The effects of dissolved organic carbon on pathways of energy flow, resource availability, and consumer biomass in nutrient-poor boreal lakes

By Joseph Tonin

A thesis submitted to the Faculty of Graduate Studies of

The University of Manitoba

In partial fulfillment of the requirement of the degree of

Master of Science

Department of Entomology

University of Manitoba

Winnipeg, Manitoba

Copyright © 2019 by Joseph Tonin

i

Abstract

Over the past few decades, terrestrially derived dissolved organic carbon (DOC) has been recognized as a fundamental driver of food web productivity in nutrient poor lakes. The mechanisms that underlie these effects remain poorly understood, particularly for higher trophic levels including zooplankton, benthic invertebrates, and fish. In a survey of eight lakes in northwestern Ontario, I determined consumer biomass and used stable isotopes of carbon, nitrogen, and hydrogen to investigate relationships between DOC and pathways of energy flow, resource and habitat availability, and consumer biomass. Using Bayesian stable isotope mixing models, I found that hypolimnetic phytoplankton were an important resource for zooplankton in low-DOC lakes. With increased DOC concentrations, light attenuation increased and chlorophyll a concentrations below the thermocline were reduced relative to epilimnetic concentrations. At higher DOC concentrations, zooplankton acquired proportionately more energy from low quality terrestrial sources. Zooplankton biomass also declined with increasing utilization of terrestrial sources (allochthony), suggesting that terrestrial organic matter suppresses zooplankton productivity through simultaneous limitations on habitat and resource availability and quality. Based on biomass, the dominant fish species across my study lakes was

White Sucker (Catostomus commersonnii). Bayesian mixing models indicated that allochthony by White Suckers increased with DOC and that greater allochthony was related to lower White

Sucker biomass measured as catch-per-unit-effort (bCPUE). Both White Sucker bCPUE and chironomid biomass were positively related to mean light irradiance, with the highest biomasses of fish and chironomids occurring in lakes with a higher proportion of their volume in the photic zone. White Sucker bCPUE was strongly and positively correlated with chironomid biomass,

ii suggesting that DOC-mediated resource limitation may influence fish productivity via reduced prey availability.

iii Acknowledgments

First, to Mike Paterson. I cannot thank you enough for providing me with this opportunity. I am grateful for everything you have done for me – from sending me to conferences and workshops to providing constructive comments and criticism. You always allowed me freedom in my work, but were always there to steer me in the right direction, and for that, I am thankful. Your passion for science rubs off on everyone you meet, and your patience and encouragement throughout this process has made me a better scientist. It has been a pleasure working with you, and I look forward to continuing our work together.

Thank you to my committee members, Dr. Mike Rennie, and Dr. Neil Holliday for keeping this project on track.

This thesis work is part of a large collaborative project Photons to Fish: Ecosystem

Indicators of fish productivity (PHISH) – special thank you to all collaborators that have made this project such a success and a joy to be a part of. Specifically, Mike Paterson, Mike Rennie, and Scott Higgins, who were instrumental in getting this project off the ground. To the

University of Waterloo, Rachel Henderson in particular, who supported everything isotope related. To Bryanna Sherbo, who was always there to provide field support and discuss the future of the project. Lastly, to Justin Budyk, thank you for your unrelenting positivity and hard-work in the field. Most importantly, thank you for being a great friend.

This thesis is largely of my own work, but does rely on contributions from others, which are summarized in Table at the end of the acknowledgements. This project was supported by

Manitoba Hydro, the IISD Experimental Lakes Area (IISD-ELA) graduate fellowship fund,

Mitacs, and Fish Futures Inc.

iv I was fortunate to be able to conduct my field-work at the world renowned IISD-ELA research facility. Not only did IISD-ELA provide a platform for me to conduct research, it provided me with an opportunity to explore and appreciate nature. I encourage anyone and everyone to consider donating to IISD-ELA to ensure it is well funded in perpetuity. The science conducted at IISD-ELA has a global impact, and in the face of large scale ecosystem change, there is a need for long-term monitoring programs and whole-ecosystem science to ensure the ongoing conservation and preservation of fresh water resources.

Of course, part of what makes IISD-ELA such a fantastic place to do freshwater science is the people. Without the help of IISD-ELA staff, this project would not have been possible.

Thank you to Lee Hrenchuk, Chandra Rodgers, Lauren Hayhurst, and Mike Rennie for always being there for anything and everything fish related. To Sonya Havens and Kelli-Nicole

Croucher, thank you for always being accommodating to my sampling schedule and for generating high quality data. To Ken Sandilands and Paul Fafard, you have always been generous to me, whether lending me equipment or sending me data, and for that I thank you.

Thank you to Mike Paterson, Scott Higgins, Vince Palace, and Craig Emmerton for always being there to talk about science – the conversations we had over a coffee break or a game of hoops were the source of inspiration for many of the ideas presented in this thesis or ideas I hope to explore in the future. Special thanks to operations staff, Roger Mollot and John Neal, for keeping camp running like a well-oiled machine, and to Jesse and Frank, who kept me well fed. Also, thank you to Stephen Paterson, who always made sure there was a bed available for me, the first aid room was well stocked, and for always saying yes to a canoe trip.

In addition to staff, there are countless summer students to thank. IISD-ELA runs on the backs of students, and each and every one of you has contributed to making this thesis happen.

v Whether it be collecting data for the long-term monitoring program, joining me on a canoe trip, or playing volleyball at the beach. My time in Winnipeg has been an enjoyable experience thanks in large part to the friendships I gained while working at IISD-ELA. Thank you all for making graduate school a little easier.

To my family (Team Tonin) – Mom, Dad, John, Brett and Ben. You all have been so supportive throughout this process even though you all have no idea what it is I do.

Lastly, I need to thank my partner, Ooma. This thesis is dedicated to you. Your love and encouragement has been a steadying presence in my life for the last six years, and without your support, I’m not sure I how I could have gotten through this. And for that, I thank you.

vi Attribution of work done by J. Tonin and others in this thesis:

Performed by J. Tonin Contributions from Others

• Monthly collection of stable isotope samples • Analysis of stable isotope samples

from all 8 study lakes in 2016 and 2017 (Environmental Isotope Laboratory,

(water, seston, periphyton, zooplankton, University of Waterloo for carbon and

benthic macroinvertebrates) nitrogen samples, and Colorado Plateau

• Processing of stable isotope samples Stable Isotope Laboratory, Northern Arizona

• Monthly collection of samples for University for hydrogen samples)

zooplankton, eplimnetic chemistry, • Collection of zooplankton, chemistry,

chlorophyll a, light, temperature and oxygen chlorophyll a, light, temperature and oxygen

profiles from Lakes 164, 658 in 2016 and profiles from Lakes 223, 224, 239, 373, 442,

2017, as well as zooplankton in Lakes 442 626 (IISD-ELA Hydro-Lim field crew)

and 373 in 2017 • Analysis of all water chemistry samples

• Collection of benthic macroinvertebrate (IISD-ELA Chemistry Laboratory)

samples from 5 lakes in 2016, and all lakes • Collection of fish from Lakes 223, 224, 239,

in 2017 373 and 626 for biomass catch per unit effort

• Processing, counting, and identification of all estimates (IISD-ELA Fish Crew); J. Tonin

benthic macroinvertebrate samples assisted with collection in 2016, and for

• Collection of fish from Lakes 164, 442, 658 L239 in 2017

for biomass catch per unit effort estimates • Counting and Identification of zooplankton

• Collection of fish stable isotopes from 6 of 8 (Plankton R Us)

lakes in 2017 • Chlorophyll a profiles (Bryanna Sherbo)

• Analysis and interpretation of all data

vii

Table of Contents ABSTRACT ...... ii ACKNOWLEDGMENTS ...... IV TABLE OF CONTENTS ...... VIII LIST OF TABLES ...... X LIST OF FIGURES ...... XI CHAPTER 1: GENERAL INTRODUCTION AND LITERATURE REVIEW ...... 1 INTRODUCTION ...... 1 BACKGROUND ...... 1 WHAT IS TERRESTRIAL ORGANIC MATTER, AND WHERE DOES IT COME FROM? ...... 4 EFFECTS OF TERRESTRIAL ORGANIC MATTER ON ECOSYSTEM STRUCTURE AND FUNCTION ...... 5 Lake Metabolism ...... 5 Food webs ...... 6 IS TERRESTRIAL OM A RESOURCE SUBSIDY FOR CONSUMERS IN NUTRIENT POOR BOREAL LAKES? ...... 9 Evidence of subsidy...... 9 Evidence of Suppression ...... 11 RESEARCH GAPS ...... 15 IMPORTANCE OF T-DOM RELATED RESEARCH ...... 16 CHAPTER 2: DOC-MEDIATED EFFECTS ON HABITAT STRUCTURE INFLUENCES BASAL RESOURCE USE AND BIOMASS OF ZOOPLANKTON ...... 24 ABSTRACT ...... 24 INTRODUCTION ...... 24 METHODS ...... 28 Study sites and limnological sampling ...... 28 Stable isotope ratios ...... 30 Data analysis ...... 32 RESULTS...... 36 DISCUSSION ...... 40 Evidence of deep phytoplankton resource use ...... 40 Effects of DOC on zooplankton biomass ...... 43 Limitations ...... 46 Conclusions ...... 48 CONNECTING STATEMENT ...... 60 CHAPTER 3: DOC-MEDIATED EFFECTS ON RESOURCE AVAILABILITY SUPPRESSES BENTHIVOROUS FISH BIOMASS ...... 61 ABSTRACT ...... 61 INTRODUCTION ...... 61 METHODS ...... 65 Study Sites and sample collection ...... 65 Fish and invertebrate biomass ...... 66 Resource Use ...... 67 Bayesian Mixing Model ...... 69 Statistical analysis ...... 71 RESULTS...... 71 Lake characteristics ...... 71

viii Invertebrate and White Sucker biomass ...... 72 Resource use...... 73 DISCUSSION ...... 75 CHAPTER 4: GENERAL DISCUSSION ...... 89 INSIGHTS ...... 89 FUTURE DIRECTIONS ...... 92 The need for multiple approaches to assess effects of DOC on lake food webs...... 92 The need for production estimates ...... 93 The emergence of novel food web tracers ...... 94 REFERENCES...... 97 APPENDIX 2A: CONTRIBUTIONS OF DIETARY WATER TO ZOOPLANKTON – SENSITIVITY ANALYSIS ...... 125 APPENDIX 2B: DATA USED IN ZOOPLANKTON BAYESIAN MIXING MODELS ...... 129 APPENDIX 2C: SEASONAL VERTICAL CHLOROPHYLL A PROFILES OF SURVEY LAKES IN 2018 ...... 139 APPENDIX 2D: ZOOPLANKTON BASAL RESOURCE USE BAYESIAN MIXING MODEL OUTPUTS ...... 143 APPENDIX 2E: VERTICAL Δ13C-DIC PROFILES ...... 145 APPENDIX 3A: BASAL ENERGY SOURCES AND LAKE CONSUMERS ISOTOPE SUMMARY ...... 146 APPENDIX 3B: SUMMARY OF LAKE SPECIFIC BENTHIC INVERTEBRATE BIOMASS AND DENSITY ESTIMATES ...... 149 APPENDIX 3C: SUMMARY OF MIXSIAR MIXING MODEL OUTPUTS ...... 152 APPENDIX 3D: PROFUNDAL CHIRONOMID DIETARY ESTIMATES ...... 159

ix

List of Tables

Table 1.1: Summary of potential effects of increasing terrestrially derived dissolved organic carbon on physical and chemical characteristics of lakes and corresponding effects on biota. Summarizes information from sources cited in text of literature review.……………………… 18

Table 1.2: Summary of whole-lake studies linking consumer productivity indicators to DOC concentration..……………………………………………………………………………..……. 19

Table 2.1: Select morphometric parameters and open-water averages of physical, chemical and biological characteristics of the study lakes, including: lake area, mean depth (Zmean), maximum depth (Zmax), epilimnetic dissolved organic carbon (DOC), light attenuation (Kd), photic depth, epilimnetic total phosphorus (TP), and epilimnetic chlorophyll a. Values in parenthesis are standard deviations…….………………………………………………………………………... 49

Table 2.2: Pairwise Pearson correlation coefficients of select chemical, physical and biological variables from Table 2.1, excluding fish species. Statistical significance is indicated by * (p £ 0.05; n = 8).……………………………………………………………………………………... 50

Table 3.1: Summary of benthic invertebrate, chironomid and, White Sucker biomass estimates among lakes. Littoral averages were weighted by area of each sampling depth < 4 m, and profundal was weighted by area of sampling depths below the thermocline…………………... 83

x List of Figures

Fig. 1.1: Hierarchical regulation of DOC concentration as described in Sobek et al. (2007)… 22

Fig. 1.2: Pathways of energy flow that support lake food webs. Heterotrophic and terrestrial organic matter energy pathways are drawn in brown, autotrophic pathways are drawn in green. Dashed lines represent pathways of energy flow that are only partly supported in the literature. After Creed et al. (2018)……………………………………………………………………….. 23

Fig. 2.1: Relationship between DOC concentration and (a) light attenuation, (b) total phosphorus, (c) epilimnetic chlorophyll a, and (d) mean light irradiance…………………………………... 51

Fig. 2.2: Patterns of habitat structure and carbon isotopes in relation to DOC concentration among the study lakes. (a) Differences between photic depth and thermocline depth decline as epilimnetic DOC concentration increase. (b) δ13C values of epilimnetic phytoplankton (grey triangles), hypolimnetic phytoplankton (black squares), terrestrial organic matter (black line), and zooplankton (open circles) in relation to epilimnetic DOC concentration in July of 2017. (c) The difference between δ13C of zooplankton and δ13C of surface phytoplankton (black circles) and difference between δ13C of hypolimnetic phytoplankton and surface phytoplankton in July 2017 (grey triangles) both become more similar to surface phytoplankton with increasing DOC concentration…………………………………………………………………………………… 52

Fig. 2.3: The vertical distribution of chlorophyll a across the study lakes in July 2017. Lakes are arranged in order of DOC concentration, from low to high (3.5 – 9.2 DOC mg/L). Black circles represent chlorophyll a, solid black lines represent the thermocline depth, dashed lines represent the photic depth, and solid red lines represent the depth of maximum dissolved oxygen concentration.……………………………………………………………………..……………. 54

Fig. 2.4: (a-e) Vertical distribution of suspended C : P ratios (by mass) and (f-j) chlorophyll a : suspended C (by mass) during the open water season (May – October) in 5 of the 8 study lakes. (k & l) Average epilimnetic suspended C : P and chlorophyll a : suspended C ratios among lakes. Boxplots represent the 5th, 25th, 50th, 75th and 95th percentiles of ratios throughout the open water season. Non-overlapping notches indicates significant difference at 95% confidence limits. Black points represent outliers. Lakes are arranged from low to high DOC concentration (left to right; 3.5 – 7.4 mg/L). Dashed lines indicate the photic depth. Note, data for L626 is from 2016 because a profile was not taken in 2017……………………………………………………….. 55

Fig. 2.5: Zooplankton allochthony (utilization of terrestrial resources) increases with epilimnetic th DOC concentration (Spearman rank correlation, rs = 0.79, p = 0.03). Boxplots represent the 5 , 25th, 50th, 75th and 95th percentiles of the posterior distribution from MixSIAR models…….... 57

Fig. 2.6: (a) Total zooplankton biomass declines with DOC concentration (r = –0.92, p = 0.001), and (c) increases with photic : thermocline depth ratio (r = 0.74, p = 0.03). (b) Copepod biomass declines with DOC concentration (r = –0.87, p = 0.004). (d) The relative abundance of zooplankton (% of biomass), arranged from low to high epilimnetic DOC concentration. Relative

xi abundance of each group did not vary predictably with DOC concentration (p > 0.05). Note, y- axis on panels a-c are log10 transformed……………………………………………………… 58

Fig. 2.7: Average annual zooplankton biomass of IISD–ELA lakes (n = 30) declines with epilimnetic DOC concentration (r = –0.62, p < 0.001). Black diamonds and grey circles are data from the extended analysis representing data collected by net or tube sampler, respectively. White triangles are data from the eight lake spatial survey……………………………………. 59

Fig. 3.1: DOC concentration and mean depth influence (a) mean light irradiance (mean light irradiance = 0.61 – 0.033 DOC – 0.024 Mean Depth; R2 = 0.93, p < 0.001) and (b) the proportion of lake area above the photic depth…………………………………………………………….. 84

Fig. 3.2: The relative abundance of benthic invertebrate (a) taxa and (b) functional feeding groups (FFG) at shallow sites (< 4 m) among lakes (arranged left to right by DOC concentration from low to high).……………………..……………………………………………………….. 85

Fig 3.3: (a) Average whole-lake chironomid biomass in relation to mean light irradiance (r = 0.85, p = 0.008), and (b) log10 White Sucker biomass catch-per-unit-effort in relation to average whole-lake chironomid biomass (r = 0.81, p = 0.01). Note: y-axis on panel b is log10 transformed ………………………………………….……………………………………………………… 86

Fig. 3.4: Patterns of allochthony for (a) littoral benthic invertebrate and (b) white sucker in relation to DOC concentration. Boxplots represent the 5th, 25th, 50th, 75th and 95th percentiles of the posterior distribution from MixSIAR models. (c) The relationship between White Sucker biomass catch per unit effort (bCPUE) and allochthony (r = –0.82, p = 0.01). Allochthony estimates are presented as medians from the posterior distribution of the Bayesian mixing model. Note: the y-axis is log10 transformed…………………………………………………………... 87

xii Chapter 1: General Introduction and Literature Review

Introduction

Inputs of terrestrial dissolved organic matter (t-DOM) to nutrient-poor lakes fundamentally shape their physical and chemical characteristics and biological interactions

(Jones, 1992; Karlsson et al., 2009; Prairie, 2008; Solomon et al., 2015). Terrestrial carbon can account for upwards of 95% of the DOM pool (Wilkinson et al., 2013), and t-DOM inputs are increasing in many lakes in the northern hemisphere coinciding with changes in temperature, hydrological processes, land use, and reductions in acid deposition (Emmerton et al., 2018;

Evans et al., 2005; Finstad et al., 2016; Monteith et al., 2007; Paterson et al., 2019). While t-

DOM inputs have been increasingly recognized as a driver of food web productivity in nutrient poor lakes, the mechanisms by which these inputs influence food webs are poorly understood.

The main objective of this thesis was to understand the extent to which inputs of t-DOM may affect consumer productivity in recipient ecosystems and the mechanisms by which t-DOM affects productivity. Specifically, I investigated how t-DOM concentration influenced the relative importance of different pathways of energy flow (i.e. how important are terrestrial inputs to consumers?) and investigated the linkages between proxies of fish and invertebrate productivity and their relative utilization of different energy pathways (i.e. does terrestrial organic matter subsidize or suppress consumer productivity?).

Background

Freshwater fish production provides humans with important ecosystem services, and creates nutritional, economic, cultural, and recreational value throughout the world (Brooks et al., 2016; Holmlund & Hammer, 1999). These benefits are particularly apparent in developing

1 nations and indigenous communities, where freshwater fish are a vital food source, and are capable of driving economic development (Beard et al., 2011; Lynch et al., 2016; Smith et al.,

2005). In developed nations, the significance of freshwater fish shifts towards recreation (Cooke

& Cowx, 2004; Post et al., 2002; Post, 2013), where, in Canada for example, recreational fisheries contribute an average of $8.8 billion in revenue each year (Brownscombe et al., 2014).

However, many freshwater fisheries are in decline, coinciding with the ongoing expansion of industrial, urban and agricultural activities (Schindler, 2001). There is a need for better understanding of what factors drive variations in fish productivity, both within and among lakes, to allow for the development of better predictive models of the effects of potential human activities on aquatic ecosystems.

Fish production in lakes is ultimately dependent upon the amount of energy available at the base of the food web, and the efficiency of energy transfer up the food chain (Brander, 2007;

Downing & Plante, 1993). Nutrient limitation has been viewed as the main driver of lake productivity (Carpenter, 2008; Schindler, 1977; Sterner, 2008), where increases in nutrient inputs increase primary production (phytoplankton), with cascading effects up to higher trophic levels.

This paradigm is strongly supported in the literature, where research suggests that the strongest predictors of lake productivity, from phytoplankton to fish, are nutrient concentrations (Downing et al., 1990; Downing & Plante, 1993; Hanson & Leggett, 1982). However, when testing the applicability of the nutrient limitation paradigm in small, nutrient-poor lakes (< 30 µg/L total phosphorus), Karlsson et al. (2009) found no relationship between nutrients and fish production.

Increasing evidence suggests that terrestrial organic matter may play a previously underappreciated role in governing lake productivity in small, nutrient-poor lakes (Karlsson et al., 2009).

2 Terrestrial organic matter inputs are primarily composed of humic substances that strongly absorb photosynthetically active radiation (PAR) and the amount of light available to primary producers (Jones, 1992). The majority of the lakes worldwide are small and nutrient poor, with low phytoplankton biomass (Downing & Duarte, 2006). Here, littoral algae can dominate whole-lake primary production and littoral energy can substantially support production at higher trophic levels (Karlsson & Byström, 2005; Vadeboncoeur et al., 2008; Vadeboncoeur et al., 2002; Vander Zanden & Vadeboncoeur, 2002). Light-limitation induced by terrestrial organic matter may suppress benthic and pelagic primary production, with concomitant effects on higher trophic levels (Karlsson et al., 2009).

The effects of terrestrial organic matter on lake productivity is not straightforward. In addition to altering littoral primary production, terrestrial organic matter may also influence food web productivity through its effects on habitat availability or may be used directly as an alternative source of energy (Carpenter et al., 2005; Craig et al., 2015). Prairie (2008) argues that terrestrial organic matter is a “modulator,” a variable that can modify the effect of other variables. It affects biota and chemistry directly and indirectly through its effects on the physical environment (Table 1.1). The complex and interacting effects of terrestrial organic matter are poorly understood, particularly its effects on fish populations (Bartels et al., 2016).

Inputs of terrestrial organic matter into lake ecosystems have been increasing in many boreal lakes for decades (Evans et al., 2005; Finstad et al., 2016; Monteith et al., 2007) and many human activities, such as forestry and hydroelectric development, affect these inputs (Adams et al., 1983; Kreutzweiser et al., 2008; Paterson et al., 2019). Hence, it is imperative to understand how changes in these inputs affect the productivity of lake food webs.

3 What is terrestrial organic matter, and where does it come from?

The largest input of terrestrial organic matter to most lakes is in the dissolved form

(hereafter referred to as t-DOM) and accounts for upwards of 95% of the dissolved organic matter pool in the majority of lakes (Wilkinson et al., 2013). Terrestrial-DOM is formed during the decomposition of plant material and is comprised of a complex mix of different organic materials including cellulose and lignin. These materials are modified by physical and biological interactions in the soil, with factors such as temperature, pH and moisture influencing solubility and microbial activity (Solomon et al., 2015). These interactions alter the chemical composition of t-DOM so that it comprises of a suite of substances that differ in size and bioavailability (Neff

& Asner, 2001). The majority of t-DOM is humic and fulvic acids that strongly absorb ultraviolet and photosynthetic active radiation, giving water a stained, brown colour, that affects light penetration and heat retention (Jones, 1992). These substances are resistant to microbial breakdown due to high carbon to nitrogen (C:N) and carbon to phosphorus (C:P) ratios and are primarily broken down by photochemical processes (Berggren et al., 2018).

Terrestrial-DOM is exported from watersheds to lakes primarily by surface or subsurface runoff, which is intensified during precipitation events (Cole et al., 2007; Hinton et al., 1997).

The amount of water entering a watershed is an important determinant of the amount of t-DOM available for export, where wetter soils lead to greater organic matter accumulation and production of leachable organic compounds (Freeman et al., 2001). In addition to hydrology, watershed characteristics such as vegetation and soil type are important contributors influencing t-DOM production (Hinton et al., 1997; Xenopoulos et al., 2003).

Concentrations of t-DOM in lakes can vary considerably both spatially and temporally and are regulated in a hierarchical manner (Fig. 1.1; Sobek et al., 2007). At a regional level,

4 climate and topography set the range of possible t-DOM concentrations in lakes by affecting terrestrial organic matter production and mobilization, as well as water yield. For example, the cool climate of boreal regions promotes t-DOM export because terrestrial productivity is high, and soils are often organic-rich and waterlogged, whereas in Arctic regions, terrestrial productivity is low and the mobility of t-DOM is limited by permafrost (Sobek et al., 2007). At the individual lake scale, catchment and lake properties dictate t-DOM concentration. Factors such as the proportion of wetlands in the catchment, drainage ratio, water retention time, and lake size are strong predictors of t-DOM concentration in individual lakes (Laudon et al., 2011;

Sobek et al., 2007; Xenopoulos et al., 2003). Climate change may cause increases of t-DOM export in boreal lakes at the regional scale due to its effects on hydrologic cycling and temperature, and through increased terrestrial primary production (Larsen et al., 2011; Wu et al.,

2011).

Effects of terrestrial organic matter on ecosystem structure and function

Lake Metabolism

The humic and fulvic compounds in t-DOM strongly influence light attenuation by absorbing solar radiation, reducing the amount of available light at a given depth (Fee et al.,

1996; Williamson et al., 1996b). Consequently, t-DOM affects where and how much primary production can be supported in a given lake (Godwin et al., 2014; Karlsson et al., 2009;

Vadeboncoeur et al., 2008). Several authors have demonstrated that t-DOM induced light- limitation has profound effects on benthic primary production, but its effect on pelagic primary production is less certain (Ask et al., 2009a; Godwin et al., 2014; Karlsson et al., 2009). This is because t-DOM has the potential to stimulate pelagic primary production through its effects on

5 nutrient availability (Kissman et al., 2013), in addition to its negative effects on light penetration.

Inputs of limiting nutrients in boreal lakes are closely associated with t-DOM inputs, and addition of these nutrients may stimulate primary producers (Dillon & Molot, 2005; Vasconcelos et al., 2018). Further, t-DOM may alleviate competition for nutrients between pelagic and littoral primary producers, as light-limited benthic communities have a lower capacity to intercept nutrients diffusing from sediment (Vasconcelos et al., 2016; Vasconcelos et al., 2018).

Terrestrial-DOM can also screen primary producers from harmful ultraviolet radiation, potentially preventing cell damage (Williamson & Rose, 2010; Williamson et al., 1996b).

While t-DOM may benefit pelagic primary producers by providing a nutrient subsidy and screening ultra-violet radiation, t-DOM also serves as a substrate for bacteria, which may outcompete phytoplankton for limiting nutrients (Blomqvist et al., 2001; Vadstein, 2000). As t-

DOM concentrations increase, dominance of pelagic energy mobilization may shift from phytoplankton to heterotrophic bacteria. Consequently, lakes may shift from being carbon sinks to carbon emitters with increasing t-DOM (Ask et al., 2009a, 2009b; Karlsson et al., 2012).

Terrestrial-DOM modifies thermal regimes in lakes, with more heat being concentrated at the surface, producing shallower, more stable epilimnia, and declines in whole-lake average temperature (Prairie, 2008; Read & Rose, 2013; Tanentzap et al., 2008). Such changes to thermal regimes can also influence the vertical distribution of oxygen and nutrients, which can affect biochemical reactions and habitat availability for aerobic organisms (Craig et al., 2015; Orihel et al., 2017; Prairie, 2008).

Food webs

In addition to altering the relative availability of pelagic and benthic primary production, t-DOM itself is an energetic input (Fig. 1.2). In many boreal lakes, inputs of t-DOM may be

6 several orders of magnitude greater than energy generated within the lake (Algesten et al., 2003;

Jansson et al., 2008). Over the past few decades, several studies using stable isotopic approaches have demonstrated that significant portions of aquatic consumer biomass can be supported by terrestrially derived sources (Cole et al., 2006, 2011; Grey et al., 2001; Karlsson et al., 2015;

Solomon et al., 2011; Tanentzap et al., 2014, 2017). This has led many researchers to suggest terrestrial organic matter may be a resource subsidy in lakes, evidenced by increasing utilization of terrestrial resources by consumers with increasing t-DOM concentration (Carpenter et al.,

2005; Cole et al., 2006; Polis et al., 1997; Solomon et al., 2011; Tanentzap et al., 2014).

However, a resource subsidy is defined as a cross-ecosystem input of energy that increases consumer productivity within the recipient ecosystem (Jones et al., 2012; Kelly et al., 2014; Polis et al., 1997). While it is clear bacterial productivity benefits from t-DOM inputs (Ask et al.,

2009a; Karlsson et al., 2012), evidence for t-DOM as a subsidy to higher consumers is contradictory (Brett et al., 2017; Tanentzap et al., 2017) and some evidence even suggests t-

DOM may suppress productivity (Karlsson et al., 2015; Kelly et al., 2014). Studies linking t-

DOM to consumer productivity are summarized in Table 1.2.

These contradictory interpretations have been difficult to resolve because most studies consider only the relative utilization of t-DOM by consumers (allochthony, i.e. percent contribution of terrestrial organic matter to biomass) rather than absolute effects; that is, the effects of t-DOM on consumer productivity. The main objective of this thesis was to understand the extent to which inputs of t-DOM may affect consumer productivity in recipient ecosystems and the mechanisms by which t-DOM affects variations in productivity.

My project is part of a larger interdisciplinary project at IISD-ELA that is assessing drivers of fish productivity with the ultimate goal of developing and evaluating tools for cost effective

7 determination of ecosystem health and fish productivity. I investigated the effects of t-DOM on secondary consumers, including invertebrates and fish. My specific objectives were:

1. Determine the relative importance of different sources of carbon (C) supporting

invertebrate and fish populations in IISD-Experimental Lakes Area (IISD-ELA) lakes

(Fig.1.2). Using stable isotopes of H, C and N with Bayesian mixing models, the relative

utilization) of terrestrial organic matter by consumers within each lake can be discerned

(i.e. allochthony; percent contribution of terrestrial organic matter to consumer biomass

and littoral vs planktonic primary production (autochthony))

2. Determine linkages between proxies of fish and invertebrate productivity to their relative

utilization of different energy pathways (i.e. does allochthony predict consumer

productivity?). If consumer productivity increases with allochthony, then terrestrial

organic matter would be considered a resource subsidy.

To address these objectives, several questions were posed relating to how food web structure and productivity change with t-DOM concentration:

• Is resource use by invertebrates and fish correlated with the relative availability of basal

resources?

• How is the utilization of different C sources affected by t-DOM, lake morphometry and

water chemistry?

• What are the mechanistic links between t-DOM concentration and fish that drive

variations in productivity?

• How does t-DOM affect habitat availability?

• Is t-DOM a resource subsidy to consumers in boreal lakes?

8 To investigate these questions, stable isotopes of H, C and N were used to investigate resource use and were compared to proxies of food web productivity among lakes spanning a gradient of dissolved organic carbon (DOC) at the IISD-Experimental Lakes Area (IISD-ELA) in northwestern Ontario. Because DOC in boreal lakes is predominately from terrestrial ecosystems

(Cole et al., 2007; Wilkinson et al., 2013), I assume that this DOC gradient is representative of the amount of terrestrial organic matter potentially available to consumers. This relationship has been previously established in studies by Karlsson et al. (2012) and Karlsson et al. (2015).

Is terrestrial OM a resource subsidy for consumers in nutrient poor boreal lakes?

