, and A. A. Perrone. 2007. Trophic relations in a stream food

web: importance of fatty acids for macroinvertebrate consumers. Journal of the

North American Benthological Society 26: 509-522.

Tourtelot, H. A. 1979. Black shale—its deposition and diagenesis. Clays and Clay

Minerals 27: 313–321.

Tranvik, L. J. 1992. Allochthonous dissolved organic matter as an energy source for

pelagic bacteria and the concept of the microbial loop. Hydrobiologia 229: 107–

114.

Tranvik, L. J., J. A. Downing, J. B. Cotner, and others. 2009. Lakes and reservoirs as

regulators of carbon cycling and climate. Limnology and Oceanography 54:

2,298–2,314.

Trimmer, M., J. Grey, C. M. Heppell, A. G. Hildrew, K. Lansdown, H. Stahl, and G.

Yvon-Durocher. 2012. River bed carbon and nitrogen cycling: state of play and

some new directions. Science of the Total Environment 434: 143–158.

214

Trimmer, M., A. G. Hildrew, M. C. Jackson, J. L. Pretty, and J. Grey. 2009. Evidence for

the role of methane-derived carbon in a free-flowing, lowland river food web.

Limnology and oceanography 54: 1,541–1,547.

Trumbore, S. E. 1997. Potential responses of soil organic carbon to global environmental

change. Proceedings of the National Academy of Sciences 94: 8,284–8,291.

Trumbore, S. 2000. Age of soil organic matter and soil respiration: radiocarbon

constraints on belowground C dynamics. Ecological Applications 10: 399–411.

Trumbore, S. 2009. Radiocarbon and soil carbon dynamics. Annual Review of Earth and

Planetary Sciences 37:47–66.

USDA NASS, 2012 Census of Agriculture, Ag Census Web Maps. Available at:

www.agcensus.usda.gov/Publications/2012/Online_Resources/Ag_Census_Web_

Maps/Overview/.

US EPA (2015). BASINS 4.1 (Better Assessment Science Integrating point & Non-point

Sources) Modeling Framework. National Exposure Research Laboratory, RTP,

North Carolina. https://www.epa.gov/exposure-assessment-models/basins. 19

November 2016.

Vander Zanden, M. J., M. K. Clayton, E. K. Moody, C. T. Solomon, and B. C. Weidel.

2015. Stable Isotope Turnover and Half-Life in Animal Tissues: A Literature

Synthesis. PLOS ONE 10:e0116182. doi:10.1371/journal.pone.0116182.

Vander Zanden, M. J., Y. Vadeboncoeur, M. W. Diebel, and E. Jeppesen. 2005. Primary

consumer stable nitrogen isotopes as indicators of nutrient source. Environmental

Science & Technology 39: 7,509–7,515. 215

Vannote, R. L., G. W. Minshall, K. W. Cummins, J. R. Sedell, and C. E. Cushing. 1980.

The river continuum concept. Canadian Journal of Fisheries and Aquatic Sciences

37: 130–137. van Vliet, M. T., W.H. Franssen, J.R. Yearsley, F. Ludwig, I. Haddeland, D.P.

Lettenmaier, and P. Kabat. 2013. Global river discharge and water temperature

under climate change. Global Environmental Change 23: 450–464.

Vengosh, A., R. B. Jackson, N. Warner, T. H. Darrah, and A. Kondash. 2014. A critical

review of the risks to water resources from unconventional shale gas development

and hydraulic fracturing in the United States. Environmental Science &

Technology 48: 8,334–8,348.

Vitosh, M. L., J. W. Johnson, and D. B. Mengel. 1995. TH-state Fertilizer

Recommendations for Corn, Soybeans, Wheat and Alfalfa.

https://www.extension.purdue.edu/extmedia/AY/AY-9-32.pdf

Vonk, J. E., P. J. Mann, S. Davydov, and others. 2013. High biolability of ancient

permafrost carbon upon thaw. Geophysical Research Letters 40: 2,689–2,693.

Wagenhoff, A., C. R. Townsend, and C. D. Matthaei. 2012. Macroinvertebrate responses

along broad stressor gradients of deposited fine sediment and dissolved nutrients:

a stream mesocosm experiment. Journal of Applied Ecology 49:892–902.

Wallace, J. B., S. L. Eggert, J. L. Meyer, and J. R. Webster. 1997. Multiple trophic levels

of a forest stream linked to terrestrial litter inputs. Science 277: 102–104.

216

Wallace, J. B., S. L. Eggert, J. L. Meyer, and J. R. Webster. 2015. Stream invertebrate

productivity linked to forest subsidies: 37 stream-years of reference and

experimental data. Ecology 96: 1,213-1,228.

Walters, D. M., K. M. Fritz, and D. L. Phillips. 2007. Reach-scale geomorphology affects

organic matter and consumer δ13C in a forested Piedmont stream. Freshwater

Biology 52: 1,105–1,119.

Wang, Y., B. Gu, M. K. Lee, S. Jiang, and Y. Xu. 2014. Isotopic evidence for

anthropogenic impacts on aquatic food web dynamics and mercury cycling in a

subtropical wetland ecosystem in the US. Science of The Total Environment 487:

557–564.

Wang, X., H. Ma, R. Li, Z. Song, and J. Wu. 2012. Seasonal fluxes and source variation

of organic carbon transported by two major Chinese Rivers: The Yellow River

and Changjiang (Yangtze) River. Global Biogeochemical Cycles 26:

doi:10.1029/2011GB004130.

Warren, D. R., S. M. Collins, E. M. Purvis, M. J. Kaylor, and H. A. Bechtold. 2016.

Spatial Variability in Light Yields Colimitation of Primary Production by Both

Light and Nutrients in a Forested Stream Ecosystem. Ecosystems

doi:10.1007/s10021-016-0024-9.

Wassenaar, L. I., and K. A. Hobson. 2003. Comparative equilibration and online

technique for determination of non-exchangeable hydrogen of keratins for use in

animal migration studies. Isotopes in environmental and health studies 39: 211–

217.

217

Watanabe, K., and T. Kuwae. 2015. Radiocarbon isotopic evidence for assimilation of

atmospheric CO2 by the seagrass Zostera marina. Biogeosciences 12: 6,251–

6,258.

Webb, B. W., D. M. Hannah, R. D. Moore, L. E. Brown, and F. Nobilis. 2008. Recent

advances in stream and river temperature research. Hydrological Processes 22:

902–918.

Weber, A. M., J. E. Bauer and G. T. Watters. Assessment of nutritional subsidies to

freshwater mussels using a multiple natural abundance isotope approach.

Freshwater Biology (in press).

Wellard Kelly, H. A., E. J. Rosi-Marshall, T. A. Kennedy, R. O. Hall Jr, W. F. Cross, and

C. V. Baxter. 2013. Macroinvertebrate diets reflect tributary inputs and turbidity-

driven changes in food availability in the Colorado River downstream of Glen

Canyon Dam. Freshwater Science 32: 397–410.

Wetzel, R. G. 1990. Land-water interfaces: metabolic and limnological regulators.

Internationale Vereinigung fuer Theoretische und Angewandte Limnologie.

Verhandlunge 24: 6-24.

Whiticar, M. J., E. Faber, and M. Schoell. 1986. Biogenic methane formation in marine

and freshwater environments: CO2 reduction vs. acetate fermentation—Isotope

evidence. Geochimica et Cosmochimica Acta 50: 693–709.

Wilkinson, G. M., M. L. Pace, and J. J.Cole. 2013. Terrestrial dominance of organic

matter in north temperate lakes. Global Biogeochemical Cycles 27: 43-51.

218

Wilkinson, G. M., J. J. Cole, and M. L. Pace. 2015. Deuterium as a food source tracer:

Sensitivity to environmental water, lipid content, and hydrogen exchange.

Limnology and Oceanography Methods 13: 213–223.

Williams, C. J., Y. Yamashita, H. F. Wilson, R. Jaffé, M. A. Xenopoulos, and others.

2010. Unraveling the role of land use and microbial activity in shaping dissolved

organic matter characteristics in stream ecosystems. Limnology and

Oceanography 55: 1,159-1,171.

Williamson, C. E., W. Dodds, T. K. Kratz, and M. A. Palmer. 2008. Lakes and streams as

sentinels of environmental change in terrestrial and atmospheric processes.

Frontiers in Ecology and the Environment 6: 247–254.

Wilson, H. F., and M. A. Xenopoulos. 2009. Effects of agricultural land use on the

composition of fluvial dissolved organic matter. Nature Geoscience 2: 37–41.

Wozniak, A. S., J. E. Bauer, and R. M. Dickhut. 2011. Fossil and contemporary aerosol

particulate organic carbon in the eastern United States: Implications for deposition

and inputs to watersheds. Global Biogeochemical Cycles 25:

doi:10.1029/2010GB003855.

Wozniak, A. S., J. E. Bauer, R. M. Dickhut, L. Xu, and A. P. McNichol. 2012 (a).

Isotopic characterization of aerosol organic carbon components over the eastern

United States. Journal of Geophysical Research: Atmospheres, 117:

doi:10.1029/2011JD017153.

219

Wozniak, A. S., J. E. Bauer, and R. M. Dickhut. 2012 (b). Characteristics of water-

soluble organic carbon associated with aerosol particles in the eastern United

States. Atmospheric Environment 46: 181–188.

Wu, Y., T. Eglinton, L. Yang, B. Deng, D. Montlucon, and J. Zhang. 2013. Spatial

variability in the abundance, composition, and age of organic matter in surficial

sediments of the East China Sea. Journal of Geophysical Research:

Biogeosciences 118: 1,495–1,507.

Xu, B., J. Cao, J. Hansen, and others. 2009. Black soot and the survival of Tibetan

glaciers. Proceedings of the National Academy of Sciences 106: 22,114–22,118.

Yule, C. M., L. Boyero, and R. Marchant. 2010. Effects of sediment pollution on food

webs in a tropical river (Borneo, Indonesia). Marine and Freshwater Research 61:

204–213.

Zaimes, G. N., R. C. Schultz, and T. M. Isenhart. 2004. Stream bank erosion adjacent to

riparian forest buffers, row-crop fields, and continuously-grazed pastures along

Bear Creek in central Iowa. Journal of Soil and Water Conservation 59: 19–27.

Zeug, S. C., and K. O. Winemiller. 2008. Evidence supporting the importance of

terrestrial carbon in a large-river food web. Ecology 89: 1,733–1,743.

Zhou, T., J. Wu, and S. Peng. 2012. Assessing the effects of landscape pattern on river

water quality at multiple scales: a case study of the Dongjiang River watershed,

China. Ecological Indicators 23: 166–175.

220

Zhu, Y., A. Vieth-Hillebrand, F. D. Wilke, and B. Horsfield. 2015. Characterization of

water-soluble organic compounds released from black shales and coals.

International Journal of Coal Geology 150: 265–275.

Zigah, P. K., E. C. Minor, J. P. Werne, and S. L. McCallister. 2011. Radiocarbon and

stable carbon isotopic insights into provenance and cycling of carbon in Lake

Superior. Limnology and Oceanography 56: 867–886.

Zigah, P. K., E.C. Minor, J. P. Werne, and S. L. McCallister. 2012 (a). An isotopic (Δ14C,

δ13C, and δ15N) investigation of the composition of particulate organic matter and

zooplankton food sources in Lake Superior and across a size-gradient of aquatic

systems. Biogeosciences 9: 3,663–3,678.

Zigah, P. K., E. C, Minor, and J. P. Werne. 2012 (b). Radiocarbon and stable-isotope

geochemistry of organic and inorganic carbon in Lake Superior. Global

Biogeochemical Cycles 26: doi: 10.1029/2011GB004132.

Zigah, P. K., E. C. Minor, H. A. Abdulla, J. P. Werne, and P.G. Hatcher. 2014. An

investigation of size-fractionated organic matter from Lake Superior and a

tributary stream using radiocarbon, stable isotopes and NMR. Geochimica et

Cosmochimica Acta 127: 264–284.

Zeug, S. C., and K. O. Winemiller. 2008. Evidence supporting the importance of

terrestrial carbon in a large-river food web. Ecology 89: 1,733–1,743.

221

Appendix A: Supplementary Information for Chapter 1

222

Section 1

Table A.1. Genera and functional feeding groups of macroinvertebrate consumers identified from the six Mohawk-Hudson subwatershed study sites. Site abbreviations are as follows: Schoharie Tributary- ST, Otsquago Creek- OC, Wampecack Creek- WC, Fly Branch- FB, Nowadega Creek- NC, Couch Hollow Branch- CH. Genera Functional feeding group Sample collection sites Ceratopsyche* Filtering collector ST, OC, WC, FB, NC, CH Ephemerella* Collector gatherer ST, NC Habrophebiodes* Scraper; collector gatherer ST Psephenus* Scraper ST, WC, FB Wormaldia* Filtering collector ST, FB, CH Amphinemura* Shredder ST Alloperla* Predator ST Procleon* Collector gatherer ST, OC Rhyacophila* Predator ST, FB Agnetina Predator ST, FB, CH Epeorus* Scraper ST, CH Hexatoma* Predator ST, FB Pteronarcys Shredder ST Boyeria Predator ST, NC Nixe* Scraper OC Chironomidae (family)* Multiple OC, WC, FB Simuliidae (family)* Filtering collector OC Macrostemum Filtering collector OC Lype Scraper OC Muscidae (family) Predator OC Anacroneuria Predator OC, WC Ancyronyx* Collector gatherer WC Maccaffertium* Scraper WC, FB Isonychia* Filtering collector WC, FB Drunella* Scraper FB Pycnopsyche Shredder FB Ephemera Collector gatherer FB, NC Gammaridae (family)* Collector gatherer FB, NC Nigronia Predator FB Hetaerina Predator FB, CH Stylogomphus Predator FB Caenis* Collector gatherer NC Laccophilus* Predator NC Hydrophilus Predator NC Corixidae (family)* Piercer NC Argia Predator NC Neoperla Predator CH Acroneuria Predator CH Beloneuria Predator CH Cordulegaster Predator CH Eccoptura* Predator CH *denotes when individuals were pooled to increase sample size for isotopic analyses.

223

Section 2

Potential nutritional sources used in the Bayesian isotopic model included aquatic algae, terrestrial vegetation, and soil. Soil-derived OM appeared to isotopically constrain the organisms better than aquatic sediment OM in two-dimensional space. We recognize that while sediment-derived OM could be a limited nutritional resource for macroinvertebrates, sediments in small streams and rivers are often dominated by OM derived from soils (i.e., non-living OM of varying ages derived initially from terrestrial vegetation) and recently living terrestrial vegetation (McConnachie and Petticrew 2006,

Grunsky et al. 2009). Additionally, microbial alteration and processing of this terrestrial

OM in soil generates distinct isotopic signatures for terrestrial vegetation and soil (Fogel and Tuross 1999, Hoefs 2008). Therefore, we treated each as a separate source in our isotopic mixing models. We chose to use bulk soil δ2H values in our mixing models, but bulk soil δ2H data should be interpreted with caution because it may be derived from both

OM and hydrated minerals (Ruppenthal et al. 2010, 2015). Low OM content in both soils and sediments can prevent the measurement of the true δ2H value of the OM fraction due to higher relative contributions from hydrated minerals (Ruppenthal et al. 2010, 2015).

Our terrestrial vegetation δ2H values were more 2H-depleted compared to soil, suggesting that there was some influence from mineral water 2H.

We ran mixing models two ways: one using all three sources (algae, terrestrial vegetation, and soil) and one using only two sources (algae and terrestrial vegetation).

We then compared estimates of autochthony for each FFG between the three- and two- source models. We found that for 85% of the model runs, there was less than a 10%

224 difference in estimates of autochthony between the three- and two-source models (Figure

A.1A, B). Using least squares regression for the data in Figure A.1A yielded an r2-value of 0.82, a slope of 0.69, and a p-value < 0.001. This comparison between the three and two-source model runs would suggest that while the data comparison did not yield a 1:1 relationship, there was little difference in the measure of autochthony using the three and two-source models. The use of soil OM as a potential nutritional resource may have underestimated contributions of terrestrial vegetation to macroinvertebrate biomass.

Trophic fractionation values for δ13C and δ15N (0.4 ± 1.3‰ and 3.4 ± 1.0‰, respectively) were taken from literature (Post 2002). Trophic fractionation for δ2H was assumed to be 0‰ as previous studies have found trophic fractionation of 2H to be small or negligible (Doucett et al. 2007, Solomon et al. 2009). Dietary water may also influence organism δ2H values because a portion of 2H in organism tissues is derived from the aqueous environment. The fraction of dietary water (ω) in macroinvertebrate primary consumer biomass and predators was assumed to be 0.20 and 0.36, respectively, based on field estimates and a recent literature review (Solomon et al. 2009, Wilkinson et al.

2015). Macroinvertebrate δ2H values were corrected using the appropriate value of ω and

2 2 the δ H-H2O values measured for stream waters of our study sites. Macroinvertebrate δ H was not corrected for lipids.

Small sample sizes prohibited a separate model to be run for each site. Models would not converge and uncertainty around proportional contribution estimates was large. Therefore, the stable isotope data (δ13C, δ15N, and δ2H) for organisms and their potential nutritional resources were treated as follows: Method 1- isotope data for

225 samples were pooled across all sites, and isotope averages of each FFG from each site were used; Method 2- individual organisms and nutritional sources were grouped by amount of agriculture (high vs. low) in the site’s watershed; and Method 3- ∆14C data was included, and individuals and nutritional sources were grouped by the presence of high-

OM shale in the watershed vs. no shale in the watershed. The third approach was used only for Fly Branch and Nowadega Creek, the only two sites for which ∆14C measurements were made. Using Markov chain Monte Carlo (MCMC) methods, models were initially run with 50,000 MCMC chains (25,000 burn-in) to ensure that the model parameters were specified correctly, and if necessary, models were run again with

300,000 (200,000 burn-in) or 1,000,000 (500,000 burn-in) MCMC chains in order for models to converge. Gelman-Rubin and Geweke diagnostics were evaluated in order to ensure that models had converged.

For the third mixing model approach, samples sizes were low because we only used stable isotope (δ13C, δ15N, and δ2H) and radiocarbon (∆14C) data from two of the six sites that we sampled. We also estimated the ∆14C values for terrestrial vegetation from

14 atmospheric CO2 ∆ C values (Levin et al. 2013) rather than directly analyzing them, as terrestrial vegetation fixes atmospheric CO2. This approach has been validated by other studies (Randerson et al. 2002, Pataki et al. 2010). We only analyzed two terrestrial vegetation and soil samples from both Nowadega Creek and Fly Branch. The average

δ13C of terrestrial vegetation for each site (Nowadega Creek: -29.3‰ and Fly Branch: -

29.9‰) closely matched the average across all sites (total of 12 terrestrial vegetation samples: -29.0‰), suggesting that even though our sample sizes were low for the model

226 runs using 14C data, the values that we used accurately represented the δ13C values of terrestrial vegetation in the region that we sampled. The δ13C values of our terrestrial vegetation samples were also consistent with values previously reported for Betula, Tilia, and Acer spp. (Garten and Taylor 1992, Balesdent et al. 1993, Diefendorf et al. 2010).

δ15N values of terrestrial vegetation were more variable for Nowadega Creek (4.1

± 3.4‰) and Fly Branch (2.7 ± 0.4‰) compared to δ15N values across all six sites (2.5 ±

2.9‰). Riparian plants obtain nitrogen from a number of sources including stream waters, the surrounding landscape, and the atmosphere; therefore, δ15N values for riparian vegetation can be highly variable even across small spatial scales (Hall et al. 2015).

Therefore, we determined it was more appropriate to use the data from the small sample sizes that we had for the mixing models for Nowadega Creek and Fly Branch, rather than using data from the literature. Our measured soil stable and radiocarbon isotopic values for both Nowadega Creek and Fly Branch also agreed with values previously presented in literature (Pessenda et al. 1996, Ehleringer et al. 2000, Bowling et al. 2008, Jenkinson et al. 2008, Hossler and Bauer 2012).

Algal samples (which primarily consisted of the filamentous algae, Cladophora spp.) were only present in large enough quantities for collection and isotopic analyses for four of the six sites (Nowadega Creek, Fly Branch, Wampecack Creek, and Otsquago

Creek). We determined that estimated algal isotope values should be used for all sites in the mixing models, rather than using estimated algal isotope values for two sites and

13 actual algal isotope values for the other four sites. δ C values of CO2(aq) were estimated using the measured δ13C values of DIC, pH, and water temperature at each sampling site

227

13 according to Mook et al. (1974). The fractionation of δ C-CO2 by algae was determined

15 using the calculated aqueous concentration of CO2 according to Finlay (2004). δ N values of algae were estimated using the δ15N of POM and terrestrial vegetation according to Francis et al. (2011). Algal δ2H values were estimated from the measured

δ2H of water for each site, and the fractionation of water by algae according to Hondula et al. (2014). Algal ∆14C values were assumed to be the same as ∆14C values of DIC as the ∆14C of algae is the same as the ∆14C of the inorganic C fixed by them (Broecker and

Walton 1959, Caraco et al. 2010).

A Bayesian isotopic mixing model was run for the four sites that had algae present using the measured algal isotope values and using the estimated algal isotope values following Method 1 above, in order to ensure that source contribution estimates from model runs using algal isotope estimates were comparable to those model runs using actual algal isotope values (Table A.2). Estimates of the proportional contributions of nutritional sources to each FFG followed similar patterns using actual and estimated algal isotope values (i.e., the importance of each nutritional resource did not change relative to the other nutritional resources). The maximum differences in median proportional contributions estimated from each model for algae, soil, and terrestrial vegetation were 26%, 10%, and 15% respectively. Terrestrial vegetation appeared to contribute more to macroinvertebrate diet when actual algal isotope values were used

(Tables A.3 A, B, Figure A.2), but this is likely an artifact due to the difficulty associated with collecting a pure, algal sample in the field (i.e., it is probable that small amounts of detritus or sediment were collected and processed along with the algae). Therefore, it was

228 assumed that the estimated algal isotope values were appropriate for use in Bayesian mixing model runs. Isotopic values used in each mixing model scenario are provided in

Tables A.4-A.8.

Table A.2. Stable isotope data used in mixing model to examine the effect of using estimated algal isotope values vs. measured algal isotope values. See Appendix A Section 2 for details. Nutritional source δ15N ± S.D. (‰) δ13C ± S. D. (‰) δ2H ± S.D. (‰) n Algae – estimated 5.5 ± 0.7 -31.4 ± 1.7 -249.2 ± 3.1 4 Algae – measured 5.7 ± 0.9 -31.1 ± 4.3 -290.7 ± 16.9 8 Terrestrial vegetation 2.5 ± 2.0 -28.9 ± 1.1 -134.8 ± 15.7 8 Soil OM 3.2 ± 1.1 -24.6 ± 2.2 -112.0 ± 5.8 6

229

Table A.3. Mixing model nutritional source contribution estimates (5%, 50%, and 95% posterior probabilities) using A) estimated algal isotope values and B) measured algal isotope values. See Appendix A Section 2 for details. A) FFG and nutritional source 5% 50% 95% Chironomid/algae 0.135 0.628 0.926 Chironomid/soil OM 0.001 0.144 0.56 Chironomid/terrestrial vegetation 0.001 0.115 0.762 Collector gatherer/algae 0.137 0.667 0.944 Collector gatherer/soil OM 0.003 0.173 0.691 Collector gatherer/terrestrial vegetation 0.000 0.069 0.575 Filtering collector/algae 0.659 0.911 0.999 Filtering collector/soil OM 0.000 0.029 0.196 Filtering collector/terrestrial vegetation 0.000 0.032 0.220 Scraper/algae 0.599 0.899 0.999 Scraper/soil OM 0.000 0.030 0.219 Scraper/terrestrial vegetation 0.000 0.038 0.300 Predator/algae 0.067 0.473 0.884 Predator/soil OM 0.014 0.198 0.682 Predator/terrestrial vegetation 0.014 0.226 0.708

B) FFG and nutritional source 5% 50% 95% Chironomid/algae 0.274 0.480 0.672 Chironomid/soil OM 0.025 0.219 0.527 Chironomid/terrestrial vegetation 0.032 0.266 0.612 Collector gatherer/algae 0.296 0.515 0.720 Collector gatherer/soil OM 0.030 0.225 0.534 Collector gatherer/terrestrial vegetation 0.025 0.220 0.554 Filtering collector/algae 0.491 0.655 0.816 Filtering collector/soil OM 0.017 0.137 0.344 Filtering collector/terrestrial vegetation 0.024 0.178 0.407 Scraper/algae 0.462 0.644 0.814 Scraper/soil OM 0.014 0.130 0.339 Scraper/terrestrial vegetation 0.028 0.192 0.443 Predator/algae 0.104 0.540 0.766 Predator/soil OM 0.012 0.154 0.536 Predator/terrestrial vegetation 0.035 0.276 0.679

230

Table A.4. Isotope data used in mixing model with FFG averages collected from all sites. See Appendix A Section 2 for details. Nutritional source δ15N ± S.D. (‰) δ13C ± S. D. (‰) δ2H ± S.D. (‰) n Algae-estimated 4.8 ± 1.9 -31.1 ± 1.8 -249.1 ± 2.5 6 Terrestrial vegetation 2.5 ± 2.9 -29.0 ± 1.1 -135.4 ± 13.4 12 Soil OM 3.2 ± 1.2 -25.0 ± 1.9 -110.9 ± 6.1 9 Sediment OM* 3.3 ± 1.8 -23.2 ± 6.8 -112.2 ± 9.1 6 *Sediment OM was not included in the mixing model but is shown here to demonstrate isotopic overlap between soil and sediment.

Table A.5. Isotope data used in mixing model for streams from high agriculture watersheds. See Appendix A Section 2 for details. Nutritional source δ15N ± S.D. (‰) δ13C ± S. D. (‰) δ2H ± S.D. (‰) n Algae-estimated 5.5 ± 0.7 -31.4 ± 1.7 -249.2 ± 3.1 4 Terrestrial vegetation 2.5 ± 2.0 -28.9 ± 1.1 -134.8 ± 15.7 8 Soil OM 3.2 ± 1.1 -24.6 ± 2.2 -112.0 ± 5.8 6

Table A.6. Isotope data used in mixing model for streams from low agriculture watersheds. See Appendix A Section 2 for details. Nutritional source δ15N ± S.D. (‰) δ13C ± S. D. (‰) δ2H ± S.D. (‰) n Algae-estimated 3.3 ± 3.1 -30.6 ± 2.7 -248.8 ± 1.4 2 Terrestrial vegetation 2.7 ± 4.6 -29.2 ± 1.3 -137.0 ± 5.8 4 Soil OM 3.1 ± 1.7 -25.9 ± 0.7 -109.1 ± 7.0 2

Table A.7. Isotope data used in mixing model for Fly Branch (shale absent from watershed). See Appendix A Section 2 for details. Nutritional source δ15N ± S.D. (‰) δ13C ± S. D. (‰) δ2H ± S.D. (‰) ∆14C ± S.D. (‰) n Algae-estimated 6.3 ± 0 -33.7 ± 0.4 -247.3 ± 0.4 -72 ± 5 2 Terrestrial vegetation 2.7 ± 0.4 -29.9 ± 0.4 -130.2 ± 3.7 39 ± 2 2 Soil OM 4.2 ± 0.3 -24.1 ± 1.5 -109.8 ± 2.4 -722 ± 94 2

Table A.8. Isotope data used in mixing model for Nowadega Creek (shale present in watershed). See Appendix A Section 2 for details. Nutritional source δ15N ± S.D. (‰) δ13C ± S. D. (‰) δ2H ± S.D. (‰) ∆14C ± S.D. n (‰) Algae-estimated 5.3 ± 0.3 -31.1 ± 0 -248.7 ± 1.8 -137 ± 5 2 Terrestrial vegetation 4.1 ± 3.4 -29.3 ± 0.8 -152.1 ± 11.8 39 ± 2 2 Soil OM 3.1 ± 1.1 -26.8 ± 0 -108.2 ± 3.4 -261 ± 223 2

231

232

100 100

80 A 80 B

60 60

40 40

20 20

Percent Contribution Percent

0 0

scraper scraper predator predator Algae chironomid chironomid Ter Veg filtering collector collector gathererfiltering collector collector gatherer Soil