Evidence of subsidy

Terrestrial resources can be incorporated into food webs directly by consuming particulate terrestrial organic matter, or indirectly via microbial assimilation of t-DOM (Berglund et al., 2007). Much of the evidence supporting the t-DOM subsidy hypothesis has focused on extreme clear-water systems (Finstad et al., 2014; Kissman et al., 2013; Tanentzap et al., 2014) or lakes following a disturbance to and/or manipulation of t-DOM inputs (Bertolo & Magnan,

2007; Kelly et al., 2016; Paterson et al., 1997). Terrestrial-DOM may subsidize consumers under two scenarios: (1) by providing an additional energy source, or (2) by subsidizing pelagic primary producers by increasing the supply of limiting nutrients (Vasconcelos et al., 2018).

Under both scenarios, the mechanism regulating food web productivity is related to basal energy supply; a process termed ‘trophic upsurge.’

In a study of eight freshwater deltas, bacterial productivity increased with t-DOM concentration, leading to higher zooplankton biomass and faster growth rates in young-of-year

Yellow Perch (Tanentzap et al., 2014). Terrestrial resource use by young-of-year (YOY) perch

9 also increased with t-DOM availability. Similarly, in an experimental reservoir at the IISD-ELA, bacterial biomass increased dramatically with t-DOM inputs, with concomitant effects on

Daphnia production (Paterson et al., 1997). Production of Daphnia often exceeded that of phytoplankton, indicating that consumption of allochthonous resources via bacteria acted as an important resource subsidy. In a study by Bertolo and Magnan (2007), logging-induced changes in t-DOM concentration stimulated the microbial loop, which had a positive effect on YOY

Yellow Perch abundance. In each of these examples, t-DOM appeared to drive a trophic upsurge.

Terrestrial-DOM induced nutrient subsidies stimulate lake food webs as would be predicted by the nutrient limitation paradigm; where an increase in nutrient supply (via t-DOM) will stimulate primary production, with concomitant effects on higher trophic levels. In a microcosm experiment simulating the effects of increased t-DOM inputs on alpine lakes and a temperate whole-lake t-DOM manipulation, t-DOM bound nutrients stimulated phytoplankton production, leading to increased zooplankton productivity (Kelly et al., 2016; Kissman et al.,

2013).

In addition to providing increasing basal energy supply, t-DOM inputs may stimulate productivity through a number of indirect mechanisms. First, t-DOM may stimulate productivity by screening harmful UV radiation (Williamson & Rose, 2010). Indeed, t-DOM can protect phytoplankton and fish eggs from harmful UV-radiation (Stasko et al., 2012; Williamson &

Rose, 2010; Williamson et al., 1996b) and reduce physiological, UV induced stress (Manek et al., 2014). Second, t-DOM decreases water transparency and can affect the foraging ability of visual feeders. This has been demonstrated to benefit walleye, a light-sensitive predator, where they acquire more energy from littoral habitats with decreasing visual conditions, ultimately leading to higher biomass (Tunney et al., 2018).

10 Evidence supporting the t-DOM subsidy hypothesis is primarily focused on pelagic energy pathways, despite the fact that littoral resources can support substantially higher trophic level productivity (Vadeboncoeur et al., 2002; Vander Zanden & Vadeboncoeur, 2002; Vander

Zanden et al., 2011), and littoral energy pathways are more susceptible to t-DOM mediated light- limitation than pelagic energy pathways (Godwin et al., 2014; Hanson et al., 2003;

Vadeboncoeur et al., 2008). Thus, a major limitation of the t-DOM subsidy hypothesis is a lack of dedicated studies linking t-DOM inputs to littoral consumer productivity (but see Craig et al.,

2015). Additionally, both nutrient inputs and light extinction are positively related to t-DOM inputs. This creates a t-DOM threshold, below which t-DOM bound nutrients may increase primary production, and above which t-DOM induced light-limitation reduces primary production (Seekell et al., 2015a, 2015b). It is unclear where the studies listed above fall within this threshold because they are typically in extreme-clear water systems. Further, it is unclear how thresholds will differ in manipulated ecosystems that are in a non-equilibrium state following perturbation (Turgeon et al., 2016).

Evidence of Suppression

Although t-DOM potentially provides additional energy for consumers, several studies suggest that t-DOM inputs can suppress zooplankton (Brett et al., 2009; Kelly et al., 2014), benthic invertebrates (Craig et al., 2015; Jones et al., 2012), and fish productivity (Benoît et al.,

2016; Finstad et al., 2014; Karlsson et al., 2009, 2015; Rask et al., 2014). These reductions in food web productivity have been attributed mainly to three mechanisms: effects of t-DOM on resource availability, resource quality, and habitat availability.

11 Several studies suggest that t-DOM can impose resource limitations on secondary consumers through its effects on basal primary production (Karlsson et al., 2009). Increasing concentrations of t-DOM may suppress in-lake primary production by attenuating light and by affecting nutrient dynamics, particularly for producers in littoral zones (Ask et al., 2009a;

Godwin et al., 2014; Karlsson et al., 2009; Seekell et al., 2015a; Vasconcelos et al., 2016). It is unclear whether higher inputs of terrestrial resources can compensate for the loss of autochthonous energy pathways.

Reliance on terrestrial resources is low relative to their availability within lakes, indicating that terrestrial resources cannot fully replace autochthonous resources and further suggesting that there may be an upper limit to the degree to which terrestrial organic matter can support freshwater food webs (Karlsson et al., 2012). Terrestrial resources may operate as a

“metabolic lifeboat,” where consumers use these resources for survival during times of autochthonous resource deficiency (Wetzel, 1995). Indeed, zooplankton reliance on terrestrial resources increases during winter when autochthonous resource availability is low (Berggren et al., 2015) and several laboratory studies have also shown that terrestrial resources can support consumers when phytoplankton resources are low, albeit resulting in lower growth and reproduction rates (Brett et al., 2009; Hiltunen et al., 2017; McMeans et al., 2015; Taipale et al.,

2016a). Accordingly, increased utilization of terrestrial resources (allochthony) has been linked to declines in zooplankton and fish productivity based on spatial t-DOM gradient surveys

(Karlsson et al., 2015; Kelly et al., 2014). Littoral algae are an important resource for benthic invertebrates (Hecky & Hesslein, 1995) and t-DOM mediated reductions in the strength of littoral energy pathways have been linked to suppression of benthic invertebrate productivity

(Karlsson et al., 2009). Benthic invertebrates are, directly or indirectly, the primary prey for most

12 fish species (Vander Zanden & Vadeboncoeur, 2002), therefore, factors affecting benthic invertebrate productivity can have consequential effects on fish production. Indeed, productivity of several fish species has been linked to the strength of the littoral energy pathways (Craig et al.,

2015; Guzzo et al., 2017; Karlsson et al., 2009).

The effect of t-DOM mediated resource limitation is further magnified by its effects on resource quality and trophic transfer efficiency. Littoral energy is mobilized more efficiently up the food web than pelagic and terrestrial resources, in part, due to larger prey and shorter food chain lengths (Guzzo et al., 2017; Karlsson & Byström, 2005; Pimm & Kitching, 1987).

Terrestrial resources have high respiratory losses during energy transfer because t-DOM requires bacterial conditioning before it can be assimilated into the food web (Berglund et al., 2007).

Bacterial conditioning is inefficient because terrestrial carbon is recalcitrant and resistant to microbial degradation, thus much of the processed carbon is respired rather than assimilated.

Further, bacteria conditioning adds additional trophic steps between t-DOM and consumers, increasing food chain lengths (Berglund et al., 2007; Kritzberg et al., 2005). As such, the trophic transfer efficiency of terrestrial resources is less efficient than for autochthonous resources (Brett et al., 2017). Thus, suppression of littoral energy pathways and increased utilization of terrestrial and pelagic resources with increasing t-DOM concentration can decrease the efficiency in which energy is transferred up the food web.

Fish and invertebrates are strongly reliant on the fatty acids eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) to maintain nutritional physiology and growth (Brett &

Muller-Navarra, 1997; Muller-Navarra et al., 2000). These animals are unable to synthesize EPA and DHA fatty acids, and must acquire them through their diet (Sargent et al., 1999). In general, phytoplankton are rich in EPA and DHA, whereas terrestrial organic matter is nearly devoid of

13 these fatty acids (Brett et al., 2009; Taipale et al., 2012). Trophic upgrading (i.e. enhancing quality of resource) of t-DOM through the microbial food web can increase the availability of

EPA and DHA, but not in sufficient quantities to replace phytoplankton (Hiltunen et al., 2017).

Consequently, consumers tend to have lower proportions of these essential fatty acids in their tissues with increasing allochthony (Taipale et al., 2016a), and t-DOM concentration (Taipale et al., 2016c).

The concertation of t-DOM can also affect the amount of suitable habitat for lake consumers. As noted above, t-DOM concentrates heat at the lakes surface, leading to stronger and shallower epilimnia that are resistant to mixing and may affect the vertical distribution of oxygen and temperature. Greater resistance to mixing coupled with increased microbial respiration often leads to longer and more frequent periods of anoxia in high t-DOM lakes

(Brothers et al., 2014; Couture et al., 2015). Indeed, loss of optimal oxythermal habitat for Lake

Trout (Salvelinus namaycush) is more prevalent in higher t-DOM lakes (Dillon et al., 2003;

Guzzo & Blanchfield, 2017). For benthic invertebrates, t-DOM mediated effects on oxygenated habitat had stronger effects on secondary productivity than that of resource limitation in a study by Craig et al. (2015). In an extreme example, two consecutive flooding years resulted in t-DOM concentration increasing from 12 to 56 mg/L, leading to severe anoxic conditions, with near complete loss of fish and benthic invertebrate populations (Brothers et al., 2014). While certain biota are capable of surviving during prolonged periods of anoxia, this occurs at the expense of growth, thus reducing productivity (Heinis & Davids, 1993; Jonasson, 1984).

14 Research gaps

The extent to which terrestrial organic matter subsidizes or suppresses lake food webs remains poorly understood. The majority of studies consider only the relative usage of terrestrial organic matter by consumers (i.e. allochthony), rather than absolute effects on productivity.

Alternatively, studies that measure secondary productivity often do not concurrently estimate allochthony directly, instead relying on proxies such as light extinction and/or t-DOM concentration (e.g. Benoît et al., 2016). There is a need to determine linkages between estimates of consumer productivity and their relative utilization of energy pathways (i.e. does allochthony predict consumer productivity?) to further understand how upper trophic levels may be affected by large scale changes to t-DOM inputs (but see: Craig et al., 2015; Karlsson et al., 2015; Kelly et al., 2014).

Because the availability of benthic prey is an important link between t-DOM concentration and fish productivity, it is surprising how few studies have assessed the effects of t-DOM on benthic invertebrate productivity. The first study to consider this mechanistic link found that t-DOM suppressed benthos biomass via its effects on benthic primary production, ultimately leading to declines in fish production (Karlsson et al., 2009). However, these results were based on rudimentary estimates of benthic productivity (3 samples per lake, taken at fixed depths), that did not account for their patchy distribution, or many abiotic factors (e.g. lake morphometry, dissolved oxygen, etc.) known to affect their productivity. A second study, with more robust benthic estimates found that t-DOM suppresses benthic invertebrates via its effect on oxygenated habitat, with concomitant declines in benthivorous fish catch-per-unit-effort

(Craig et al., 2015). However, these catch data are based on six lakes with an unknown amount of effort and did not account for other potential drivers of fish productivity such as habitat

15 availability, resource use, and fishing pressure. Therefore, given the shortcomings of these two studies, there is a need for more robust estimates of both benthic invertebrate and fish productivity and models to account for the several mechanisms that may drive variations in productivity.

Importance of t-DOM related research

While concentrations of t-DOM are increasing in many boreal and temperate lakes

(Finstad et al., 2016; Monteith et al., 2007), there are contradictory interpretations as to the effects of increasing t-DOM on lake productivity (Brett et al., 2017; Tanentzap et al., 2017).

Understanding how t-DOM influences fish productivity has important management implications to ensure conservation and protection of fish and their habitat. First, this research can guide monitoring programs and fisheries managers to better assess the long-term impacts of increasing t-DOM inputs on fish communities. For example, if fish productivity declines with increasing t-

DOM, sustainable catch limits may need to be reduced in order to ensure the rate of exploitation does not exceed that of fish growth and recruitment. Further, t-DOM inputs may affect the quality of fish for human consumption; fish in high t-DOM lakes have lower concentrations of essential fatty acids, and tend to have higher concentrations of contaminants, such as mercury

(Lescord et al., 2018; Taipale et al., 2016c). Inputs of t-DOM may have substantial effects on valuable commercial, recreational and sustenance fisheries, and a poor understanding of the mechanisms that drive variations in productivity makes it difficult to plan for and manage in a changing environment.

Second, a better understanding of where and how increasing t-DOM inputs modify freshwater ecosystems will improve our ability to predict how land-use and climate change may

16 affect fish productivity. An understanding of the dependence of aquatic food webs on different resources in different systems is fundamental to almost all predictions concerning the effects of natural and anthropogenic activities. There is a need to provide policy-relevant science to help conserve and protect fish and their habitat, in the face of large-scale ecosystem change.

17 Table 1.1: Summary of potential effects of increasing terrestrially derived dissolved organic

carbon on physical and chemical characteristics of lakes and corresponding effects on biota.

Information sources are cited in the text of literature review.

Physical and Consequences of increasing DOC Corresponding effects on biota Chemical characteristics • Shallower thermocline • Changes in thermal habitat • Higher epilimnion temperature; • Changes in invertebrate and fish more heat being retained in surface metabolism Thermal structure waters • More stable epilimnia; greater resistance to mixing • Lower hypolimnetic temperatures caused by a shorter period of overturn in spring

• Depleted hypolimnetic oxygen • Loss of oxygenated habitat for O2 caused by more prolonged sensitive taxa; habitat gain for Oxygen stratification and a shorter period of methanotrophic bacteria overturn in spring • Release of nutrients bound in • Greater bacterial respiration due to sediments, potentially stimulating increased organic matter inputs primary producers

• Increased absorption of light and • Reduced benthic primary production, UV radiation and potentially pelagic producers • Declines in essential fatty acids available to consumers because of reduced primary producer biomass Light • Loss of macrophyte cover for spawning, refuge and/or feeding • Changes in feeding efficiency for visual feeders • Loss of hypolimnetic primary production • Reduction in UV-induced physiological stress

• Greater concentrations of organic • Change in phytoplankton stoichiometry carbon; replacement of (higher C:P, and C:N ratios); poorer autochthonous carbon with food quality for consumers Nutrients allochthonous carbon • Increased epilimnetic primary • Increased total nutrient pool but production reduction in bioavailable nutrients

18 Table 1.2: Summary of whole-lake studies linking consumer productivity indicators to DOC

concentration.

Study Design Study organism(s) DOC related effects

Positive response to DOC

Manipulation Zooplankton Post flooding, zooplankton biomass, in Impoundment of small boreal lake particular, cladocerans, increased 10-fold. resulting in flooding of surrounding Zooplankton production exceeded that of peatland and increases in DOC primary production, suggesting allochthonous Paterson et al., 1997 carbon inputs were responsible for this increase in zooplankton biomass.

Spatial Survey Perca flavescens Logging of watersheds led to increases in DOC 9 boreal lakes with varying degree concentration. Recruitment of young-of-year of watershed area logged (1-78%) Yellow Perch increased proportionately to DOC range: 4.1-14.9 mg/L changes in DOC concentration. Bertolo & Magnan, 2007

Spatial Survey Zooplankton, The authors found support for the trophic 8 freshwater deltas in Daisy Lake, Perca flavescens upsurge hypothesis, where increasing Ontario Canada terrestrial organic matter supported greater DOC range: 2.4 – 3.4 mg/L bacteria biomassand larger zooplankton, which Tanentzap et al., 2014 ultimately enhanced fish growth.

Manipulation Zooplankton Over the duration of the manipulation, Divided small temperate lake with zooplankton density slightly increased in impermeable curtain – elevated response to an increase in DOC concentration, DOC concentration from 8 to 11 which was likely a result of increased primary mg/L production. Kelly et al., 2016

Manipulation Micropterus salmoides Led to increased density of Largemouth Bass, Divided small temperate lake with however, there were no observed changes in impermeable curtain – elevated Bass productivity, growth or condition. DOC concentration from 8 to 10.2 mg/L Koizumi et al., 2018

Negative response to DOC

Spatial survey Salvelinus alpinus, In this 12-lake spatial survey, fish production 12 boreal lakes in Northern Sweden Perca fluviatilis was positively related with benthic primary DOC range: 1.5 – 16.8 mg/L production and benthos biomass, which was & correlated with mean light irradiance. 33 lakes in Sweden and Finland Karlsson et al., 2009

19 In a 33-lake spatial survey, similar correlations between fish production and light were found, highlighting how DOC-mediated effects on light, coupled with lake depth can influence fish production in nutrient poor lakes.

Spatial Survey Perca flavescens Abundance of Walleye declined with 59 lakes in Southern Quebec, Sander vitreus increasing DOC concentrations, and possibly Canada Salvelinus namaycush for the other two species. Early growth rates of DOC range: 2.6 – 14 mg/L both Walleye and Lake Trout were negatively Benoît et al., 2016 related to DOC concentration, but not Yellow Perch.

Spatial Survey Zooplankton Zooplankton production declined with 10 temperate lakes on Wisconsin- increasing DOC concentrations, light Michigan border, USA, attenuation, and epilimnion depth. The residual DOC range: 5.9 – 25.9 mg/L variation from zooplankton production and Kelly et al., 2014 light attenuation models was correlated with allochthony, indicating terrestrial organic matter was not a resource subsidy.

Time-series Macroinvertebrates A five-fold increase in DOC resulted in A small eutrophic temperate lake Leucaspius delineatus increased primary production and reduced light underwent 5-fold increase in DOC penetration, which promoted hypolimnetic concentration over two years. anoxia. As DOC inputs increased, the DOC range: 24.6-53 mg/L proportion of anoxic sediments did as well, Brothers et al., 2014 which resulted in a near complete loss of fish and invertebrates over two years.

Time-series Perca fluviatilis Perch recruitment and growth was variable A small boreal lake was evaluated over the 20-year time series. Early growth rates over a 20-year duration of Perch were lower in years with higher than Rask et al., 2014 average DOC concentrations.

Spatial Survey Benthic Invertebrates, DOC concentration negatively affected benthic 10 temperate lakes on Wisconsin- Benthivorous fish invertebrate productivity by reducing the Michigan border, USA, DOC availability of suitable oxythermal habitat. ranged from: 5.3 – 23 mg/L Average-whole lake benthos productivity was Craig et al., 2015 also negatively related to DOC concentration. CPUE of benthivorous fish was positively related to benthos productivity

Spatial Survey Perca fluviatilis, Fish productivity and catch-per-unit-effort 13 lakes in northern Sweden Salmo trutta declined with increasing DOC concentration. DOC range: 7 – 22.1 mg/L Fish productivity was also negatively related to Karlsson et al., 2015 allochthony.

Spatial Survey Lepomis macrochirus Maximum attainable size, inferred from Von 11 temperate lakes on Wisconsin- Bertalanffy models, declined with increasing Michigan border, USA DOC concentration. Consequently, total life- DOC range: 3 – 24 mg/L

20 Craig et al., 2017 time fecundity in Bluegill Sunfish was lower in high DOC lakes.

Spatial Survey Perca fluviatilis, Biomass of fish communities declined with 16 boreal lakes in Sweden, with Kd Rutilus rutilus, increasing light attenuation and the effect of ranging from Esox lucius, light attenuation on biomass was more negative 0.4 – 2.64 m-1 Lota lota for deeper lakes relative to shallower lakes Seekell et al., 2018

Unimodal response to DOC

Spatial Survey Salmo trutta Brown Trout biomass, inferred from 168 lakes across Norway standardized catch data, was unimodally Finstad et al., 2014 related to DOC concertation. At concentrations below ~1.5 mg/L DOC, fish biomass increased with DOC concentration, and thereafter biomass declined. The threshold of DOC concentration was moderated by lake depth, where the negative effects of DOC occurred at lower concentrations in deeper lakes.

21 Topography Climate

Regulates

REGIONAL Terrestrial vegetation Soils Hydrology

Sets range of possible DOC concentrations

Proportion of wetlands Drainage ratio Water retention time Lake size

LOCAL

Sets DOC concentration within individual lake

Fig. 1.1: Hierarchical regulation of DOC concentration as described in Sobek et al. (2007).

22 Fish

Zooplankton

Heterotrophic Phytoplankton Bacteria

DOM pool Littoral benthic Profundal chironomids invertebrates

Terrestrial organic matter Methane oxidizing Benthic primary Benthic bacteria production bacteria t-DOM

t-POM Methane

Fig. 1.2: Pathways of energy flow that support lake food webs. Heterotrophic and terrestrial organic matter energy pathways are drawn in brown, autotrophic pathways are drawn in green.

Dashed lines represent pathways of energy flow that are only partly supported in the literature.

After Creed et al. (2018).

23 Chapter 2: DOC-mediated effects on habitat structure influences basal resource use and biomass of zooplankton

Abstract

In a survey of eight boreal lakes spanning a gradient of dissolved organic carbon (DOC),

I assessed how DOC-mediated effects on light availability influenced vertical habitat structure and basal resource use and biomass of zooplankton. Using Bayesian stable isotope mixing models, I show that hypolimnetic phytoplankton were an important resource for zooplankton in low-DOC lakes. As DOC concentrations increased, chlorophyll a concentrations in the hypolimnion were reduced relative to the epilimnion and zooplankton acquired proportionately more energy from terrestrial sources. The potential food quality of seston for zooplankton was determined using C:P and chlorophyll a : C ratios and was found to be better at depth relative to surface waters. Zooplankton biomass declined with increasing DOC concentration and with increasing reliance on terrestrial sources. These results suggest that terrestrial organic matter can suppress zooplankton productivity through simultaneous effects on habitat and resource availability and quality.

Introduction

The productive potential of secondary consumers in freshwater ecosystems is ultimately dependent upon the amount of energy at the base of the food web and the efficiency with which energy is mobilized to higher trophic levels (Brander, 2007; Downing & Plante, 1993).

Formerly, in-lake primary production has been viewed as the main driver of secondary productivity in lakes (Carpenter, 2008; Schindler, 1977; Sterner, 2008), but there is increasing evidence that terrestrial organic matter may play a role in small, oligotrophic systems (Finstad et

24 al., 2014; Karlsson et al., 2009; Seekell et al., 2015a). Terrestrial organic matter can dominate the carbon pool in such lakes (Wilkinson et al., 2013) and inputs have been increasing over time in many temperate and boreal lakes (Emmerton et al., 2018; Evans et al., 2005; Finstad et al., 2016;

Monteith et al., 2007). As a result, there is a need to understand how consumers utilize these inputs and resultant consequences on consumer productivity.

The bulk of terrestrial organic matter in boreal lakes enters in the form of dissolved organic carbon (DOC; Wilkinson et al., 2013), which is primarily composed of humic and fulvic acids (Jones, 1992). These substances give water a brown colour and strongly absorb solar radiation, affecting the vertical distribution of light and heat (Fee et al., 1996; Tanentzap et al.,

2008; Williamson et al., 1996b). Consequently, DOC influences where and how much primary production can be supported in a given lake (Jones, 1992; Karlsson et al., 2009). Additionally, the effects of DOC on light and temperature may lead to shallower thermoclines and changes in light availability for phytoplankton and oxythermal habitat suitable for fish and invertebrates

(Craig et al., 2015; Jones, 1992).

Terrestrial organic matter is an energetic input at the base of the food web and has been viewed as a resource subsidy (Carpenter et al., 2005; Craig et al., 2015; Karlsson, 2007;

Karlsson et al., 2003). However, terrestrial inputs are considered a poor quality resource relative to autochthonous production due to high C:P and C:N ratios and low essential fatty acid content

(Brett et al., 2009; Elser et al., 2000) and there is increasing evidence that terrestrial resources can suppress consumer productivity (Brett et al., 2009; Karlsson et al., 2015; Kelly et al., 2014).

Zooplankton are an important trophic link and have been associated with enhanced fish recruitment and growth (Carpenter et al., 1985; Mehner & Thiel, 1999). Several studies have used analyses of the stable isotopic composition of zooplankton (C, N, H) to assess the

25 utilization by zooplankton of terrestrial resources. Most have found that assimilation of terrestrial resources (allochthony) by zooplankton is widespread and variable (Berggren et al., 2014; Cole et al., 2011; Francis et al., 2011; Karlsson et al., 2003; Kelly et al., 2014; Pace et al., 2007;

Solomon et al., 2011; Tanentzap et al., 2017). In shallow, subarctic lakes of northern Finland, allochthony varied between 9% and 77%, with nearly half of zooplankton biomass supported by terrestrial resources, on average (Karlsson et al., 2003). Similar ranges in allochthony were observed in northern Midwestern lakes of the United States (Kelly et al., 2014; Solomon et al.,

2011). In a global survey of 147 lakes, zooplankton allochthony averaged 42%, and was greatest in lakes with organic matter rich watersheds (Tanentzap et al., 2017). In deep, clear lakes or large boreal lakes, zooplankton allochthony appears to be negligible (Francis et al., 2011; Galloway et al., 2014).

A major limitation of many studies that quantify contributions of terrestrial organic matter to zooplankton is the failure to consider hypolimnetic phytoplankton as a potential resource (Francis et al., 2011; Wilkinson et al., 2014). Deep chlorophyll maxima (DCM), an indicator of hypolimnetic phytoplankton, are widespread in oligotrophic lakes (Fee, 1976) and are often of higher quantity and quality than epilimnetic phytoplankton (Williamson et al.,

1996a; Winder et al., 2004). When hypolimnetic phytoplankton is present, zooplankton have been shown to heavily rely upon it, so that it contributes between 40% and 80% of their biomass

(Francis et al., 2011; but see Wilkinson et al., 2014). Zooplankton that heavily utilize this resource tend to have higher growth and reproduction rates (DeMott et al., 2004; Williamson et al., 1996a; Winder et al., 2003; but see Cole et al., 2002b). Exclusion of hypolimnetic phytoplankton as a source in isotopic mixing models may lead to an overestimation of terrestrial

26 contributions to consumers (Francis et al., 2011) and influence how we interpret the consequences of allochthony on zooplankton productivity.

The potential for hypolimnetic primary production is strongly affected by the vertical distribution of temperature and light (Camacho, 2006; Fee et al., 1996; Fee, 1976; Leach et al.,

2017), as well as zooplankton grazing (Pannard et al., 2015; Pilati & Wurtsbaugh, 2003). In clear lakes where the photic depth extends below the thermocline, the potential for hypolimnetic primary production is high. Alternatively, in eutrophic or high DOC lakes, shading can reduce light penetration, restricting primary production to the upper mixed layer (Jones, 1992). As DOC concentrations in lakes are projected to increase (Evans et al., 2005; Finstad et al., 2016;

Monteith et al., 2007), thereby decreasing light penetration, there will likely be a direct effect on the vertical distribution and production of hypolimnetic phytoplankton, with possible concomitant effects on zooplankton productivity.

While it has been established that zooplankton can be supported by terrestrial resources and that zooplankton may utilize hypolimnetic phytoplankton, there is little empirical data linking these patterns of resource utilization to overall zooplankton biomass and/or productivity

(but see Kelly et al., 2014). Here, I present results from an eight lake survey showing that DOC- mediated effects on light availability influence the potential of hypolimnetic primary production with consequent effects on zooplankton resource use and biomass. As DOC concentrations increased, the potential for hypolimnetic primary production was diminished. Using Bayesian isotopic mixing models, I show that hypolimnetic phytoplankton are an important resource for zooplankton in low-DOC lakes and that utilization of terrestrial resources increases with DOC concentrations. Seston quality, as determined using chlorophyll a : C and C : P ratios was better for zooplankton at depth relative to surface waters, suggesting loss of hypolimnetic

27 phytoplankton with increasing DOC concentration coincided with a decline in resource quality.

Consequently, zooplankton biomass declined with increasing DOC concentration and with greater reliance on terrestrial resources. These results add to a growing body of literature suggesting that terrestrial organic matter can suppress limnetic food web productivity through simultaneous effects on the vertical distribution, availability, and quality of resources used by consumers.

Methods

Study sites and limnological sampling

I studied eight boreal lakes spanning a gradient of DOC concentrations at the

International Institute for Sustainable Development Experimental Lakes Area (IISD–ELA) in northwestern Ontario. The lakes are all dimictic and small (8.4 – 54 ha, mean depth 4.9 – 10.7 m), with negligible macrophyte growth and watersheds dominated by coniferous forest (Table

2.1). All lakes contain populations of White Sucker (Catostomus commersonii), at least one species of piscivore, either Lake Trout (Salvelinus namaycush), Northern Pike (Esox lucius) or both (e.g. L239), and various cyprinid species and/or Yellow Perch (Perca flavescens). Larger planktonic invertebrate predators include Chaoborus spp. and Mysis diluviana.

Lakes were sampled in 2017 approximately every 2 – 4 weeks during the open water season for water chemistry (dissolved and particulate nutrients) and chlorophyll a, as well as vertical profiles of dissolved oxygen, temperature and photosynthetically active radiation (PAR) at the deepest point of each lake. Samples for water chemistry were collected in the epilimnion by water grabs or a peristaltic pump and were analyzed by the IISD–ELA Chemistry Laboratory following the methods of Stainton et al. (1977). Four lakes (L224, L239, L373 and L442) were also profiled at discrete depths monthly for water chemistry. All dissolved nutrient samples were

28 filtered through a GF/C filter within four hours of collection, and nitrogen and phosphorus samples were analyzed within 48 hours. Chlorophyll a samples were frozen and analyzed fluorometrically, unless stated otherwise. DOC was analyzed following the ‘automated’ method outlined in Stainton et al. (1977).

Light attenuation (Kd) was calculated by regressing the natural logarithm of PAR vs. depth. Using Kd and mean depth (Zm), mean light irradiance of the whole lake volume (Im; mean light irradiance) was estimated using the equation of Karlsson et al. (2009):

()×+, !" = (1 − ' )/(/0 × 1")

Mean light irradiance represents the average proportion of PAR at the surface throughout the water column.

Zooplankton were collected by vertically hauling a 50 µm net through the whole water column during the day at the deepest point of the lake (n = 6 – 12 samples per lake).

Zooplankton were preserved in 4% formalin in the field and then enumerated and identified in the lab. Counts were converted to biomass following the methods of Paterson et al. (2010). I estimated White Sucker and forage fish abundance (cyprinids and Yellow Perch) using catch- per-unit-effort (CPUE) from littoral trap net surveys (further detailed in Chapter 3 methods) to investigate relationships between fish predation and zooplankton biomass.

To evaluate DCM presence, a single vertical profile of chlorophyll a was collected in each lake over a two week period at the end of June to early July 2017 and monthly in 2018 (B.

Sherbo, unpublished data). Relative fluorescence units (RFU) were measured in situ using a YSI

6025 chlorophyll probe, and I converted RFU to chlorophyll a concentrations using a regression of RFU with known chlorophyll concentrations (R2 = 0.77, y = 1.0301 + 1.0736x).