233

Section 3

Table A.9. Means and standard deviations of isotopic values for macroinvertebrates grouped by functional feeding group (FFG) from the streams sampled from the six Mohawk-Hudson subwatersheds. Isotopic values presented here are raw data uncorrected for trophic fractionation. See main text Results section “δ13C, δ15N, δ2H, and ∆14C Values of Macroinvertebrate Consumers” for details. FFG (n) δ13C (‰) δ15N (‰) δ2H (‰) ∆14C (‰)* Nowadega Creek Filtering collector (3) -28.5 ± 0.4 10.4 ± 0.3 -238 ± 4.2 -114 ± 0 Collector gatherer (4) -27.1 ± 0.9 8.7 ± 0.3 -202.9 ± 30.1 -81 ± 7 Hydrophilus spp (2) -33.1 ± 2.6 7.4 ± 8.4 -156.0 ± 8.0 -0.5 Predator (4) -26.8 ± 1.1 10.3 ± 0.2 -177.1 ± 15.3 -83 ± 0 Fly Branch Filtering collector (5) -29.6 ± 0.9 7.4 ± 0.5 -197.6 ± 18.3 -44 ± 4 Collector gatherer (6) -27.2 ± 0.3 6.9 ± 1 -160.2 ± 5.6 -28 ± 3 Shredder (1) -27.1 4.8 -132.3 23 Chironomidae (2) -27.4 ± 0.1 5.8 ± 0.2 -154.2 ± 3.6 -26 Scraper (3) -28.8 ± 1.3 6.4 ± 0.7 -194.9 ± 30.2 -43 ± 17 Predator (10) -29.5 ± 1.4 8.7 ± 0.6 -194.3 ± -22.0 -57 ± 4 Schoharie Tributary Filtering collector (2) -28.3 ± 0.7 4.4 ± 0.7 -212.5 ± 12.2 ND Collector gatherer (1) -26.4 2.5 -222.6 ND Scraper -33.8 ± 0.3 2.9 ± 0.3 -242.8 ± 0.1 ND Shredder (1) -28.9 3.1 -183.3 ND Predator (8) -28.1 ± 1.1 5.9 ± 0.7 -191.4 ± 24.6 ND Otsquago Creek Scraper (2) -30.1 ± 0.3 6.9 ± 0.02 -234.2 ± 3.7 ND Filtering Collector (2) -30.8 ± 0.4 9.3 ± 0.4 -241.6 ± 1.3 ND Chironomidae (2) -27.5 ± 2.2 8.6 ± 0.8 -213.4 ± 26.3 ND Wampecack Creek Scraper (1) -32.6 8.5 -241.2 ND Chironomidae (1) -27.6 9.1 -166.7 ND Filtering collector (4) -30.4 ± 0.4 9.0 ± 0.3 -206.1 ± 6.5 ND Predator (1) -31.2 10.2 -215.3 ND Couch Hollow Branch Scraper (2) -32.6 ± 0.8 6.2 ± 0.6 -198.6 ± 6.5 ND Filtering collector (2) -30.0 ± 1.0 6.5 ± 0.6 -178.4 ± 34.1 ND Predator (5) -30.1 ± 0.9 7.8 ± 0.7 -169.3 ± 14.4 ND *Sample replicate numbers (n) for ∆14C measurements do not match those for stable isotope measurements presented in FFG column and are as follows: Nowadega Creek: filtering collectors: n=3, collector gatherers: n=3, predators: n=2; Fly Branch: filtering collectors: n=2, collector gatherers: n=4, shredders: n=1, scrapers: n=2, predators: n=2.

234

Table A.10. ANOSIM results of δ13C, δ15N, and δ2H values of A) macroinvertebrates from each site and B) macroinvertebrate FFGs pooled across sites. See main text Results section “δ13C, δ15N, δ2H, and ∆14C Values of Macroinvertebrate Consumers” for details. A) ANOSIM R: 0.369, p = 0.001 Pairwise Test R Statistica Significance Levelb Fly Branch, Nowadega Creek 0.37 0.001 Fly Branch, Schoharie Tributary 0.304 0.001 Fly Branch, Otsquago Creek 0.353 0.001 Fly Branch, Wampecack Creek 0.172 0.025 Fly Branch, Couch Hollow 0.163 0.03 Nowadega Creek, Schoharie Tributary 0.566 0.001 Nowadega Creek, Otsquago Creek 0.348 0.014 Nowadega Creek, Wampecack Creek 0.394 0.006 Nowadega Creek, Couch Hollow 0.74 0.001 Schoharie Tributary, Otsquago Creek 0.338 0.008 Schoharie Tributary, Wampecack 0.492 0.001 Creek Schoharie Tributary, Couch Hollow 0.345 0.001 Otsquago Creek, Wampecack Creek 0.271 0.022 Otsquago Creek, Couch Hollow 0.667 0.001 Wampecack Creek, Couch Hollow 0.431 0.003

B) ANOSIM R: 0.206, p = 0.001 Pairwise Test R Statistica Significance Levelb Filtering collector, chironomid 0.347 0.011 Filtering collector, scraper 0.212 0.012 Filtering collector, collector-gatherer 0.387 0.001 Filtering collector, predator 0.098 0.027 Chironomid, scraper 0.382 0.011 Chironomid, collector-gatherer -0.021 0.49 Chironomid, predator 0.093 0.207 Scraper, collector-gather 0.474 0.002 Scaper, predator 0.336 0.001 Collector-gatherer, predator 0.081 0.172

235

Section 4

Table A.11. Mixing model source contribution estimates (5%, 50%, and 95% posterior probabilities) for FFG averages across all sites. See main text Results section “Estimation of Source Contributions to Macroinvertebrate Nutrition” for details. FFG and nutritional source 5% 50% 95% Chironomid/algae 0.288 0.626 0.891 Chironomid/soil OM 0.005 0.205 0.538 Chironomid/terrestrial vegetation 0.001 0.082 0.555 Collector gatherer/algae 0.404 0.733 0.936 Collector gatherer/soil OM 0.006 0.159 0.474 Collector gatherer/terrestrial vegetation 0.000 0.047 0.369 Filtering collector/algae 0.689 0.855 0.981 Filtering collector/soil OM 0.000 0.059 0.216 Filtering collector/terrestrial vegetation 0.000 0.049 0.239 Scraper/algae 0.769 0.922 0.999 Scraper/soil OM 0.000 0.027 0.148 Scraper/terrestrial vegetation 0.000 0.027 0.162 Predator/algae 0.321 0.738 0.903 Predator/soil OM 0.007 0.099 0.416 Predator/terrestrial vegetation 0.009 0.130 0.494

Table A.12. Mixing model source contribution estimates (5%, 50%, and 95% posterior probabilities) for FFGs from high agriculture sites. See main text Results section “Estimation of Source Contributions to Macroinvertebrate Nutrition” for details. FFG and nutritional source 5% 50% 95% Chironomid/algae 0.370 0.590 0.788 Chironomid/soil OM 0.037 0.239 0.477 Chironomid/terrestrial vegetation 0.004 0.132 0.455 Collector gatherer/algae 0.457 0.613 0.779 Collector gatherer/soil OM 0.128 0.305 0.480 Collector gatherer/terrestrial vegetation 0.001 0.061 0.237 Filtering collector/algae 0.788 0.880 0.952 Filtering collector/soil OM 0.001 0.043 0.128 Filtering collector/terrestrial vegetation 0.005 0.064 0.171 Scraper/algae 0.552 0.757 0.909 Scraper/soil OM 0.001 0.060 0.223 Scraper/terrestrial vegetation 0.011 0.154 0.406 Predator/algae 0.702 0.811 0.902 Predator/soil OM 0.008 0.079 0.191 Predator/terrestrial vegetation 0.009 0.096 0.246

236

Table A.13. Mixing model source contribution estimates (5%, 50%, and 95% posterior probabilities) for FFGs from low agriculture sites. See main text Results section “Estimation of Source Contributions to Macroinvertebrate Nutrition” for details. FFG and nutritional source 5% 50% 95% Collector gatherer/algae 0.012 0.959 1.000 Collector gatherer/soil OM 0.000 0.004 0.572 Collector gatherer/terrestrial vegetation 0.000 0.003 0.686 Filtering collector/algae 0.245 0.775 1.000 Filtering collector/soil OM 0.000 0.054 0.446 Filtering collector/terrestrial vegetation 0.000 0.038 0.597 Scraper/algae 0.736 0.986 1.000 Scraper/soil OM 0.000 0.001 0.126 Scraper/terrestrial vegetation 0.000 0.002 0.185 Predator/algae 0.634 0.745 0.842 Predator/soil OM 0.020 0.127 0.257 Predator/terrestrial vegetation 0.011 0.115 0.282

Table A.14. Mixing model source contribution estimates (5%, 50%, and 95% posterior probabilities) for FFGs collected from Fly Branch. See main text Results section “Estimation of Source Contributions to Macroinvertebrate Nutrition” for details. FFG and nutritional source 5% 50% 95% Collector gatherer/algae 0.279 0.419 0.571 Collector gatherer/soil OM 0.004 0.028 0.052 Collector gatherer/terrestrial vegetation 0.420 0.552 0.679 Filtering collector/algae 0.470 0.722 0.848 Filtering collector/soil OM 0.000 0.004 0.040 Filtering collector/terrestrial vegetation 0.150 0.274 0.495 Predator/algae 0.046 0.386 0.834 Predator/soil OM 0.012 0.213 0.721 Predator/terrestrial vegetation 0.023 0.282 0.758

237

Table A.15. Mixing model source contribution estimates (5%, 50%, and 95% posterior probabilities) for FFGs collected from Nowadega Creek. See main text Results section “Estimation of Source Contributions to Macroinvertebrate Nutrition” for details. FFG and nutritional source 5% 50% 95% Collector gatherer/algae 0.472 0.623 0.735 Collector gatherer/soil OM 0.000 0.042 0.316 Collector gatherer/terrestrial vegetation 0.002 0.328 0.475 Filtering collector/algae 0.839 0.934 1.000 Filtering collector/soil OM 0.000 0.001 0.024 Filtering collector/terrestrial vegetation 0.000 0.059 0.153 Predator/algae 0.080 0.531 0.813 Predator/soil OM 0.007 0.129 0.640 Predator/terrestrial vegetation 0.037 0.279 0.701

238

Section 5

Streambed Substrate Composition Estimates

We used Wohlman pebble counts in the field in order to characterize the compositions of the streambeds in which we sampled. Data for each stream are provided below in Table A.16.

Estimation of Composition of Particulate Organic Matter (POM) in Nowadega Creek

We used a simple mass balance approach and a system of equations to estimate percent contributions of algae/phytoplankton, terrestrial vegetation, soil, and shale to particulate organic matter (POM) in Nowadega Creek (fossil-shale-containing watershed). We used the ∆14C and C:N values of each potential source and suspended

POM, and these values are presented in the table below (Table A5.2). The system of equations is as follows:

pA + pTV + pS = 1 (1)

14 14 14 14 pA*∆ CA + pTV*∆ CTV + pS*∆ CS = ∆ CPOM (2)

pA*C:NA + pTV*C:NTV + pS*C:NS = C:NPOM (3) where A represents algae/phytoplankton, TV represents terrestrial vegetation, and S represents an aged OM endmember, either soil or shale. ∆14C represents the ∆14C (‰) of a source or the POM, and p represents the proportion of a source. We used only one aged end member at a time for our source contribution estimates. When soil was used as the aged end member, we found that terrestrial vegetation did not contribute to POM (0%), and algae/phytoplankton and soil contributed 51% and 53%, respectively. When shale

OM was used as the aged end member, there was still 0% contribution of terrestrial

239 vegetation, and algae/phytoplankton and shale contributed 80% and 19%, respectively.

Assuming a 50:50 mixture of surface and 20-cm-depth soil did not yield a ∆14C value low enough to explain the ∆14C of the POM.

Table A.16. Characterization of streambed substrate for each stream sampled. Site % Silt/Clay % Sand % Gravel % Cobble % Boulder % Bedrock (<0.06 mm) (0.06-2 mm) (2-64 mm) (65-225 mm) (256->2048 mm) CH 7 10 66 5 2 0 FB 5 11 28 17 34 0 WC 10 10 25 27 30 0 ST 6 0 26 18 50 0 NC 12 6 37 18 8 19 OC 0 0 0 0 0 100 CH-Couch Hollow Branch, FB-Fly Branch, WC-Wampecack Creek, ST-Schoharie Tributary, NC-Nowadega Creek, OC-Otsquago Creek.

Table A.17. C:N and ∆14C values of algae/phytoplankton, terrestrial vegetation, soil, and shale used to estimate their contributions to suspended particulate organic matter (POM) composition. Source C:N ∆14C (‰) Algae/phytoplankton 6.8* -119 Terrestrial vegetation 16.4 39 Soil OM (20 cm depth) 8.6 -426 Soil OM (50:50 mix of 9.3 -261 surface and 20 cm depths) Shale OM 18** -977** Suspended POM 7.3 -288 *Value from Francis et al. 2011 **Value from Longworth et al. 2007

240

Appendix B: Supplementary Information for Chapter 2

241

Section 1

Estimates of Percent Agricultural Land Use

Percent agricultural land use (including cropland and pasture) in each watershed was estimated by first delineating the area of each stream’s watershed, and then using the

GIRAS land use/land cover dataset in the EPA BASINS environmental analysis system, version 4.1 (US EPA 2015). Our sampling sites were located in Bradford (Cowanesque

River, Crooked Creek, Lamb’s Creek, Elk Run, North Elk Run), Tioga (Towanda Creek), and Lycoming (Little Muncy Creek) counties in Pennsylvania. Of the land designated as agricultural in these counties, on average ~52% is used for cropland and 10% for pastureland. The remaining 38% of land designated as agricultural is woodland/other uses

(USDA-NASS 2012; Table B.1). There are also large concentrations of livestock (cattle, swine, and chickens) in the counties that we sampled (USDA-NASS 2012; M. Shank, personal communication).

Table B.1. Percentage of land use within farms by county for the subwatersheds of this study in the upper and middle Susquehanna River network, Pennsylvania, USA. County % Crop % Pasture % Woodland % Other Bradford 53.0 10.7 27.3 9.0 Lycoming 50.3 6.5 35.1 8.2 Tioga 53.1 13.6 26.1 7.2 All data were obtained from the USDA 2012 Census of Agriculture.