29

Stable isotope ratios

I used stable isotopes of hydrogen (H), carbon (C), and nitrogen (N) to estimate the relative contributions of deep phytoplankton, epilimnetic phytoplankton, and terrestrial organic matter to zooplankton biomass. Using H isotopes in combination with C and N isotopes is particularly useful for tracing terrestrial contributions to consumers because of strong isotopic separation between terrestrial and aquatic primary producers (Doucett et al., 2007).

The isotopic composition of terrestrial organic matter was estimated from samples of leaf material (n = 27) collected from dominant tree species (Pinus banksiana, Picea mariana, Larix laricina, and Betula papyrifera) from each of the eight watersheds. There were no statistically significant differences in stable isotope signatures among tree species or watersheds (ANOVA; p

> 0.05), so samples were pooled to represent a single terrestrial end member in mixing models.

Because phytoplankton are difficult to isolate from bulk particulate organic matter

(POM), I used indirect methods to estimate their isotopic H, C and N composition. For these

2 13 15 indirect methods, I used determinations of d H in water, and d C and d N in POM. Water and

POM for epilimnetic phytoplankton were sampled monthly (n = 6 per lake), and samples for deep phytoplankton once in mid-summer slightly above the photic depth (depth at which 99% of surface PAR is attenuated), which, in general, was where DCM occurred. Water for d2H analysis was filtered through a 0.2 µm filter and stored at 4 °C in glass scintillation vials with no headspace. POM was collected by filtering water onto pre-combusted (550 °C for 2 h) 47 mm

QMA filters (2.2 µm pore size) and frozen. Additionally, samples for dissolved inorganic carbon

(DIC) d13C were collected monthly from the epilimnion, and once mid-summer slightly above

30 the photic depth, by storing water without headspace in small serum bottles with stoppers and preserved with magnesium chloride.

Phytoplankton d2H was estimated by multiplying a published photosynthetic fractionation factor (-160.9 ± 17; Wilkinson et al., 2015) by d2H of water. Phytoplankton d13C and d15N was estimated by correcting bulk particulate organic matter (POM) isotope signatures for the terrestrial fraction (2T) of POM, which was estimated using a C:N ratio algebraic mixing model (Francis et al., 2011; Yang et al., 2014):

(5: 789: − 5: 7<) ∅4 = (5: 74 − 5: 7<)

where C:NPOM is the C:N ratio of POM, C:NT is the C:N ratio of terrestrial organic matter

(measured directly), and C:NA is C:N ratio of phytoplankton. For, C:NA, I used a literature average of 6.8 (Vuorio et al., 2006; Yang et al., 2014). Once the proportion of terrestrial organic

13 15 matter in POM was determined (2T), d C and d N were solved for algebraically (Yang et al.,

2014):

AB AB AJ AJ >? @ CDEFG(∅H × @ CH) >I @ KDEFG(∅H × @ KH× L.?) = 5< = or = 7< = (>G∅H) (>G∅H× L.?)

13 15 13 15 13 where, d CA or d NA is the d C and d N signature of phytoplankton, and d CT or

15 13 15 15 d NT is the d C and d N value of terrestrial organic matter. In the d NA model, a value of 0.3 was used to correct for potential overrepresentation of phytoplankton molecular N in POM C : N

31 ratios, as recommended by Yang et al. (2014). Zooplankton for stable isotope analyses were collected monthly (n = 6 times per lake) by vertically hauling a 150 µm net through the whole water column at the deepest point in the lake. Zooplankton d13C values were mathematically lipid corrected using the model of Post et al. (2007). All organic samples were freeze dried or oven dried at 60 °C, homogenized if necessary, and stored in a desiccator until analysis.

Analyses of d13C and d15N were carried out at the University of Waterloo Environmental

Isotope Laboratory using a Finnegan Deltaplus XL-EA mass spectrometer for solid samples, and a

Micromass IsoChrom continuous-flow mass spectrometer for d13C-DIC. All d13C and d15N were expressed as the per mil (‰) deviation from universal standards CO2 in PeeDee limestone or N2 gas, respectively, as follows:

>? >I = 5 NO = 7 (‰) = [(RST"UVW/RSXTY0TZ0 ) − 1] × 1000

Where R is the 13C/12C or 15N/14N of either samples or standards. d2H was determined by the Colorado Plateau Stable Isotope Laboratory, Northern Arizona University following the methods of Doucett et al. (2007) for organic samples, including correction of exchangeable H using bench-top equilibrium, and d2H water samples were analyzed by cavity-ring-down laser spectroscopy.

Data analysis

I estimated the proportional reliance of zooplankton on terrestrial organic matter, epilimnetic phytoplankton, and deep phytoplankton using three-isotope (H,C,N) Bayesian mixing models with the r-package ‘MixSIAR’, which accounts for multiple sources of model

32 uncertainty using a Markov Chain Monte Carlo approach (Stock et al., 2018). In three of the study lakes (L164, L239, L658), deep phytoplankton and epilimnetic phytoplankton were averaged because photic depths were close to thermocline depths (< 1 m), and sources were similar isotopically (< 2‰ d13C, < 5‰ d2H). Models did not converge (did not achieve a stationary distribution) when deep and epilimnetic phytoplankton were input as individual sources in these three lakes.

MixSIAR uses three matrices: (1) a ‘consumer’ matrix of observed isotopic ratios of consumers; (2) a ‘source’ matrix of the means (+SD) of dietary resources and (3) a

‘discrimination factor’ matrix of the means (+SD) of changes in isotope ratios from source to consumer. Zooplankton replicates were input in the consumer matrix (n = 6 per lake). Mean terrestrial organic matter isotope values and seasonal means of epilimnetic phytoplankton (+SD) were input in the source matrix. Hypolimnetic phytoplankton were included as a source in 6 of 9 lakes (rationale provided above) and I assumed the same seasonal variability as that of epilimnetic phytoplankton. I rationalized this assumption because the variability in dissolved inorganic carbon (DIC) isotopes in the epilimnion and deeper in the water column were similar

(data not shown), and d13C–DIC is a strong determinant of phytoplankton d13C during fractionation (Taipale et al., 2016b).

For each model, I ran three Markov chains with uninformed Dirichlet priors for 300,000 iterations, with 200,000 iteration burn-in, at a thinning rate of 100. Model convergence was assessed using Gelman-Rubin test within the MixSIAR package, where a scale reduction factor

<1.1 indicates an acceptable model. Posterior distributions were skewed for some mixing models, so medians were used as point estimates.

33 I used trophic discrimination factors of 0.0 ± 1‰ for δ13C to capture the range of variability in these assumptions for invertebrates (Vander Zanden & Rasmussen, 2001) and 3.91

± 0.91‰ for δ15N based on the average of published discrimination factors for zooplankton

(Vanderklift & Ponsard, 2003). The enrichment of δ2H across trophic levels does not occur in the traditional sense, rather it is caused by progressive incorporation of dietary water, an isotopically heavy source of hydrogen (Solomon et al., 2009). The influence of dietary water on consumer

δ2H was corrected for using a model from Solomon et al. (2009):

X ]X^X = 1 − (1 − ])

Where ]X^X is the proportion of H from dietary water in consumer tissue at trophic level t and ] is the per-trophic-level contribution of dietary water to consumer H. Allochthony estimates are sensitive to changes in ], so I used a conservative value of 0.2, based on the recommendations of Wilkinson et al. (2015), and validated by a sensitivity analysis (Appendix

2A). I assumed zooplankton occupy trophic position 2 (Karlsson et al., 2015). Zooplankton δ2H was corrected for dietary water by subtracting the δ2H enrichment factor from raw δ2H values.

The dietary water enrichment correction factor was calculated following the equation outlined in

Berggren et al. (2014):

_ _ _ _ = ` 'aObcℎe'af = = `ST"UVW − (= `ST"UVW − ]X^X × = `gTXWZ)/(1 − ]X^X )

_ 2 _ Where = `gTXWZ is the δ H of water, and = `ST"UVW is the sample being corrected for.

All data used in the Bayesian mixing models is summarized in Appendix 2B.

34 I evaluated how seston quality differs among the epilimnia of lakes, and within lakes at depth using ratios of chlorophyll a : suspended carbon (chl a : C) and suspended carbon : suspended phosphorus ratios (C : P). For seston quality comparisons at depth, data were only available for L224, L239, L373 and L442 in 2017, and L626 in 2016. Differences in seston quality among lakes were assessed visually with notched-boxplots using 95% confidence limits about the median. Seston is comprised of allochthonous and autochthonous material that forms the resource base for planktonic primary consumers, and the efficiency by which this material is utilized for biomass production is related to the concentration of potentially limiting nutrients

(Elser et al., 2002). Lower C : P ratios and higher chlorophyll a : C ratios indicate higher quality food.

I used Pearson correlation (r) models to assess how zooplankton resource use and biomass was related to habitat availability, resource availability, and predation. I used DOC, Kd, maximum depth, and photic depth : thermocline depth ratio as indicators of habitat availability.

Photic depth : thermocline depth ratio is an indicator of the potential for deep primary production, where a ratio £ 1 indicates primary production is restricted to the upper mixed layer, and ³ 1 indicates potential for hypolimnetic primary production (Francis et al. 2011). Photic depth is defined as the depth at which 99% of surface PAR is attenuated, and thermocline depth is defined as the depth at which maximum temperature change occurs. I used chlorophyll a and

DOC as predictors of resource availability, under the assumption that DOC is representative of the amount of terrestrial organic matter potentially available to consumers (Karlsson et al.,

2015). Direct measurement of terrestrial organic matter concentration in surface waters is difficult, and the majority of DOC in boreal lakes is of terrestrial origin (Wilkinson et al. 2013), thus DOC is a suitable proxy. Notably, zooplankton cannot use DOC directly, rather it requires

35 microbial assimilation before it can be incorporated into the food web (Berglund et al., 2007), however zooplankton can consume terrestrial particulates. I also used allochthony estimates as an indicator of resource availability for zooplankton biomass models. White Sucker (Catostomus commersonii) catch-per-unit-effort (CPUE), forage fish (cyprinids and Yellow Perch) CPUE were used as indicators of predation. Model assumptions were assessed visually by comparing quantile-quantile plots, and histograms of residuals. White Sucker CPUE, zooplankton biomass, and copepod biomass was log10 transformed to meet model assumptions. If model assumptions were not met after log10 transformations, I used Spearman rank correlations (rs). Pearson correlations were also used to assess relationships among predictor variables.

I also extended the analysis of DOC-zooplankton biomass relationships to all unmanipulated IISD-ELA lakes where both variables were measured (n = 30 lakes). In this data set, zooplankton were collected by two different methods; either by replicate vertical water column hauls with a 53 µm zooplankton net at the deepest point of the lake or at various pelagic stations with a volume calibrated tube sampler (Paterson et al., 2010). Zooplankton biomass and

DOC concentrations were averaged across all years of available data for each lake. There was no significant difference in the slope or intercept of zooplankton-DOC regressions between the two collection methods (ANCOVA; F1,28 = 0.044, p > 0.05), so data were treated equivalently.

Results

Average concentrations of epilimnetic DOC throughout the open water season ranged from 3.5 – 9.2 mg/L (Table 2.1) and were not significantly correlated with lake area, mean or maximum depth as indicated by Pearson correlations (p > 0.05; Table 2.2). DOC was significantly positively correlated with open water season averages of light attenuation (Fig. 2.1;

36 r = 0.99, p < 0.001), total phosphorus (r = 0.91, p = 0.002), and negatively correlated with photic depth (depth at which 99% of surface PAR is attenuated; r = –0.94, p < 0.001) and average thermocline depth (depth of maximum temperature change; r = –0.93, p = 0.002). Epilimnetic chlorophyll a concentrations were positively correlated with DOC concentration (r = 0.93, p<0.0001), but volume weighted whole-lake chlorophyll a concentrations in July of 2017 were not (p > 0.05). Both volume weighted hypolimnetic oxygen concentration and the proportion of lake area with dissolved oxygen less than 1 mg/L in late August were not related to DOC concentration (p > 0.05; data not shown). These relationships highlight DOC’s fundamental role in structuring the physical and chemical properties of my study lakes.

Using δ2H, there was strong separation of allochthonous and autochthonous sources, where average terrestrial organic matter δ2H values were heavier than epilimnetic phytoplankton by 46‰ (± 4.3‰; Appendix 2B). There was also strong isotopic separation between the δ15N signatures of sources, with terrestrial organic matter being more depleted (–5.5‰ ± 1.6‰) than both epilimnetic (1.8‰ ± 1.7‰) and deep phytoplankton (2.0‰ ± 1.8‰). Terrestrial organic matter and epilimnetic phytoplankton had similar δ13C values, but deep phytoplankton were more depleted in δ13C than both of these sources (range 1.5‰ to 6.5‰), with the greatest differences occurring in low DOC lakes (Fig. 2.2; Appendix 2B). Based on the isotopic separation among sources, mixing model outputs of autochthonous vs allochthonous resource use were primarily driven by hydrogen isotopes, whereas model outputs regarding epilimnetic vs deep phytoplankton resource use were primarily determined by carbon isotopes.

The potential for metalimnetic and hypolimnetic primary production is affected by DOC- mediated effects on light penetration (Fig. 2.2a). In my study lakes, the photic zone extended below the thermocline in lakes with high light penetration and low DOC concentrations. As

37 DOC concentrations increased and light penetration decreased, the differences in thermocline and photic zone depths was diminished. Accordingly, the photic depth : thermocline depth ratio was strongly, negatively related to DOC concentration (r = –0.89, p = 0.003). In clearer lakes, the depth of maximum chlorophyll a concentration in July usually fell between the photic and thermocline depths (Fig. 2.3a-f), whereas the chlorophyll a maximum occurred in the upper mixed layer in the two darkest lakes (Fig. 2.3g-h). The depth of maximum oxygen concentration also coincided with these deep chlorophyll a peaks, indicating primary production is probably occurring in the hypolimnion. More complete seasonal sampling in 2018 from June to September confirmed these patterns, where maximum chlorophyll a concentrations were below the thermocline in the clearest lakes (Appendix 2C; B. Sherbo, unpublished data).

Where data exist, the vertical distribution of chlorophyll a coincided with changes in

POM quality. Suspended C : P ratios varied with depth, and, in general, ratios deeper in the water column were slightly lower than ratios in the epilimnion (Fig. 2.4a-e). Chlorophyll a : suspended C ratios also varied vertically in the water column, with significantly different ratios occurring near the photic depth relative to the surface waters in the four clearest lakes based on

95% confidence limits about the median (Fig. 2.4f-j). There was no relationship between epilimnetic suspended C : P ratios among lakes (Fig. 2.4k; p > 0.05), but epilimnetic chlorophyll a : suspended carbon ratios increased significantly with DOC concentration (Fig. 2.4l r = 0.88, p

= 0.004).

In lakes with algal peaks below the thermocline, stable isotopic Bayesian mixing models suggested that zooplankton relied upon them heavily, contributing between 22% – 72% of their assimilated diet (Appendix 2D). Accordingly, the difference between average δ13C of zooplankton and δ13C of epilimnetic phytoplankton diminished with increasing DOC

38 concentration (Fig. 2.2b & c), where zooplankton from the lakes without hypolimnetic phytoplankton maxima more closely resembled δ13C of epilimnetic phytoplankton. Similarly, the difference in δ13C between epilimnetic and hypolimnetic phytoplankton also diminished with increasing DOC concentration (Fig. 2.2b & c), and δ13C–DIC became more depleted at depth

(Appendix 2E). This pattern was also evident in the δ13C values of zooplankton, which became more enriched with increasing DOC concentration (Fig. 2.2b).

Zooplankton allochthony, as estimated by the mixing model, was positively correlated

with DOC concentration (rs = 0.79, p = 0.03; Fig. 2.5) and light attenuation (rs = 0.76, p = 0.04).

In lakes with DCM and low DOC, terrestrial organic matter contributed < 10% of assimilated diet, on average. In lakes with higher DOC and more chlorophyll in the upper mixed layer, terrestrial organic matter contributed an average of 19% – 26% to zooplankton C. Zooplankton

allochthony was also positively related to epilimnetic chlorophyll a concentration (rs = 0.88, p =

0.007), but not whole-lake volume weighted chlorophyll a (p > 0.05).

Zooplankton biomass ranged from 0.13 to 0.54 g/m2 and was significantly negatively correlated to DOC (r = –0.92, p = 0.001; Fig. 2.6a), light attenuation (r = –0.89, p = 0.002), allochthony (r = –0.76, p = 0.03) and epilimnetic chlorophyll a (r = –0.75, p = 0.04), and positively correlated to photic depth : thermocline ratio (r = 0.73, p = 0.04; Fig. 2.6b). Similarly, copepod biomass (ranging from 0.038 to 0.46 g/m2) declined with DOC (r = –0.87, p = 0.004;

Fig. 2.6c), light attenuation (r = –0.85, p = 0.007), and epilimnetic chlorophyll a (r = –0.75, p =

0.03), but is not correlated with photic depth : thermocline ratio (p = 0.14). Variation in copepod biomass was primarily related to variations in Cyclopoid biomass, which decreased with increasing DOC (r = –0.82, p = 0.01). There were no observed relationships between Calanoid or Cladoceran biomass and DOC among lakes (p > 0.05). Zooplankton biomass was also

39 positively correlated with White Sucker catch-per-unit-effort (CPUE; r = 0.82, p = 0.01), and weakly related to forage fish CPUE, although this relationship was not significant (p = 0.055; data not shown). In general, copepods dominated zooplankton biomass in my lakes, but I did not find any significant correlations between the relative biomass of cladocerans or copepods and

DOC or light attenuation (Fig. 2.6d; p > 0.05).

Similar to the eight-lake analysis, zooplankton biomass declined with DOC in the larger dataset of IISD-ELA lakes (r = –0.62, p < 0.001; Fig. 2.7).

Discussion

I observed strong declines in zooplankton biomass with increasing DOC concentration, which were associated with changes in thermocline and photic depths, with simultaneous changes in resource availability and quality. As DOC concentrations increased, so too did the utilization of terrestrial organic matter by zooplankton. In low DOC lakes, stable isotopic evidence suggested that zooplankton relied heavily on hypolimnetic phytoplankton, which were of higher quality than epilimnetic resources based on ratios of chlorophyll a : C and C : P.

Evidence of deep phytoplankton resource use

My results highlight that hypolimnetic phytoplankton are an important resource to zooplankton, and their importance is diminished with increasing DOC concentration. Deep chlorophyll maxima are widespread (Fee, 1976; Williamson et al., 1996a), and experimental evidence suggests zooplankton are capable of locating and feeding on phytoplankton at DCMs.

(Kessler & Lampert, 2004; Lampert & Grey, 2003). Further, zooplankton populations often coincide vertically in the water column with these peaks (e.g. Kettle et al., 1987; Paterson et al.,

40 2010). Despite this, inclusion of hypolimnetic phytoplankton resources is often overlooked when quantifying the relative importance of allochthonous and autochthonous energy pathways

(Francis et al., 2011), under the assumption that primary production is restricted to the upper mixed layer (Karlsson et al., 2003; Kelly et al., 2014; Solomon et al., 2011).

I found that the δ13C values of zooplankton were more depleted than that of epilimnetic phytoplankton and/or POM, as has been observed in previous work (del Giorgio & France, 1996;

Francis et al., 2011; Jones et al., 1999), and these differences decreased with increasing DOC concentration. The mean δ13C of terrestrial organic matter in my system (-29.7 ‰) is either equivalent to or slightly more enriched than my estimates for epilimnetic phytoplankton δ13C, therefore terrestrial organic matter cannot be the alternative resource driving the depleted isotopic values in zooplankton. Thus, zooplankton must rely on an additional, more δ13C depleted resource, which varies consistently with DOC.

Hypolimnetic phytoplankton tend be depleted in δ13C relative to the upper mixed layer

(del Giorgio & France, 1996), because they are thermally isolated from sources of

13 atmospherically enriched δ C – CO2 and utilize heterotrophically respired CO2 that is depleted in δ13C (Lennon et al., 2006). In my lakes, δ13C – DIC became more depleted with depth

(Appendix 2E). In high light, low DOC conditions, the difference between δ13C zooplankton and epilimnetic phytoplankton in my study lakes was high (~4 – 6 ‰), and these differences decreased with increasing DOC concentration. Differences between δ13C of deep phytoplankton and epilimnetic phytoplankton followed a similar pattern. These δ13C patterns suggest that zooplankton were utilizing hypolimnetic phytoplankton in the lower DOC lakes of my study.

While utilization of hypolimnetic phytoplankton by zooplankton is the most likely explanation for these results, it is possible that I failed to include an additional δ13C depleted

41 resource, such as methane oxidising bacteria (MOB). However, I did not observe any relationships between hypolimnetic oxygen concentration and zooplankton δ13C, which would be expected given that anoxic conditions favour methane production (Jones & Grey, 2011). In addition, there is limited evidence that zooplankton consume MOB directly (Jones & Lennon,

2009; Lennon et al., 2006). Another alternative explanation is that selective feeding by zooplankton on δ13C depleted phytoplankton decreased with DOC. Copepods have been shown to be selective grazers (DeMott, 1988), whereas cladocerans are commonly considered to be less discriminate filter feeders (Jürgens, 1994). I did not analyze the isotopic values of copepods and cladocerans individually, and thus cannot infer the degree to which selective consumption of

δ13C depleted resources occurred in my lakes. However, the relative biomass of copepods and cladocerans did not change consistently among lakes, thus my bulk isotope values of zooplankton should not be biased towards a particular group. Furthermore, epilimnetic chlorophyll a : C ratios increased with DOC, suggesting that reliance of indiscriminate filter feeders on phytoplankton should have increased with DOC, rather than decreased, as we observed.

Stable isotopic mixing models suggested reliance on hypolimnetic phytoplankton was high (22% – 77% of diet), with limited reliance on terrestrial organic matter (< 7% of diet) in lakes with deep chlorophyll peaks. The importance of terrestrial organic matter increased (19% –

26% of diet) as primary production became restricted to the upper mixed layer. The increased reliance on terrestrial organic matter, and small differences between δ13C of zooplankton and epilimnetic phytoplankton in high DOC lakes suggest zooplankton feeding is restricted to the upper mixed layer in these lakes. Indeed, zooplankton that specialize in epilimnetic feeding tend to utilize more terrestrial resources (Matthews & Mazumder, 2006). My estimates of terrestrial

42 resource use by zooplankton are lower than several other spatial surveys (Karlsson et al., 2003;

Kelly et al., 2014; Solomon et al., 2011), likely caused by the omission of deep primary production as a potential resource in these other studies (Francis et al., 2011; but see Wilkinson et al., 2014).

Effects of DOC on zooplankton biomass

The decline in zooplankton biomass with DOC was not linked to declines in epilimnetic chlorophyll a concentration. In fact, I observed the opposite, where higher DOC lakes had greater concentrations of epilimnetic chlorophyll a. Because zooplankton biomass decreased in lakes where both chlorophyll a and terrestrial organic matter were higher, there was no evidence that terrestrial inputs were a resource subsidy for zooplankton, similar to conclusions reached by

Kelly et al. (2014).

Rather, declines in zooplankton biomass in high DOC lakes were associated with the loss of hypolimnetic phytoplankton resources, where food was of higher quality and chlorophyll a was most concentrated spatially. Zooplankton reliance on hypolimnetic phytoplankton appears advantageous, where consumption of hypolimnetic phytoplankton leads to higher growth and reproduction rates in zooplankton (DeMott et al., 2004; Williamson et al., 1996a; Winder et al.,

2003; but see Cole et al., 2002b). Indeed, fitness of Daphnia galeata appears to be optimal when occupying mid-waters, despite lower temperatures (Winder et al., 2004).

Decreases of zooplankton in higher DOC lakes may have arisen for several reasons.

Terrestrial organic matter is a poor quality resource relative to phytoplankton, as it is nearly devoid of essential fatty acids (Brett et al., 2009; Taipale et al., 2015), and the quality of dietary resources in regards to their fatty acid composition has been shown to directly influence

43 zooplankton growth and reproduction (Brett & Muller-Navarra, 1997; Galloway et al., 2014).

Accordingly, I found that zooplankton biomass declined with higher utilization of terrestrial resources, and that reliance on terrestrial resources was associated with epilimnetic feeding.

In addition to changes in the distribution of chlorophyll a with depth, there are accompanying changes in seston quality that may influence zooplankton biomass. Suspended C

: P ratios, and chlorophyll a : suspended C ratios indicated seston quality for zooplankton was greater at deeper depths. Feeding at depth, where C : P ratios are lower, zooplankton are exposed to an enhanced phosphorus supply that may help alleviate phosphorus deficiency (DeMott et al.,

2004; Sterner & Schwalbach, 2001). Experimental evidence suggests seston with low C : P ratios are assimilated more efficiently by zooplankton (Urabe et al., 2002), which can ultimately enhance zooplankton biomass (Hessen, 2006).

Similar patterns have been observed where seston quality is greater at depth based on stoichiometric arguments (Barbiero & Tuchman, 2004; DeMott et al., 2004; Matthews &

Mazumder, 2006; Rothhaupt, 1991; Winder et al., 2003), and there is also evidence suggesting that the proportion of edible algae available to zooplankton increases with depth (Winder et al.,

2003). Additionally, heterotrophic bacteria biomass, and bacterial quality (C : P and C : N ratios) tend to be greatest deeper in the water column (Bennett et al., 1990; Hessen & Andersen, 1990).

Seston quality may be greater at depth due to the balance between light and nutrients

(Sterner et al., 1997), where, under high light and low nutrient conditions, there is a skewed uptake of carbon relative to limiting nutrients, such as phosphorus (Hessen, 2006; Sterner et al.,

1997; Urabe et al., 2002; Urabe & Sterner, 1996). Thus, nutrient limitation should be greatest in surface waters where light availability is high and decline with depth as light becomes limiting.

In support of this hypothesis, Urabe et al. (2002) demonstrated that shading increased the relative

44 availability of nutrients, resulting in lower seston C : P ratios. Concentrations of limiting nutrients also tend to increase with depth (Fee, 1976), and utilization of hypolimnetic nutrients can lead to greater carbon assimilation efficiency at low light conditions (Fee, 1973). Similarly,

Cloern et al. (1995) found that algal growth efficiency, inferred from chlorophyll a : C ratios, increased with nutrient availability under low light conditions. This may explain why I observed greater chlorophyll a : C ratios at depth within lakes, and increasing chlorophyll a : C ratios within epilimnia among lakes with increasing DOC concentration.

Hypolimnetic phytoplankton may also be advantageous because it provides feeding opportunities for zooplankton while also providing a spatial refuge from predators (Meester et al., 1995; Zaret & Suffern, 1973). In the absence of hypolimnetic phytoplankton, zooplankton are more likely to occupy surface water habitat during the day, where predation risk is greater

(Dodson, 1990; Winder et al., 2003, 2004). In my lakes, zooplankton utilized more epilimnetic resources in higher DOC lakes, evidenced by δ13C isotope patterns. However, in high DOC lakes, zooplankton often vertically migrate less, and are more likely to occupy surface waters because predation risk is reduced to due poorer visual feeding conditions for fish (Wissel et al.,

2003). Thus, zooplankton occupying surface waters in my high DOC lakes may not be subject to greater predation because the visual environment should be similar to that of deep-water refuges in low DOC lakes. In my lakes, zooplankton biomass was positively related to White Sucker

CPUE, and unrelated to forage fish CPUE. Further, my six clearest lakes have Lake Trout populations, that likely predate upon zooplankton, yet, the lakes with Lake Tout had greater zooplankton biomass than lakes without. This suggests that predation was not a major contributing factor to observed variations in zooplankton biomass.

In my extended analysis investigating zooplankton biomass among all unmanipulated

45 IISD-ELA lakes, I observed similar declines with increasing DOC concentration as in the eight lake spatial survey, albeit with greater variability. In this extended analysis, zooplankton biomass followed a wedge-shaped pattern, where zooplankton biomass was more variable at low DOC concentrations, with consistently low biomass at high concentrations. Similar wedge-shaped patterns have been observed for primary producers, zooplankton, and fish across DOC concentration gradients in other studies (Craig et al., 2017; Karlsson et al., 2015; Kelly et al.,

2014; Seekell et al., 2015a). While zooplankton biomass was consistently low at higher DOC concentrations, many factors may affect biomass at lower DOC, including lake morphometry, nutrient availability, predation, etc. The wedge-shaped pattern was not present in my eight-lake spatial survey, likely because I selected lakes with similar bathymetry and food web structure in order to reduce potential variability.

Limitations

My study has several important limitations. First, I suggest that terrestrial organic matter is not a resource subsidy (i.e. cross ecosystem flux of matter that enhances productivity of the recipient ecosystem) because zooplankton biomass declined with increasing allochthony. While zooplankton biomass declines with allochthony and DOC concentration, productivity may not.

Discrepancies between biomass and productivity estimates may arise due to differences in zooplankton species composition and water temperatures where zooplankton communities primarily occur (Dolbeth et al., 2012). Second, there are several assumptions incorporated in my mixing models that could influence model outcomes. For example, the value selected to estimate the contribution of dietary water (w) to consumer tissues resulted in allochthony estimates ranging from 100% to 0% (Appendix 2A). I selected a conservative dietary water value of 0.2,

46 and my sensitivity analysis indicated that this was reasonable, however, changing this value by

0.05 can change allochthony estimates by 15 – 25%. Another potential limitation is my assumption that zooplankton occupy a trophic position of two. Previous studies have outlined that zooplankton trophic position differs among taxa (Berggren et al., 2015). However, because my zooplankton samples were measured in bulk, I was unable to apply taxa specific trophic positions. Further, trophic position is used to correct for the proportion of H in consumer tissue, and assuming a trophic position of three caused my δ2H corrected zooplankton values to fall outside of the isotopic mixing space for basal resources. Thus, an assumed trophic position of two appeared appropriate based on evidence from literature and model outcomes. Third, because zooplankton isotopes were measured in bulk, I was unable to generate taxa specific allochthony estimates. Cladoceran allochthony has been shown to increase with DOC concentration, whereas calanoid copepod allochthony is negligible (Berggren et al., 2014; Karlsson et al., 2015). Thus, it is possible that my estimates of zooplankton allochthony could be related to differences in species composition, however, I found no consistent changes in the relative proportion of cladocerans among lakes suggesting this is unlikely. Fourth, it has been suggested that epilimnetic grazing by zooplankton, coupled with downward sedimentation of nutrients may also contribute to the persistence of DCM (Pannard et al., 2015; Pilati & Wurtsbaugh, 2003).

However, the stable isotope evidence in my study suggested that zooplankton relied heavily upon hypolimnetic phytoplankton when present. Determinations of the vertical distribution of zooplankton in IISD–ELA lakes also indicate that zooplankton populations often occur in coincidence with these peaks (Kettle et al., 1987; Paterson et al., 2010), suggesting that light and nutrients are the predominate drivers of deep chlorophyll peaks in my lakes, not grazing.

47 Conclusions

These results from an eight lake spatial survey highlight how DOC-mediated effects on light attenuation influences the vertical distribution of chlorophyll a, with concomitant effects on resource availability and quality, and ultimately zooplankton biomass and resource use. I demonstrate that hypolimnetic phytoplankton are an important resource to zooplankton and need to be accounted for in future attempts to assess resource linkages between aquatic and terrestrial habitats. Loss of hypolimnetic phytoplankton and greater reliance on terrestrial resources was related to declines in zooplankton biomass. My results support a growing body of literature that suggest that DOC can suppress within lake secondary productivity through simultaneous reductions in habitat availability and resource quality. Zooplankton represent an important resource to fish, and can be linked to greater fish recruitment and growth (Carpenter et al., 1985;

Mehner & Thiel, 1999). As concentrations of DOC are anticipated to increase in boreal lakes

(Evans et al., 2005; Finstad et al., 2016; Monteith et al., 2007), corresponding declines in zooplankton biomass may be anticipated, which may negatively influence the productive potential of inland fisheries.

48 Table 2.1: Select morphometric parameters and open-water averages of physical, chemical and biological characteristics of the study

lakes, including: lake area, mean depth (Zmean), maximum depth (Zmax), epilimnetic dissolved organic carbon (DOC), light attenuation

(Kd), photic depth, epilimnetic total phosphorus (TP), and epilimnetic chlorophyll a (chl a). Values in parenthesis are standard

deviations. LT and NP represent Lake Trout and Northern Pike, respectively.

Lake Area Zmean Zmax DOC Kd Photic Therm- TP Epilim- Mean % Lake Pisc- (ha) (m) (m) (mg/L) (m-1) depth ocline (µg/L) netic light area ivores (m) depth chl a irrad- above (m) (µg/L) iance photic (Im) depth L224 25.9 11.6 27.4 3.5 (0.3) 0.35 14.8 7.8 5.4 (1.1) 1.0 (1) 0.24 63.6 LT L373 27.3 10.8 20.8 4.3 (0.3) 0.45 11 7.6 5.3 (1.2) 1.3 (0.6) 0.20 46.5 LT L223 27.3 7.15 14.4 4.8 (0.1) 0.52 9.7 6.0 5.1 (1.1) 1.2 (0.3) 0.26 67.2 LT L626 25.9 6.8 11.2 5.1 (0.3) 0.47 10.2 6.4 6.6 (1.3) 1.8 (0.4) 0.30 88.3 LT L442 16 8.1 17.8 6.8 (0.4) 0.64 6.9 5.7 6.4 (1.7) 1.6 (0.5) 0.19 32.5 LT L239 56.1 10.5 30.4 7.4 (0.3) 0.81 6 5.6 6.4 (1.2) 2.3 (1.1) 0.12 33 LT, NP L658 8.4 7.4 13.2 9.2 (0.5) 1.01 5.59 5.0 8.0 (1.4) 2.4 (1.1) 0.13 33.8 NP L164 20.3 4.94 7.1 9.2 (0.5) 1.02 5.15 4.6 8.6 (1.3) 2.5 (0.5) 0.20 36.5 NP Note: Photic depth is defined as the depth at which 99% of surface PAR is attenuated.

49 Table 2.2: Pairwise Pearson correlation coefficients of select chemical, physical and biological variables from Table 2.1, excluding fish species. Statistical significance is indicated by * (p £ 0.05; n = 8).

Area Zmax Zmean DOC Kd Photic Therm- TP Chla Mean (ha) (m) (m) (mg/L) (m-1) depth ocline (µg/L) (µg/L) light (m) depth irrad- (m) iance (Im) Zmax 0.67 (m)

Zmean 0.47 0.93* (m) DOC -0.21 -0.42 -0.59 (mg/L)

Kd -0.15 -0.37 -0.54 0.99* (m-1) Photic depth 0.05 0.37 0.58 -0.94* -0.91* (m) Thermocline 0.17 0.53 0.76* -0.93* -0.91* 0.93* depth (m) TP -0.35 -0.57 -0.67 0.91* 0.88* -0.75* -0.81* (µg/L) Chlorophyll a 0.04 -0.31 -0.49 0.93* 0.93* -0.89* -0.85* 0.88* (µg/L) Mean light -0.19 -0.34 -0.20 -0.66 -0.69 0.64 0.44 -0.40 -0.63 irradiance (Im) Area above photic 0.02 -0.16 -0.05 -0.67 -0.69 0.64 0.46 -0.43 -0.54 0.91* depth (%)

50

Fig. 2.1: Relationships between DOC concentration and (a) light attenuation, (b) total phosphorus, (c) epilimnetic chlorophyll a, and (d) mean light irradiance.

51

Fig. 2.2: Patterns of habitat structure and carbon isotopes in relation to DOC concentration among the study lakes. (a) Differences between photic depth and thermocline depth decline as epilimnetic DOC concentration increase. (b) δ13C values of epilimnetic phytoplankton (grey triangles), hypolimnetic phytoplankton (black squares), terrestrial organic matter (black line), and zooplankton (open circles) in relation to epilimnetic DOC concentration in July of 2017. (c)

The difference between δ13C of zooplankton and δ13C of surface phytoplankton (black circles)

52 and difference between δ13C of hypolimnetic phytoplankton and surface phytoplankton in July

2017 (grey triangles) both become more similar to surface phytoplankton with increasing DOC concentration.

53

Fig. 2.3: The vertical distribution of chlorophyll a in the study lakes in July 2017. Lakes are arranged in order of DOC concentration, from low to high (3.5 – 9.2 DOC mg/L). Black circles represent chlorophyll a, solid black lines represent the thermocline depth, dashed lines represent the photic depth, and solid red lines represent the depth of maximum dissolved oxygen concentration.

54

Fig. 2.4: (a-e) Vertical distribution of suspended C : P ratios (by mass) and (f-j) chlorophyll a : suspended C (by mass) during the open water season (May – October) in 5 of the 8 study lakes.

(k & l) Average epilimnetic suspended C : P and chlorophyll a : suspended C ratios among lakes. Boxplots represent the 5th, 25th, 50th, 75th and 95th percentiles of ratios throughout the open

55 water season. Non-overlapping notches indicates significant difference at 95% confidence limits.

Black points represent outliers. Lakes are arranged from low to high DOC concentration (left to right; 3.5 – 7.4 mg/L). Dashed lines indicate the photic depth. Note, data for L626 is from 2016 because a profile was not taken in 2017.

56

Fig. 2.5: Zooplankton allochthony (utilization of terrestrial resources) increased with epilimnetic

th DOC concentration (Spearman rank correlation, rs = 0.79, p = 0.03). Boxplots represent the 5 ,

25th, 50th, 75th and 95th percentiles of the posterior distribution from MixSIAR models.

57

Fig. 2.6: (a) Total zooplankton biomass declines with DOC concentration (r = –0.92, p = 0.001), and (c) increases with photic : thermocline depth ratio (r = 0.73, p = 0.04). (b) Copepod biomass declines with DOC concentration (r = –0.87, p = 0.004). (d) The relative abundance of zooplankton (% of biomass), arranged from low to high epilimnetic DOC concentration. Relative abundance of each group did not vary predictably with DOC concentration (p > 0.05). Note, y- axis on panels a-c are log10 transformed.

58

Fig. 2.7: Average annual zooplankton biomass of IISD–ELA lakes (n = 30) declined with epilimnetic DOC concentration (r = –0.62, p < 0.001). Black diamonds and grey circles are data from the extended analysis representing data collected by net or tube sampler, respectively.

White triangles are data from the eight lake spatial survey.

59 Connecting statement

In the previous thesis chapter, I outlined how DOC-mediated effects on vertical habitat structure influenced zooplankton basal resource use and biomass. Using Bayesian stable isotope mixing models, I showed that hypolimnetic algal peaks were an important resource for zooplankton in low-DOC lakes. As DOC concentrations increased, hypolimnetic chlorophyll a concentrations were reduced in prominence relative to epilimnetic concentrations and zooplankton acquired proportionately more energy from terrestrial resources. Seston quality as indicated by C : P and chlorophyll a : C ratios was better for zooplankton at depth relative to surface waters. Zooplankton biomass declined with DOC concentration and with reliance on terrestrial resources. Given that zooplankton represent an important resource to fish, and concentrations of DOC are anticipated to increase in boreal lakes (Evans et al., 2005; Finstad et al., 2016; Monteith et al., 2007), corresponding declines in zooplankton biomass may negatively influence the productive potential of inland fisheries.

In my next thesis chapter, I describe how benthic invertebrate and benthivorous fish biomass (White Sucker, Catostomus commersonii) and basal resource utilization varied with

DOC concentration among eight boreal lakes. While declines in fish biomass with DOC concertation are well documented in the literature, mechanisms by which DOC affects fish productivity are poorly understood. The aim of this chapter is to evaluate potential mechanisms that link declining fish biomass with DOC, specifically, looking at how DOC influences benthic invertebrate prey availability and basal resource use.

60 Chapter 3: DOC-mediated effects on resource availability suppresses benthivorous fish biomass

Abstract

Inputs of terrestrially derived dissolved organic carbon (DOC) to aquatic ecosystems have been increasing over the past few decades in the northern hemisphere and these changes have potential to influence fish and invertebrate productivity. I surveyed eight stratified boreal lakes to evaluate how benthic invertebrate and benthivorous fish biomass (White Sucker,

Catostomus commersonii) and basal resource utilization varied with DOC concentration. Using a three-isotope Bayesian mixing model, I determined that utilization of terrestrial organic matter

(allochthony) by White Sucker increased positively with DOC concentration, but greater allochthony was related to lower White Sucker biomass as catch-per-unit-effort (bCPUE). Both

White Sucker bCPUE and chironomid biomass were positively related to mean light irradiance, with the highest biomasses of fish and chironomids occurring in lakes with a higher proportion of their volume in the photic zone. White Sucker bCPUE was also strongly positively correlated with chironomid biomass. This suggests that DOC-mediated resource limitation may influence fish productivity. My results provide mechanistic links between declines in fish biomass with increasing DOC concentration.

Introduction

Fish production is among the most important ecosystem services that freshwater habitats provide (Brooks et al., 2016; Holmlund & Hammer, 1999) and factors influencing the productive potential of inland fisheries are of both basic (e.g. understanding energetic pathways and efficiency in ecosystems) and applied interest (e.g. improving production for commercial and recreational harvest). Inputs of terrestrial DOC to nutrient-poor lakes fundamentally shape their

61 physical and chemical characteristics and biological interactions in these systems (Jones, 1992;

Karlsson et al., 2015; Prairie, 2008; Solomon et al., 2015). Terrestrial carbon can account for upwards of 95% of the DOC pool (Wilkinson et al., 2013), and terrestrial DOC inputs are increasing in many lakes in the northern hemisphere coinciding with changes in temperature, hydrological processes, land use, and reductions in acid deposition (Emmerton et al., 2018;

Evans et al., 2005; Finstad et al., 2016; Monteith et al., 2007; Paterson et al., 2019). As a result, there is a need to understand how changes in these inputs may affect fish productivity.

Terrestrial DOC can influence the productivity of all lake trophic levels, from primary producers to fish. In boreal lakes, high concentrations of DOC often stain water brown, resulting in shading of the water column and alteration of the vertical distribution of light, heat, oxygen and nutrients (Jones, 1992; Read & Rose, 2013; Solomon et al., 2015; Tanentzap et al., 2008).

DOC-mediated effects on light availability can suppress benthic and pelagic primary production

(Ask et al., 2009a; Godwin et al., 2014), which in turn can limit resources for benthic and pelagic invertebrates that are key dietary items for many fish species (Karlsson et al., 2009). Further, the

DOC concentration required to diminish the photic depth to less than lake depth is greater in shallow lakes than deep (Finstad et al., 2014). As such, the proportion of energy generated in benthic habitats declines more rapidly in deeper lakes with increasing DOC concentration (Ask et al., 2009a). Higher DOC concentrations may also decrease the foraging efficiency of visual predators, such as fish (Solomon et al., 2015; Weidel et al., 2017).

DOC may also affect habitat availability for consumers through its effects on dissolved oxygen (DO) and temperature. Terrestrial DOC absorbs light, concentrating heat at the lake surface, resulting in shallower, more stable epilimnia (Houser et al., 2008; Read & Rose, 2013;

Tanentzap et al., 2008). As a result, a greater proportion of lake volume and sediment area may

62 fall within the hypolimnion, where temperatures are lower and water is isolated from atmospheric exchange. DOC-mediated effects on oxy-thermal habitat availability can have consequences for fish and invertebrate productivity (Craig et al., 2015; Stasko et al., 2012). Loss of optimal oxy-thermal habitat volume for Lake Trout occurred more frequently in high DOC lakes (Guzzo & Blanchfield, 2017) and benthic invertebrate productivity was supressed in high

DOC lakes due to a reduction in oxygenated habitat (Craig et al., 2015).

In small, nutrient poor lakes, terrestrial organic carbon can be several orders of magnitude more abundant than carbon produced by within-lake primary producers (Jansson et al., 2008; Karlsson et al., 2012; Wilkinson et al., 2013). Using food web tracers such as stable isotopes and fatty acids, it has also been established that fish directly and indirectly utilize terrestrial organic matter as a resource and that relative utilization of terrestrial organic matter increases with DOC (Karlsson et al., 2015; Solomon et al., 2011). Despite these findings, it is unclear whether increases in terrestrial organic matter can compensate for decreases in autochthonous primary producers (Ask et al., 2009a; Karlsson et al., 2015). Terrestrial organic matter is often considered a low quality resource for invertebrate and fish consumers (Karlsson et al., 2015), in part because it lacks essential fatty acids, and has a lower trophic transfer efficiency

(Berglund et al., 2007; Brett et al., 2009). Two correlative studies, as well as Chapter 2 of this thesis, found that increased utilization of terrestrial organic matter was associated with decreased biomass production in zooplankton and fish (Karlsson et al., 2015; Kelly et al, 2014), and others have observed lower fish biomass and growth rates with increasing DOC concentration (Benoît et al., 2016; Craig et al., 2017; Karlsson et al., 2009; Seekell, 2018).

In contrast to these studies, there is also evidence to suggest that inputs of terrestrial DOC can subsidize lake foods webs (Carpenter et al., 2005; Polis et al., 1997; Tanentzap et al., 2014).

63 In a study of eight freshwater deltas, greater DOC inputs resulted in a trophic upsurge, whereby zooplankton biomass was stimulated by terrestrial DOC, leading to faster growth rates in young- of-year Yellow Perch (Tanentzap et al., 2014). Terrestrial organic matter can also subsidize lake food webs by increasing nutrient availability (Vasconcelos et al., 2018). Inputs of nutrients and

DOC to boreal lakes are closely associated with each other and addition of these nutrients can stimulate primary producers, with concomitant effects on higher trophic levels (Dillon & Molot,

2005; Vasconcelos et al., 2018). Indeed, in a microcosm experiment simulating the effects of increased DOC inputs on alpine lakes and in a temperate whole-lake DOC manipulation, DOC- bound nutrients stimulated phytoplankton production, leading to increased zooplankton productivity and/or density (Kelly et al., 2016; Kissman et al., 2013). However, because DOC both increases nutrient supply and suppresses light availability, there appears to be a DOC threshold below which DOC-bound nutrients can increase primary production, and above which

DOC-induced light-limitation reduces primary production (Seekell et al., 2015a; Seekell et al.,

2015b). Similar effects have been observed for fish, where biomass increased with DOC concentrations to a threshold and then declined (Finstad et al., 2014).

I used a survey of eight stratified boreal lakes to investigate how the biomass of invertebrates and a dominant fish consumer, White Sucker (Catostomus commersonii), varied with DOC concentration and how DOC-mediated effects on basal resource availability. Using a three-isotope Bayesian mixing model, I determined that utilization of terrestrial organic matter

(allochthony) by White Sucker increased with DOC concentration. However, White Sucker biomass catch-per-unit-effort (bCPUE) was inversely related to allochthony, suggesting that terrestrial organic matter does not provide a resource subsidy. Both White Sucker bCPUE and benthic invertebrate biomass was positively related to average light climate and the proportion of

64 lake area above the photic depth, where the highest biomass for fish and invertebrates occurred in lakes with a higher proportion of their volume in the photic zone. These results provide a mechanistic link between declines in fish biomass with increasing DOC concentration and provide insight into the potential consequences of increasing DOC concentrations on lake food webs.

Methods

Study Sites and sample collection

I used a space-for-time approach to assess the effects of DOC on lake food webs with eight boreal lakes ranging from 3.5 – 9.2 mg/L DOC at the International Institute for Sustainable

Development Experimental Lakes Area (IISD-ELA) field station in northwestern Ontario. The lakes are small (8.4 –56 ha), dimictic, and nutrient poor (5 – 9 µg/L total phosphorus), with negligible macrophyte growth (Chapter 2, Table 2.1). The watersheds are dominated by coniferous forest and are minimally impacted by development. All lakes contain populations of

White Sucker, which dominate fish biomass in IISD-ELA lakes, similar to other boreal lakes across Ontario (Chu et al., 2016; Mills et al., 1987; Trippel & Harvey, 1987). Each lake also has at least one species of piscivore, either Northern Pike or Lake Trout , and variable forage fish communities that include Yellow Perch and cyprinids (e.g. Fathead Minnow (Pimephales promelas), Pearl Dace (Margariscus margarita), Northern Redbelly Dace (Chrosomus eos). In

IISD–ELA lakes, there are no observable differences in White Sucker growth rates among lakes with differing piscivore predators (Chalanchuck, 1998).

Epilimnetic water samples were collected either bi-weekly (4 lakes) or monthly (4 lakes) for DOC, particulate and dissolved nutrients (N and P) and chlorophyll a, with concurrent profiling for DO, temperature and light as outlined in Chapter 2 methods. Light attenuation (Kd)

65 and mean light irradiance of the whole lake volume (Im; mean light irradiance) were calculated as described in Chapter 2.

Fish and invertebrate biomass

I estimated White Sucker biomass by deploying 2 – 3 trap nets (Beamish, 1973; Guzzo et al., 2014; Mills et al., 1987) in the littoral zone (< 3 m depth) of each lake for 4 – 6 weeks during the spawning period of spring 2017 (May to mid-June). White Sucker congregate during spawning, so nets were set at known spawning sites. Nets were checked every few days and

White Sucker were enumerated, weighed and measured for total and fork length. Mean biomass catch-per-unit effort (bCPUE) was calculated by dividing the total weight of catch by ‘net days’

(number of traps per lake multiplied by number of days set). Thus, bCPUE represents the mass of White Sucker caught per net per day (Guzzo et al., 2014).

Benthic invertebrates were collected from each lake over a three-week period in August of 2017 using a modified version of the sampling design of Craig et al. (2015). Samples were collected with a gravity corer (7 cm diameter, 38.48 cm2; Aquatic Research Instruments Gravity

NLA corer, Hope, ID, USA) at 4-6 depths in each lake, along five replicate transects. Every lake was sampled at 0.5 m, 1 m, 4 m, and 8 m, except L164, where the deepest sample was collected at 7 m. In deeper lakes, samples were also taken at 8 m and 12 m (except L626, where the deepest sample was 11.2 m). When the maximum depth exceeded 20 m, samples were also collected at 18 m. The top 5 cm of each sediment core was retained for invertebrate analysis and samples were sieved through a 500 µm mesh and preserved in 70% ethanol. Benthic invertebrates were sorted under a dissecting microscope and identified. I evaluated sorting efficiency by performing spot checks on 10% of the samples. Less than 10% of the total number

66 of organisms per sample were missed, indicating acceptable sorting efficiency (Environment

Canada, 2012).

Benthic invertebrates were identified under a dissecting microscope to genus, except

Diptera, which were identified to family, following the keys of Merritt et al. (2008) and

Holsinger (1972). Each individual was photographed using a digital microscope camera and body length or shell width measured using ImageJ software (Natural Institutes of Health, U.S.A).

Length measurements were converted to mass using published length-weight relationships

(Baumgärtner & Rothhaupt, 2003; Benke et al., 1999). I expressed benthos biomass in four ways; as depth-specific averages and as whole-lake averages weighted by the area of each depth zone, based on morphometric maps, and littoral and profundal averages, which were weighted by area of each sampling depth < 4 m, and weighted by area of sampling depths below the thermocline, respectively. I also estimated average whole-lake, littoral and profundal biomass estimates for chironomids alone.

Zooplankton were collected bi-weekly or monthly by vertically hauling a 50 µm net through the whole water column during the day at the deepest point of the lake (n = 6 – 12 collections per lake) and then counted and converted to biomass following the methods of

Paterson et al. (2010).

Resource Use

I used stable isotopes of hydrogen (H), carbon (C), and nitrogen (N) to estimate the relative dietary contributions of pelagic primary production, benthic primary production, and terrestrial organic matter to White Sucker and benthic invertebrates.

67 Terrestrial organic matter and pelagic primary production (i.e. phytoplankton) stable isotopes were sampled according the methods outlined in Chapter 2. Stable isotopes in benthic primary producers were determined from periphyton growing on unglazed ceramic tiles that were suspended at 1 m along the shore of each lake. Tiles were initially set out in May and each month thereafter two tiles were removed and colonized periphyton were scraped off, pooled, and frozen (n = 5 replicates per lake). Macrophytes were excluded as a potential benthic resource because macrophyte growth is negligible in my study system.

Littoral benthic invertebrates (benthos) for stable isotope analyses were sampled 2 – 4 times per lake through June-August at ~1 m depth with a D-framed net and separated by order.

Profundal chironomids were collected using an Ekman grab on 2 – 4 occasions below the thermocline, or at the deepest portion of the lake. Zooplankton were collected monthly (n = 6 per lake) by vertically hauling a 150 µm net through the whole water column at the deepest point in the lake. White Sucker were collected by gill-nets and/or trap nets and a single pelvic fin ray was taken from each individual for analysis. To account for potential intra-fin variability, only the tip of each fin ray was used for isotope analysis (Hayden et al., 2015). All fish samples for stable isotope analyses were collected in September-early October to account for the assimilated diet during the open water season based on tissue turnover (Vander Zanden et al., 2015). All White

Sucker analyzed were > 150 mm in length, and assumed to be adults with respect to their feeding ontogeny (Hamilton Stewart, 1926). All organic samples were freeze dried or oven-dried at 60 °C, homogenized if necessary, and stored in a desiccator until analysis.

Analyses of d13C and d15N were carried out at the University of Waterloo Environmental

Isotope Laboratory using a Finnegan Deltaplus XL-EA mass spectrometer. d2H was determined by the Colorado Plateau Stable Isotope Laboratory, Northern Arizona University following the

68 methods of Doucett et al. (2007) for organic samples, including correction of exchangeable H using bench-top equilibrium, and d2H water samples were analyzed by cavity-ring-down laser spectroscopy.

Bayesian Mixing Model

I used a Bayesian Mixing model, ‘Bayesian Mixing models in R’ (MixSIAR; Stock et al.,

2018), to estimate the proportional use of pelagic primary production, benthic primary production, and terrestrial organic matter by White Sucker and benthic invertebrates. This model uses three matrices: (1) the consumer matrix, which is the isotopic values of White Sucker or benthic invertebrate replicates; (2) the source matrix, where I used lake specific averages (±SD) of benthic (i.e. periphyton) and pelagic (i.e. phytoplankton) primary producers, and an average of terrestrial organic matter among all watersheds (± SD); and (3) a matrix of mean (±SD) discrimination factor of the per trophic level enrichment of isotopes. For each model, I ran three

Markov chains with uninformed Dirichlet priors for 1,000,000 iterations, with 500,000 iteration burn-in, at a thinning rate of 500 (Stock et al., 2018). Model convergence was assessed using

Gelman-Rubin test within the MixSIAR package, where a scale reduction factor < 1.1 indicates an acceptable model. Posterior distributions were skewed for some mixing models, so medians were used as point estimates. Isotopic values were similar among littoral benthic invertebrate orders, so I fitted a single model using an average value for all samples within each lake

(Appendix 3A).

I used δ13C trophic discrimination factors of 0.4 ± 1.3‰ for White Sucker (Post, 2002), and 0.0 ± 1‰ for invertebrates to capture the range of variability in these assumptions (Vander

Zanden & Rasmussen, 2001), and 2.98 ± 0.97‰ for δ15N for both fish and invertebrates based on

69 the average of published trophic discrimination factors for freshwater organisms (Vanderklift &

Ponsard, 2003). The enrichment of δ2H across trophic levels occurs by progressive incorporation of environmental water and was corrected using the model of Solomon et al. (2009):

" !"#" = 1 − (1 − !)

Where !"#" is the proportion of H from environmental water in consumer tissue, t is the trophic position of a consumer above primary producers, and ! is the per-trophic-level contribution of dietary water to consumer H. I used a conservative ! value of 0.2 (Wilkinson et al., 2015), and a ! value of 0.39 for profundal chironomids (Brett et al., 2018). Trophic position

(t) was estimated using the following equation from Post (2002):

./ ./ ) = + + (- 0123#45678 3#419:27 − - 0;612)/∆4

15 where l is the trophic position of the organism used to estimate δ Nbase (i.e. periphyton),

15 δ Nsecondary consumer is the isotopic signature of either White Sucker or benthos, and Dn is the trophic discrimination factor for δ15N. Once the proportion of environmental water in a consumer’s diet (!"#") and trophic position is calculated, a correction factor for environmental water enrichment can be derived (Berggren et al., 2014):

> > > > - ? @ABCDℎF@A) = - ?16:GH2 − (- ?16:GH2 − !"#" × - ?J6"27)/(1 − !"#" )

> 2 > Where - ?J6"27 is the δ H of water, and - ?16:GH2 is the sample being corrected for.

70 I also ran additional mixing models to estimate the extent by which White Sucker relied on littoral benthic invertebrates, profundal chironomids and zooplankton, and replaced basal primary producers as ‘sources’ with the isotopic values (± SD) of the potential prey items, adjusted for trophic discrimination. All data used for isotopic mixing models are summarized in

Appendix 3A.

Statistical analysis

I used Pearson correlations to evaluate hypothesized drivers of White Sucker bCPUE and invertebrate biomass, including resource availability, resource utilization, and habitat availability. Chlorophyll a and DOC were used as predictors of resource availability. I assumed

DOC concentration is representative of the amount of terrestrial organic matter potentially available to consumers (Karlsson et al., 2015). I also used average whole-lake chironomid biomass and zooplankton biomass as indicators of potential resource availability in White Sucker bCPUE models. Allochthony was used as an indicator of basal resource utilization, as well as the relative contribution of prey items for White Sucker models. Mean light irradiance, volume- weighted hypolimnetic DO concentration, proportion of lake area above the photic depth, and mean depth were used as indicators of habitat availability. Model assumptions were assessed visually by comparing quantile-quantile plots, and histograms of residuals. White Sucker bCPUE was log10 transformed to meet model assumptions.

Results

Lake characteristics

Although my lakes were selected to span a gradient of DOC concentrations, they also covered gradients of other limnological characteristics. Light attenuation ranged from 0.35 to

71 1.02 m-1 and was highly positively correlated with DOC (r = 0.99, p < 0.001), as well as total phosphorus (r = 0.91, p = 0.002), and epilimnetic chlorophyll a (r = 0.93, p < 0.0001; Chapter 2,

Table 2.1 & 2.2), which were positively correlated with each other. Concentrations of DOC also influenced the proportion of lake area above the photic depth (depth of 1% surface PAR) when included in a multiple regression model with mean depth (% area above photic depth = 164.9 –

10.1 DOC – 6.1 mean depth; R2 = 0.76, p = 0.03), where DOC and mean depth account for 46% and 30% of model variance, respectively (Fig. 3.1). Lake area above the photic depth and mean light irradiance were also positively correlated (r = 0.91, p = 0.002). DOC concentration was not significantly correlated with mean or maximum depth, lake area or volume, or volume-weighted hypolimnetic DO (p > 0.05). All other correlations were not statistically significant.

Invertebrate and White Sucker biomass

The average whole-lake depth-weighted biomass of benthic invertebrates ranged from

0.16 to 0.94 g/m2 (Table 3.1), and average densities ranged from 1498 to 4215 individuals/m2 among lakes (Appendix 3B). Chironomids were the dominant taxa in all lakes, accounting for

43% to 87% of individuals in shallow sites (Fig. 3.2a), and nearly 100% of individuals at deep sites, with the exception of L239, where I encountered high densities of the amphipod

Monoporeria spp. at 18 m. At littoral sites, functional feeding groups were dominated by collector-gatherers (Fig. 3.2b), and the proportion of shredders increased with DOC concentration (r = 0.82, p = 0.01), from 1% to 18% of individuals. Average among-lake benthos biomass was not significantly related to DOC concentration, light attenuation, average light climate, maximum or mean depth, area, chlorophyll a, nutrients, or allochthony (p > 0.05).

Depth-specific biomass estimates were also not linearly related to DO or temperature measured on the day of collection or averages for the open water season (Appendix 3B, Fig. 3B.1; p >

72 0.05). Benthic invertebrate biomass was variable at high DO concentrations but was exclusively low at concentrations < 2 mg/L. Despite this, I found no statistically significant relationships between profundal invertebrate biomass and hypolimnetic DO concentration (p = 0.6).

Average whole-lake chironomid biomass ranged from 0.08 to 0.37 g/m2 and was positively correlated to mean light irradiance (Fig. 3.3a; r = 0.82, p = 0.01) and the proportion of lake area above the photic depth (r = 0.77, p = 0.02). The biomass of chironomids was similar among lakes in the littoral zone (Appendix 3B) and the greatest differences in chironomid biomass occurred below the thermocline (profundal zone). Zooplankton biomass ranged from

0.13 to 0.54 g/m2 and was significantly negatively related to DOC concentration (r = –0.92, p =

0.001; Chapter 2).

White Sucker bCPUE ranged from 0.13 to 12.8 kg/net/day and was positively related to the average whole-lake chironomid biomass (Fig. 3.3b; r = 0.74, p = 0.04), mean light irradiance

(r = 0.90, p = 0.002), and the proportion of lake area above the photic depth (r = 0.88, p = 0.004).

Nutrients, epilimnetic chlorophyll a, mean depth, and average benthic invertebrate biomass or zooplankton biomass were not univariately correlated with White Sucker bCPUE.

Resource use

Terrestrial organic matter δ2H values were heavier than periphyton by 22‰ (± 3.6‰) and phytoplankton by 46‰ (± 4.3‰), on average (±SD) (Appendix 3B). There was also strong isotopic separation using δ15N signatures, with t-OM being most depleted on average (–5.5‰ ±

1.6‰) versus 1.1‰ ± 1.3 δ15N for periphyton, and 1.9‰ ± 1.5 δ15N for phytoplankton.

Terrestrial organic matter and epilimnetic phytoplankton had similar δ13C values on average, but both were both more negative than periphyton (by 2 – 9 ‰).

73 Median littoral benthic invertebrate allochthony ranged from 27% to 44% (median of posterior estimate) among lakes (Appendix 3C) and was positively correlated with DOC concentration (Fig. 3.4a; r = 0.83, p = 0.01) and light attenuation (r = 0.80, p = 0.02). In the majority of my lakes, profundal chironomids were more depleted in δ13C than epilimnetic phytoplankton and terrestrial organic matter (Appendix 3A), as well as metalimnetic POM

(Chapter 2). As a result, I was unable to generate meaningful mixing model results and no reliable solution was feasible. Profundal chironomid δ13C values ranged from –30.4‰ to –41.9‰

(Appendix 3A), and were positively correlated with late summer volume-weighted hypolimnetic

DO concentrations (Appendix 3D, Fig. 3D.1; r = 0.84, p = 0.009), suggesting possible contributions from methanotrophic bacteria as seen by Jones et al. (2008; Appendix 3D).