242

Table B.2. Genera and functional feeding groups identified in upper and middle Susquehanna and Chemung River watershed sites. Site abbreviations are as follows: Towanda Creek- TC, Little Muncy Creek- LMC, Crooked Creek- CrCr, Cowanesque River- CR, North Elk Run- NER, Elk Run- ER, Lamb’s Creek- LC. Genera Functional feeding group Sites collected from Glossosoma* Scraper TC, LMC Caenis* Collector-gatherer TC, CrCr, CR, NER Ephemerella* Collector-gatherer TC, LC, NER, CrCr, ER, CR Isonychia* Filtering collector TC, LMC, CR, CrCr Gammaridae (family)* Collector-gatherer TC, LMC, NER, ER Liodessus* Predator TC, CrCr, CR Cheumatopsyche* Filtering collector TC, ER, NER, LC, CR Stenacron* Scraper TC, ER, CR Maccaffertium* Scraper TC Stenonema femoratum* Scraper TC, CrCr, CR Leucrocuta* Scraper TC, CR Argia* Predator TC, CrCr, NER Chironomidae (family)* Multiple TC, ER, LMC, CR, CrCr, LC Chimarra* Filtering collector CrCr, CR, LC, ER, TC Ceratopsyche* Filtering collector TC, ER, CrCr, LMC, CR, NER, LC Lestes Predator TC, NER, CrCr, ER Hetaerina Predator TC, CrCr, LMC, NER, LC, CR Ranatra Predator TC Boyeria Predator TC, CrCr, NER, ER Drunella* Scraper ER, CrCr, LC, CR Neoperla* Predator ER, CR Simuliidae (family)* Filtering collector CrCr, CR, LC, NER Cernotina* Predator CrCr, LC Helicopsyche* Scraper CrCr Tipula Shredder CrCr, LMC, NER Hexatoma Predator CrCr, CR, LC Stenelmis* Scraper CrCr, LMC, ER,CR, TC Atherix* Predator CrCr, CR, ER, TC Optioservus* Collector-gatherer CR, TC Peltodytes* Shredder CrCr, CR, NER, ER Psephenus* Scraper CrCr, CR, ER, LMC, NER, TC Gyretes Predator CrCr Agnetina Predator CrCr, CR, TC, ER Ophiogomphus Predator CrCr Gomphus Predator CrCr, CR, ER, NER, LC continued 243

Table B.2, continued Heptagenia* Scraper LMC, LC, ER, CR Paraleptophlebia* Collector-gatherer LMC, CR Serratella* Collector-gatherer LMC, CR, NER Hydropsyche* Filtering collector LMC, CR, LC, ER, TC Promoresia Omnivore CR Agapetus* Scraper LMC Dolophilodes* Filtering collector LMC, LC Paragnetina Predator LMC Helichus* Shredder LMC Culicidae (family)* Collector-gatherer CR, NER Eccoptura Predator CR Beloneuria Predator CR, LC Agabetes Predator NER Trichocorixa* Predator NER, ER Neoporus* Predator NER Pycnopsyche Shredder LC, CrCr, CR Rhyacophila Predator LC Chauliodes Predator LC, CR Acroneuria Predator CR, LC, ER, CrCr, TC Leuctra* Shredder LC Stylogomphus Predator LC, CrCr Erpetogomphus Predator CrCr Baetis* Collector-gatherer NER, ER Corydalus Predator NER Neophylax* Scraper LC Pteronarcys Shredder LC Basiaeschna Predator TC Triaenodes* Shredder NER Strophopteryx* Scraper NER Ilybius Predator NER, CrCr Cinygmula* Scraper ER, CR Perlesta Predator CR Cordulegaster Predator LC Ameletus* Collector-gatherer LC Epeorus* Scraper LC Goera* Scraper LC Polycentropus Predator LC, CR *Denotes cases in which multiple individuals were pooled to increase sample size within a genus.

244

Section 2

Additional details about the Bayesian mixing models and how data were prepared to use in the mixing models are presented in this section. Potential nutritional sources used in the model included aquatic algae, terrestrial vegetation, and the combination of soil and sediment. There was substantial overlap in the δ13C, δ15N, and δ2H values of the soil and sediment endmembers, and therefore the isotopic values of both were pooled as one potential nutritional resource. Even though soil and sediment OM is derived from terrestrial vegetation, microbial alteration and processing generates distinct isotopic signatures for terrestrial vegetation and soil/sediment (Fogel and Tuross 1999, Hoefs

2008). Therefore, we treated each as a separate source in our isotopic mixing models.

Biofilm and stream POM were not considered as potential nutritional sources because both are mixtures of undifferentiated autochthonous and allochthonous OM, and using them in mixing models would not have provided resolution of the original starting forms of OM supporting macroinvertebrate production (Goni et al. 2013, Ishikawa et al. 2015,

Cummins 2016). Trophic fractionation values for δ13C and δ15N (0.4 ± 1.3‰ and 3.4 ±

1‰, respectively) were taken from (Post 2002).

Trophic fractionation for δ2H was assumed to be 0‰ as previous studies have found trophic fractionation of 2H to be small or negligible (Doucett et al. 2007, Solomon et al. 2009). Dietary water may also influence organism δ2H values because a portion of

2H (deuterium) in organism tissues is derived from environmental water. The fraction of dietary water (ω) for macroinvertebrate primary consumers and predators was assumed to be 0.20 and 0.36, respectively and were based on field estimates and a literature review

245

(Solomon et al. 2009, Wilkinson et al. 2015). Macroinvertebrate δ2H values were corrected using the ω value appropriate to the organism’s assumed trophic level and the

δ2H values of stream waters measured for each site in the present study.

Macroinvertebrate δ2H was measured for whole organisms and was not corrected for potential 2H-depletion associated with lipids (Soto et al. 2013, Wilkinson et al. 2015).

While we used the δ2H of bulk soil and sediments in our mixing models, these data should be interpreted with caution. Hydrogen content in soils and sediments comes from both organic matter (OM) and hydrated minerals (Ruppenthal et al. 2010, 2015). Soil and sediments have low OM content, and the δ2H measurement is not a true measure of the

2H content of the OM fraction alone (Ruppenthal et al. 2010, 2015). Our soil δ2H values were consistently more 2H-enriched compared to terrestrial vegetation, suggesting that mineral water 2H contaminated our samples. Nevertheless, the δ2H data helped to differentiate between autochthonous and allochthonous nutritional resources and were therefore useful to include in the models and statistical analyses.

In order to determine how the use of soil and sediment OM in the mixing models influenced our estimates of autochthony, we ran mixing models in two different ways: i) using all three sources (algae, terrestrial vegetation, and soil) and ii) using only two sources (algae and terrestrial vegetation). We then compared estimates of autochthony between the three- and two- sources models and found that 94% of the time the difference in estimates of percent autochthony between the three and two-source models was less than 10% (Figure B.1 A and B). Least squares regression for the data in Figure

B.1A yielded an r2-value of 0.91, and a slope of 0.9989 and a p-value < 0.0001. This

246 comparison between the three- and two-source model runs was close to a 1:1 realtionship, suggesting that there was very little difference in the measure of autochthony using the three- and two-source models. By including soil OM in our mixing models, we may have overestimated its importance as a nutritional resource, and concomitantly underestimated the contribution of terrestrial vegetation.

Mixing models were run using four different approaches. Approach 1 assessed the diet of macroinvertebrates at each site, and a separate mixing model was run for each site using stable isotope data (δ13C, δ15N, and δ2H) pooled across all sampling times in order to increase sample size. Approach 2 assessed the effect of agriculture on diet, and macroinvertebrates and their potential nutritional resource stable isotope data (δ13C, δ15N, and δ2H) were grouped according to the amount of agriculture in each site’s watershed

(high agriculture: >50% and low agriculture: <50%). Approach 3 assessed the overall global macroinvertebrate diet for the Susquehanna watershed, and isotope averages for the combined FFGs from each site (except Lamb’s Creek) were determined for a global model run, potential nutritional resource isotope values were averaged across the six sites used, and ∆14C data were included in the model along with stable isotope data. For approach 4, a separate mixing model was run for macroinvertebrates collected from Little

13 2 14 Muncy Creek using δ C, δ H, and ∆ C data, and including fossil-aged methane (CH4) as an endmember, to assess the differences in nutritional resource utilization by macroinvertebrates during different sampling times for this site. We did not include δ15N data in this model as methane does not contain nitrogen. Using Markov chain Monte

Carlo (MCMC) methods, models were initially run with 50,000 MCMC chains (25,000

247 burn-in) to ensure that the model parameters were specified correctly, and, if necessary, models were run again with 300,000 (200,000 burn-in) or 1,000,000 (500,000 burn-in)

MCMC chains in order for models to converge. Gelman-Rubin and Geweke diagnostics were evaluated in order to ensure that models had converged. Means and standard deviations (SD) of the stable isotopic values measured for algae, terrestrial vegetation, and soil & sediment OM used in each mixing model run are provided in Tables B.3-B.13.

Table B.3. Lamb’s Creek Sources δ15N ± S.D. (‰) δ13C ± S.D. (‰) δ2H ± S.D. (‰) n Algae 1.4 ± 1.7 -36.1 ± 8.1 -234.0 ± 22.8 4 Terrestrial vegetation -0.6 ± 1.0 -31.4 ± 2.5 -127.5 ± 10.0 4 Soil & sediment OM 3.0 ± 0.7 -26.6 ± 1.9 -106.1 ± 4.7 7

Table B.4. North Elk Run Sources δ15N ± S.D. (‰) δ13C ± S.D. (‰) δ2H ± S.D. (‰) n Algae 10.8 ± 2.4 -30.3 ± 4.6 -226.9 ± 34.8 3 Terrestrial vegetation 2.1 ± 0.8 -28.4 ± 2.0 -126.5 ± 17.0 5 Soil & sediment OM 4.9 ± 1.4 -27.0 ± 0.9 -108.0 ± 5.9 8

Table B.5. Little Muncy Creek Sources δ15N ± S.D. (‰) δ13C ± S.D. (‰) δ2H ± S.D. (‰) n Algae 4.0 ± 2.9 -25.1 ± 9.1 -222.4 ± 25.3 4 Terrestial vegetation -2.0 ± 0.8 -29.7 ± 1.4 -122.7 ± 24.0 2 Soil & sediment OM 2.3 ± 2.5 -27.1 ± 1.4 -104.3 ± 10.3 5

Table B.6. Elk Run Sources δ15N ± S.D. (‰) δ13C ± S.D. (‰) δ2H ± S.D. (‰) n Algae 5.2 ± 1.9 -23.8 ± 5.0 -259.8 ± 34.2 5 Terrestial vegetation 2.2 ± 0.4 -28.3 ± 0.9 -142.0 ± 20.0 3 Soil & sediment OM 3.1 ± 1.7 -26.6 ± 1.2 -106.7 ± 3.9 7

248

Table B.7. Crooked Creek Sources δ15N ± S.D. (‰) δ13C ± S.D. (‰) δ2H ± S.D. (‰) n Algae 6.6 ± 2.6 -25.6 ± 4.1 -255.3 ± 20.4 8 Terrestrial vegetation 3.7 ± 1.3 -29.8 ± 2.0 -147.9 ± 24.4 4 Soil & sediment OM 3.2 ± 1.0 -28.1 ± 1.7 -113.0 ± 4.5 8

Table B.8. Towanda Creek Sources δ15N ± S.D. (‰) δ13C ± S.D. (‰) δ2H ± S.D. (‰) n Algae 10.6 ± 1.5 -23.9 ± 4.1 -272.3 ± 13.1 7 Terrestrial vegetation 0.3 ± 1.8 -30.1 ± 1.8 -127.2 ± 9.4 6 Soil & sediment OM 3.5 ± 1.5 -26.6 ± 1.1 -103.4 ± 3.6 8

Table B.9. Cowanesque River Sources δ15N ± S.D. (‰) δ13C ± S.D. (‰) δ2H ± S.D. (‰) n Algae 6.8 ± 1.3 -27.0 ± 3.3 -296.8 ± 18.8 5 Terrestrial vegetation 2.0 ± 1.7 -29.5 ± 1.3 -143.6 ± 19.8 7 Soil & sediment OM 3.2 ± 1.8 -27.5 ± 1.4 -111.4 ± 4.5 6

Table B.10. High agriculture sites Sources δ15N ± S.D. (‰) δ13C ± S.D. (‰) δ2H ± S.D. (‰) n Algae 7.8 ± 0.9 -24.9 ± 1.5 -261.8 ± 16.8 27 Terrestial vegetation 1.9 ± 1.7 -29.3 ± 1.7 -137.7 ± 19.3 25 Soil & sediment OM 3.6 ± 1.6 -27.2 ± 1.4 -107.4 ± 5.3 37

Table B.11. Low agriculture sites Sources δ15N ± S.D. (‰) δ13C ± S.D. (‰) δ2H ± S.D. (‰) n Algae 2.9 ± 1.5 -34.1 ± 1.3 -229.8 ± 14.9 8 Terrestrial vegetation -1.1 ± 1.1 -30.8 ± 2.2 -125.9 ± 13.5 6 Soil & sediment OM 2.7 ± 1.7 -26.8 ± 1.6 -105.1 ± 7.9 12

Table B.12. Global mixing model including stable isotope and radiocarbon data Sources δ15N ± S.D. (‰) δ13C ± S.D. (‰) δ2H ± S.D. (‰) ∆14C ± S.D. (‰) n Algae 7.8 ± 1.4 -27.1 ± 1.0 -261.7 ± 4.5 -44 ± 12 16 Terr. veg. 1.6 ± 2.0 -29.3 ± 1.6 -136.3 ± 19.6 41 ± 1 27 Soil & 3.1 ± 2.3 -26.7 ± 1.4 -106.1 ± 8.2 -207 ± 206 16 sediment OM

249

Table B.13. Little Muncy Creek mixing model including δ13C, δ2H, and 14C data Sources δ13C ± S.D. (‰) δ2H ± S.D. (‰) ∆14C ± S.D. (‰) n Algae -29.7 ± 11.0 -217.1 ± 33.9 -78 ± 117 4 Terrestrial vegetation -29.7 ± 1.3 -122.7 ± 24.0 41 ± 1 2 Soil & sediment OM -27.0 ± 1.4 -104.3 ± 10.3 -143 ± 230 5 * CH4 -35.0 ± 15.0 -214.1 ± 25.9 -1000 ± 10 2 *Isotopic values for CH4 from Osborn and McIntosh (2010).

250

Section 3

Means and standard deviations (SD) of isotope values for macroinvertebrate functional feeding groups collected across all sites and sampling dates are presented in the table below.

Table B.14. Means and standard deviations (SD) of isotopic values for functional feeding groups (FFG) collected in 2011, 2012, 2013, and 2014 from the middle and upper Susquehanna River watershed. Isotopic values presented here are uncorrected for trophic fractionation. ND: no data. FFG (n) δ13C (‰) δ15N (‰) δ2H (‰) Δ14C (‰) (n) 2011 Cowanesque River Filtering collectors (3) -29.2 ± 0.4 9.9 ± 1.3 -211.9 ± 5.4 -92 ± 14 (2) Predators (1) -22.5 5.3 -148.9 ND Crooked Creek Filtering collectors (1) -28.0 11.4 -214.6 ND Scrapers (1) -26.4 9.7 -187.8 ND Predators (2) -28.4 ± 0.5 11.6 ± 0.1 -166.4 ± 6.2 ND Elk Run Scrapers (1) -25.0 7.9 -178.4 ND Predators (4) -25.0 ± 0.6 9.2 ± 0.6 -166.2 ± 11.7 ND Lamb’s Creek Filtering collectors (1) -29.8 6.0 -146.1 ND Scrapers (2) -30.7 ± 0.2 3.4 ± 1.6 -128.9 ± 5.0 ND Shredders (1) -28.6 2.6 -143.1 ND Predators (1) -28.6 7.6 -128.7 ND Little Muncy Creek Filtering collectors (4) -26.5 ± 1.5 7.3 ± 0.3 -147.9 ± 8.3 -175 ± 132 (3) Scrapers (2) -30.8 ± 11.5 5.7 ± 1.5 -165.8 ± 51.1 -196 ± 271 (3) Predators (1) -36.5 7.0 -98.9 -405 North Elk Run Scrapers (2) -29.9 ± 0.3 9.2 ± 0.2 -166.1 ± 32.3 ND Predators (1) -27.7 9.6 -126.2 ND 2012 Cowanesque River Filtering collectors (5) -26.3 ± 0.5 8.7 ± 0.3 -191.0 ± 7.5 -57 ± 2 (2) Scrapers (2) -22.7 ± 0.3 7.7 ± 0.7 -217.1 -32 (1) Predators (1) -26.0 12.3 -174.1 ND

continued 252

Table B3.1, continued Crooked Creek Scrapers (3) -23.9 ± 2.8 9.2 ± 1.2 -191.9 ± 30.6 -129 (1) Chironomids (1) -24.7 9.5 -152.7 -64 Predators (3) -24.4 ± 0.7 11.0 ± 0.6 -161.9 ± 17 ND Elk Run Filtering collectors (1) -23.1 7.9 -184.4 -42 Scrapers (1) -24.2 7.2 -223.2 ND Chironomids (2) -21.1 ± 0.0 9.7 ± 0.7 -147.9 ± 0.3 -31 (1) Predators (1) -23.7 11.1 -166.7 ND Lamb’s Creek Shredders (4) -28.1 ± 0.2 3.1 ± 0.2 -137.9 ± 2.3 ND Predators (2) -30.6 ± 0.3 6.6 ± 0.0 -153.2 ± 5.8 ND North Elk Run Collector-gatherers (2) -28.6 ± 2.7 12.9 ± 1.7 -156.0 ± 22.4 ND Shredders (1) -31.8 4.9 -184.2 ND Towanda Creek Filtering collectors (4) -24.9 ± 0.7 11.7 ± 1.1 -180.6 ± 9.7 -6 (1) Chironomids (2) -22.2 ± 0.1 12.3 ± 0.1 -181.2 ± 21.8 ND Shredders (1) -23.5 11.1 -212.8 ND Predators (2) -25.6 ± 2.6 13.9 ± 0.4 -149.0 ± 5.6 ND 2013 Cowanesque River Filtering collectors (4) -20.1 ± 1 10.5 ± 0.6 -218.2 ± 16.3 -40 ± 5 (3) Scrapers (5) -21.6 ± 3.8 9.0 ± 1.3 -211.4 ± 10.4 -33 (1) Collector-gatherers (1) -18.8 8.3 -220.9 ± 12.9 (2) -38 (1) Shredder (1) -21.3 9.2 -195.4 -18 Predators (5) -21.3 ± 0.6 11.9 ± 1.1 -178.2 ± 24.3 -29 (1) Crooked Creek Filtering collectors (3) -23.2 ± 0.2 8.7 ± 0.1 -230.8 ± 0.6 ND Scrapers (7) -22.5 ± 2.2 7.8 ± 1.1 -209.7 ± 7.1 -78 ± 63 (2) Collector-gatherers (1) -22.2 7.9 -212.4 -34 Predators (3) -24.2 ± 1.3 10.1 ± 0.3 -160.8 ± 12.4 ND Elk Run Filtering collectors (2) -23.5 ± 0.2 8.3 ± 0.1 -193.8 ± 0.5 -21 (1) Scrapers (1) -22.7 6.6 -181.0 ND Collector-gatherers (2) -21.0 ± 0.1 7.7 ± 0.0 -199.5 ± 43.9 -26 (1) Predators (1) -23.5 8.8 -176.1 ND Lamb’s Creek Filtering collectors (2) -30.2 ± 0.2 4.9 ± 0.1 -184.5 ± 7.2 ND Scrapers (2) -35.7 ± 1.1 2.9 ± 0.1 -206.5 ± 12.9 ND Shredders (4) -27.5 ± 0.3 2.9 ± 0.4 -116.4 ± 3.6 ND Predators (5) -30.9 ± 1.4 5.9 ± 0.6 -165.0 ± 23.4 ND continued 253

Table B3.1, continued Little Muncy Creek Filtering collectors (5) -23.1 ± 0.4 4.8 ± 0.2 -182.1 ± 2.8 -5 ± 6 (2) Scrapers (2) -20.2 ± 1.6 4.4 ± 0.1 -199.8 ± 16.6 ND Collector-gatherers (2) -23.0 ± 3.4 3.6 ± 0.2 -159.5 ± 37.5 ND Predators (4) -22.2 ± 0.3 6.7 ± 0.6 -140.8 ± 12.4 ND North Elk Run Filtering collector (1) ND ND ND -6 Scrapers (1) -32.3 11.0 -226.3 ND Collector-gatherers (2) -26.3 ± 1.3 9.8 ± 0.6 -181.5 ± 33.2 ND Towanda Creek Filtering collectors (1) -21.6 11.9 -219.4 -21 Scrapers (2) -22.7 ± 2.0 9.4 ± 2.0 -203.1 ± 4.4 ND Collector-gatherers (1) -21.4 10.1 -171.8 ND Predators (3) -22.0 ± 0.8 12.2 ± 0.4 -175.8 ± 23.7 ND 2014 Cowanesque River Filtering collectors (2) -21.8 ± 0. 11.0 ± 0.0 -225.8 ± 3.3 -53 ± 3 (1) Scrapers (1) -20.6 10.4 -201.1 -39 Collector-gatherers (2) -20.9 ± 0.5 10.5 ± 0.9 -191.1 ± 3.8 ND Predators (2) -21.4 ± 0.5 11.6 ± 0.0 -204.0 ± 12.7 -54 (2) Crooked Creek Scrapers (3) -24.6 ± 2.5 8.7 ± 0.2 -248.8 ± 14.9 -57 (1) Collector-gatherers (1) -26.1 7.4 -216.4 ND Shredders (1) -25.4 6.4 -165.7 -22 Predators (10) -24.9 ± 0.7 9.6 ± 0.5 -187.6 ± 16.6 -46 ± 1 (4) Elk Run Filtering collectors (5) -23.8 ± 0.3 9.4 ± 0.2 -201.7 ± 12.7 -44 ± 2 (2) Scrapers (3) -24.1 ± 0.2 9.9 ± 0.7 -224.9 ± 3.9 -42 ± 1 (2) Predators (2) -22.2 ± 0.4 11.4 ± 0.1 -176.8 ± 4.0 -30 (1) Lamb’s Creek Filtering collectors (4) -29.5 ± 0.5 5.3 ± 0.4 -181.6 ± 5.8 ND Scrapers (2) -34.5 ± 4.3 4.2 ± 0.0 -208.7 ± 17.8 ND Shredders (3) -27.7 ± 0.5 2.8 ± 0.6 -117.1 ± 13.4 ND Predators (5) -30.6 ± 3.3 5.7 ± 0.6 -172.6 ± 26.2 ND Little Muncy Creek Filtering collectors (4) -22.4 ± 1 5.3 ± 0.3 -177.2 ± 17.9 -13 ± 9 (3) Scrapers (2) -20.7 ± 2.5 4.9 ± 1.9 -207.7 ± 5.8 -32 ± 1 (2) Collector-gatherers (1) -22.8 4.0 -180.6 ND Shredders (2) -25.3 ± 0.7 4.0 ± 0.6 -106.7 ± 1.4 ND Predators (5) -21.2 ± 0.6 6.4 ± 0.7 -173.4 ± 22.1 -12 ± 8 (3)

continued 254

Table B3.1, continued North Elk Run Filtering collectors (3) -28.4 ± 0.3 11.4 ± 0.5 -160.1 ± 6.8 -10 ± 1 (2) Collector-gatherers (1) -27.2 9.4 -157.5 ND Shredders (1) -27.3 7.2 -136.4 ND Predators (5) -26.5 ± 0.4 10.8 ± 0.5 -144.6 ± 20.0 1 ± 6 (3) Towanda Creek Filtering collectors (6) -22.7 ± 0.3 11.7 ± 0.1 -215.9 ± 10.0 -24 ± 3 (3) Scrapers (6) -23.5 ± 1.6 10.5 ± 0.7 -208.4 ± 23.6 -28 ± 5 (3) Collector-gatherers (1) -26.2 10.2 -213.3 ND Predators (7) -23.2 ± 0.7 12.2 ± 0.3 -169.4 ± 16.5 -17 ± 1 (3)

255

Section 4

Results of one-way ANOSIMs that were performed to assess differences in isotopic composition (δ13C, δ15N, δ2H, and ∆14C) of macroinvertebrates across site, functional feeding group, and year are provided in the following tables.

Table B.15. One-way ANOSIM results based on the stable isotopic values (i.e., δ13C, δ15N, and δ2H) of macroinvertebrates across sites, combined across all sampling years. The ANOSIM R value is the test statistic for the overall ANOSIM test, and is shown at the top of the table along with the associated p-value. ANOSIM R = 0.497, p = 0.001 Pairwise Test R Statistica Significance Levelb Crooked Creek, Cowanesque River 0.127 0.001 Crooked Creek, Little Muncy Creek 0.506 0.001 Crooked Creek, Towanda Creek 0.227 0.001 Crooked Creek, Elk Run -0.023 0.738 Crooked Creek, North Elk Run 0.365 0.001 Crooked Creek, Lamb’s Creek 0.788 0.001 Cowanesque River, Little Muncy Creek 0.575 0.001 Cowanesque River, Towanda Creek 0.134 0.001 Cowanesque River, Elk Run 0.079 0.023 Cowanesque River, North Elk Run 0.554 0.001 Cowanesque River, Lamb’s Creek 0.849 0.001 Little Muncy Creek, Towanda Creek 0.778 0.001 Little Muncy Creek, Elk Run 0.430 0.001 Little Muncy Creek, North Elk Run 0.636 0.001 Little Muncy Creek, Lamb’s Creek 0.546 0.001 Towanda Creek, Elk Run 0.234 0.001 Towanda Creek, North Elk Run 0.609 0.001 Towanda Creek, Lamb’s Creek 0.932 0.001 Elk Run, North Elk Run 0.562 0.001 Elk Run, Lamb’s Creek 0.811 0.001 North Elk Run, Lamb’s Creek 0.562 0.001 aThe R statistic is the test statistic for pairwise comparisons and indicates the amount of separation between pairs. bDifferences between pairs were significant when p<0.05 or when R>0.8.

256

Table B.16. One-way ANOSIM results based on the stable isotopic values (i.e., δ13C, δ15N, and δ2H) of macroinvertebrate functional feeding groups (FFGs), combined across all sampling years. The ANOSIM R value is the test statistic for the overall ANOSIM test, and is shown at the top of the table along with the associated p-value. ANOSIM R = 0.221, p = 0.001 Pairwise Test R Statistica Significance Levelb predator, filtering collector 0.120 0.001 predator, scraper 0.265 0.001 predator, chironomid -0.114 0.89 predator, collector-gatherer 0.209 0.003 predator, shredder 0.487 0.001 filtering collector, scraper 0.051 0.027 filtering collector, chironomid 0.128 0.054 filtering collector, collector-gatherer 0.104 0.049 filtering collector, shredder 0.528 0.001 scraper, chironomid 0.209 0.052 scraper, collector-gatherer 0.059 0.189 scraper, shredder 0.498 0.001 chironomid, collector-gatherer 0.026 0.334 chironomid, shredder 0.603 0.001 collector-gatherer, shredder 0.471 0.001 aThe R statistic is the test statistic for pairwise comparisons and indicates the amount of separation between pairs. bDifferences between pairs were significant when p<0.05 or when R>0.8.

Table B.17. One-way ANOSIM results based on the stable isotopic values (i.e., δ13C, δ15N, and δ2H) of macroinvertebrates across sampling year. The ANOSIM R value is the test statistic for the overall ANOSIM test, and is shown at the top of the table along with the associated p-value. ANOSIM R = 0.064, p = 0.002 Pairwise Test R Statistica Significance Levelb 2011, 2012 0.087 0.018 2011, 2013 0.175 0.01 2011, 2014 0.127 0.015 2012, 2013 0.066 0.04 2012, 2014 -0.009 0.565 2013, 2014 0.036 0.018 aThe R statistic is the test statistic for pairwise comparisons and indicates the amount of separation between pairs. bDifferences between pairs were significant when p<0.05 or when R>0.8.