White Sucker allochthony ranged from 26% to 48% (medians of posterior distribution;

Appendix 3C) and was positively related with DOC concentration (Fig. 3.4b; r = 0.86, p =

0.007), light attenuation (r = 0.88, p = 0.004), and negatively correlated with mean light irradiance (r = –0.79, p = 0.02). White Sucker allochthony was also positively related to littoral benthic invertebrate allochthony (r = 0.83, p = 0.01; Appendix 3C, Fig. 3C.1). Utilization of littoral primary producers ranged from 6% to 55% but was not related to DOC concentration or light attenuation (p > 0.05; Appendix 3C). The main dietary items supporting White Sucker in my lakes were benthic invertebrates (profundal + littoral), which accounted for 56% to 88% of

White Sucker diet (Appendix 3C, Table 3C.3). Although chironomid biomass was a strong predictor of White Sucker bCPUE, I found no relationship between bCPUE and reliance on benthic invertebrates (p = 0.17; profundal + littoral) or between bCPUE and reliance on profundal chironomids (p = 0.12), and White Sucker reliance on benthic invertebrates was not related to average whole-lake chironomid biomass (p = 0.33; Appendix 3C, Fig. 3C.4). However,

74 bCPUE was negatively correlated with White Sucker allochthony (Fig. 3.4c; r = –0.82, p = 0.01).

Zooplankton accounted for 12% to 44% of White Sucker diet, and the proportion of zooplankton in White Sucker diet was not significantly correlated with DOC (p = 0.086), although consumption of zooplankton was greatest in the darkest lakes (Appendix 3C).

Discussion

The results of my survey highlight how mean light irradiance, which is a function of

DOC concentration and lake depth, affects variations in White Sucker biomass in eight nutrient- poor, boreal lakes. The results of this study support other among-lake surveys that have demonstrated declines in fish biomass associated with increasing DOC concentrations, presumably through the negative effects of DOC on light and in-lake primary production (Benoît et al., 2016; Finstad et al., 2014; Karlsson et al., 2009, 2015). The effect of DOC on mean light irradiance is shaped by lake depth, where deeper lakes are comparatively more vulnerable to increased DOC loading because a greater proportion of lake area falls below the photic depth, relative to shallower lakes of the same DOC concentration (Finstad et al., 2014). Indeed, fish biomass was lowest in lakes with a high proportion of their volume outside of the photic zone. In addition to finding correlations between fish biomass and mean light irradiance, I found support for two potential mechanisms that may affect these relationships. First, mean light irradiance was correlated with allochthony, where in low light (high DOC) lakes a greater proportion of fish biomass was supported by terrestrial organic matter, which is a relatively poor-quality resource compared to autochthonous production. Second, chironomid biomass was positively correlated with mean light irradiance, suggesting that resource limitation of benthivorous fish may be greater in low light lakes.

75 White Sucker derived a considerable proportion of their energy from terrestrial organic matter (26 – 48%), and utilization of this resource increased with DOC concentration and light attenuation. Allochthony also decreased with higher mean light irradiance, which accounts for both changes in both light attenuation and lake morphometry. Two previous comparative studies had similar findings, where consumer allochthony increased with DOC and/or light attenuation, and demonstrated a similar range of allochthony among different fish species (Karlsson et al.,

2015; Solomon et al., 2011). White Sucker allochthony was also positively related to benthic invertebrate allochthony (Appendix 3C). Benthic invertebrates were the primary prey of White

Sucker (based on stable isotopes), suggesting indirect utilization of terrestrial carbon through invertebrate consumption. Although White Sucker may also consume terrestrial organic matter directly, this is only observed when invertebrate prey are scarce (Ahlgren, 1990). In a few lakes, utilization of benthic primary producers (periphyton) is low relative to basal dietary estimates of benthivorous fish from other studies (e.g. Karlsson et al., 2009). This may be related to greater predation in the IISD-ELA lakes by White Suckers upon profundal chironomids (10% to 57% of diet; Appendix 3C) that likely rely more on pelagic energy pathways (Appendix 3D). Because methane-oxidizing bacteria (MOB) were not included as a potential energy pathway in the mixing models, allochthony estimates may be biased towards artificially high phytoplankton use.

To examine this possibility, I explored the potential contribution of MOB using a series of hypothetical models and the methods of Ravinet et al. (2010) (Appendix 3D). These models suggest that the potential contribution of methane oxidising bacteria to White Sucker biomass via profundal chironomids was likely less than 11% (Appendix 3D, Table 3D.3; Ravinet et al.,

2010).

76 Similar to the findings of Karlsson et al. (2015), I observed declines in White Sucker bCPUE with increasing allochthony, suggesting that terrestrial organic matter is not a resource subsidy, despite being widely used by secondary consumers. There are several possible explanations for declining White Sucker biomass with increasing allochthony. DOC-mediated effects on light can influence where and how much primary production can be supported within a lake (Godwin et al., 2014; Karlsson et al., 2009; Vadeboncoeur et al., 2008). With increasing

DOC, basal production switches from autotrophic dominance to production dominated by bacterial heterotrophs primarily supported by terrestrial organic matter.

There is also considerable evidence that the quality of terrestrial organic matter is poor and that terrestrial organic matter cannot completely compensate for the loss of autochthonous production (Ask et al., 2009a). In a small boreal lake in northern Sweden, there appeared to be an upper limit to the extent to which terrestrial organic matter could support top consumers (~60%), despite terrestrial inputs being three-orders of magnitude greater than that of autochthonous production (Karlsson et al., 2012). My allochthony estimates for White Sucker (26 – 48%) fall below this proposed ~60% upper limit. This upper limit may arise in part because a large fraction of terrestrial inputs is highly recalcitrant, and likely never mobilized into the food web (Cole et al., 2002a; Koehler et al., 2012). Further, terrestrial organic matter is nearly devoid of essential fatty acids required for growth and reproduction in fish (Brett et al., 2009; Brett & Muller-

Navarra, 1997), and terrestrial energy pathways have high respiratory losses as energy is mobilized up the food web (Berglund et al., 2007; Kritzberg et al., 2005). Consequently, these high respiratory losses and lack of essential fatty acids result in trophic transfer efficiency of terrestrial resources being 5 –10 times less efficient than autochthonous resources (Brett et al.,

2017).

77 In my study lakes, the decline in White Sucker biomass with increasing DOC concentrations occurred with concomitant decreases in benthic invertebrates, which are favoured prey of White Suckers (Hamilton, 1971; Hamilton Stewart, 1926; Mills et al., 1987; Misra et al.,

1995; Tremblay & Magnan, 1991; Trippel & Harvey, 1987). Lakes with higher DOC concentrations often have lower benthic invertebrate biomass because DOC-mediated light limitation may supress benthic primary production supporting benthic invertebrates (Godwin et al., 2014; Karlsson et al., 2009). Additionally, DOC affects stratification and the vertical distribution of oxygen and temperature, which may affect the availability of suitable oxythermal habitat (Craig et al., 2015). I found that chironomid biomass was strongly, positively correlated with mean light irradiance and White Sucker bCPUE was, in turn, correlated with chironomid biomass. Specifically, lakes with a higher proportion of their lake volume in the photic zone supported more chironomid biomass, and consequently, more White Sucker biomass. This suggests that increases in DOC concentrations and light attenuation may impose greater resource limitation on White Sucker via the availability of benthic invertebrate prey. White Sucker are strongly linked to the benthic food web and other studies also suggest that variations in benthic invertebrate biomass and/or chironomid density can affect White Sucker biomass and growth

(Munkittrick et al., 1991; Trippel & Harvey, 1987). Following experimental acidification of

L223 in the 1970s and 1980s, increases in White Sucker abundance were linked to greater availability of chironomid prey (Mills et al., 1987; Schindler et al., 1985). Chironomids numerically dominate the benthic communities of IISD-ELA lakes, and are the primary prey for

White Sucker (Hamilton, 1971; Hamilton Stewart, 1926; Mills et al., 1987; Misra et al., 1995;

Tremblay & Magnan, 1991; Trippel & Harvey, 1987). In this study, depth-specific benthic biomass was not linearly related with temperature or dissolved oxygen concentration (Appendix

78 3B) as found by Craig et al. (2015). In general, benthic invertebrate biomass decreased with increasing depth within lakes, and biomass was low at low DO concentrations (< 2 mg O2 / L) and variable at higher DO concentrations (Appendix 3B), suggesting that DO concentration may influence invertebrate biomass, but not linearly. Chironomid biomass, however, tended to be greater at deep sites relative to shallow sites, and I found no relationship between profundal chironomid biomass and volume-weighted hypolimnetic oxygen. Because chironomids can tolerate periods of anoxia and may have lower oxygen thresholds than many fish species, it is possible that low dissolved oxygen concentrations may have provided a refuge from predation

(Jonasson, 1984). Chironomid biomass may also be lower at shallow sites due to greater exposure to ultra violet radiation (Donahue et al., 2003).

At littoral sites, utilization of benthic primary production by benthos declined with increasing DOC concentrations, but littoral benthic invertebrate biomass was not related to utilization of this resource or DOC concentration. The lack of a connection between benthic invertebrate biomass and reliance on benthic primary production may reflect changes in community composition. Benthic invertebrates are a diverse group with various feeding strategies (Merritt et al., 2008), and loss of benthic primary production with increasing DOC may drive changes in community composition that favour detrital, or other feeding strategies, while maintaining biomass. Indeed, in my lakes, the proportion of benthic invertebrates at littoral sites belonging to the functional feeding group ‘shredders’ increased with DOC concentration, from

1-18% of individuals.

In many lakes, the majority of chironomid biomass occurred at sites below the photic depth, where benthic primary producers are light limited, and chironomids presumably cannot rely on this basal resource. Indeed, profundal chironomid δ13C values were always more negative

79 than benthic primary producers, and in several lakes more depleted than phytoplankton and terrestrial organic matter. Profundal chironomid δ13C was positively related to volume-weighted hypolimnetic oxygen, indicating potential dietary contributions from methanotrophic bacteria

(Bastviken et al., 2003; Jones et al., 2008; Appendix 3D). Similar relationships between hypolimnetic oxygen and profundal chironomid δ13C have been observed in other systems,

13 where appreciable C-deleption can occur at oxygen concentrations around 2 mg O2 / L (Jones et al., 2008). Nonetheless, these findings suggest that benthic primary production alone is not driving the variability in benthic invertebrate biomass among my lakes.

Declines in fish and benthic invertebrate biomass with reductions in mean light irradiance may be related to decreases in whole-lake primary production (both benthic + pelagic). Ask et al.

(2009a) found that whole-lake energy mobilization declines with increasing light attenuation and increasing depth, because a greater proportion of lake area and volume falls below the photic depth in deep lakes relative to shallow lakes with similar light attenuation (Finstad et al., 2014).

In my lakes, the proportion of lake area above the photic depth was related to both DOC and mean depth and strongly related to mean light irradiance. Consequently, whole-lake average chironomid biomass and White Sucker bCPUE were positively correlated with the proportion of lake area above the photic depth. These results support the findings of Karlsson et al. (2009) and

Craig et al. (2015), where DOC-mediated effects on benthos availability influences benthivorous fish biomass.

This study has several important limitations. First, my findings are based on a limited number of small, dimictic boreal lakes and focus on a single species of fish, White Suckers.

Lakes of different sizes, mixing patterns, and fish species may respond differently to inputs of terrestrial organic matter. For example, shallow lakes with high concentrations of DOC may not

80 become light limited if light can penetrate to the bottom across the entire lake. White Sucker are a species that is tightly linked to benthic food webs (Johnston et al., 2018) and fish species that are more flexible in their feeding ecology or piscivorous fishes, may not respond in a similar manner to declines in benthic prey. However, concentrations of DOC have also been shown to suppress zooplankton biomass (Kelly et al., 2014), thus DOC-mediated prey limitation may still limit the productive potential of fish with more flexible feeding ecologies. In my lakes, I also observed declines in zooplankton biomass with increasing DOC concentrations (Chapter 2).

Second, I could not account for all of the potential dietary sources for profundal chironomids, suggesting possible contributions from methanotrophic bacteria. Because White Sucker consumed profundal chironomids (10 – 57% of diet), they may also have been partially supported by methane oxidizing bacteria, although my analysis shows this contribution was likely small (< 11% of biomass; Appendix 3D; Ravinet et al., 2010). Third, my benthic invertebrate biomass estimates were determined from a single survey and does not account for temporal variations in biomass. Further, my benthic sampling probably did not efficiently collect larger mobile invertebrate taxa. Chironomids numerically dominate the benthic communities of

IISD-ELA lakes however, and are the primary prey for White Sucker (Hamilton, 1971; Hamilton

Stewart, 1926; Mills et al., 1987; Misra et al., 1995; Tremblay & Magnan, 1991; Trippel &

Harvey, 1987). Lastly, I used White Sucker bCPUE and benthic invertebrate standing stock biomass as indicators of productivity, however, biomass and productivity may be decoupled

(Dolbeth et al., 2012). In a hypothetical example, two fish communities with equal biomass but large differences in density (i.e. a few large fish vs. many small fish) may differ substantially in their productivity because growth rates for smaller fish are generally faster, and therefore, have a greater rate of biomass turnover. However, in my lakes, White Sucker density (measured as

81 catch per unit effort) declined with DOC concentration, and the size distribution of White Sucker was more right skewed in low DOC lakes (data not shown). This indicates that a greater proportion of White Sucker biomass in low DOC lakes is made up of smaller individuals, which may be more productive.

These results support a growing body of literature indicating that DOC-mediated light limitation results in lower fish biomass in nutrient poor lakes (Karlsson et al., 2009; Prairie,

2008), and that the effects of DOC on fish biomass is also influenced by lake depth (Finstad et al., 2014; Seekell et al., 2018). This research provides evidence for two, non-mutually exclusive mechanisms that link declines in fish biomass with DOC-mediated light limitation. First, I found support for the conceptual model where mean light irradiance influences availability of benthic invertebrate prey, thereby affecting fish biomass (Karlsson et al., 2009; Craig et al., 2015).

Second, I found that White Sucker biomass declined with increasing allochthony (Karlsson et al.,

2015), suggesting that terrestrial organic matter is not a resource subsidy. Allochthony is primarily controlled by DOC concentration, but is also inversely related to benthic invertebrate density, suggesting that declines of benthos with increasing DOC concentration can also influence how top consumers incorporate energy, with consequences on growth. Increased DOC export from watersheds into lakes in response to climate change and land use (Monteith et al.,

2007; Paterson et al., 2019) are likely to have important effects on the productive potential of inland fisheries and further investigation is required to determine the mechanisms by which DOC suppresses the productivity of upper trophic levels.

82

Table 3.1: Summary of benthic invertebrate, chironomid and White Sucker biomass estimates among lakes. Littoral averages were weighted by area of each sampling depth < 4 m, and profundal was weighted by area of sampling depths below the thermocline.

Benthic invertebrate biomass (g/m2) Chironomid biomass (g/m2) White Sucker bCPUE Lake Littoral Profundal Whole-lake Littoral Profundal Whole-lake (kg/net/day ± SD) L224 0.17 0.34 0.28 0.08 0.34 0.27 1.27 ± 1.54 L373 0.26 0.06 0.16 0.13 0.06 0.08 1.84 ± 1.44 L223 0.25 0.50 0.42 0.10 0.46 0.32 1.54 ± 0.86 L626 0.54 0.53 0.53 0.07 0.53 0.37 12.84 ± 13.25 L442 0.31 0.14 0.20 0.03 0.14 0.11 0.23 ± 0.09 L239 1.05 0.88 0.94 0.06 0.09 0.08 0.13 ± 0.21 L658 1.14 0.14 0.44 0.09 0.14 0.11 0.24 ± 0.29 L164 0.54 0.49 0.50 0.10 0.49 0.30 1.00 ± 2.07

83

Fig. 3.1: DOC concentration and mean depth influence (a) mean light irradiance (mean light irradiance = 0.61 – 0.033 DOC – 0.024 Mean Depth; R2 = 0.93, p < 0.001) and (b) the proportion of lake area above the photic depth.

84

Fig. 3.2: The relative abundance of benthic invertebrate (a) taxa and (b) functional feeding groups (FFG) at shallow sites (< 4 m) among lakes (arranged left to right by increasing DOC concentration).

85

Fig. 3.3: (a) Average whole-lake chironomid biomass in relation to mean light irradiance (r =

0.82, p = 0.01), and (b) log10 White Sucker biomass catch-per-unit-effort (bCPUE) in relation to average whole-lake chironomid biomass (r = 0.74, p = 0.04). Note: y-axis on panel b is log10 transformed.

86

Fig. 3.4: Patterns of allochthony for (a) littoral benthic invertebrate and (b) White Sucker in relation to DOC concentration. Boxplots represent the 5th, 25th, 50th, 75th and 95th percentiles of

87 the posterior distribution from MixSIAR models. (c) The relationship between White Sucker biomass catch per unit effort (bCPUE) and allochthony (r = –0.82, p = 0.01). Allochthony estimates are presented as medians from the posterior distribution of the Bayesian mixing model.

Note: the y-axis is log10 transformed.

88 Chapter 4: General Discussion Insights

The factors that influence the productive potential of inland fisheries are of basic and applied interest, in part, because fish production is among the most important ecosystem services freshwater habitats provide (Brooks et al., 2016; Holmlund & Hammer, 1999). Over the past few decades, terrestrially derived dissolved organic carbon (DOC) has been increasingly recognized as a fundamental driver of food web productivity in nutrient poor lakes (Karlsson et al., 2009;

Prairie, 2008). Despite this recognition, the mechanisms that drive these effects on productivity, particularly for higher trophic levels, remain poorly understood. Further pressing the need to understand how DOC impacts higher trophic level productivity, concentrations of DOC have been increasing in many lakes in the northern hemisphere over the past few decades (Evans et al., 2005; Finstad et al., 2016; Monteith et al., 2007). A better understanding of how terrestrial

DOC modifies food web processes will improve our ability to predict how lakes will respond to increasing DOC inputs.

A resource subsidy is defined as a cross-ecosystem flux of a resource (energy or matter) that increases consumer productivity within the recipient ecosystem (Jones et al., 2012; Polis et al., 1997). Terrestrial inputs are commonly viewed as a resource subsidy, evidenced by increasing utilization of these resources by consumers as concentrations of terrestrial resources increase (Carpenter et al., 2005; Polis et al., 1997; Weidel et al., 2008). Inputs of terrestrial organic matter to lakes may be several orders of magnitude greater than from within-lake primary production (Jansson et al., 2008), and can support upwards of 60% of fish and 80% of invertebrate biomass (Karlsson et al., 2012, 2015; Solomon et al., 2011). While it is clear that lake consumers may rely heavily upon terrestrial resources, there is little empirical data linking

89 terrestrial resource use to consumer productivity (but see: Karlsson et al., 2015; Kelly et al.,

2014).

The two main chapters of this thesis reject the notion that terrestrial organic matter is a resource subsidy for zooplankton, benthic invertebrates or fish the in the boreal lakes under study. Using a survey of eight lakes varying in DOC concentrations and a three-isotope Bayesian mixing model approach, I found that reliance on terrestrial organic matter (i.e. allochthony) by consumers increased with DOC concentration, but increased allochthony was associated with reductions in zooplankton and White Sucker biomass. This is only the second field study to link allochthony with declines in zooplankton and fish biomass (Karlsson et al., 2015; Kelly et al.,

2014). Similar to the findings of Craig et al. (2015), I did not observe any relationship between allochthony and littoral benthic invertebrate biomass, despite allochthony increasing with DOC.

My results still do not support the subsidy hypothesis as there was no detectable increase in benthic invertebrate biomass with allochthony.

These findings contribute to the understanding of how lake consumers will respond to an increase in terrestrial inputs and further support the suggestion that terrestrial organic matter suppresses and does not subsidize lake food webs. These findings are also important because they outline a potential mechanism by which DOC can suppress food web productivity.

Terrestrial organic matter is a poor-quality resource relative to within lake primary production. It is nearly devoid of fatty acids and has high respiratory losses as it mobilizes up the food web

(Berglund et al., 2007), As terrestrial inputs increase and augment within-lake primary production, reliance on terrestrial resources increase and reduce consumer productivity.

During periods of autochthonous resource deficiency, reliance on terrestrial resources may increase in order to maintain basic metabolic function within secondary consumers (Wetzel,

90 1995). Indeed, zooplankton reliance on terrestrial resources increases during winter when autochthonous resource availability is low (Berggren et al., 2015) and several laboratory studies have shown terrestrial resources can support consumers when phytoplankton resources are low, albeit resulting in lower growth and reproduction rates (Brett et al., 2009; Hiltunen et al., 2017;

McMeans et al., 2015; Taipale et al., 2016a). In my lakes, increased zooplankton allochthony corresponded with a loss of high-quality phytoplankton resources deep in the water column.

White Sucker allochthony was strongly correlated with benthic invertebrate allochthony, suggesting indirect utilization of terrestrial carbon through invertebrate consumption, although

White Sucker may also consume terrestrial organic matter directly (Ahlgren, 1990).

In addition to determining linkages between allochthony and consumer biomass, I found that DOC-mediated effects on habitat can influence how consumers acquire energy. For zooplankton, allochthony was low (< 8% of diet) when hypolimnetic phytoplankton was present, which they relied upon heavily. As DOC concentrations and light attenuation increased, primary production became restricted to the upper mixed layer, which coincided with a loss of high- quality phytoplankton below the thermocline. Consequently, in lakes without hypolimnetic phytoplankton, zooplankton allochthony increased to 19 – 27% of the diet. This finding highlighted how DOC-mediated effects on habitat can influence zooplankton energy acquisition, with potential consequences on biomass. Additionally, this work demonstrated that hypolimnetic phytoplankton can be an important resource for zooplankton, and this energy pathway needs to be accounted for in future attempts to assess resource linkages between terrestrial and aquatic habitats.

This research supports the notion that DOC can impact fish productivity through effects on prey availability. I found that chironomid biomass increased with mean light irradiance,

91 which is a function of DOC concentration and mean depth. In other words, lakes with a higher proportion of their lake volume in the photic zone supported greater chironomid biomass.

Consequently, White Sucker bCPUE was correlated with chironomid biomass, and with mean light irradiance. This matches findings by Craig et al. (2015) and Karlsson et al. (2009) who also observed declines in fish biomass as benthic invertebrate biomass declined. This work also supports the idea that lake depth can shape how consumers respond to increasing concentrations of DOC (Finstad et al., 2014; Seekell et al., 2018). I found that the proportion of lake sediment above the photic depth was greatest in low DOC lakes with relatively shallower mean depths, which was also positively correlated with benthic invertebrates and fish biomass.

Future Directions

The need for multiple approaches to assess effects of DOC on lake food webs

While this thesis provided new evidence for mechanisms that link increasing DOC concentration to declines of food web productivity, research is still needed to better understand how terrestrial inputs modify lake food webs. In particular, there is a need to incorporate multiple approaches to assess the effects of DOC on lake food webs. As in the current study, the majority of studies on the effects of DOC on lake food webs utilize spatial DOC gradients

(Benoît et al., 2016; Craig et al., 2015, 2017; Finstad et al., 2014; Karlsson et al., 2009, 2015;

Kelly et al., 2014; Seekell et al., 2018). While these spatial surveys have been useful in describing patterns along environmental gradients and identifying the potential consequences of increasing DOC inputs, they cannot be used to describe how an ecosystem may respond to change over time, and can be confounded by factors that co-vary among lakes in a spatial gradient (e.g. differences in fish communities, lake depth, lake size, etc.; Carpenter, 1998). There

92 is a need for long-term, whole-ecosystem experiments and long-term monitoring programs where many of the limitations of spatial surveys can be addressed, most notably, how ecosystems respond to change over time (Leach et al., 2019). Over the past five years, whole-lake experiments studying the effects of DOC on lake food webs have found contrary results to many of the spatial surveys, where increased DOC concentrations resulted in greater densities of zooplankton and fish (Kelly et al., 2016; Koizumi et al., 2018). Similarly, long-term monitoring studies can alleviate many issues associated with spatial surveys, as well as potential issues with whole-lake experiments. Nonetheless, to fully understand the consequences of increasing DOC concentration on temperate and boreal lakes, there is a need for more whole-lake experiments and long-term monitoring programs to supplement the spatial surveys, mesocosm experiments, and modelling approaches.

The need for production estimates

Despite productivity being a foundational theme in ecology, estimates of fish production in the literature are rare, which, in part, can be attributed to the difficulty of obtaining biomass estimates (Rypel & David, 2017). Several researchers have recommended a resurgence of fish production studies because they are better suited to assess the flow and accumulation of biomass in a population than standing stock biomass alone, and production estimates are a powerful tool for assessing anthropogenic and environmental stress (Dolbeth et al., 2012; Rypel et al., 2015;

Valentine-Rose et al., 2011). Despite this, only one study has evaluated the effects of DOC on fish populations by measuring productivity directly (Karlsson et al., 2015). In general, the majority of studies investigating the effects of DOC on lake food webs rely upon indicators of

93 productivity, including the work in this thesis. While these indicators are widely used, they do not always correlate with production (Bartley et al., 2015) and thus can give spurious results.

There is also a need for more life-history based approaches to evaluate the effects of

DOC on fish. Life-history approaches can provide insight into how changes in DOC affect individual growth rates, age at maturity, life-time and realized fecundity, and maximum theoretical size and weight, etc. that can be used to assess potential environmental stressors and population vulnerability to anthropogenic stressors (Craig et al., 2017). These types of data can complement production estimates because they inform researchers on the mechanisms by which environmental/anthropogenic stressors impact productivity. Further, depending on the method, productivity estimates may require life-history based approaches.

There is also a need for studies that link benthic invertebrate production to DOC concentration. Currently, only Craig et al. (2015) has collected enough data to generate productivity estimates across a DOC gradient. The work in this thesis, as well as work by

Karlsson et al. (2009), have attempted to link benthic invertebrate biomass to DOC concentration, but these efforts were based on single surveys and are missing essential temporal information required to estimate production. Given that benthic invertebrates are essential prey for many fish species, it is surprising how little attention benthic invertebrate production studies have garnered.

The emergence of novel food web tracers

Following recent methodological advancements, the utility of hydrogen stable isotopes as a dietary tracer has become well established (Wassenaar & Hobson, 2003, Vander Zanden et al.

2016). Hydrogen isotopes are particularly useful at tracing terrestrial contributions to aquatic

94 ecosystems and provide greater isotopic separation between terrestrial and aquatic primary production than C isotopes (Doucett et al., 2007). Accordingly, the use of hydrogen isotopes for studies tracing terrestrial contributions to aquatic food webs are now widely used (Cole et al.,

2011; Karlsson et al., 2012, 2015; Kelly et al., 2014; Solomon et al., 2011; Tanentzap et al.,

2017; Wilkinson et al., 2013).

However, using hydrogen isotopes to trace terrestrial contributions to aquatic food webs is dependent on several poorly defined assumptions that strongly influence resource use estimates (Brett et al., 2018; Wilkinson et al., 2015). One of the most important assumptions regards the contribution of dietary-water to consumer tissues. In my dataset, changing the value of dietary water contributions from 0.0 to 0.3 resulted in terrestrial resource estimates ranging from 100% to 0% of zooplankton assimilated diet (Appendix 2A). Similarly, the utility of carbon and nitrogen stable isotopes for tracing dietary contributions also depends upon a series of assumptions (e.g. trophic discrimination factors), and while more defined than assumptions for hydrogen isotopes, can still influence model outcomes (Gulka et al., 2017).

Further confounding the utility of stable isotopes for aquatic food web studies is the difficulty of obtaining stable isotope ratios of phytoplankton. Phytoplankton are difficult to separate from seston, so indirect methods are often used to estimate stable isotope values. Most studies use photosynthetic fractionation factors, where researchers assume a fractionation value relative to the isotopic value of CO2 or water for carbon and hydrogen isotopes, respectively.

Carbon fractionation rates can be strongly influenced by temperature, growth rate, cell size, nutrients etc. and the transferability of a photosynthetic fraction factor may not be applicable from lake to lake (Marty & Planas, 2008). However, unlike carbon, there is no known physiological effect on photosynthetic fractionation factors for hydrogen isotopes, thus

95 fractionation factors appear to be more applicable among systems (Cole et al., 2011). Further, using stable isotope approaches, phytoplankton are often represented in bulk, despite many phytoplankton species having distinct fractionation factors resulting in variable carbon isotope values that are often not accounted for in stable isotope mixing models (Taipale et al., 2016b).

This is especially important because some zooplankton grazers may select for different members of the phytoplankton community.

Over the past few years, the toolbox of dietary tracers available to researchers has grown considerably (Nielsen et al., 2018), including fatty acids, compound-specific isotope analysis,

DNA-based approaches, and additional isotopes (e.g. oxygen, sulfur and iron). Further methodological development, application, and integration of multiple dietary tracing methods will improve our ability to more accurately discern pathways of energy flow in the future.

96 References

Adams, S. M., Kimmel, B. L., & Ploskey, G. R. (1983). Sources of organic matter for reservoir

fish production: a trophic-dynamics analysis. Canadian Journal of Fisheries and Aquatic

Sciences, 40, 1480–1495.

Ahlgren, M. (1990). Diet selection and the contribution of detritus to the diet of the juvenile

White Sucker (Catostomus commeronsii). Canadian Journal of Fisheries and Aquatic

Science, 47, 41–48.

Algesten, G., Sobek, S., Bergström, A. K., Agren, A., Tranvik, L. J., & Jansson, M. (2003). Role

of lakes for organic carbon cycling in the boreal zone. Global Change Biology, 10, 141–

147.

Ask, J., Karlsson, J., Persson, L., Ask, P., Byström, P., & Jansson, M. (2009a). Terrestrial

organic matter and light penetration: Effects on bacterial and primary production in lakes.

Limnology and Oceanography, 54, 2034–2040.

Ask, J., Karlsson, J., Persson, L., Ask, P., Byström, P., & Jansson, M. (2009b). Whole-lake

estimates of carbon flux through algae and bacteria in benthic and pelagic habitats of clear-

water lakes. Ecology, 90, 1923–1932.

Barbiero, R. P., & Tuchman, M. L. (2004). The deep chlorophyll maximum in Lake Superior.

Journal of Great Lakes Research, 30, 256–268.

Bartels, P., Hirsch, P. E., Svanback, R., & Eklöv, P. (2016). Dissolved organic carbon reduces

habitat coupling by top predators in lake ecosystems. Ecosystems, 19, 955–967.

Bartley, D. M., De Graaf, G. J., Valbo-Jørgensen, J., & Marmulla, G. (2015). Inland capture

fisheries: Status and data issues. Fisheries Management and Ecology, 22, 71–77.

97 Bastviken, D., Ejelrtsson, J., Sundh, I., & Tranvik, L. J. (2003). Methane as a source of carbon

and energy for chironomid larvae. Ecology, 84, 969–981.

Baumgärtner, D., & Rothhaupt, K. O. (2003). Predictive length-dry mass regressions for

freshwater invertebrates in a pre-alpine lake littoral. International Review of Hydrobiology,

88, 453–463.

Beamish, R. J. (1973). Design of a trapnet with interchangeable parts for the capture of large and

small fishes from varying depths. Journal of the Fisheries Research Board of Canada, 30,

587–590.

Beard, T. D., Arlinghaus, R., Cooke, S. J., McIntyre, P. B., De Silva, S., Bartley, D., & Cowx, I.

G. (2011). Ecosystem approach to inland fisheries: research needs and implementation

strategies. Biology Letters, 7, 481–483.