257

Table B.18. One-way ANOSIM results based on the stable isotopic (δ13C, δ15N, and δ2H) and radiocarbon (∆14C) values of macroinvertebrates across six sites (Lamb’s Creek excluded), combined across all sampling years. The ANOSIM R value is the test statistic for the overall ANOSIM test, and is shown at the top of the table along with the associated p-value. ANOSIM R = 0.308, p = 0.001 Pairwise Test R Statistica Significance Levelb Crooked Creek, Cowanesque River 0.107 0.077 Crooked Creek, Little Muncy Creek 0.248 0.001 Crooked Creek, Towanda Creek 0.324 0.001 Crooked Creek, Elk Run -0.081 0.988 Crooked Creek, North Elk Run 0.621 0.001 Cowanesque River, Little Muncy Creek 0.456 0.001 Cowanesque River, Towanda Creek 0.068 0.129 Cowanesque River, Elk Run 0.036 0.263 Cowanesque River, North Elk Run 0.572 0.001 Little Muncy Creek, Towanda Creek 0.517 0.001 Little Muncy Creek, Elk Run 0.246 0.012 Little Muncy Creek, North Elk Run 0.380 0.028 Towanda Creek, Elk Run 0.246 0.025 Towanda Creek, North Elk Run 0.731 0.001 Elk Run, North Elk Run 0.797 0.001 aThe R statistic is the test statistic for pairwise comparisons and indicates the amount of separation between pairs. bDifferences between pairs were significant when p<0.05 or when R>0.8.

Table B.19. One-way ANOSIM results based on the stable isotopic (δ13C, δ15N, and δ2H) and radiocarbon (∆14C) values of macroinvertebrates combined across all sampling years. The ANOSIM R value is the test statistic for the overall ANOSIM test, and is shown at the top of the table along with the associated p-value. ANOSIM R = 0.211, p = 0.003 Pairwise Test R Statistica Significance Levelb 2012, 2013 0.293 0.013 2012, 2014 -0.097 0.801 2012, 2011 0.080 0.168 2013, 2014 0.014 0.404 2013, 2011 0.581 0.001 2014, 2011 0.611 0.001 aThe R statistic is the test statistic for pairwise comparisons and indicates the amount of separation between pairs. bDifferences between pairs were significant when p<0.05 or when R>0.8.

258

Section 5

Bayesian mixing model source contribution estimates (5%, 50%, and 95% posterior probabilities) of nutritional source contributions to each functional feeding group (FFG) for each of the mixing model approached used.

Table B.20. Lamb’s Creek using δ13C, δ15N, and δ2H FFG/nutritional source 5% 50% 95% filtering collector/algae 0.407 0.577 0.828 filtering collector/soil & sediment OM 0.077 0.337 0.520 filtering collector/terrestrial vegetation 0.000 0.051 0.283 scraper/algae 0.393 0.645 0.936 scraper/soil & sediment OM 0.000 0.017 0.163 scraper/terrestrial vegetation 0.021 0.312 0.592 shredder/algae 0.000 0.073 0.169 shredder/soil & sediment OM 0.000 0.061 0.253 shredder/terrestrial vegetation 0.623 0.853 1.000 predator/algae 0.439 0.59 0.760 predator/soil & sediment OM 0.010 0.083 0.235 predator/terrestrial vegetation 0.107 0.309 0.505

Table B.21. North Elk Run using δ13C, δ15N, and δ2H FFG/nutritional source 5% 50% 95% collector-gatherer/algae 0.383 0.584 0.921 collector-gatherer/soil & sediment OM 0.002 0.195 0.507 collector-gatherer/terrestrial vegetation 0.000 0.161 0.461 filtering collector/algae 0.318 0.526 0.875 filtering collector/soil & sediment OM 0.004 0.246 0.579 filtering collector/terrestrial vegetation 0.000 0.160 0.501 scraper/algae 0.383 0.649 0.993 scraper/soil & sediment OM 0.000 0.103 0.417 scraper/terrestrial vegetation 0.000 0.171 0.507 shredder/algae 0.000 0.379 0.738 shredder/soil & sediment OM 0.000 0.096 0.477 shredder/terrestrial vegetation 0.027 0.458 0.995 predator/algae 0.076 0.254 0.441 predator/soil & sediment OM 0.014 0.148 0.482 predator/terrestrial vegetation 0.277 0.575 0.775

259

Table B.22. Little Muncy Creek using δ13C, δ15N, and δ2H FFG/nutritional source 5% 50% 95% chironomid/algae 0.247 0.511 0.804 chironomid/soil & sediment OM 0.032 0.263 0.588 chironomid/terrestrial vegetation 0.008 0.167 0.528 collector-gatherer/algae 0.327 0.606 0.885 collector-gatherer/soil & sediment OM 0.013 0.178 0.466 collector-gatherer/terrestrial vegetation 0.008 0.160 0.506 filtering collector/algae 0.442 0.567 0.716 filtering collector/soil & sediment OM 0.129 0.316 0.465 filtering collector/terrestrial vegetation 0.005 0.092 0.322 scraper/algae 0.600 0.786 0.960 scraper/soil & sediment OM 0.005 0.091 0.235 scraper/terrestrial vegetation 0.005 0.095 0.291 shredder/algae 0.115 0.256 0.440 shredder/soil & sediment OM 0.069 0.458 0.802 shredder/terrestrial vegetation 0.009 0.263 0.641 predator algae 0.260 0.454 0.671 predator/soil & sediment OM 0.007 0.078 0.263 predator/terrestrial vegetation 0.184 0.454 0.664

Table B.23. Elk Run using δ13C, δ15N, and δ2H FFG/nutritional source 5% 50% 95% chironomid/algae 0.281 0.487 0.763 chironomid/soil & sediment OM 0.061 0.442 0.689 chironomid/terrestrial vegetation 0.001 0.055 0.280 collector-gatherer/algae 0.577 0.738 0.961 collector-gatherer/soil & sediment OM 0.005 0.137 0.319 collector-gatherer/terrestrial vegetation 0.002 0.081 0.295 filtering collector/algae 0.511 0.621 0.760 filtering collector/soil & sediment OM 0.026 0.228 0.401 filtering collector/terrestrial vegetation 0.006 0.133 0.366 scraper/algae 0.547 0.672 0.830 scraper/soil & sediment OM 0.008 0.180 0.345 scraper/terrestrial vegetation 0.005 0.124 0.362 predator/algae 0.325 0.456 0.598 predator/soil & sediment OM 0.019 0.161 0.398 predator/terrestrial vegetation 0.091 0.372 0.601

260

Table B.24. Crooked Creek using δ13C, δ15N, and δ2H FFG/nutritional source 5% 50% 95% chironomid/algae 0.000 0.444 0.862 chironomid/soil & sediment OM 0.000 0.152 0.726 chironomid/terrestrial vegetation 0.001 0.185 0.999 collector-gatherer/algae 0.545 0.822 1.000 collector-gatherer/soil & sediment OM 0.000 0.049 0.274 collector-gatherer/terrestrial vegetation 0.000 0.061 0.366 filtering collector/algae 0.701 0.890 1.000 filtering collector/soil & sediment OM 0.000 0.030 0.184 filtering collector/terrestrial vegetation 0.000 0.038 0.242 scraper/algae 0.667 0.808 0.981 scraper/soil & sediment OM 0.000 0.074 0.222 scraper/terrestrial vegetation 0.000 0.070 0.282 predator/algae 0.396 0.542 0.661 predator/soil & sediment OM 0.023 0.165 0.339 predator/terrestrial vegetation 0.060 0.289 0.542

Table B.25. Towanda Creek using δ13C, δ15N, and δ2H FFG/nutritional source 5% 50% 95% chironomid/algae 0.514 0.613 0.693 chironomid/soil & sediment OM 0.138 0.295 0.437 chironomid/terrestrial vegetation 0.009 0.086 0.227 collector-gatherer/algae 0.495 0.603 0.690 collector-gatherer soil & sediment OM 0.113 0.284 0.431 collector-gatherer/terrestrial vegetation 0.011 0.108 0.284 filtering collector/algae 0.594 0.652 0.715 filtering collector/soil & sediment OM 0.124 0.256 0.359 filtering collector/terrestrial vegetation 0.010 0.089 0.212 scraper/algae 0.556 0.620 0.684 scraper/soil & sediment OM 0.085 0.259 0.381 scraper/terrestrial vegetation 0.013 0.119 0.288 predator/algae 0.435 0.496 0.557 predator/soil & sediment OM 0.092 0.260 0.429 predator/terrestrial vegetation 0.083 0.244 0.404

261

Table B.26. Cowanesque River using δ13C, δ15N, and δ2H FFG/nutritional source 5% 50% 95% chironomid/algae 0.432 0.594 0.689 chironomid/soil & sediment OM 0.211 0.355 0.516 chironomid/terrestrial vegetation 0.003 0.046 0.180 collector-gatherer/algae 0.536 0.618 0.706 collector-gatherer/soil & sediment OM 0.193 0.330 0.428 collector-gatherer/terrestrial vegetation 0.003 0.045 0.165 filtering collector/algae 0.557 0.619 0.692 filtering collector/soil & sediment OM 0.196 0.327 0.405 filtering collector/terrestrial vegetation 0.003 0.045 0.170 scraper/algae 0.565 0.634 0.721 scraper/soil & sediment OM 0.181 0.309 0.399 scraper/terrestrial vegetation 0.003 0.046 0.163 shredder/algae 0.482 0.605 0.706 shredder/soil & sediment OM 0.192 0.339 0.463 shredder/terrestrial vegetation 0.003 0.047 0.175 predator/algae 0.468 0.571 0.684 predator/soil & sediment OM 0.065 0.305 0.451 predator/terrestrial vegetation 0.005 0.105 0.356

Table B.27. High agriculture sites using δ13C, δ15N, and δ2H FFG/nutritional source 5% 50% 95% chironomid/algae 0.381 0.489 0.605 chironomid/soil & sediment OM 0.338 0.483 0.599 chironomid/terrestrial vegetation 0.000 0.019 0.110 collector-gatherer/algae 0.608 0.679 0.751 collector-gatherer/soil & sediment OM 0.097 0.276 0.367 collector-gatherer/terrestrial vegetation 0.001 0.030 0.219 filtering collector/algae 0.678 0.722 0.768 filtering collector/soil & sediment OM 0.168 0.249 0.304 filtering collector/terrestrial vegetation 0.000 0.022 0.104 scraper/algae 0.685 0.734 0.789 scraper/soil & sediment OM 0.008 0.152 0.267 scraper/terrestrial vegetation 0.003 0.109 0.270 shredder/algae 0.098 0.291 0.474 shredder/soil & sediment OM 0.002 0.176 0.632 shredder/terrestrial vegetation 0.009 0.528 0.851 predator/algae 0.461 0.506 0.557 predator/soil & sediment OM 0.005 0.053 0.166 predator/terrestrial vegetation 0.314 0.438 0.509

262

Table B.28. Low agriculture sites using δ13C, δ15N, and δ2H FFG/nutritional source 5% 50% 95% chironomid/algae 0.211 0.525 0.839 chironomid/soil & sediment OM 0.012 0.216 0.614 chironomid/terrestrial vegetation 0.004 0.160 0.627 collector-gatherer/algae 0.275 0.630 0.904 collector- gatherer/soil & sediment OM 0.002 0.082 0.420 collector-gatherer/terrestrial vegetation 0.011 0.220 0.636 filtering collector/algae 0.545 0.634 0.739 filtering collector/soil & sediment OM 0.198 0.306 0.410 filtering collector/terrestrial vegetation 0.002 0.049 0.143 scraper/algae 0.629 0.759 0.872 scraper/soil & sediment OM 0.000 0.015 0.087 scraper/terrestrial vegetation 0.099 0.215 0.355 shredder/algae 0.057 0.130 0.215 shredder/soil & sediment OM 0.013 0.150 0.339 shredder/terrestrial vegetation 0.503 0.715 0.894 predator/algae 0.344 0.462 0.579 predator/soil & sediment OM 0.002 0.024 0.091 predator/terrestrial vegetation 0.380 0.508 0.622

Table B.29. Global model using δ13C, δ15N, δ2H, and ∆14C data FFG/nutritional source 5% 50% 95% chironomid/algae 0.230 0.384 0.608 chironomid/soil & sediment OM 0.069 0.154 0.348 chironomid/terrestrial vegetation 0.221 0.436 0.633 collector-gatherer/algae 0.598 0.743 0.853 collector-gatherer/soil & sediment OM 0.013 0.047 0.122 collector-gatherer/terrestrial vegetation 0.094 0.204 0.336 filtering collector/algae 0.469 0.592 0.706 filtering collector soil/sediment OM 0.059 0.120 0.242 filtering collector/terrestrial vegetation 0.140 0.277 0.420 scraper/algae 0.631 0.757 0.842 scraper/soil & sediment OM 0.027 0.066 0.131 scraper/terrestrial vegetation 0.088 0.173 0.290 shredder/algae 0.364 0.547 0.713 shredder soil & sediment OM 0.028 0.081 0.200 shredder/terrestrial vegetation 0.187 0.359 0.553 predator/algae 0.373 0.546 0.704 predator/soil & sediment OM 0.048 0.120 0.280 predator terrestrial vegetation 0.136 0.318 0.508

263

Table B.30. Little Muncy Creek primary consumers using δ13C, δ2H and ∆14C data Year and nutritional source Mean SD 5% 50% 95% 2011 algae 0.358 0.182 0.057 0.349 0.680 2013 algae 0.671 0.235 0.277 0.669 1.000 2014 algae 0.757 0.193 0.413 0.768 1.000 2011 methane 0.069 0.094 0.000 0.041 0.291 2013 methane 0.016 0.031 0.000 0.005 0.089 2014 methane 0.014 0.022 0.000 0.005 0.069 2011 soil & sediment OM 0.388 0.187 0.037 0.403 0.724 2013 soil & sediment OM 0.072 0.100 0.000 0.027 0.344 2014 soil & sediment OM 0.052 0.072 0.000 0.021 0.246 2011 terrestrial vegetation 0.184 0.199 0.000 0.123 0.719 2013 terrestrial vegetation 0.241 0.231 0.000 0.182 0.740 2014 terrestrial vegetation 0.177 0.182 0.000 0.123 0.609

Table B.31. Little Muncy Creek predators using δ13C, δ2H and ∆14C data Year and nutritional source Mean SD 5% 50% 95% 2011 algae 0.092 0.239 0.000 0.000 0.839 2014 algae 0.770 0.344 0.010 0.988 1.000 2011 methane 0.172 0.303 0.000 0.002 0.985 2014 methane 0.022 0.106 0.000 0.000 0.074 2011 soil & sediment OM 0.607 0.423 0.000 0.823 1.000 2014 soil & sediment OM 0.044 0.154 0.000 0.000 0.245 2011 terrestrial vegetation 0.128 0.276 0.000 0.000 0.949 2014 terrestrial vegetation 0.164 0.287 0.000 0.000 0.895

264

Appendix C: Supplementary Information for Chapter 3

265

Section 1

Macroinvertebrates were collected in November 2012 and 2013, and in May and

July of 2013 (Table C.1). At each collection, potential nutritional resource materials was collected: algae, terrestrial vegetation, soil organic matter, sediment organic matter, suspended particulate organic matter, and biofilm. The δ13C, δ15N, δ2H and were measured on all collected samples and the ∆14C was measured on a subset of the samples.

Details on the isotopic analyses are given in the main body of the text. The proportionate contribution of the various sources to the macroinvertebrate diet was determined using

Bayesian isotopic mixing models.

Mixing Model Methods

Potential nutritional sources used in the Bayesian mixing model included aquatic algae, terrestrial vegetation, and soil organic matter (OM: see Results and Discussion section of the main text). While soil is derived from terrestrial vegetation, alteration and processing of soil OM by microbes leads to distinct isotopic signatures for terrestrial vegetation and soil (Fogel & Tuross 1999; Hoefs 2008), and we treated each as a separate source in our isotopic mixing models. We chose to not include as potential nutritional resources suspended particulate OM (POM) and biofilm in the model because both POM and biofilm consist of mixtures of different sources, and would provide little information about the basal (i.e., ultimate starting) nutritional resources contributing to macroinvertebrate biomass (Goni et al. 2013; Ishikawa et al. 2015; Cummins 2016).

Trophic fractionation values for δ13C and δ15N (0.4 ± 1.3‰ and 3.4 ± 1‰, respectively) were taken from Post (2002). Trophic fractionation for δ2H was assumed to

266 be 0‰ as previous studies have found trophic fractionation of δ2H to be small or negligible (Doucett et al. 2007; Solomon et al. 2009). Dietary water can also influence organism δ2H values because a portion of the 2H in organism tissues is derived from environmental water. The fraction of dietary water (ω) for macroinvertebrate primary consumers and predators was assumed to be 0.20 and 0.36, respectively, based on field estimates and a literature review by Wilkinson et al. (2015) (Solomon et al. 2009;

Wilkinson, Cole & Pace 2015). Macroinvertebrate δ2H values were corrected using the

2 appropriate ω value and the δ H-H2O literature values for Upper Twin Creek, Ohio, which is in close proximity to Paint Creek and a part of the Paint Creek watershed

(Coplen & Kendall 2000). We did not correct macroinvertebrate δ2H for the effect of lipids, which may have resulted in underestimating the contributions of allochthonous

OM to macroinvertebrate biomass.

While we used the δ2H of bulk soil in our mixing models, these data should be interpreted with caution because the hydrogen content of soils derives from both OM and hydrated minerals. When OM content is low, as for most soils, the δ2H measurement is not an accurate measure of the 2H content of the OM fraction (Ruppenthal, Oelmann &

Wilcke 2010; Ruppenthal et al. 2015). Our soil δ2H values were consistently more 2H- enriched compared to terrestrial vegetation (Figure 3B, D) suggesting that there was some contamination from mineral water 2H.

In order to assess the influence of using bulk soil δ2H on mixing model outcomes, we ran mixing models two different ways. The first used all three sources (algae, terrestrial vegetation, and soil OM) and the second used only two sources (algae and

267 terrestrial vegetation). We then compared estimates of autochthony for each FFG using each model. We found that 71% of the time the difference in estimates of percent autochthony between the three and two source models was less than 10% (Figure C.1 A and B). Least squares regression for the data in Figure C.1A yielded an r2-value of 0.39, and a slope of 1.08 and a p-value of 0.0175. This comparison between the three and two- source model runs did not follow the 1:1 relationship, suggesting that there was some difference in the measure of autochthony using the three and two-source models. The largest differences in estimates of autochthony were for the models that included stable isotope (δ13C, δ15N, and δ2H) and radiocarbon (∆14C) data. One possible explanation for this is that in models using three nutritional resources, low ∆14C values of both algae and soil OM can contribute to the low ∆14C values observed in macroinvertebrate consumers.

In contrast, for the two-source models only low algal ∆14C values contributed to the low

∆14C values observed for macroinvertebrates. Consequently, larger contributions of algae to macroinvertebrate biomass (autochthony) are observed from two-source mixing model outcomes. By including soil OM in our mixing models, we may therefore have overestimated its importance as a nutritional resource, and concomitantly underestimated contributions of terrestrial vegetation.

In order to determine the proportional contribution of each potential nutritional resource to macroinvertebrate biomass, we ran mixing models using three different approaches. Even though one-way ANOSIM suggested that there were significant differences in the isotopic composition of macroinvertebrates between most sampling dates, we chose to combine data from November (2012 and 2013), and from spring and

268 summer (May and July 2013) in order to have large enough samples sizes to optimize mixing model resolution. In addition, one-way ANOSIM indicated that difference in the isotopic composition of macroinvertebrates collected in November 2013 from the shale site and the site upstream of the shale outcrop were not significant (p = 0.06, R = 0.136).

Therefore, for purposes of the statistical and mixing model analyses, we combined macroinvertebrates from these two sites. Approach 1 consisted of separate models for

November 2012/2013, and for May/July 2013, and only macroinvertebrates and algae samples from those specific time periods were included. Approach 2 was a global model for which all macroinvertebrates and potential nutritional resources were pooled together across all time periods. Approach 3 was a model for macroinvertebrates where both stable isotope and natural 14C data were available, and was pooled across all time periods.

Using Markov chain Monte Carlo (MCMC) methods, models were initially run with 50,000 MCMC chains (25,000 burn-in) to ensure that the model parameters were specified correctly, and if necessary, models were run again with 300,000 (200,000 burn- in) or 1,000,000 (500,000 burn-in) MCMC chains in order for models to converge.

Gelman-Rubin and Geweke diagnostics were evaluated in order to ensure that models had converged. Isotopic data for potential nutritional resources that were used in each model are provided in Tables C.2-C.9 below.

Algal biomass in quantities adequate for isotopic analyses were not available at the sampling site in July or November 2013. In these instances, we used estimated algal

13 isotopic values for the mixing models. δ C values of CO2 (aq) were estimated using the measured δ13C values of DIC, pH, and water temperature at each sampling site according

269

13 to Mook et al. (1974). The fractionation of δ C-CO2 by algae was determined using the calculated aqueous concentration of CO2 and a regression equation according to Finlay

(2004). The δ15N values of algae were estimated using the δ15N of suspended POM and terrestrial vegetation according to Francis et al. (2011). Algal δ2H values were determined from the estimated δ2H of water for each site (Coplen & Kendall 2000), and the 2H fractionation of water by algae according to Hondula et al. (2014). Since algal material was not available in significant quantities in November 2013, algal ∆14C values for that sampling date were assumed to be the same as ∆14C values of DIC because ∆14C of primary producers (both terrestrial and aquatic) is the same as the ∆14C of the inorganic C that is fixed (Broecker & Walton 1959; Caraco et al. 2010). Measured algal

∆14C values were available November 2012 and May 2013.

270

Table C.1. Genera and functional feeding groups identified in Paint Creek, OH. Genera Functional Feeding Group Collection dates Lestes Predator Nov. 2012 Argia Predator Nov. 2012, July 2013, Nov. 2013 Hetaerina Predator Nov. 2012, July 2013 Agnetina Predator Nov. 2012, Nov. 2013 Corydalus Predator Nov. 2012, July 2013 Chimarra* Filtering collector Nov. 2012, Nov. 2013 Ceratopsyche* Filtering collector Nov. 2012, Nov. 2013 Anthopotamus* Filtering collector Nov. 2012, May 2013. Nov. 2013 Isonychia* Filtering collector Nov. 2012, May 2013, Nov. 2013 Hydropsyche* Filtering collector Nov. 2012, May 2013 Stenelmis* Scraper Nov. 2012, July 2013 Basiaeschna Predator Nov. 2012 Pleurocerid snails* Scraper Nov. 2012 Physid snails* Scraper Nov. 2012 Anacroneuria Predator May 2013, July 2013 Cheumatopsyche* Filtering collector May 2013, July 2013, Nov. 2013 Heptagenia* Scraper May 2013 Stenacron* Scraper May 2013, July 2013 Cinygmula* Scraper May 2013 Macronychus* Collector gatherer July 2013 Macromia Predator July 2013, Nov. 2013 Peltodytes* Shredder, piercer July 2013 Gyretes Predator July 2013, Nov. 2013 Simuliidae* Filtering collector July 2013 Maccaffertium* Scraper Nov. 2013 Asellidae* Collector gatherer Nov. 2013 Neocorixa* Predator Nov. 2013 Neoperla* Predator Nov. 2013 Chironomidae* None assigned Nov. 2013 Tipula Shredder Nov. 2013 Luricta Predator Nov. 2013 Belostoma Predator Nov. 2013 *Denotes that in some cases multiple individuals were pooled to increase sample size within a genus.

271

Table C.2. November 2012 potential nutritional resource stable isotope data. Source Mean δ15N ± S.D. Mean δ13C ± S.D. Mean δ2H ± S.D. N Algae 9.1 ± 1.7‰ -31.2 ± 0.6‰ -238.4 ± 28.1‰ 3 Terrestrial vegetation 4.2 ± 1.4‰ -30.2 ± 1.9‰ -129.7 ± 15.2‰ 5 Sediment OM 1.9 ± 0.3‰ -29.2 ± 0.7‰ -116.4 ± 2.4‰ 4 Soil OM 3.6 ± 1.2‰ -22.4 ± 1.3‰ -105.7 ± 8.8‰ 2 Suspended POM 9.8 ± 4.0‰ -32.7 ± 0.8‰ -152.5 ± 7.1‰ 2 Biofilm 13.5 ± 1.8‰ -27.2 ± 1.9‰ -187.0 ± 27.8‰ 2

Table C.3. May 2013 potential nutritional resource stable isotope data. Source Mean δ15N ± S.D. Mean δ13C ± S.D. Mean δ2H ± S.D. N Algae 5.0 ± 1.5‰ -30.6 ± 2.5‰ -211.4 ± 27.8‰ 3 Terrestrial vegetation 4.2 ± 1.4‰ -30.2 ± 1.9‰ -129.7 ± 15.2‰ 5 Sediment OM 1.9 ± 0.3‰ -29.2 ± 0.7‰ -116.4 ± 2.4‰ 4 Soil OM 3.6 ± 1.2‰ -22.4 ± 1.3‰ -105.7 ± 8.8‰ 2 Suspended POM 5.7 ± 4.0‰ -30.9 ± 0.8‰ -169.6 ± 7.1‰ 2

Table C.4. July 2013 potential nutritional resource stable isotope data. Source Mean δ15N ± S.D. Mean δ13C ± S.D. Mean δ2H ± S.D. N Algae 10.3 ± 1.5‰ -32.3 ± 2.5‰ -231.2 ± 27.8‰ 2 Terrestrial vegetation 4.2 ± 1.4‰ -30.2 ± 1.9‰ -129.7 ± 15.2‰ 5 Soil OM 3.6 ± 1.2‰ -22.4 ± 1.3‰ -105.7 ± 8.8‰ 2 Biofilm 12.2 ± 0.1‰ -26.1 ± 2.7‰ -198.8 ± 10.4‰ 2 Suspended POM 10.1 ± 0.6‰ -32.6 ± 0.3‰ ND 2 Sediment OM 1.9 ± 0.3‰ -29.2 ± 0.7‰ -116.4 ± 2.4‰ 4

Table C.5. November 2013 potential nutritional resource stable isotope data. Source Mean δ15N ± S.D. Mean δ13C ± S.D. Mean δ2H ± S.D. N Algae 15.1 ± 0.5‰ -31.1 ± 0.8‰ -231.2 ± 28.1‰ 2 Terrestrial vegetation 4.2 ± 1.4‰ -30.2 ± 1.9 -129.7 ± 15.2‰ 5 Soil OM 3.6 ± 1.2‰ -22.4 ± 1.3‰ -105.7 ± 8.8‰ 2 Sediment OM 1.9± 0.3‰ -29.2 ± 0.7‰ -116.4 ± 2.4‰ 4 Suspended POM 15.0 ± 0.6‰ -32.8 ± 0.3‰ -161.0 ± 1.7‰ 2 Biofilm 9.8 ± 1.6‰ -28.8 ± 0.1‰ -152.5 ± 27.3‰ 2

272

Table C.6. May and July 2013 mixing model sources. Source Mean δ15N ± S.D. Mean δ13C ± S.D. Mean δ2H ± S.D. N Algae 7.7 ± 3.8‰ -31.4 ± 1.2‰ -221.3 ± 14.0‰ 4 Terrestial veg. 4.2 ± 1.4‰ -30.2 ± 1.9‰ -129.7 ± 15.2‰ 5 Soil OM 3.6 ± 1.2‰ -22.4 ± 1.3‰ -105.7 ± 8.8‰ 2

Table C.7. November 2012 and 2013 mixing model sources. Source Mean δ15N ± S.D. Mean δ13C ± S.D. Mean δ2H ± S.D. N Algae 12.1 ± 4.3‰ -31.2 ± 0.1‰ -234.8 ± 5.1 4 Terrestrial veg. 4.2 ± 1.4‰ -30.2 ± 1.9‰ -129.7 ± 15.2‰ 5 Soil OM 3.6 ± 1.2‰ -22.4 ± 1.3‰ -105.7 ± 8.8‰ 2

Table C.8. Global mixing model sources including stable isotope and 14C data. Source Mean δ15N ± Mean δ13C ± Mean δ2H ± Mean ∆14C ± n S.D. S.D. S.D. S.D. Algae 9.0 ± 4.3‰ -31.0 ± 1.5‰ -228.2 ± 25.5‰ -52 ± 8‰ 8 Terr. veg. 4.2 ± 1.4‰ -30.2 ± 1.9‰ -129.7 ± 15.2‰ 39 ± 2‰ 5 Sediment 1.9 ± 0.3‰ -29.2 ± 0.7‰ -116.4 ± 2.4‰ -768 ± 211‰ 4 OM Soil OM 3.6 ± 1.2‰ -22.4 ± 1.3‰ -105.7 ± 8.8‰ -316 ± 121‰ 2 Suspended 11.4 ± 4.5‰ -32.3 ± 0.9‰ -161.0 ± 7.1‰ -82 ±13‰ 4 POM Biofilm 11.0 ± 2.4‰ -28.3 ± 0.9‰ -164.0 ± 27.7‰ -63 ± 8‰ 3

Table C.9. Global mixing model sources using stable isotope data only. Source Mean δ15N ± S.D. Mean δ13C ± S.D. Mean δ2H ± S.D. N Algae 9.2 ± 4.1‰ -31.0 ± 1.5‰ -228.2 ± 25.5‰ 9 Terrestrial veg. 4.2 ± 1.4‰ -30.2 ± 1.9‰ -129.7 ± 15.2‰ 5 Soil OM 3.6 ± 1.2‰ -22.4 ± 1.3‰ -105.7 ± 8.8‰ 2

273

274

Section 2

Results of one-way ANOSIMs that were performed to assess differences in isotopic composition (δ13C, δ15N, δ2H, and ∆14C) of macroinvertebrates across time and functional feeding group are provided in the following tables (Tables C.10-C.12).