Benke, A. C., Huryn, A. D., Smock, L. A., & Wallace, J. B. (1999). Length-mass relationships

for freshwater macroinvertebrates in North America with particular reference to the

southeastern United States. Journal of the North American Benthological Society, 18, 308–

343.

Bennett, S. J., Sanders, R. W., & Porter, K. G. (1990). Heterotrophic, autotrophic, and

mixotrophic nanoflagellates: Seasonal abundances and bacterivory in a eutrophic lake.

Limnology and Oceanography, 35, 1821–1832.

Benoît, P., Beisner, B. E., & Solomon, C. T. (2016). Growth rate and abundance of common

fishes is negatively related to dissolved organic carbon concentration in lakes. Canadian

Journal of Fisheries and Aquatic Science, 73, 1230–1236.

98 Berggren, M., Bergström, A. K., & Karlsson, J. (2015). Intraspecific autochthonous and

allochthonous resource use by zooplankton in a humic lake during the transitions between

winter, summer and fall. PLoS ONE, 10, e0120575.

Berggren, M., Klaus, M., Selvam, B. P., Ström, L., Laudon, H., Jansson, M., & Karlsson, J.

(2018). Quality transformation of dissolved organic carbon during water transit through

lakes: contrasting controls by photochemical and biological processes. Biogeosciences, 15,

457–470.

Berggren, M., Ziegler, S. E., St-Gelais, N. F., Beisner, B. E., & del Giorgio, P. A. (2014).

Contrasting patterns of allochthony among three major groups of crustacean zooplankton in

boreal and temperate lakes. Ecology, 95, 1947–1959.

Berglund, J., Müren, U., Båmstedt, U., & Andersson, A. (2007). Efficiency of a phytoplankton-

based and a bacterial-based food web in a pelagic marine system. Limnology and

Oceanography, 52, 121–131.

Bertolo, A., & Magnan, P. (2007). Logging-induced variations in dissolved organic carbon affect

Yellow Perch (Perca flavescens) recruitment in Canadian Shield lakes. Canadian Journal of

Fisheries and Aquatic Sciences, 64, 181–186.

Blomqvist, P., Jansson, M., Drakare, S., Bergström, A. K., & Brydsten, L. (2001). Effects of

additions of DOC on pelagic biota in a clearwater system: Results from a whole lake

experiment in northern Sweden. Microbial Ecology, 42, 383–394.

Brander, K. M. (2007). Global fish production and climate change science. Proceedings of the

National Academy of Sciences, 104, 19709–19714.

99 Brett, M. T., Bunn, S. E., Chandra, S., Galloway, A. W. E., Guo, F., Kainz, M. J., … Taipale, S.

J. (2017). How important are terrestrial organic carbon inputs for secondary production in

freshwater ecosystems? Freshwater Biology, 62, 833–853.

Brett, M. T., Holtgrieve, G. W., & Schindler, D. E. (2018). An assessment of assumptions and

uncertainty in deuterium-based estimates of terrestrial subsidies to aquatic consumers.

Ecology, 99, 1073–1088.

Brett, M. T., Kainz, M. J., Taipale, S. J., & Seshan, H. (2009). Phytoplankton, not allochthonous

carbon, sustains herbivorous zooplankton production. Proceedings of the National Academy

of Sciences, 106, 21197–21201.

Brett, M. T., & Muller-Navarra, D. C. (1997). The role of highly unsaturated fatty acids in

aquatic foodweb processes. Freshwater Biology, 38, 483–499.

Brooks, E. G. E., Holland, R. A., Darwall, W. R. T., & Eigenbrod, F. (2016). Global evidence of

positive impacts of freshwater biodiversity on fishery yields. Global Ecology and

Biogeography, 25, 553–562.

Brothers, S., Ko, J., Attermeyer, K., Grossart, H. P., Mehner, T., Meyer, N., … Hilt, S. (2014). A

feedback loop links brownification and anoxia in a temperate, shallow lake. Limnology and

Oceanography, 59, 1388–1398.

Brownscombe, J. W., Bower, S. D., Bowden, W., Nowell, L., Midwood, J. D., Johnson, N., &

Cooke, S. J. (2014). Canadian recreational fisheries: 35 years of social, biological, and

economic dynamics from a national survey. Fisheries, 39, 251–260.

Camacho, A. (2006). On the occurence and ecological features of deep chlorophyll maxima

(DCM) in Spanish stratified lakes. Limnetica, 25, 453–478.

100 Carpenter, S.R. (1998). The need for large-scale and experiments to assess and predict the

response of ecosystems to perturbation. Success, Limitations, and Frontiers in Ecosystem

Science (Pace, M. L. et al., eds), pp. 287-312, Springer-Verlag.

Carpenter, S. R. (2008). Phosphorus control is critical to mitigating eutrophication. Proceedings

of the National Academy of Sciences, 105, 11039–11040.

Carpenter, S. R., Cole, J. J., Pace, M. L., Van De Bogert, M., Bade, D. L., Bastviken, D., …

Kritzberg, E. S. (2005). Ecosystem subsidies: Terrestrial support of aquatic food webs from

13C addition to contrasting lakes. Ecology, 86, 2737–2750.

Carpenter, S. R., Kitchell, J. F., & Hodgson, J. R. (1985). Cascading trophic interactions and lake

productivity. BioScience, 35, 634–639.

Chalanchuck, S. M. (1998). Growth of White Sucker, Catostomus commersonii, in thirty-one

lakes at the Experimental Lakes Area, in northwestern Ontario. Canadian Technical Report

of Fisheries and Aquatic Sciences, 2207.

Chu, C., Nester, N. P., Giacomini, H., Shuter, B. J., & Jackson, D. A. (2016). Catch-per-unit-

effort and size spectra of lake fish assemblages reflect underlying patterns in ecological

conditions and anthropogenic activities across regional and local scales. Canadian Journal

of Fisheries and Aquatic Sciences, 73, 1–12.

Cloern, J. E., Grenz, C., & Vidergar-Lucas, L. (1995). An empirical model of the phytoplankton

chloropyll : carbon ratio - the conversion factor between productivity and growth rate.

Limnology and Oceanography, 40, 1313–1321.

Cole, J. J, Carpenter, S. R., Kitchell, J. F., & Pace, M. L. (2002a). Pathways of organic carbon

utilization in small lakes: Results from a whole-lake 13C addition and coupled model.

Limnology and Oceanography, 47, 1664–1675.

101 Cole, J. J., Carpenter, S. R., Kitchell, J., Pace, M. L., Solomon, C. T., & Weidel, B. C. (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, 1975–1980.

Cole, J. J., Carpenter, S. R., Pace, M. L., Van De Bogert, M. C., Kitchell, J. L., & Hodgson, J. R.

(2006). Differential support of lake food webs by three types of terrestrial organic carbon.

Ecology Letters, 9, 558–568.

Cole, J. J., Prairie, Y. T., Caraco, N. F., McDowell, W. H., Tranvik, L. J., Striegl, R. G., …

Melack, J. (2007). Plumbing the global carbon cycle: Integrating inland waters into the

terrestrial carbon budget. Ecosystems, 10, 171–184.

Cole, P. C., Luecke, C., Wurtsbaugh, W. A., & Burkart, G. (2002b). Growth and survival of

Daphnia in epilimnetic and metalimnetic water from oligotrophic lakes: the effects of food

and temperature. Freshwater Biology, 47, 2113–2122.

Cooke, S. J., & Cowx, I. G. (2004). The role of recreational fishing in global fish crises.

Bioscience, 54, 857–859.

Couture, R. M., De Wit, H. A., Tominaga, K., Kiuru, P., & Markelov, I. (2015). Oxygen

dynamics in a boreal lake responds to long-term changes in climate, ice phenology, and

DOC inputs. Journal of Geophysical Research: Biogeosciences, 120, 2441–2456.

Craig, N., Jones, S. E., Weidel, B. C., & Solomon, C. T. (2015). Habitat, not resource

availability, limits consumer production in lake ecosystems. Limnology and Oceanography,

60, 2079–2089.

102 Craig, N., Jones, S. E., Weidel, B. C., & Solomon, C. T. (2017). Life history constraints explain

negative relationship between fish productivity and dissolved organic carbon in lakes.

Ecology and Evolution, 7, 6201–6209.

Creed, I. F., Trick, C. G., Grimm, N. B., Hessen, D. O., Kidd, K. A., Kritzberg, E., …

Weyhenmeyer, G. A. (2018). Global change-driven effects on dissolved organic matter

composition: Implications for food webs of northern lakes. Global Change Biology, 24,

3692–3714. del Giorgio, P. A., & France, R. L. (1996). Ecosystem-specific patterns in the relationship

between zooplankton and POM or microplankton d13C. Limnology and Oceanography, 41,

359–365.

DeMott, W. R. (1988). Discrimination between algae and artificial particles by freshwater and

marine copepods. Limnology and Oceanography, 33, 397–408.

DeMott, W. R., Edington, J. R., & Tessier, A. J. (2004). Testing zooplankton food limitation

across gradients of depth and productivity in small stratified lakes. Limnology and

Oceanography, 49, 1408–1416.

Dillon, P. J, Clark, B. J., Molot, L. A., & Evans, H. E. (2003). Predicting the location of optimal

habitat boundaries for Lake Trout (Salvelinus namaycush) in Canadian Shield lakes.

Canadian Journal of Fisheries and Aquatic Science, 970, 959–970.

Dillon, P. J, & Molot, L. (2005). Long-term trends in catchment export and lake retention of

dissolved organic carbon, dissolved organic nitrogen, total iron, and total phosphorus: The

Dorset, Ontario, study, 1978 –1998. Journal of Geophysical Research, 110, G01002.

Dodson, S. (1990). Predicting diel vertical migration of zooplankton. Limnology and

Oceanography, 35, 1195–1200.

103 Dolbeth, M., Cusson, M., Sousa, R., Pardal, M. a., & Prairie, Y. T. (2012). Secondary production

as a tool for better understanding of aquatic ecosystems. Canadian Journal of Fisheries and

Aquatic Sciences, 69, 1230–1253.

Donahue, W. F., Turner, M. A., Findlay, D. L., & Leavitt, P. R. (2003). The role of solar

radiation in structuring the shallow benthic communities of boreal forest lakes. Limnology

and Oceanography, 48, 31–47.

Doucett, R. R., Marks, J. C., Blinn, D. W., Caron, M., & Hungate, B. A. (2007). Measuring

terrestrial subsidies to aquatic food webs using stable isotopes of hydrogen. Ecology, 88,

1587–1592.

Downing, J. A., & Duarte, C. M. (2006). Abundance and size distribution of lakes, ponds and

impoundments. Reference Module in Earth Systems and Environmental Sciences, 51, 2388–

2397.

Downing, J. A., Plante, C., & Lalonde, S. (1990). Fish production correlated with primary

productivity, not the morphoedaphic Index. Canadian Journal of Fisheries and Aquatic

Sciences, 47, 1929–1936.

Downing, J. A., & Plante, C. (1993). Production of fish population in lakes. Canadian Journal of

Fisheries and Aquatic Sciences, 50, 110-120.

Elser, J. J., Chrzanowski, T. H., Guildford, S. J., Sterner, R. W., & Fee, E. J. (2002). The light :

nutrient ratio in lakes: the balance of energy and materials affects ecosystem structure and

process. The American Naturalist, 150, 663–684.

Elser, J. J., Fagan, W. F., Denno, R. F., Dobberfuhl, D. R., Folarin, A., Huberty, A., … Sterner,

R. W. (2000). Nutritional constraints in terrestrial and freshwater food webs. Nature, 480,

578–580.

104 Emmerton, C. A., Beaty, K. G., Casson, N. J., Graydon, J. A., Hesslein, R. H., Higgins, S. N., …

Tardif, J. C. (2018). Long-term responses of nutrient budgets to concurrent climate-related

stressors in a boreal watershed. Ecosystems, 22, 363–378.

Environment Canada (2012). Metal Mining Technical Guidance for Environmental Effects

Monitoring. ISBN 978-1-100-20496-3.

Evans, C. D., Monteith, D. T., & Cooper, D. M. (2005). Long-term increases in surface water

dissolved organic carbon: Observations, possible causes and environmental impacts.

Environmental Pollution, 137, 55–71.

Fee, E. J. (1973). A numerical model for determining integral primary production and its

application to Lake Michigan. Fisheries Research Board of Canada, 30, 1447–1467.

Fee, E. J. (1976). The vertical and seasonal distribution of chlorophyll in lakes of the

Experimental Lakes Area, northwestern Ontario: Implications for primary production

estimates. Limnology and Oceanography, 21, 767–783.

Fee, E. J., Hecky, R. E., Kasian, S. E. M., & Cruikshank, D. R. (1996). Effects of lake size, water

clarity, and climatic variability on mixing depths in Canadian Shield lakes. Limnology and

Oceanography, 41, 912–920.

Finstad, A. G., Andersen, T., Larsen, S., Tominaga, K., Blumentrath, S., de Wit, H. A., …

Hessen, D. O. (2016). From greening to browning: Catchment vegetation development and

reduced S-deposition promote organic carbon load on decadal time scales in Nordic lakes.

Scientific Reports, 6, 31944.

Finstad, A. G., Helland, I. P., Ugedal, O., Hesthagen, T., & Hessen, D. O. (2014). Unimodal

response of fish yield to dissolved organic carbon. Ecology Letters, 17, 36–43.

105 Francis, T. B., Schindler, D. E., Holtgrieve, G. W., Larson, E. R., Scheuerell, M. D., Semmens,

B. X., & Ward, E. J. (2011). Habitat structure determines resource use by zooplankton in

temperate lakes. Ecology Letters, 14, 364–372.

Freeman, C., Evans, C. ., Monteith, D. T., Reynolds, B., & Fenner, N. (2001). Export of organic

carbon from peat soils. Nature, 412, 785.

Galloway, A. W. E., Taipale, S. J., Hiltunen, M., Peltomaa, E., Strandberg, U., Brett, M. T., &

Kankaala, P. (2014). Diet-specific biomarkers show that high-quality phytoplankton fuels

herbivorous zooplankton in large boreal lakes. Freshwater Biology, 59, 1902–1915.

Godwin, S. C., Jones, S. E., Weidel, B. C., & Solomon, C. T. (2014). Dissolved organic carbon

concentration controls benthic primary production: Results from in situ chambers in north-

temperate lakes. Limnology and Oceanography, 59, 2112–2120.

Grey, J., Jones, R. I., & Sleep, D. (2001). Seasonal changes in the importance of the source of

organic matter to the diet of zooplankton in Loch Ness, as indicated by stable isotope

analysis. Limnology and Oceanography, 46, 505–513.

Gulka, J., Carvalho, P. C., Jenkins, E., Johnson, K., Maynard, L., & Davoren, G. K. (2017).

Dietary niche shifts of multiple marine predators under varying prey availability on the

northeast Newfoundland coast. Frontiers in Marine Science, 4, e324.

Guzzo, M. M., & Blanchfield, P. J. (2017). Climate change alters the quantity and phenology of

habitat for Lake Trout (Salvelinus namaycush) in small boreal shield lakes. Canadian

Journal of Fisheries and Aquatic Science, 74, 871–884.

Guzzo, M. M., Blanchfield, P. J., & Rennie, M. D. (2017). Behavioral responses to annual

temperature variation alter the dominant energy pathway, growth, and condition of a cold-

water predator. Proceedings of the National Academy of Sciences, 114, 9912–9917.

106 Guzzo, M. M., Rennie, M. D., & Blanchfield, P. J. (2014). Evaluating the relationship between

mean catch per unit effort and abundance for littoral cyprinids in small boreal shield lakes.

Fisheries Research, 150, 100–108.

Hamilton, A. L. (1971). Zoobenthos of fifteen lakes in the Experimental Lakes Area, north-

western Ontario. Journal of the Fisheries Research Board of Canada, 28, 257–263.

Hamilton Stewart, N. (1926). Development, growth, and food habits of the White Sucker,

Catostomus commersonii Lesueur. Bulletin of the Bureau of Fisheries, 42, 146–184.

Hanson, J. M., & Leggett, W. C. (1982). Empirical prediction of fish biomass and yield.

Canadian Journal of Fisheries and Aquatic Sciences, 39, 257–263.

Hanson, P. C., Bade, D. L., Carpenter, S. R., & Kratz, T. K. (2003). Lake metabolism:

Relationships with dissolved organic carbon and phosphorus. Limnology and

Oceanography, 48, 1112–1119.

Hayden, B., Soto, D. X., Jardine, T. D., Graham, B. S., Cunjak, R. A., Romakkaniemi, A., &

Linnansaari, T. (2015). Small tails tell tall tales - intra-individual variation in the stable

isotope values of fish fin. PLoS ONE, 10, e0145154.

Hecky, R. E., & Hesslein, R. H. (1995). Contributions of benthic algae to lake food webs as

revealed by stable isotope analysis. Journal of the North American Benthological Society,

14, 631-653.

Heinis, F., & Davids, C. (1993). Factors governing the spatial and temporal distribution of

chironomid larvae in the Maarsseveen lakes with special emphasis on the role of oxygen

conditions. Netherlands Journal of Aquatic Ecology, 27, 21–34.

Hessen, D. O. (2006). Determinants of seston C : P-ratio in lakes. Freshwater Biology, 51, 1560–

1569.

107 Hessen, D. O., & Andersen, T. (1990). Bacteria as a source of phosphorus for zooplankton.

Hydrobiologia, 206, 217–223.

Hiltunen, M., Honkanen, M., Taipale, S., Strandberg, U., & Kankaala, P. (2017). Trophic

upgrading via the microbial food web may link terrestrial dissolved organic matter to

Daphnia. Journal of Plankton Research, 39, 861–869.

Hinton, M. J., Schiff, S. L., & English, M. C. (1997). The significance of storms for the

concentration and export of dissolved organic carbon from two Precambrian Shield

catchments. Biogeochemistry, 36, 67–88.

Holmlund, C. M., & Hammer, M. (1999). Ecosystem services generated by fish populations.

Ecological Economics, 29, 253–268.

Holsinger, J. R. (1972). The freshwater amphipod crustaceans (Gammaridae) of North America.

Biota of Freshwater Ecosystems, Identification Manual 5. U.S. EPA, 1-89.

Houser, J. N., Bade, D. L., Cole, J. J., & Pace, M. L. (2008). The dual influences of dissolved

organic carbon on hypolimnetic metabolism: organic substrate and photosynthetic

reduction. Biogeochemistry, 64, 247–270.

Jansson, M., Hickler, T., Jonsson, A., & Karlsson, J. (2008). Links between terrestrial primary

production and bacterial production and respiration in lakes in a climate gradient in

subarctic Sweden. Ecosystems, 11, 367–376.

Johnston, T. A., Stasko, A. D., Hamilton, G. L., Nugent, B. K., Cott, P. A., & Gunn, J. M.

(2018). Plenty of room at the bottom: Niche variation and segregation in large-bodied

benthivores of boreal lakes. Canadian Journal of Fisheries and Aquatic Sciences, 76,

e1139.

108 Jonasson, P. M. (1984). Oxygen demand and long-term changes of profundal zoobenthos.

Hydrobiologia, 115, 121–126.

Jones, R. I. (1992). The influence of humic substances on lacustrine planktonic food chains.

Hydrobiologia, 229, 73–91.

Jones, R. I., Grey, J., Sleep, D., & Arvola, L. (1999). Stable isotope analysis of zooplankton

carbon nutrition in humic lakes. Oikos, 86, 97–104.

Jones, R. I., Carter, C. E., Kelly, A., Ward, S., Kelly, D. J., & Grey, J. (2008). Widespread

contribution of methane-cycle bacteria to the diets of lake profundal chironomid larvae.

Ecology, 89, 857–864.

Jones, R. I., & Grey, J. (2011). Biogenic methane in freshwater food webs. Freshwater Biology,

56, 213–229.

Jones, S. E., & Lennon, J. T. (2009). Evidence for limited microbial transfer of methane in a

planktonic food web. Aquatic Microbial Ecology, 58, 45–53.

Jones, S. E., Solomon, C. T., & Weidel, B. C. (2012). Subsidy or subtraction: How do terrestrial

inputs influence consumer production in lakes? Freshwater Reviews, 5, 37–49.

Jürgens, K. (1994). The impact of Daphnia on microbial food webs - a review. Marine Microbial

Food Webs, 8, 295–324.

Karlsson, J. (2007). Different carbon support for respiration and secondary production in

unproductive lakes. Oikos, 116, 1691–1696.

Karlsson, J., Berggren, M., Ask, J., Byström, P., Jonsson, A., Laudon, H., & Jansson, M. (2012).

Terrestrial organic matter support of lake food webs: Evidence from lake metabolism and

stable hydrogen isotopes of consumers. Limnology and Oceanography, 57, 1042–1048.

109 Karlsson, J., Bergström, A. K., Byström, P., Gudasz, C., Rodriguez, P., & Hein, C. (2015).

Terrestrial organic matter input suppresses biomass production in lake ecosystems. Ecology,

96, 2870–2876.

Karlsson, J., & Byström, P. (2005). Littoral energy mobilization dominates energy supply for top

consumers in subarctic lakes. Limnology and Oceanography, 50, 538–543.

Karlsson, J., Byström, P., Ask, J., Ask, P., Persson, L., & Jansson, M. (2009). Light limitation of

nutrient-poor lake ecosystems. Nature, 460, 506–509.

Karlsson, J., Jonsson, A., Meili, M., & Jansson, M. (2003). Control of zooplankton dependence

on allochthonous organic carbon in humic and clear-water lakes in northern Sweden.

Limnology and Oceanography, 48, 269–276.

Kelly, P. T., Craig, N., Solomon, C. T., Weidel, B. C., Zwart, J. A., & Jones, S. E. (2016).

Experimental whole-lake increase of dissolved organic carbon concentration produces

unexpected increase in crustacean zooplankton density. Global Change Biology, 22, 2766–

2775.

Kelly, P. T., Solomon, C. T., Weidel, B. C., & Jones, S. E. (2014). Terrestrial carbon is a

resource, but not a subsidy, for lake zooplankton. Ecology, 95, 1236–1242.

Kessler, K., & Lampert, W. (2004). Depth distribution of Daphnia in response to a deep-water

algal maximum: The effect of body size and temperature gradient. Freshwater Biology, 49,

392–401.

Kettle, W. D., Moffett, M. F., & deNoyelles Jr., F. (1987). Vertical distribution of zooplankton in

an experimentally acidified lake containing a metalimnetic phytoplankton peak. Canadian

Journal of Fisheries and Aquatic Sciences, 44, 91–95.

110 Kissman, C. E. H., Williamson, C. E., Rose, K. C., & Saros, J. E. (2013). Response of

phytoplankton in an alpine lake to inputs of dissolved organic matter through nutrient

enrichment and trophic forcing. Limnology and Oceanography, 58, 867–880.

Koehler, B., Von Wachenfeldt, E., Kothawala, D., & Tranvik, L. J. (2012). Reactivity continuum

of dissolved organic carbon decomposition in lake water. Journal of Geophysical Research:

Biogeosciences, 117, G01024.

Koizumi, S., Nicola, C., Zwart, J., Weidel, B. С., Jones, S. E., & Solomon, C. T. (2018).

Experimental whole-lake dissolved organic carbon increase alters fish diet but not growth or

productivity. Canadian Journal of Fisheries and Aquatic Sciences, 75, 1859–1867.

Kreutzweiser, D. P., Hazlett, P. W., & Gunn, J. M. (2008). Logging impacts on the

biogeochemistry of boreal forest soils and nutrient export to aquatic systems: a review.

Environmental Reviews, 16, 157–179.

Kritzberg, E. S., Cole, J. J., Pace, M. M., & Granéli, W. (2005). Does autochtonous primary

production drive variability in bacterial metabolism and growth efficiency in lakes

dominated by terrestrial C inputs? Aquatic Microbial Ecology, 38, 103–111.

Lampert, W., & Grey, J. (2003). Exploitation of a deep-water algal maximum by Daphnia: A

stable-isotope tracer study. Hydrobiologia, 500, 95–101.

Larsen, S., Andersen, T., & Hessen, D. O. (2011). Climate change predicted to cause severe

increase of organic carbon in lakes. Global Change Biology, 17, 1186–1192.

Laudon, H., Berggren, M., Anneli, Å., Buffam, I., Bishop, K., Grabs, T., … Stephan, K. (2011).

Patterns and dynamics of dissolved organic carbon (DOC) in boreal streams: The role of

processes, connectivity, and scaling. Ecosystems, 14, 880–893.

111 Leach, T. H., Beisner, B. E., Carey, C. C., Pernica, P., Rose, K. C., Huot, Y., … Verburg, P.

(2017). Patterns and drivers of deep chlorophyll maxima structure in 100 lakes: The relative

importance of light and thermal stratification. Limnology and Oceanography, 63, 628–646.

Leach, T. H., Winslow, L. A., Hayes, N. M., & Rose, K. C. (2019). Decoupled trophic responses

to long-term recovery from acidification and associated browning in lakes. Global Change

Biology, 25, 1779–1792.

Lennon, J. T., Faiia, A. M., Feng, X., & Cottingham, K. L. (2006). Relative importance of CO2

recycling and CH4 pathways in lake food webs along a dissolved organic carbon gradient.

Limnology and Oceanography, 51, 1602–1613.

Lescord, G. L., Emilson, E. J. S., Johnston, T. A., Bran, B. A., & Gunn, J. M. (2018). Optical

properties of dissolved organic matter and their relation to mercury concentrations in water

and biota across a remote freshwater drainage basin. Environmental Science & Technology,

52, 3344–3353.

Lynch, A. J., Cooke, S. J., Deines, A. M., Bower, S. D., Bunnell, D. B., Cowx, I. G., … Beard,

T. D. (2016). The social, economic, and environmental importance of inland fish and

fisheries. Environmental Reviews, 24, 115–121.

Manek, A. K., Ferrari, M. C. O., Chivers, D. P., & Niyogi, S. (2014). Dissolved organic carbon

ameliorates the effects of UV radiation on a freshwater fish. Science of the Total

Environment, 490, 941–946.

Marty, J., & Planas, D. (2008). Comparison of methods to determine algal δ13C in freshwater.

Limnology and Oceanography: Methods, 6, 51–63.

Matthews, B., & Mazumder, A. (2006). Habitat specialization and the exploitation of

allochthonous carbon by zooplankton. Ecology, 87, 2800–2812.

112 McMeans, B. C., Koussoroplis, A., & Michael, T. (2015). Terrestrial dissolved organic matter

supports growth and reproduction of Daphnia magna when algae are limiting. Journal of

Plankton Research, 37, 1201–1209.

Meester, L. D., Weider, L., & Tollrian, R. (1995). Altnerative antipredator defences and genetic

polymorphism in a pelagic predator-prey system. Nature, 378, 483–485.

Mehner, T., & Thiel, R. (1999). A review of predation impact by 0+ fish on zooplankton in fresh

and brackish waters of the temperate northern hemisphere. Environmental Biology of

Fishes, 56, 169–181.

Merritt, R. W., Cummins K. W., and Berg, M. N. (2008). An introduction to the aquatic insects

of North America, 4th ed. Kendall Hunt.

Mills, K. H., Chalanchuk, S. M., Mohr, L. C., & Davies, I. J. (1987). Responses of fish

populations in Lake 223 to 8 years of experimental acidification. Canadian Journal of

Fisheries and Aquatic Sciences, 44, 114–125.

Misra, R. F., Davies, I. J., Watson, N. H. F., & Uthe, J. F. (1995). Investigation of temporal

variations in relative abundance of macroinvertebrates in Lake 224 of the Experimental

Lakes Area by a mulitvariate method. Canadian Technical Report of Fisheries and Aquatic

Sciences, 2026.

Monteith, D. T., Stoddard, J. L., Evans, C. D., de Wit, H. A, Forsius, M., Høgåsen, T., …

Vesely, J. (2007). Dissolved organic carbon trends resulting from changes in atmospheric

deposition chemistry. Nature, 450, 537–540.

Muller-Navarra, D. C., Brett, M. T., Liston, A., & Goldman, C. (2000). A highly unsaturated

fatty acid predicts carbon transfer between primary producers and consumers. Nature, 403,

74–77.

113 Munkittrick, K. R., Miller, P. A., Barton, D. R., & Dixon, D. G. (1991). Altered performance of

White Sucker populations in the Manitouwadge chain of lakes is associated with changes in

benthic macroinvertebrate communities as a result of copper and zinc contamination.

Ecotoxicology and Environmental Safety, 21, 318–326.

Neff, J. C., & Asner, G. P. (2001). Dissolved organic carbon in terrestrial ecosystems: Synthesis

and a model. Ecosphere, 4, 29–48.

Nielsen, J. M., Kratina, P., Clare, E. L., Hayden, B., & Brett, M. T. (2018). Diet tracing in

ecology: Method comparison and selection. Methods in Ecology and Evoluation, 9, 278–

291.

Orihel, D. M., Baulch, H. M., Casson, N. J., North, R. L., Parsons, C. T., Seckar, D. C. M., &

Venkiteswaran, J. J. (2017). Internal phosphorus loading in Canadian fresh waters: a critical

review and data analysis. Canadian Journal of Fisheries and Aquatic Sciences, 74, 2005–

2029.

Pace, M. L., Carpenter, S. R., Cole, J. J., Coloso, J. J., Kitchell, J. F., Hodgson, J. R., … Weidel,

B. C. (2007). Does terrestrial organic carbon subsidize the planktonic food web in a clear-

water lake? Limnology and Oceanography, 52, 2177–2189.

Pannard, A., Planas, D., & Beisner, B. E. (2015). Macrozooplankton and the persistence of the

deep chlorophyll maximum in a stratified lake. Freshwater Biology, 60, 1717–1733.

Paterson, M. J., Beaty, K. G., Findlay, D. L., Findlay, W. J., Schiff, S. L., St. Louis, V. L., &

Venkiteswaran, J. J. (2019). Long-term changes in nutrient dynamics and plankton

communities following the creation of a new reservoir. Canadian Journal of Fisheries and

Aquatic Sciences, 76, 1459–1469.

114 Paterson, M. J., Findlay, D., Beaty, K., Findlay, W., Schindler, E. U., Stainton, M., &

McCullough, G. (1997). Changes in the planktonic food web of a new experimental

reservoir. Canadian Journal of Fisheries and Aquatic Sciences, 54, 1088–1102.

Paterson, M. J., Podemski, C. L., Findlay, W. J., Findlay, D. L., & Salki, A. G. (2010). The

response of zooplankton in a whole-lake experiment on the effects of a cage aquaculture

operation for Rainbow Trout (Oncorhynchus mykiss). Canadian Journal of Fisheries and

Aquatic Sciences, 67, 1852–1861.

Pilati, A., & Wurtsbaugh, W. A. (2003). Importance of zooplankton for the persistence of a deep

chlorophyll layer: A limnocorral experiment. Limnology and Oceanography, 48, 249–260.

Pimm, S., & Kitching, R. L. (1987). The determinants of food chain lengths. Oikos, 50, 302–307.

Polis, G. A., Anderson, W. B., & Holt, R. D. (1997). Toward an integration of landscape and

food web ecology: The dynamics of spatially subsidized food webs. Annual Review of

Ecology and Systematics, 28, 289–316.

Post, D. M. (2002). Using stable isotopes to estimate trophic position: model, methods, and

assumptions. Ecology, 83, 703–718.

Post, D. M., Layman, C. A., Arrington, D. A., Takimoto, G., Quattrochi, J., & Montana, C. G.

(2007). Getting to the fat of the matter: models, methods and assumptions for dealing with

lipids in stable isotope analyses. Oecologia, 152, 179–189.

Post, J. R. (2013). Resilient recreational fisheries or prone to collapse? A decade of research on

the science and management of recreational fisheries. Fisheries Management and Ecology,

20, 99–110.

115 Post, J. R., Sullivan, M., Cox, S., Lester, N., Walters, C., Parkinson, E. A., … Shuter, B. J.

(2002). Canada’s recreational fisheries: The invisible collapse? Fisheries Management, 27,

6–17.

Prairie, Y. T. (2008). Carbocentric limnology: looking back, looking forward. Canadian Journal

of Fisheries and Aquatic Sciences, 65, 543–548.

Rask, M., Sairanen, S., Vesala, S., Arvola, L., Estlander, S., & Olin, M. (2014). Population

dynamics and growth of perch in a small, humic lake over a 20-year period — importance

of abiotic and biotic factors. Boreal Environment Research, 19, 112–123.

Ravinet, M., Syva, J., Jones, R. I., & Grey, J. (2010). A trophic pathway from biogenic methane

supports fish biomass in a temperate lake ecosystem. Oikos, 119, 409–416.

Read, J. S., & Rose, K. C. (2013). Physical responses of small temperate lakes to variation in

dissolved organic carbon concentrations. Limnology and Oceanography, 58, 921–931.

Rothhaupt, K. O. (1991). Variations on the zooplankton menu - A reply to the comment by

Smith. Limnology and Oceanography, 36, 824–827.

Rypel, A. L., & David, S. R. (2017). Pattern and scale in latitude-production relationships for

freshwater fishes. Ecosphere, 8, e01660.

Rypel, A. L., Goto, D., Sass, G. G., & Vander Zanden, M. J. (2015). Production rates of Walleye

and their relationship to exploitation in Escanaba Lake. Canadian Journal of Fisheries and

Aquatic Sciences, 844, 834–844.

Sanseverino, A. M., Bastviken, D., Sundh, I., Pickova, J., & Enrich-Prast, A. (2012). Methane

carbon supports aquatic food webs to the fish level. PLoS One, 7, e42723.

Sargent, J., Bell, G., McEvoy, L., Tocher, D., & Estevez, A. (1999). Recent developments in the

essential fatty acid nutrition of fish. Aquaculture, 177, 191–199.

116 Schindler, D. W., Mills, K. H., Malley, D. F., Findlay, D. L., Shearer, J. A., Davies, I. ., …

Cruikshank, D. R. (1985). Long-term ecosystem stress: The effects of years of experimental

acidification on a smal lake. Science, 228, 1395–1401.

Schindler, D. W. (1977). Evolution of phosphorus limitation in lakes. Science, 195, 260–262.

Schindler, D. W. (2001). The cumulative effects of climate warming and other human stresses on

Canadian freshwaters in the new millennium. Canadian Journal of Fisheries and Aquatic

Sciences, 58, 18–29.

Seekell, D. A., Byström, P., & Karlsson, J. (2018). Lake morphometry moderates the

relationship between water color and fish biomass in small boreal lakes. Limnology and

Oceanography, 65, 2171–2178.

Seekell, D.A., Lapierre, J. F., Ask, J., Bergström, A. K., Deininger, A., Rodriguez, P., &

Karlsson, J. (2015a). The influence of dissolved organic carbon on primary production in

northern lakes. Limnology and Oceanography, 60, 1276–1285.

Seekell, D. A., Lapierre, J., & Karlsson, J. (2015b). Trade-offs between light and nutrient

availability across gradients of dissolved organic carbon concentration in Swedish lakes:

implications for patterns in primary production. Canadian Journal of Fisheries and Aquatic

Sciences, 72, 1663–1671.

Smith, L. E. D., Khoa, S. N., & Lorenzen, K. (2005). Livelihood functions of inland fisheries:

policy implications in developing countries. Water Policy, 7, 359–383.

Sobek, S., Tranvik, L. J., Prairie, Y. T., & Cole, J. J. (2007). Patterns and regulation of dissolved

organic carbon: an analysis of 7,500 widely distributed lakes. Limnology and

Oceanography, 52, 1208–1219.

117 Solomon, C. T., Carpenter, S. R., Clayton, M. K., Cole, J. J., Coloso, J. J., Pace, M. L., …

Weidel, B. С. (2011). Terrestrial, benthic, and pelagic resource use in lakes: Results from a

three-isotope Bayesian mixing model. Ecology, 92, 1115–1125.

Solomon, C. T., Cole, J. J., Doucett, R. R., Pace, M. L., Preston, N. D., Smith, L. E., & Weidel,

B. C. (2009). The influence of environmental water on the hydrogen stable isotope ratio in

aquatic consumers. Oecologia, 161, 313–324.

Solomon, C. T., Jones, S. E., Weidel, B. C., Buffam, I., Fork, M. L., Karlsson, J., … Saros, J. E.

(2015). Ecosystem consequences of changing inputs of terrestrial dissolved organic matter

to lakes: Current knowledge and future challenges. Ecosystems, 18, 376–389.

Stainton, M. P., Capel, M. J., & Armstrong, F. A. J. (1977). The chemical analysis of fresh

water. Fisheries and Environment Canada Miscellaneous Special Publication, 25, 1–180.

Stasko, A. D., Gunn, J. M., & Johnston, T. A. (2012). Role of ambient light in structuring north-

temperate fish communities: potential effects of increasing dissolved organic carbon

concentration with a changing climate. Environmental Reviews, 20, 173–190.

Sterner, R. W. (2008). On the phosphorus limitation paradigm for lakes. International Review of

Hydrobiology, 93, 433–445.

Sterner, R. W., Elser, J. J., Fee, E. J., Guildford, S. J., & Chrzanowski, T. H. (1997). The light :

nutrient ratio in lakes: The balance of energy and materials affects ecosystem structure and

process. The American Naturalist, 150, 663–684.

Sterner, R. W., & Schwalbach, M. S. (2001). Diel integration of food quality by Daphnia:

Luxury consumption by a freshwater planktonic herbivore. Limnology and Oceanography,

46, 410–416.

118 Stock, B. C., Jackson, A. L., Ward, E. J., Parnell, A. C., Phillips, D. L., & Semmens, B. X.

(2018). Analyzing mixing systems using a new generation of Bayesian tracer mixing

models. PeerJ, 6:e5096.

Taipale, S. J., Brett, M. T., Pulkkinen, K., & Kainz, M. J. (2012). The influence of bacteria-

dominated diets on Daphnia magna somatic growth, reproduction, and lipid composition.

FEMS Microbiology Ecology, 82, 50–62.

Taipale, S. J., Galloway, A. W. E., Aalto, S. L., Kahilainen, K. K., Strandberg, U., & Kankaala,

P. (2016a). Terrestrial carbohydrates support freshwater zooplankton during phytoplankton

deficiency. Scientific Reports, 6, 30897.

Taipale, S. J., Kainz, M., & Brett, M. T. (2015). A low w-3 : w-6 ratio in Daphnia indicates

terrestrial resource utilization and poor nutritional condition. Journal of Plankton Research,

37, 596–610.

Taipale, S. J., Vuorio, K., Brett, M. T., Peltomaa, E., Hiltunen, M., & Kankaala, P. (2016b). Lake

zooplankton δ13C values are strongly correlated with the δ13C values of distinct

phytoplankton taxa. Ecosphere, 7, e01392.

Taipale, S. J., Vuorio, K., Strandberg, U., Kahilainen, K. K., Jarvinen, M., Hiltunen, M., …

Kankaala, P. (2016c). Lake eutrophication and brownification downgrade availability and

transfer of essential fatty acids for human consumption. Environment International, 96,

156–166.

Tanentzap, A. J., Kielstra, B. W., Wilkinson, G. M., Berggren, M., Craig, N., del Giorgio, P. A.,

… Pace, M. L. (2017). Terrestrial support of lake food webs: Synthesis reveals controls

over cross-ecosystem resource use. Science Advances, 3, e1601765.

119 Tanentzap, A. J., Szkokan-Emilson, E. J., Kielstra, B. W., Arts, M. T., Yan, N. D., & Gunn, J.

M. (2014). Forests fuel fish growth in freshwater deltas. Nature Communications, 5, 4077.

Tanentzap, A. J., Yan, N. D., Keller, B., Girard, R., Heneberry, J., Gunn, J. M., … Taylor, P.

(2008). Cooling lakes while the world warms: Effects of forest regrowth and increased

dissolved organic matter on the thermal regime of a temperate, urban lake. Limnology and

Oceanography, 53, 404–410.

Tremblay, S., & Magnan, P. (1991). lnteractions between two distantly related species, Brook

Trout (Salvelinus fontinalis) and White Sucker (Catostomus commersonii). Canadian

Journal of Fish and Aquatic Science, 48, 857–867.

Trimmer, M., Hildrew, A. G., Jackson, M. C., Pretty, J. L., & Grey, J. (2009). Evidence for the

role of methane-derived carbon in a free-flowing, lowland river food web. Limnology and

Oceanography, 54, 1541–1547.

Trippel, E. A., & Harvey, H. H. (1987). Abundance, growth, and food supply of White Suckers

(Catostomus commersonii) in relation to lake morphometry and pH. Canada Journal of.

Zoology, 65, 558–564.

Tunney, T. D., McCann, K. S., Jarvis, L., Lester, N. P., & Shuter, B. J. (2018). Blinded by the

light? Nearshore energy pathway coupling and relative predator biomass increase with

reduced water transparency across lakes. Oecologia, 186, 1031–1041.

Turgeon, K. A., Solomon, C. T., Nozais, C. H., & Gregory-Eaves, I. (2016). Do novel

ecosystems follow predictable trajectories? Testing the trophic surge hypothesis in

reservoirs using fish. Ecosphere, 7, e01617.

120 Urabe, J., Kyle, M., Makino, W., Yoshida, T., Andersen, T., & Elser, J. J. (2002). Reduced light

increases herbivore production due to stoichiometric effects of light/nutrient balance.

Ecology, 83, 619–627.

Urabe, J., & Sterner, R. W. (1996). Regulation of herbivore growth by the balance of light and

nutrients. Ecology, 93, 8465–8469.

Vadeboncoeur, Y., Peterson, G., Vander Zanden, M. J., & Kalff, J. (2008). Benthic algal

production across lake size gradients: Interactions among morphometry, nutrients, and light.

Ecology, 89, 2542–2552.

Vadeboncoeur, Y., Vander Zanden, M. J., & Lodge, D. M. (2002). Putting the lake back

together: Reintegrating benthic pathways into lake food web models. BioScience, 52, 44–

54.

Vadstein, O. (2000). Heterotrophic, planktonic bacteria and cycling of phosphorus: Phosphorus

requirements, competitive ability and food web interactions. Advances in Microbial

Ecology, 16, 115–167.

Valentine-Rose, L., Rypel, A. L., & Layman, C. A. (2011). Community secondary production as

a measure of ecosystem function: A case study with aquatic ecosystem fragmentation.

Bulletin of Marine Science, 87, 913–937.

Vander Zanden, H. B., Soto, D. X., Bowen, G. J., & Hobson, K. A. (2016). Expanding the

isotopic toolbox: Applications of hydrogen and oxygen stable isotope ratios to food web

studies. Frontiers in Ecology and Evolution, 4, 1–19.

Vander Zanden, M. J., Clayton, M. K., Moody, E. K., Solomon, C. T., & Weidel, B. C. (2015).

Stable isotope turnover and half-life in animal tissues : A literature synthesis. Plos One, 10,

e0116182.

121 Vander Zanden, M. J., & Rasmussen, J. B. (2001). Variation in δ15N and δ13C trophic

fractionation: Implications for aquatic food web studies. Limnology and Oceanography, 46,

2061–2066.

Vander Zanden, M. J., & Vadeboncoeur, Y. (2002). Fishes as integrators of benthic and pelagic

food webs in lakes. Ecology, 83, 2152–2161.

Vander Zanden, M. J., Vadeboncoeur, Y., & Chandra, S. (2011). Fish reliance on littoral-benthic

resources and the distribution of primary production in lakes. Ecosystems, 14, 894–903.

Vanderklift, M. A., & Ponsard, S. (2003). Sources of variation in consumer-diet d15N

enrichment: a meta-analysis. Oecologia, 136, 169–182.

Vasconcelos, F.R., Diehl, S., Rodriguez, P., Hederström, P., Karlsson, J., & Byrström, P. (2016).

Asymmetrical competition between aquatic primary producers in a warmer and browner

world. Ecology, 97, 2580–2592.

Vasconcelos, F.R., Diehl, S., Rodriguez, P., Hederström, P., Karlsson, J., & Byrström, P. (2018).

Effects of terrestrial organic matter on aquatic primary production as mediated by pelagic –

benthic resource fluxes. Ecosystems, 21, 1255–1268.

Vuorio, K., Meili, M., & Sarvala, J. (2006). Taxon-specific variation in the stable isotopic

signatures ( d13C and d15N ) of lake phytoplankton. Freshwater Biology, 51, 807–822.

Wassenaar, L. I., & Hobson, K. A. (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.

Weidel, B. C., Baglini, K., Jones, S. E., Kelly, P. T., Solomon, C. T., Zwart, J. A., … Zwart, J.

A. (2017). Light climate and dissolved organic carbon concentration influence species-

specific changes in fish zooplanktivory. Inland Waters, 7, 210–217.

122 Weidel, B. C., Carpenter, S. R., Cole, J., Hodgson, J., Kitchell, J., Pace, M., & Solomon, C.

(2008). Carbon sources supporting fish growth in a north temperate lake. Aquatic Sciences,

70, 446–458.

Wetzel, R. G. (1995). Death, detritus, and energy flow in aquatic ecosystems. Freshwater

Biology, 33, 83–89.

Wilkinson, G. M., Carpenter, S. R., Cole, J. J., & Pace, M. L. (2014). Use of deep autochthonous

resources by zooplankton: Results of a metalimnetic addition of 13C to a small lake.

Limnology and Oceanography, 59, 986–996.

Wilkinson, G. M., Cole, J. J., & Pace, M. L. (2015). Deuterium as a food source tracer:

Sensitivity to environmental water, lipid content, and hydrogen exchange. Limnology and

Oceanography: Methods, 13, 213–223.

Wilkinson, G. M., Pace, M. L., & Cole, J. J. (2013). Terrestrial dominance of organic matter in

north temperate lakes. Global Biogeochemical Cycles, 27, 43–51.

Williamson, C. E., & Rose, K. C. (2010). When UV meets fresh water. Science, 329, 637.

Williamson, C. E., Sanders, R. W., Moeller, R. E., & Stutzman, P. L. (1996a). Utilization of

subsurface food resources for zooplankton reproduction: Implications for diel vertical

migration theory. Limnology and Oceanography, 41, 224–233.

Williamson, C. E., Stemberger, R. S., Morris, D. P., Frost, T. M., & Paulsen, S. G. (1996b).

Ultraviolet radiation in North American lakes: Attenuation estimates from DOC

measurements and implications for plankton communities. Limnology and Oceanography,

41, 1024–1034.

123 Winder, M., Boersma, M., & Spaak, P. (2003). On the cost of vertical migration: are feeding

conditions really worse at greater depths? Freshwater Biology, 48, 383–393.

Winder, M., Spaak, P., & Mooji, W. (2004). Trade-offs in Daphnia habitat selection. Ecology,

85, 2027–2036.

Wissel, B., Boeing, W. J., & Ramcharan, C. W. (2003). Effects of water color on predation

regimes and zooplankton assemblages in freshwater lakes. Limnology and Oceanography,

48, 1965–1976.

Wu, Z., Dijkstra, P., Koch, G., Penuelas, J., & Hungate, B. A. (2011). Responses of terrestrial

ecosystems to temperature and precipitation change: a meta-analysis of experimental

manipulation. Global Change Biology, 17, 927–942.

Xenopoulos, M. A., Lodge, D. M., Frentress, J., Kreps, T. A., Bridgham, S. D., Grossman, E., &

Jackson, C. J. (2003). Regional comparisons of watershed determinants of dissolved organic

carbon in temperate lakes from the Upper Great Lakes region and selected regions globally.

Limnology and Oceanography, 48, 2321–2334.

Yang, C., Wilkinson, G. M., Cole, J. J., Macko, S. A., & Pace, M. L. (2014). Assigning

hydrogen, carbon, and nitrogen isotope values for phytoplankton and terrestrial detritus in

aquatic food web studies. Inland Waters, 4, 233–242.

Zaret, T. M., & Suffern, S. (1973). Vertical migration in zooplankton. Limnology and

Oceanography, 21, 804–813.

124 Appendix 2A: Contributions of dietary water to zooplankton – Sensitivity analysis

Hydrogen isotope analysis has become a widely used tool to estimate terrestrial organic matter contributions to aquatic consumers because of strong differentiation between terrestrial and primary producers (Doucett et al., 2007). However, use of hydrogen isotopes as a dietary tracer requires several highly influential and poorly defined assumptions that affect model outcomes, particularly, the proportional contribution of dietary water (w) to consumer tissue

(Brett et al., 2018; Vander Zanden et al., 2016; Wilkinson et al., 2015). I performed a sensitivity analysis to demonstrate how these assumptions influence model outcomes and to provide justification for my use of a dietary water value of 0.2.

First, to demonstrate how model outputs are sensitive to variations in the contribution of dietary water, I calculated zooplankton allochthony (KT); proportional contribution of terrestrial resources) using a single isotope (H), two source mixing model:

> > - ?MNNO.QNR − - ?OSTLN KL = > > - ?LURR − - ?OSTLN

2 2 where δ HZOOP.COR represents the zooplankton δ H corrected for dietary water, and

2 2 δ HPHYTO and δ HTERR represent the isotopic ratios of phytoplankton and terrestrial organic matter respectively. The correction for the influence of dietary water on consumer δ2H was determined using the model of Solomon et al. (2009):

" !"#" = 1 − (1 − !)

125 Where !"#" is the proportion of H from dietary water in consumer tissue at trophic level t, and ! is the per-trophic-level contribution of dietary water to consumer H. Although in my more complex models, I use both deep and epilimnetic phytoplankton as sources, for the purposes of this sensitivity analysis, I only used epilimnetic phytoplankton. The δ2H values of deep and epilimnetic phytoplankton were similar within lakes and tended to differ more in their

δ13C and δ15N values. To assess the effects of altering the contributions of dietary water, I used w values of 0.0, 0.05, 0.15, 0.2, 0.25 and 0.3. I assumed zooplankton occupy trophic position 2.

Zooplankton δ2H was corrected for dietary water by subtracting the δ2H enrichment factor from raw δ2H values. The dietary water enrichment correction factor was calculated following the equation outlined in Berggren et al. (2014):

> > > > - ?MNNO.QNR = - ?16:GH2 − (- ?16:GH2 − !"#" × - ?J6"27)/(1 − !"#")

> 2 > where - ?J6"27 is the δ H of water, and - ?16:GH2 is the sample being corrected for.

For samples where zooplankton δ13C values were similar to that of phytoplankton (4 lakes), I solved for the value of w using a two-source, δ2H mixing model. Assuming that zooplankton only received δ2H from either phytoplankton or water, the contribution of dietary water was determined using the equation:

> > - ?MNNO.RWX − - ?OSTLN ω = > > - ?XWLUR − - ?OSTLN

126 2 2 2 Where δ HZOOP.RAW , δ HPHYTO and δ HTERR represent the isotopic ratios of uncorrected zooplankton values, phytoplankton and terrestrial organic matter respectively.

When contributions from dietary water were low (e.g. w of 0.0), contributions of terrestrial organic matter were high, and in certain cases, exceeded 100% (Fig. 2A.1).

Alternatively, when dietary water contributions were high (e.g. w of 0.3), nearly all terrestrial contributions were < 0. Seeing as the contributions of a resource must be within 0 and 100%, these dietary water values of 0.0, 0.25 and 0.3 are not feasible, and fall outside of geometric mixing space. For the dietary water value of 0.2, I only had one lake in which allochthony was <

0 and was largely driven by a single outlier (proportion of terrestrial organic matter in diet = –

0.5). Removal of this outlier resulted in an average allochthony value of – 0.07.

For samples assumed to exclusively get their δ2H from either phytoplankton or water, the average calculated value of w was 0.203 ± 0.024 (Table 2A.1). This value is similar to the recommended w outlined in Wilkinson et al. (2015), and slightly lower than the value suggested by Brett et al. (2018). The dietary water value of 0.2 appears to be suitable for my data and represents a conservative value since these estimates are based on the assumption that terrestrial contributions are negligible.

127

Fig. 2A.1: Relationship between allochthony estimates (proportion) with varying contributions from dietary water (ranging from 0.0 to 0.3). The boxplots represent the 5th, 25th, 50th, 75th, and

95th percentiles of the averaged combined results of the two-source mixing model across the eight study lakes.

Table 2A.1: Estimates of w for lakes where allochthony was assumed to be zero 2 2 2 Lake δ H ZOOPLANKTON δ HPHYTO δ HWATER w L224 -193.71 -227.14 -64.55 0.21 L373 -201.38 -228.12 -67.12 0.17 L442 -197.31 -232.68 -71.68 0.22 L626 -193.30 -228.99 -68.00 0.22

Average -196.42 -229.23 -67.84 0.203 SD 3.76 2.42 2.95 0.024

128 Appendix 2B: Data used in zooplankton Bayesian Mixing models

Table 2B.1: Average basal end-member isotopic values for terrestrial organic matter, epilimnetic

and hypolimnetic (deep) phytoplankton, as well as average zooplankton values used in MixSIAR

Bayesian mixing models. Zooplankton d2H was corrected for dietary water, d13C values used in

the model are lipid corrected; values in parenthesis are d13C values prior to lipid correction.

Lake Sample Type d2H ± SD (‰) d13C ± SD (‰) d15N ± SD (‰) n All Terrestrial vegetation -187.60 11.18 -29.70 1.03 -5.50 1.61 27 Water -76.89 2.79 6 L164 Epilimnetic Phytoplankton -239.90 17.23 -31.28 0.82 -0.63 1.35 7

Zooplankton -215.54 6.58 -31.61 (-33.20) 1.18 3.37 1.52 6

Water -72.52 1.63 6 Epilimnetic Phytoplankton -233.52 17.15 -30.30 2.21 1.05 1.33 6 L223 Deep Phytoplankton -234.93 17.00 -34.80 2.21 -0.98 1.33 1 Zooplankton -220.91 8.35 -33.31 (-35.01) 2.94 5.47 0.95 6 Water -64.93 2.56 6 Epilimnetic Phytoplankton -225.01 17.12 -29.70 3.10 1.99 1.17 6 L224 Deep Phytoplankton -227.14 17.00 -36.30 3.10 1.78 1.17 1

Zooplankton -226.00 12.90 -32.81 (-35.38) 1.65 6.17 0.91 6

Water -74.94 1.87 6 L239 Epilimnetic Phytoplankton -236.87 17.30 -31.25 1.52 -0.54 1.81 7 Zooplankton -226.89 7.19 -32.10 (-34.40) 1.93 5.18 1.52 6 Water -67.12 1.26 5 Epilimnetic Phytoplankton -228.12 17.05 -29.39 2.54 2.29 2.17 5 L373 Deep Phytoplankton -229.82 17.00 -34.70 2.54 2.66 2.17 1 Zooplankton -234.94 14.15 -33.75 (-36.51) 1.86 6.10 0.71 5

Water -71.68 1.88 6 Epilimnetic Phytoplankton -232.68 17.10 -30.30 1.84 2.87 1.16 6 L442 Deep Phytoplankton -235.31 17.00 -36.70 1.84 3.53 1.16 1 Zooplankton -228.72 11.41 -34.33 (-36.50) 2.10 8.60 1.77 6 Water -68.00 0.84 6 Epilimnetic Phytoplankton -228.99 17.02 -27.76 4.25 3.46 2.25 6 L626 Deep Phytoplankton -230.48 17.00 -33.44 4.25 2.90 2.25 1 Zooplankton -224.63 5.54 -32.64 (-35.10) 3.56 8.10 3.10 6

Water -73.59 1.92 6 L658 Epilimnetic Phytoplankton -236.40 17.2 -31.24 2.91 3.93 1.10 7 Zooplankton -208.43 7.03 -32.58 (-34.04) 3.78 6.81 1.67 6

129

The following biplots represent the data used to generate Bayesian mixing models, with trophic discrimination factors applied. In general, at least one zooplankton sample from each lake falls outside the isotopic mixing space. In all cases, this zooplankton sample was collected in May, prior to lake stratification, or shortly after stratification occurs. Notably, removal of these samples had minimal influence on resource use estimates.

130 -28 Epilimnetic phytoplankton -29 Terrestrial vegetation Zooplankton -30

-31 C(‰) 13 d -32

-33

-34 -260 -250 -240 -230 -220 -210 -200 -190 -180 d2H (‰)

3 2 1 0 -1 -2

N(‰) -3 15 d -4 -5 -6 -7 -8 -34 -33 -32 -31 -30 -29 -28 d13C (‰)

Fig. 2B.1: Data used in mixing models for L164.

131 -20

-22 Deep phytoplankton Epilimnetic phytoplankton -24 Terrestrial vegetation -26 Zooplankton -28

-30 C(‰) 13

d -32

-34

-36

-38

-40 -255 -235 -215 -195 -175 2 d H (‰) 4

2

0

-2 N(‰) 15 d -4

-6

-8 -40 -38 -36 -34 -32 -30 -28 -26 -24 -22 -20 d13C (‰)

Fig. 2B.2: Data used in mixing models for L223.

132 -26

-28

-30

-32

-34 C(‰) 13 d -36 Deep phytoplankton -38 Epilimnetic phytoplankton

-40 Terrestrial vegetation Zooplankton -42 -255 -235 -215 -195 -175 2 d H (‰) 6

4

2

0 N(‰)

15 -2 d

-4

-6

-8 -42 -40 -38 -36 -34 -32 -30 -28 -26 d13C (‰)

Fig. 2B.3: Data used in mixing models for L224.

133 -28

-29

-30

-31

-32 C(‰) 13 d -33 Epilimnetic phytoplankton

-34 Terrestrial vegetation

-35 Zooplankton

-36 -260 -250 -240 -230 -220 -210 -200 -190 -180 d2H (‰)

6

4

2

0 N(‰)

15 -2 d

-4

-6

-8 -36 -35 -34 -33 -32 -31 -30 -29 -28 d13C (‰)

Fig. 2B.4: Data used in mixing models for L239.

134 -25

-27

-29

-31 C(‰)

13 -33 d Deep phytoplankton -35 Epilimnetic phytoplankton Terrestrial vegetation -37 Zooplankton -39 -255 -235 -215 -195 -175

2 d H (‰) 6

4

2

0 N(‰)

15 -2 d

-4

-6

-8 -39 -37 -35 -33 -31 -29 -27 -25 d13C (‰)

Fig. 2B.5: Data used in mixing models for L373.

135 -26

-28

-30

-32 C(‰)

13 Deep phytoplankton

d -34 Epilimnetic phytoplankton -36 Terrestrial vegetation -38 Zooplankton

-40 -255 -245 -235 -225 -215 -205 -195 -185 -175 d2H (‰)

10

8

6

4

2

N(‰) 0 15 d -2

-4

-6

-8 -40 -38 -36 -34 -32 -30 -28 -26 d13C (‰)

Fig. 2B.6: Data used in mixing models for L442.

136 -20 -22 -24 -26 -28 -30 C(‰) 13

d -32 Deep phytoplankton -34 Epilimnetic phytoplankton -36 Terrestrial vegetation -38 Zooplankton -40 -255 -245 -235 -225 -215 -205 -195 -185 -175 d2H (‰)

12 10 8 6 4 2 N(‰) 15

d 0 -2 -4 -6 -8 -40 -38 -36 -34 -32 -30 -28 -26 -24 -22 -20 d13C (‰)

Fig. 2B.7: Data used in mixing models for L626.

137 -25 -27 -29 -31 -33 -35 C(‰) 13

d -37 -39 Epilimnetic phytoplankton -41 Terrestrial vegetation Zooplankton -43 -45 -255 -245 -235 -225 -215 -205 -195 -185 -175 d2H (‰)

8

6

4

2

0 N(‰) 15

d -2

-4

-6

-8 -41 -39 -37 -35 -33 -31 -29 -27 -25 d13C (‰)

Fig. 2B.8: Data used in mixing models for L658.

138 Appendix 2C: Seasonal vertical chlorophyll a profiles of survey lakes in 2018

Plots are arranged in order of DOC concentration. L239 was only profiled for chlorophyll a twice (B. Sherbo, unpublished data). Note: Solid lines indicate thermocline depth on that particular sampling date.

139

140

141

142 Appendix 2D: Zooplankton basal resource use Bayesian mixing model outputs

Table 2D.1: Bayesian mixing model posterior distribution outputs of zooplankton resource use.

Deep phytoplankton is phytoplankton below the thermocline, Epi phytoplankton is phytoplankton from the epilimnion, and t-OM is terrestrial organic matter.

Note, the 50th percentile corresponds to the posterior distribution median.

Posterior Distribution Percentiles Lake Resource Mean SD 5th 25th 50th 75th 95th Proportion L224 Deep Phytoplankton 0.554 0.141 0.314 0.467 0.561 0.646 0.777 Epi Phytoplankton 0.391 0.151 0.153 0.290 0.387 0.485 0.651 t-OM 0.055 0.048 0.004 0.018 0.041 0.078 0.153 L373 Deep Phytoplankton 0.705 0.141 0.450 0.626 0.724 0.806 0.900 Epi Phytoplankton 0.216 0.147 0.025 0.104 0.193 0.300 0.491 t-OM 0.080 0.054 0.008 0.036 0.071 0.114 0.181 L223 Deep Phytoplankton 0.223 0.115 0.038 0.138 0.221 0.301 0.419 Epi Phytoplankton 0.697 0.125 0.484 0.614 0.701 0.785 0.897 t-OM 0.080 0.065 0.006 0.029 0.065 0.116 0.204 L626 Deep Phytoplankton 0.503 0.228 0.124 0.328 0.507 0.685 0.870 Epi Phytoplankton 0.415 0.221 0.062 0.244 0.412 0.575 0.789 t-OM 0.083 0.067 0.006 0.030 0.066 0.120 0.212 L442 Deep Phytoplankton 0.698 0.142 0.448 0.608 0.708 0.803 0.912 Epi Phytoplankton 0.250 0.149 0.032 0.140 0.235 0.348 0.511 t-OM 0.052 0.047 0.004 0.017 0.039 0.071 0.144 L239 Epi Phytoplankton 0.813 0.053 0.728 0.777 0.812 0.845 0.902 t-OM 0.187 0.053 0.098 0.155 0.188 0.223 0.272 L658 Epi Phytoplankton 0.779 0.100 0.601 0.718 0.787 0.850 0.933 t-OM 0.221 0.100 0.067 0.150 0.213 0.282 0.399 L164 Epi Phytoplankton 0.731 0.119 0.536 0.648 0.732 0.816 0.928 t-OM 0.269 0.119 0.072 0.184 0.268 0.352 0.464

The Bayesian mixing model posterior is a probability distribution that must sum to 1. My models use Markov Chain Monte Carlo, where dietary proportions are repeatedly “guessed,” and those that are not probabilistically consistent with the data are discarded. Each new “guess” is required to be closer to the older “guesses,” generating a Markov chain. If “guesses” cannot find a high-

143 probability value close to the previous “guess,” then the Markov chains cannot converge (i.e. did not achieve a stationary distribution of high probability guesses), as was the case when deep phytoplankton was included in models for L164, L239 and L658. Thus, the posterior distribution represents a probable range of resource use estimates given the data provided (and their associated errors). A graphical example of the posterior distribution resources estimates can be seen in Fig. 2D.1, and corresponds to the posterior distribution outlined in Table 2D.1, for L224.