Table C.10. One-way ANOSIM test of macroinvertebrates (δ13C, δ15N, and δ2H) at each sampling time point. ANOSIM R = 0.39, p = 0.001 Pairwise Testa R Statisticb Significance Levelc Nov. 2012/May 2013 0.390 0.002 Nov. 2012/July 2013 -0.079 0.828 Nov. 2012/Nov. 2013 0.278 0.002 May 2013/July 2013 0.608 0.001 May 2013/Nov. 2013 0.617 0.001 July 2013/Nov. 2013 0.353 0.003

Table C.11. One-way ANOSIM test of macroinvertebrates (δ13C, δ15N, and δ2H) between functional feeding groups (FFGs). ANOSIM R = 0.38, p = 0.001 Pairwise Testa R Statisticb Significance Levelc Predator/filtering collector 0.448 0.001 Predator/Anthopotamus 0.037 0.345 Predator/scraper 0.047 0.192 Filtering collector/Anthopotamus 0.500 0.001 Filtering collector/scraper 0.573 0.001 Anthopotamus/scraper -0.269 0.994

Table C.12. One-way ANOSIM test of macroinvertebrates (δ13C, δ15N, and δ2H) with the combined factor of functional feeding group and sampling date. ANOSIM R = 0.736, p=0.001 Pairwise Testa R Statisticb Significance Levelc Nov. 2012 predator/Nov. 2012 filtering collector 0.892 0.008 Nov. 2012 predator/Nov. 2012 Anthopotamus 0.400 0.190 Nov. 2012 predator/Nov. 2012 scraper 0.680 0.167 Nov. 2012 predator/May 2013 filtering collector 1.000 0.001 Nov. 2012 predator/May 2013 Anthopotamus 0.582 0.048 Nov. 2012 predator/May 2013 scraper 1.000 0.048 Nov. 2012 predator/July 2013 predator -0.164 0.714 Continued 275

Table C.12, continued Nov. 2012 predator/July 2013 scraper 0.272 0.107 Nov. 2012 predator/July 2013 scraper 0.727 0.048 Nov. 2012 predator/Nov. 2013 scraper 0.140 0.151 Nov. 2012 predator/Nov. 2013 Anthpotamus 0.000 0.333 Nov. 2012 predator/Nov. 2013 filtering collector 0.989 0.001 Nov. 2012 predator/Nov. 2013 predator 0.154 0.117 Nov. 2012 filtering collector/Nov. 2012 Anthpotamus 1.000 0.048 Nov. 2012 filtering collector/Nov. 2012 scraper 1.000 0.167 Nov. 2012 filtering collector/May 2013 filtering 0.888 0.001 collector Nov. 2012 filtering collector/May 2013 Anthpotamus 0.818 0.048 Nov. 2012 filtering collector/May 2013 Scraper 0.818 0.048 Nov. 2012 filtering collector/July 2013 predator 0.964 0.048 Nov. 2012 filtering collector/July 2013 scraper 0.846 0.018 Nov. 2012 filtering collector/July 2013 filtering 0.109 0.238 collector Nov. 2012 filtering collector/Nov. 2013 scraper 0.768 0.008 Nov. 2012 filtering collector/Nov. 2013 Anthpotamus 0.960 0.167 Nov. 2012 filtering collector/Nov. 2013 filtering 0.940 0.001 collector Nov. 2012 filtering collector/Nov. 2013 predator 0.340 0.016 Nov. 2012 Anthpotamus/Nov. 2012 scraper 1.000 0.333 Nov. 2012 Anthpotamus/May 2013 filtering collector 1.000 0.028 Nov. 2012 Anthopotamus/May 2013 Anthopotamus 1.000 0.333 Nov. 2012 Anthpotamus/May 2013 scraper 1.000 0.333 Nov. 2012 Anthpotamus/July 2013 predator 1.000 0.333 Nov. 2012 Anthpotamus, July 2013 scraper -0.167 0.800 Nov. 2012 Anthpotamus/July 2013 filtering collector 1.000 0.333 Nov. 2012 Anthpotamus/Nov. 2013 scraper -0.127 0.667 Nov. 2012 Anthpotamus/Nov. 2013 Anthpotamus 1.000 0.333 Nov. 2012 Anthpotamus/Nov. 2013 filtering collector 1.000 0.013 Nov. 2012 Anthpotamus/Nov. 2013 predator 0.130 0.333 Nov. 2012 scraper/May 2013 filtering collector 1.000 0.125 Nov. 2012 scraper/May 2013 Anthopotamus 1.000 0.333 Nov. 2012 scraper/May 2013 scraper 1.000 0.333 Nov. 2012 scraper/July 2013 predator 1.000 0.333 Nov. 2012 scraper/July 2013 scraper 0.111 0.500 Nov. 2012 scraper/July 2013 filtering collector 1.000 0.333 Nov. 2012 scraper/Nov. 2013 scraper 0.400 0.167 Continued 276

Table C.12, continued Nov. 2012 scraper/Nov. 2013 filtering collector 1.000 0.083 Nov. 2012 scraper/Nov. 2013 predator 0.660 0.125 May 2013 filtering collector/May 2013 Anthopotamus 1.000 0.028 May 2013 filtering collector/May 2013 scraper 1.000 0.028 May 2013 filtering collector/May 2013 predator 1.000 0.028 May 2013 filtering collector/July 2013 scraper 0.972 0.008 May 2013 filtering collector/July 2013 filtering collector 0.974 0.028 May 2013 filtering collector/Nov. 2013 Anthopotamus 1.000 0.125 May 2013 filtering collector/Nov. 2013 filtering 1.000 0.001 collector May 2013 filtering collector/Nov. 2013 predator 0.755 0.004 May 2013 Anthopotamus/May 2013 scraper 1.000 0.333 May 2013 Anthopotamus/July 2013 predator 1.000 0.333 May 2013 Anthopotamus/July 2013 scraper 0.000 0.5 May 2013 Anthopotamus/July 2013 filtering collector 0.750 0.333 May 2013 Anthopotamus/Nov. 2013 scraper 0.182 0.333 May 2013 Anthopotamus/Nov. 2013 Anthopotamus 1.000 0.333 May 2013 Anthopotamus/Nov. 2013 filtering collector 1.000 0.013 May 2013 Anthopotamus/Nov. 2013 predator 0.279 0.083 May 2013 scraper/July 2013 predator 1.000 0.333 May 2013 scraper/July 2013 scraper 0.667 0.200 May 2013 scraper/July 2013 filtering collector 1.000 0.333 May 2013 scraper/Nov. 2013 scraper 0.673 0.048 May 2013 scraper/Nov. 2013 Anthopotamus 1.000 0.333 May 2013 scraper/Nov. 2013 filtering collector 1.000 0.013 May 2013 scraper/Nov. 2013 predator 0.351 0.139 July 2013 predator/July 2013 scraper -0.083 0.600 July 2013 predator/July 2013 filtering collector 0.750 0.333 July 2013 predator/Nov. 2013 scraper 0.018 0.429 July 2013 predator/Nov. 2013 Anthopotamus 1.000 0.333 July 2013 predator/Nov. 2013 filtering collector 1.000 0.013 July 2013 predator/Nov. 2013 predator 0.175 0.222 July 2013 scraper/July 2013 filtering collector 0.500 0.200 July 2013 scraper/Nov. 2013 scraper -0.179 0.804 July 2013 scraper/Nov. 2013 Anthopotamus -0.778 1.000 July 2013 scraper/Nov. 2013 filtering collector 0.930 0.003 July 2013 scraper/Nov. 2013 predator 0.151 0.200 July 2013 filtering collector/Nov. 2013 scraper 0.636 0.048 July 2013 filtering collector/Nov. 2013 Anthopotamus 1.000 0.333 Continued 277

Table C.12, continued July 2013 filtering collector/Nov. 2013 filtering 1.000 0.013 collector July 2013 filtering collector/Nov. 2013 predator 0.331 0.083 Nov. 2013 scraper/Nov. 2013 Anthopotamus -0.480 1.000 Nov. 2013 scraper/Nov. 2013 filtering collector 0.898 0.001 Nov. 2013 scraper/Nov. 2013 predator 0.016 0.395 Nov. 2013 Anthpotamus/Nov. 2013 filtering collector 1.000 0.083 Nov. 2013 Anthpotamus/Nov. 2013 predator -0.401 0.875 Nov. 2013 filtering collector/Nov. 2013 predator 0.633 0.001

278

Section 3

Bayesian mixing model source contribution estimates (mean and S.D. as well as

5%, 50%, and 95% posterior probabilities) for each of the mixing models run are presented in Tables C.13-C.16.

Table C.13. Model run using δ13C, δ15N, and δ2H for May and July 2013. FFG and nutritional resource Mean SD 5% 50% 95% Anthopotamus/algae 0.551 0.164 0.277 0.558 0.819 Anthopotamus/soil OM 0.265 0.161 0.018 0.252 0.557 Anthopotamus/terrestrial vegetation 0.184 0.188 0.000 0.129 0.571 Filtering collector/algae 0.750 0.106 0.572 0.752 0.921 Filtering collector/soil OM 0.028 0.037 0.000 0.015 0.102 Filtering collector/terrestrial vegetation 0.222 0.112 0.044 0.217 0.414 Scraper/algae 0.637 0.110 0.454 0.640 0.827 Scraper/soil OM 0.314 0.111 0.130 0.313 0.501 Scraper/terrestrial vegetation 0.049 0.067 0.000 0.022 0.181 Predator/algae 0.311 0.182 0.050 0.288 0.667 Predator/soil OM 0.265 0.185 0.023 0.238 0.621 Predator/terrestrial vegetation 0.424 0.220 0.069 0.427 0.794

Table C.14. Model run using δ13C, δ15N, and δ2H for November 2012 and 2013. FFG and nutritional resource Mean SD 5% 50% 95% Anthopotamus/algae 0.411 0.083 0.279 0.407 0.558 Anthopotamus/soil OM 0.382 0.116 0.207 0.375 0.587 Anthopotamus/terrestrial vegetation 0.207 0.133 0.002 0.202 0.431 Filtering/collector algae 0.568 0.041 0.499 0.568 0.633 Filtering/collector soil OM 0.014 0.016 0.000 0.008 0.045 Filtering collector/terrestrial vegetation 0.418 0.044 0.345 0.420 0.489 Scraper/algae 0.478 0.066 0.372 0.477 0.589 Scraper/soil OM 0.458 0.080 0.328 0.457 0.584 Scraper/terrestrial vegetation 0.065 0.066 0.000 0.045 0.200 Predator/algae 0.435 0.075 0.318 0.431 0.563 Predator/soil OM 0.119 0.065 0.021 0.116 0.230 Predator/terrestrial vegetation 0.446 0.104 0.270 0.451 0.609

279

Table C.15. Model run using global data set and δ13C, δ15N, and δ2H values. FFG and nutritional resource Mean SD 5% 50% 95% Anthopotamus/algae 0.504 0.078 0.382 0.501 0.640 Anthopotamus/soil OM 0.338 0.079 0.213 0.337 0.470 Anthopotamus/terrestrial vegetation 0.158 0.111 0.000 0.155 0.346 Filtering collector/algae 0.674 0.060 0.578 0.672 0.775 Filtering collector/soil OM 0.014 0.018 0.000 0.007 0.051 Filtering collector/terrestrial vegetation 0.312 0.062 0.209 0.314 0.411 Scraper/algae 0.603 0.052 0.515 0.603 0.690 Scraper/soil OM 0.374 0.052 0.289 0.374 0.460 Scraper/terrestrial vegetation 0.023 0.032 0.000 0.010 0.089 Predator/algae 0.480 0.084 0.347 0.479 0.618 Predator/soil OM 0.111 0.065 0.016 0.105 0.227 Predator/terrestrial vegetation 0.409 0.110 0.229 0.408 0.584

Table C.16. Model run using δ13C, δ15N, δ2H, and ∆14C values. FFG and nutritional resource Mean SD 5% 50% 95% Anthopotamus/algae 0.527 0.132 0.349 0.511 0.753 Anthopotamus/soil OM 0.144 0.072 0.053 0.132 0.278 Anthopotamus/terrestrial vegetation 0.328 0.117 0.111 0.339 0.499 Filtering collector/algae 0.703 0.092 0.574 0.694 0.869 Filtering collector/soil OM 0.067 0.026 0.022 0.068 0.112 Filtering collector/terrestrial vegetation 0.230 0.074 0.096 0.236 0.337

280