L224

1.00 Deep phytoplankton Epilimnetic phytoplankton Terrestrial organic matter

0.75

0.50 Scaled Posterior Density Scaled Posterior

0.25

0.00

0.00 0.25 0.50 0.75 1.0 Proportion of Diet Fig. 2D.1: Posterior distribution plot of zooplankton resource use in L224. The results show the posterior densities (y-axis) of the proportional utilization (x-axis) of the three main basal resources within L224. Note posterior densities correspond to the percentiles in Table 2D.1.

144 Appendix 2E: Vertical δ13C-DIC profiles

Fig. 2E.1: Vertical profiles of δ13C-DIC collected mid-summer, 2017. Note: data were not collected for L164. Horizontal dashed and solid lines indicate photic and thermocline depth, respectively.

145 Appendix 3A: Basal energy sources and lake consumers isotope summary

Table 3A.1: Average isotopic values for terrestrial vegetation, phytoplankton, and periphyton, and average consumer values of White Sucker and benthic invertebrates used in Bayesian mixing models. The d2H value of water is also included. Terrestrial vegetation values were pooled among all lakes and tree species.

Lake Type d2H ± SD (‰) d13C ± SD (‰) d15N ± SD (‰) n Terrestrial vegetation -187.60 -29.70 -5.50 27 All ± 11.18 ± 1.03 ± 1.61 Water -76.89 ± 2.79 6 White Sucker -198.40 ± 11.13 -33.95 ± 1.10 6.41 ± 0.37 7

Littoral Benthos -213.55 -26.90 1.82 10 L164 ± 15.10 ± 1.39 ± 1.31 Profundal Chironomids -193.55 ± 27.75 -39.70 ± 0.68 0.90 ± 0.35 3 Phytoplankton -237.89 ± 17.23 -30.45 ± 0.82 0.55 ± 1.35 6 Periphyton -207.36 -26.23 0.83 5 ± 11.33 ± 1.93 ± 3.01 Water -72.52 ± 1.63 6 White Sucker -201.00 ± 15.12 -28.95 ± 4.14 6.90 ± 1.24 9

Littoral Benthos -201.03 -26.36 2.39 9 L223 ± 18.63 ± 4.08 ± 1.11 Profundal Chironomids -193.41 ± 0.74 -41.94 ± 1.19 1.15 ± 1.50 2 Phytoplankton -233.52 ± 17.15 -30.30 ± 2.21 1.05 ± 1.33 6 Periphyton -209.15 ± 11.16 -23.27 ± 2.27 1.14 ± 0.80 5

Water -64.93 ± 2.56 6 White Sucker -191.77 ± 18.95 -24.80 ± 3.23 6.10 ± 0.63 9

Littoral Benthos -189.34 -24.14 2.16 6 L224 ± 20.80 ± 1.35 ± 1.32 Profundal Chironomids -204.60 ± 11.23 -36.23 ± 0.03 3.62 ± 0.06 2 Phytoplankton -225.01 ± 17.12 -29.7 ± 3.10 1.99 ± 1.17 6 Periphyton -211.44 16.10 -22.34 1.45 -0.10 0.22 5 ± ± ± Water -74.94 ± 1.87 6 White Sucker -208.54 ± 17.66 -23.64 ± 2.83 5.91 ± 0.82 9

Littoral Benthos -198.48 -25.50 1.50 6 L239 ± 21.35 ± 1.10 ± 1.42 Profundal Chironomids -184.48 ± 11.30 -30.45 ± 1.08 6.11 ± 1.42 2 Phytoplankton -235.94 ± 17.10 -30.79 ± 1.52 0.06 ± 1.80 6 Periphyton -213.60 12.31 -23.82 2.98 1.48 2.13 5 ± ± ± Water -67.12 ± 1.26 5 White Sucker -202.75 ± 14.10 -27.16 ± 3.26 5.99 ± 0.75 7

Littoral Benthos -185.45 -24.36 1.53 5 L373 ± 13.60 ± 1.61 ± 0.99 Profundal Chironomids -207.70 ± 12.30 -33.20 ± 1.40 4.10 ± 0.98 2

146 Phytoplankton -228.12 ± 17.05 -29.39 ± 2.54 2.29 ± 2.17 5 Periphyton -202.90 -25.03 0.43 5 ± 8.10 ± 0.42 ± 0.71 Water -71.68 ± 1.88 6 White Sucker -214.50 ± 10.43 -31.53 ± 1.30 9.24 ± 1.47 7

Littoral Benthos -191.50 -26.15 3.31 5 L442 ± 13.74 ± 0.95 ± 1.26 Profundal Chironomids -177.10 ± 6.50 -40.58 ± 0.22 5.10 ± 0.19 2 Phytoplankton -232.68 ± 17.10 -30.30 ± 1.84 2.87 ± 1.16 6 Periphyton -213.00 10.87 -22.32 1.12 1.09 0.65 5 ± ± ± Water -68.00 ± 0.84 6 White Sucker -229.64 ± 7.20 -27.28 ± 1.12 10.46 ± 0.74 8

Littoral Benthos -189.46 -23.31 3.10 5 L626 ± 20.95 ± 1.83 ± 1.35 Profundal Chironomids -190.75 ± 14.70 -32.13 ± 1.10 7.39 ± 0.98 2 Phytoplankton -228.99 ± 17.02 -27.76 ± 4.25 3.46 ± 2.25 6 Periphyton -220.80 -20.40 2.76 5 ± 9.36 ± 2.93 ± 1.78 Water -73.59 ± 1.92 6 White Sucker -201.19 ± 6.10 -31.95 ± 1.93 8.22 ± 0.36 3

Littoral Benthos -196.77 -26.75 3.48 5 L658 ± 5.22 ± 0.83 ± 0.42 Profundal Chironomids -175.89 ± 4.72 -39.17 ± 2.79 3.81 ± 0.77 2 Phytoplankton -234.59 ± 17.24 -30.49 ± 2.91 2.87 ± 1.10 6 Periphyton -217.80 ± 14.18 -22.38 ± 2.73 1.28 ± 1.12 5 Note: All consumer values have been corrected for dietary water, using a per trophic level dietary water contribution (!) of 0.2, except for profundal chironomids, which had a dietary water contribution value of 0.39.

147 -19 Terrestrial Vegetation -21 Periphyton Phytoplankton -23

-25

C (‰ ) ) (‰ C -27 13 d -29

-31

-33 -250 -240 -230 -220 -210 -200 -190 -180 -170 d2H (‰)

4 3 2 1 0 -1 -2 N (‰ ) ) (‰ N -3 15

d -4 -5 -6 -7 -8 -32 -31 -30 -29 -28 -27 -26 -25 -24 -23 -22 -21 -20 -19 d13C (‰ )

Fig. 3A.1: Biplots of basal resource (phytoplankton, periphyton, and terrestrial vegetation) isotopic values used in the Bayesian mixing models. Each point represents the open-water season average for each lake, with the exception of terrestrial vegetation, which is a grand average of all samples. The error bars are the standard deviation.

148 Appendix 3B: Summary of lake specific benthic invertebrate biomass and density estimates

Table 3B.1: Depth specific, habitat specific, and whole-lake estimates of benthic invertebrate and chironomid density and biomass. Littoral averages were weighted by area of each sampling depth < 4m, and profundal was weighted by area of sampling depths below the thermocline.

All Benthos Chironomid Density Biomass Density Biomass Lake Depth (ind. /m2) (g / m2) (ind. / m2) (g / m2) 0.5 m 4938 1.09 4574 0.10 1 m 2859 0.12 2339 0.06 4 m 2859 0.55 2651 0.16 L164 7 m 260 0.49 455 0.49 Average Littoral 3470 0.54 3088 0.10 Average Profundal 455 0.49 455 0.49 Average Whole-lake 1745 0.50 1674 0.30 0.5 m 6029 0.55 5146 0.10 1 m 7277 0.25 6289 0.16 4 m 2443 0.06 2027 0.04

9 m 2664 0.63 2209 0.58 L223 12 m 260 0.23 195 0.22 Average Littoral 5127 0.25 4383 0.10 Average Profundal 1870 0.50 1544 0.46 Average Whole-lake 3187 0.42 2690 0.32 0.5 m 7328 0.41 6029 0.12 1 m 4054 0.14 3690 0.10 4 m 2144 0.04 1819 0.03

8 m 4223 0.14 4223 0.14 L224 12m 1624 0.17 1559 0.17 18 m 910 0.48 910 0.48 Average Littoral 4208 0.17 3619 0.08 Average Profundal 1860 0.34 1848 0.34 Average Whole-lake 2677 0.28 2466 0.27 0.5 m 8316 1.07 2859 0.07 1 m 5561 0.52 2807 0.09 4 m 2274 1.75 936 0.02

8 m 2924 1.17 312 0.06 L239 12 m 1559 0.65 156 0.02

149 18 m 3248 0.76 416 0.21 Average Littoral 5393 1.05 2258 0.06 Average Profundal 2565 0.88 290 0.09 Average Whole-lake 2865 0.94 949 0.08 0.5 m 6705 0.25 6133 0.17 1 m 1975 0.09 1299 0.02 4 m 780 0.52 520 0.01

8 m 910 0.05 910 0.05 L373 12 m 585 0.08 585 0.08 18 m 0.0 0 0 0.0 Average Littoral 2780 0.26 2261 0.13 Average Profundal 372 0.06 372 0.06 Average Whole-lake 1498 0.16 1346 0.08 0.5 m 4574 0.47 2131 0.04 1 m 1715 0.21 1195 0.01 4 m 676 0.01 624 0.04

8 m 2599 0.22 2599 0.29 L442 12 m 0 0 0 0.0 Average Littoral 2578 0.31 1409 0.03 Average Profundal 1225 0.14 1225 0.14 Average Whole-lake 2444 0.20 747 0.05 0.5 m 11071 0.73 7173 0.09 1 m 9667 0.60 5301 0.08 4 m 1715 0.08 1247 0.03

8 m 5912 0.54 5782 0.54 L626 11 m 195 0.03 65 0.01 Average Littoral 6808 0.55 4249 0.06 Average Profundal 3459 0.53 3381 0.53 Average Whole-lake 4215 0.53 3135 0.37 0.5 m 7277 2.31 4574 0.11 1 m 4314 0.45 3378 0.08 4 m 728 0.07 572 0.04

8 m 650 0.24 650 0.24 L658 12 m 0 0.0 0 0.0 Average Littoral 4856 1.14 3345 0.09 Average Profundal 365 0.14 365 0.14 Average Whole-lake 1806 0.44 1317 0.11

150

Fig. 3B.1: Average depth-specific benthos biomass is not related to either (a) depth-specific dissolved oxygen concentration or (b) depth-specific temperature. Note, dissolved oxygen and temperature were measured on the day of collection, however, the relationships did not change if average depth specific temperature or dissolved oxygen for the open-water season were plotted instead.

151 Appendix 3C: Summary of MixSIAR mixing model outputs

Table 3C.1: Summary of MixSIAR Bayesian mixing model basal resource use estimates by

White Suckers. The 50th percentile of the posterior distribution represents the median.

Posterior Distribution Percentiles Resource Mean SD 5th 25th 50th 75th 95th Periphyton 0.549 0.101 0.375 0.485 0.553 0.619 0.706 L224 Phytoplankton 0.132 0.083 0.016 0.067 0.123 0.184 0.282 Terrestrial 0.319 0.060 0.224 0.279 0.317 0.356 0.418 Periphyton 0.484 0.175 0.176 0.367 0.491 0.610 0.758 L373 Phytoplankton 0.235 0.114 0.050 0.153 0.233 0.312 0.436 Terrestrial 0.281 0.093 0.130 0.217 0.279 0.343 0.438 Periphyton 0.431 0.156 0.154 0.330 0.438 0.545 0.675 L223 Phytoplankton 0.221 0.114 0.037 0.134 0.218 0.303 0.412 Terrestrial 0.348 0.099 0.201 0.282 0.339 0.404 0.521 Periphyton 0.116 0.081 0.010 0.052 0.103 0.165 0.271 L626 Phytoplankton 0.624 0.097 0.464 0.561 0.627 0.688 0.786 Terrestrial 0.260 0.073 0.133 0.213 0.264 0.311 0.374 Periphyton 0.093 0.093 0.005 0.027 0.063 0.126 0.281 L442 Phytoplankton 0.551 0.082 0.410 0.505 0.560 0.609 0.666 Terrestrial 0.356 0.068 0.240 0.315 0.356 0.399 0.462 Periphyton 0.488 0.110 0.303 0.417 0.487 0.560 0.664 L239 Phytoplankton 0.119 0.084 0.011 0.052 0.106 0.172 0.278 Terrestrial 0.393 0.088 0.240 0.336 0.399 0.454 0.527 Periphyton 0.237 0.195 0.014 0.078 0.189 0.350 0.631 L658 Phytoplankton 0.305 0.142 0.062 0.204 0.308 0.406 0.543 Terrestrial 0.458 0.165 0.170 0.337 0.476 0.579 0.708 Periphyton 0.447 0.144 0.200 0.361 0.449 0.541 0.674 L164 Phytoplankton 0.160 0.109 0.018 0.078 0.143 0.225 0.365 Terrestrial 0.392 0.098 0.222 0.335 0.400 0.455 0.536

152

Fig. 3C.1: Relationship between White Sucker allochthony and benthic invertebrate allochthony

(r = 0.83, p = 0.01).

153 Table 3C.2: Summary of MixSIAR Bayesian mixing model estimates of basal resource use by littoral benthic invertebrates. The 50th percentile of the posterior distribution represents the median.

Column1 Column2 Column3 Column4 Posterior Distribution Percentiles Resource Mean SD 5th 25th 50th 75th 95th Periphyton 0.562 0.095 0.397 0.505 0.568 0.622 0.711 L224 Phytoplankton 0.080 0.065 0.006 0.029 0.064 0.116 0.210 Terrestrial 0.358 0.088 0.208 0.301 0.359 0.414 0.498 Periphyton 0.652 0.175 0.336 0.548 0.670 0.778 0.904 L373 Phytoplankton 0.080 0.077 0.004 0.023 0.057 0.111 0.227 Terrestrial 0.268 0.148 0.038 0.156 0.259 0.372 0.519 Periphyton 0.535 0.118 0.335 0.467 0.541 0.610 0.708 L223 Phytoplankton 0.105 0.084 0.007 0.040 0.086 0.150 0.269 Terrestrial 0.360 0.086 0.175 0.217 0.307 0.363 0.496 Periphyton 0.535 0.105 0.357 0.470 0.535 0.602 0.703 L626 Phytoplankton 0.115 0.085 0.008 0.048 0.099 0.166 0.278 Terrestrial 0.351 0.095 0.187 0.295 0.354 0.414 0.501 Periphyton 0.444 0.083 0.311 0.390 0.441 0.495 0.579 L442 Phytoplankton 0.152 0.076 0.028 0.098 0.147 0.204 0.285 Terrestrial 0.404 0.073 0.289 0.354 0.402 0.451 0.529 Periphyton 0.526 0.098 0.375 0.460 0.522 0.588 0.692 L239 Phytoplankton 0.090 0.071 0.007 0.034 0.074 0.129 0.228 Terrestrial 0.384 0.103 0.203 0.322 0.390 0.455 0.544 Periphyton 0.372 0.105 0.203 0.305 0.371 0.437 0.545 L658 Phytoplankton 0.179 0.087 0.039 0.116 0.177 0.239 0.323 Terrestrial 0.449 0.086 0.327 0.391 0.440 0.497 0.606 Periphyton 0.491 0.111 0.320 0.416 0.488 0.560 0.678 L164 Phytoplankton 0.081 0.061 0.006 0.033 0.068 0.115 0.200 Terrestrial 0.428 0.104 0.243 0.364 0.434 0.500 0.585

154 Table 3C.3: MixSIAR Bayesian mixing model posterior distributions for White Sucker prey use of littoral benthic invertebrates, profundal chironomids and zooplankton. The 50th percentile of the posterior distribution represents the median.

Posterior Distribution Percentiles Prey Mean SD 5th 25th 50th 75th 95th Littoral benthic invertebrates 0.752 0.074 0.629 0.703 0.754 0.802 0.871 L224 Profundal Chironomids 0.129 0.084 0.013 0.061 0.120 0.185 0.280 Zooplankton 0.118 0.078 0.011 0.054 0.109 0.172 0.262 Littoral benthic invertebrates 0.583 0.087 0.431 0.531 0.591 0.643 0.710 L373 Profundal Chironomids 0.199 0.146 0.014 0.081 0.171 0.292 0.481 Zooplankton 0.218 0.101 0.043 0.142 0.226 0.295 0.369 Littoral benthic invertebrates 0.646 0.132 0.421 0.558 0.648 0.738 0.863 L223 Profundal Chironomids 0.076 0.067 0.003 0.025 0.056 0.108 0.210 Zooplankton 0.279 0.138 0.053 0.178 0.276 0.372 0.508 Littoral benthic invertebrates 0.289 0.125 0.093 0.199 0.283 0.369 0.515 L626 Profundal Chironomids 0.570 0.225 0.050 0.480 0.622 0.723 0.850 Zooplankton 0.141 0.141 0.008 0.040 0.092 0.188 0.459 Littoral benthic invertebrates 0.459 0.068 0.345 0.415 0.460 0.504 0.566 L442 Profundal Chironomids 0.103 0.066 0.012 0.050 0.095 0.145 0.221 Zooplankton 0.439 0.087 0.289 0.382 0.441 0.497 0.576 Littoral benthic invertebrates 0.663 0.128 0.443 0.581 0.671 0.755 0.863 L239 Profundal Chironomids 0.144 0.102 0.010 0.063 0.129 0.208 0.335 Zooplankton 0.193 0.142 0.014 0.078 0.166 0.283 0.466 Littoral benthic invertebrates 0.352 0.145 0.093 0.254 0.359 0.454 0.582 L658 Profundal Chironomids 0.317 0.132 0.098 0.227 0.313 0.404 0.542 Zooplankton 0.331 0.133 0.104 0.245 0.327 0.416 0.550 Littoral benthic invertebrates 0.199 0.124 0.016 0.096 0.191 0.297 0.409 L164 Profundal Chironomids 0.446 0.085 0.303 0.388 0.449 0.506 0.581 Zooplankton 0.355 0.189 0.043 0.206 0.364 0.506 0.647

155

Fig. 3C.2: White Sucker dietary resource use estimates for (a) littoral invertebrates (b) zooplankton and (c) profundal chironomids in relation to DOC concentration. Boxplots represent the 5th, 25th, 50th, 75th and 95th percentiles of the posterior distribution from MixSIAR models.

156

-20 Zooplankton

-25 Littoral invertebrates Profundal chironomids

-30 ‰ ) ‰ C ( C

13 -35 d

-40

-45 -240 -230 -220 -210 -200 -190 -180 -170 2 d H (‰ ) 10 9 8 7 6 ‰ ) ‰ 5 N ( N

15 4 d 3 2 1 0 -45 -40 -35 -30 -25 -20 d13C (‰ )

Fig. 3C.3: Biplots of littoral invertebrates, zooplankton and profundal chironomids used to estimate dietary resource use estimates (Table 3C.3). Each point represents the average littoral benthic invertebrate, zooplankton or profundal chironomid isotope values from each lake.

157

Fig. 3C.4: Patterns between (a) reliance on benthic invertebrates (littoral + profundal) and chironomid biomass, and (b) White Sucker bCPUE and reliance on benthic invertebrates.

158 Appendix 3D: Profundal chironomid dietary estimates

The δ13C values of profundal chironomids in a few of my lakes were more depleted than potential dietary sources (e.g. phytoplankton, terrestrial organic matter) and no plausible mixing solution was possible. The δ13C values of profundal chironomids were correlated with volume- weighted hypolimnetic oxygen, suggesting that biogenic methane may be contributing energy because anoxic conditions favour methane production (Fig. 3D.1). Biogenic methane is 13C- depleted, ranging from –60 to –80‰, and utilization of methane oxidizing bacteria (MOB) may further deplete δ13C depending on methane supply and the types of bacteria present (Jones &

Grey, 2011; Ravinet et al., 2010). This relationship between hypolimnetic oxygen and profundal chironomid δ13C has been previously established, where considerable 13C-depletion occurred in profundal chironomids when late summer hypolimnetic oxygen was < 2 mg O2 / L (Jones et al.,

2008), and profundal chironomids may ‘garden’ MOB under such conditions (Jones & Grey,

2011).

To estimate potential biogenic methane contributions to profundal chironomids, I ran a series of hypothetical mixing models using δ13C. The models were run twice for each lake, with

MOB δ13C being assigned a value of either –60‰ or –80‰ (Jones & Grey, 2011). The other end-members in this exercise were the average δ13C value of phytoplankton and terrestrial vegetation, as their δ13C values were similar (within 2‰) (Table 3D.1). Methane contributions to profundal chironomids was estimated using a simple two source, one isotope mixing model

(Vander Zanden & Vadeboncoeur, 2002):

13 13 13 13 Percent MOB contribution = (δ CC – δ CP) / (δ CM – δ CP)

159 13 13 Where δ CC is the isotope value of profundal chironomids, δ CP is the average value of

13 phytoplankton and terrestrial vegetation, and δ CM is the value for MOB.

Fig. 3D.1: Profundal chironomid δ13C in relation to volume-weighted hypolimnetic oxygen

(VWHO) (r = 0.84, p = 0.009)

160 Table 3D.1: Summary of δ13C values used in the simple mixing model and estimate of MOB contributions to profundal chironomids based on two scenarios, where MOB was assigned a δ13C value of –60‰ or –80‰.

Lake Profundal Phytoplankton Terrestrial Terrestrial and Proportion Proportion 13 Chironomids δ C (‰) Vegetation Phytoplankton MOB in MOB in 13 13 13 δ C (‰) δ C (‰) average δ C (‰) diet (-60‰) diet (-80‰) L164 -40.04 -30.45 -29.70 -30.08 0.330 0.200 L223 -41.10 -30.30 -29.70 -30.00 0.370 0.220 L224 -36.20 -29.70 -29.70 -29.70 0.210 0.130 L239 -30.45 -30.79 -29.70 -30.25 0.010 0.004 L373 -33.20 -29.39 -29.70 -29.55 0.120 0.072 L442 -40.42 -30.30 -29.70 -30.00 0.350 0.210 L626 -32.13 -27.76 -29.70 -28.73 0.110 0.066 L658 -41.14 -30.49 -29.70 -30.10 0.370 0.220 Note: I did not apply trophic discrimination factors for this analysis.

When MOB was assigned a value -60‰ for the four lakes where profundal chironomid

δ13C values were near –40‰, MOB contributions to profundal chironomids were estimated to be approximately one third (Table 3D.1). When MOB was assigned a value –80‰, the contributions of MOB in those same lakes declined, contributing nearly one fifth (~20 %). In the other four lakes, contributions from MOB were low, ranging from 1 to 21% of diet under the –

60‰ MOB scenario, to 0.4 to 13% under the –80‰ MOB scenario.

The models I used were simplified and do not allow for the estimation of terrestrial vegetation and phytoplankton dietary contributions because the number of sources cannot exceed the number of isotopes. Further, I could not incorporate uncertainty around basal resource isotope values because the model is algebraic. To overcome some of these limitations, I ran an additional set of mixing models in MixSIAR. The models are three-source (MOB, phytoplankton, and terrestrial vegetation), with three isotopes (H, C, N). Phytoplankton isotope

161 values were lake-specific averages from the open water season, and terrestrial vegetation was represented by a grand mean across all watersheds (same values used for other models, as outlined in Appendix 3A). Isotope values of MOB were acquired from Kelly et al. (2014), with slight modifications. The MOB δ13C value from Kelly et al. (2014) was –60‰ with no error, whereas MOB δ13C was set at –70‰ with an standard deviation of 10, to encompass the error around MOB estimates (Jones & Grey, 2011; Ravinet et al., 2010). For δ2H, I used a value of -

200‰ and for δ15N, I used the same lake specific averages values as of phytoplankton. I used discrimination factors of 0.4 ± 1.3‰ for δ13C (Post, 2002), and 2.98 ± 0.97‰ for δ15N based on the average of published discrimination factors for freshwater organisms (Vanderklift & Ponsard,

2003).

The enrichment of δ2H across trophic levels occurs by progressive incorporation of environmental water, and was corrected for using the model of Solomon et al. (2009):

" !"#" = 1 − (1 − !)

Where !"#" is the proportion of H from environmental water to consumer tissue, t is the trophic position of a consumer above primary producers (assumed trophic position of 2), and ! is the per-trophic-level contribution of dietary water to consumer H. I used a ! value of 0.39, which is the average value for chironomids (Brett et al., 2018). Once the proportion of environmental water in a consumer’s diet (!"#") and trophic position is calculated, a correction factor for environmental water enrichment can be derived (Berggren et al., 2014):

> > > > - ? @ABCDℎF@A) = - ?16:GH2 − (- ?16:GH2 − !"#" × - ?J6"27)/(1 − !"#" )

162

> 2 > Where - ?J6"27 is the δ H of water, and - ?16:GH2 is the sample being corrected for.

Table 3D.2: Estimates of methane oxidising bacteria (MOB), phytoplankton, and terrestrial organic matter (t-OM) resource use by profundal chironomids using a three-source, three-isotope

Bayesian mixing model. The 50th percentile of the posterior distribution represents the median.

Posterior Distribution Percentiles Lake Resource Mean SD 5th 25th 50th 75th 95th MOB 0.247 0.046 0.182 0.217 0.243 0.273 0.329 L164 Phytoplankton 0.334 0.140 0.079 0.249 0.345 0.424 0.553 t-OM 0.419 0.132 0.168 0.211 0.409 0.497 0.659 MOB 0.312 0.071 0.202 0.267 0.308 0.354 0.433 L223 Phytoplankton 0.177 0.134 0.013 0.072 0.151 0.250 0.446 t-OM 0.510 0.140 0.248 0.429 0.529 0.610 0.705 MOB 0.214 0.079 0.110 0.165 0.203 0.249 0.354 L224 Phytoplankton 0.416 0.121 0.199 0.346 0.425 0.498 0.596 t-OM 0.370 0.126 0.187 0.284 0.357 0.442 0.595 MOB 0.075 0.062 0.009 0.035 0.062 0.097 0.117 L239 Phytoplankton 0.784 0.172 0.487 0.739 0.826 0.887 0.950 t-OM 0.141 0.164 0.008 0.040 0.094 0.178 0.432 MOB 0.165 0.077 0.049 0.111 0.157 0.213 0.309 L373 Phytoplankton 0.670 0.057 0.576 0.636 0.673 0.708 0.758 t-OM 0.165 0.074 0.053 0.113 0.161 0.209 0.291 MOB 0.309 0.086 0.192 0.257 0.300 0.348 0.461 L442 Phytoplankton 0.469 0.196 0.068 0.359 0.519 0.610 0.714 t-OM 0.222 0.187 0.024 0.086 0.161 0.296 0.635 MOB 0.191 0.109 0.048 0.118 0.174 0.242 0.398 L626 Phytoplankton 0.692 0.122 0.479 0.618 0.704 0.780 0.873 t-OM 0.117 0.087 0.009 0.046 0.102 0.166 0.285 MOB 0.272 0.090 0.143 0.213 0.260 0.318 0.439 L658 Phytoplankton 0.344 0.207 0.022 0.153 0.372 0.510 0.653 t-OM 0.384 0.199 0.110 0.228 0.339 0.549 0.732

The results of the more robust mixing model approach gave similar estimates of MOB resource use as the simpler approach (Table 3D.2), ranging from 6% to 31% of profundal

163 chironomid biomass, falling within a similar range as previously published estimates for profundal benthic invertebrates MOB contributions (Ravinet et al., 2010; Sanseverino et al.,

2012; Trimmer et al., 2009). In 5 of 8 lakes, phytoplankton contributed more to profundal chironomid diet than terrestrial organic matter. In the two darkest lakes (L164 and L658), terrestrial organic matter contributed slightly more to consumer biomass than that of phytoplankton, and in L223, terrestrial organic matter supported the majority of biomass (52%).

The proportion of MOB in profundal chironomid diets was negatively correlated with volume- weighted hypolimnetic oxygen (Fig. 3D.2; r = –0.81, p = 0.02).

Fig. 3D.2: The proportion of methane oxidising bacteria (MOB) in the diet of profundal chironomids (median of posterior estimate) declines with increasing volume-weighted hypolimnetic oxygen (VWHO) concentration.

164 I did not include MOB as a potential source in the White Sucker mixing models because I did not measure these values directly. As an alternative, I estimated the potential contribution of

MOB to White Sucker under the assumption that these fish only receive MOB via profundal chironomids. In other words, I solved for MOB contributions by looking at how much MOB contribute to profundal chironomid biomass, and how much profundal chironomids contribute to

White Sucker biomass, similar to the method outlined in Ravinet et al. (2010).

Table 3D.3: Estimates of the contribution of MOB to White Sucker Proportion of profundal Proportion of MOB in Proportion of MOB in Lake chironomids in White Profundal chironomid White Sucker diet (%) Sucker diet (%) diet (%) L164 44.6 24.3 10.84 L223 7.6 30.8 2.34 L224 12.9 20.3 2.62 L239 14.4 6.2 0.89 L373 19.9 15.7 3.12 L442 10.3 30 3.1 L626 57.3 16.7 9.57 L658 31.7 26 8.2

Based on this analysis, MOB could potentially contribute approximately 1% to 11% of

White Sucker biomass (Table 3D.3), which falls within a previously established range for benthivorous fish (Ravinet et al., 2010; Sanseverino et al., 2012). In an urban Finnish lake, approximately 17% of Ruffe (Gymnocephalus cernuus) biomass was supported by methanotrophic bacteria, with profundal chironomids being the main conduit of MOB uptake

(Ravinet et al., 2010).

While the depleted isotope δ13C values of profundal chironomids suggest utilization of

MOB, these results are speculative. I do not have any direct estimates of methanotrophic bacteria abundance, methane concentrations, or methane stable isotope values, and without these, many

165 of the assumptions in the mixing models cannot be verified. As illustrated in the first series of mixing models, changing MOB δ13C values by 20‰ changed the reliance on MOB estimates by upwards of 15%. Without knowing how variable methane–δ13C is among lakes, and isotopic fractionation by MOB in my system, I may be greatly over or under-estimating methane contributions to profundal chironomids.

I also cannot speculate as to what energy pathways (allochthonous vs. autochthonous) support methane production, although the literature seems to suggest autochthonous resources are more important as they are energetically favourable for bacterial metabolism (Brett et al.,

2017). Nonetheless, given that MOB contributed < 11% in these models, and MOB likely use a mix of autochthonous and allochthonous carbon, it is unlikely that MOB use by White Sucker strongly skewed my allochthony estimates.

166