LINKING THE EFFECTS OF NITROGEN AND PHOSPHORUS ENRICHMENT TO

CONTROLS OF DETRITAL CARBON LOSS RATES FROM STREAMS

by

DAVID WILLIAM PIERCE MANNING

(Under the Direction of Amy D. Rosemond)

ABSTRACT

Human activities such as agriculture and urbanization result in mobilization of nitrogen

(N) and phosphorus (P) to aquatic ecosystems. Despite increased availability of both N and P,

little is known about the relative importance of N vs. P on detrital carbon (C) loss rates, or the

combined effects of N, P and increased temperature or dissolved organic C (DOC) due to land

use or climate change. Here, we focused on how N and P controls detrital C loss rates mediated

by microbial decomposers and/or detritivores, and the interactive effects of elevated nutrients,

temperature and DOC. To test N vs. P effects on detrital C, five streams were experimentally

enriched with crossed N and P concentration gradients and ratios of N:P. I examined naturally

occurring detritus (leaf litter, wood, and fine particles), and deployed detrital resources (four leaf

species and wood veneers) across seasonal temperature gradients, and determined how increased

N and P altered microbial and detritivore biomass, resource stoichiometry (C:nutrient content),

respiration and breakdown rates. I used nutrient and DOC additions to stream mesocosms to

determine their effects on detrital C loss. Breakdown and respiration rates of coarse detrital

substrates increased with elevated nutrients and temperature; the largest response to nutrients

was for breakdown rates (~2.8× higher with nutrients), followed by respiration (1.5× higher with

nutrients, or seasonal temperature). DOC had negligible effects on respiration or litter

decomposition. Nutrient enrichment increased nutrient content (reduced C:N, C:P) of all detritus

types; nutrient-poor detritus tended to decrease the most, such that detrital stoichiometry

converged during decay. Nutrient effects on detrital C:nutrient stoichiometry were critical

predictors of increased detrital C loss rates, and detritivore biomass. These data suggest that N

and P enrichment predictably increases detrital C loss rates, and that nutrient-altered detrital

stoichiometry is a critical mechanism for predicting the occurrence of increased detrital C loss

from streams. Mitigating excessive nutrient pollution is a key management goal for streams, and

these studies imply that detrital stoichiometry could be used as an integrative measure of nutrient

pollution and its effects on a key ecosystem function that is currently overlooked in nutrient

management policies.

INDEX WORDS: Ecosystem, Coweeta, Ecological stoichiometry, Heterotrophic, Nutrients,

Detritus, Detritivores, Leaf litter, Wood, Fine benthic organic matter,

Dissolved organic carbon, Microbial Respiration, Fungal biomass,

Threshold elemental ratio.

LINKING THE EFFECTS OF NITROGEN AND PHOSPHORUS ENRICHMENT TO

CONTROLS OF DETRITAL CARBON LOSS RATES FROM STREAMS

by

DAVID WILLIAM PIERCE MANNING

B.A. St. Olaf College, 2009

A Dissertation Submitted to the Graduate Faculty of The University of Georgia in Partial

Fulfillment of the Requirements for the Degree

DOCTOR OF PHILOSOPHY

ATHENS, GEORGIA

2015

© 2015

David William Pierce Manning

All Rights Reserved

LINKING THE EFFECTS OF NITROGEN AND PHOSPHORUS ENRICHMENT TO

CONTROLS OF DETRITAL CARBON LOSS RATES FROM STREAMS

by

DAVID WILLIAM PIERCE MANNING

Major Professor: Amy Rosemond Committee: Jonathan Benstead Alan Covich Nina Wurzburger Electronic Version Approved: Suzanne Barbour Dean of the Graduate School The University of Georgia December 2015

iv

DEDICATION

To Mom, Dad, Glenn, Walker, Meg and Jonas, for being my constant reminders to drive

safely, park legally, and watch out for trees.

v

ACKNOWLEDGEMENTS

In this 80th year since Sir Arthur Tansley put forward the idea of the ecosystem, I think it

is important to reflect on some of the many the people who have advanced this idea in ecology.

Specifically, I want to acknowledge the work of E. P. Odum, J. B. Wallace, and P. J.

Mulholland. I am fortunate that these eminent ecologists relied on the ecosystem concept to

frame their work, thus allowing me to use their work to frame my own.

An ecosystem involves complex relationships among organisms, evoking the image of a

web. My mentor Amy Rosemond once reminded me that we all exist within such tangled webs

of individuals who help and support our work along the way. She is very much at the nexus of

the unique web associated with my dissertation, and I am most grateful for her ability to expertly

walk the fine line between being my strongest critic, and most steadfast advocate. Committee

members Alan Covich, Nina Wurzburger and Jon Benstead were also important anchors of this

network. Each contributed much advice and insight throughout my program of research: Alan

Covich drove home the importance of historical and ecological context; Jon Benstead provided

invaluable criticism and commentary on experimental design, analyses and presentation; Nina

Wurzburger helped me learn the importance and meaning of scientific integrity, and the

importance of testing your ideas at the small scale. Vlad Gulis was kind enough to help me

collect all of the microbial respiration and biomas data presented here, and he also provided

much needed insight and expertise related to fungal physiology and ecology. John Maerz

provided advice on experimental design and data analyses. In addition, I am indebted to John

Kominoski; I am grateful that he taught me the ropes of leaf litter breakdown studies, and for all

vi

the memorable trips to and from Coweeta in the old Trailblazer (including changing a couple flat

tires at 1200 m asl).

Beyond these anchors, I had support from a superb crowd of people who helped me

immensely with field and lab work (unfortunately, more than can be mentioned here). Jason

Coombs and Katie Norris were the two pillars of the SNAX3 project that maintained the entire

infrastructure for the experiment and were certainly a huge help to me in the field (Jason in

particular also spent many, many hours grinding leaves and rinsing leaves over nested sieves

with me – both are not easy tasks). Tom Maddox and Emmy Deng were resources for making

sure all the analytical chemistry was done with precision and accuracy. Katie Norris was a

sounding board for the DOC addition experimental design, as was REU student Jenna Martin.

Rosemond lab members past and present were important for advice of all kinds, including Jess

Sterling, Cindy Tant, Jake Allgeier and Amy Trice. Phillip Bumpers, Kait Farrell and James

Wood were certainly among the best office/lab mates anyone could ask for, and endured sharing

an office with me, which usually meant helping me figure out some statistics or R code several

times a day. Chao Song was my go-to stats guru who was always very generous with this time.

Kait, Phillip, and James also frequently assisted me in the field, or helped process leaf bags back

in the lab.

No web of support is complete without those few people who are crucial for their

provision of love and friendship. I am particularly grateful for the friendships that were made

possible by being a part of the Odum School of Ecology and Athens, Georgia communities. I

want to thank Kyle, Alexa, Troy, Gareth, Adrienne, Alyssa, Danny, Nelson and Anna, Bud and

Mary Freeman, Dac Crossley and the Freshwater Mussels, and many, many others for watching

Jonas or our dogs in a pinch, teaching me about old time music, game nights, family dinners,

vii

dance parties, and more. This crucial part of my web also certainly includes my wife, Meghan,

who has been unwavering in her love and support throughout my graduate studies. Our son Jonas

inspires me everyday as he explores the world with new eyes. My parents and brothers continue

to take care of me in large and small ways. For instance, they still make sure that this stream

ecologist’s feet stay dry. I think of them every time I step into a stream wearing the boots they

gave to me when this amazing journey was just beginning. Thank you all.

viii

TABLE OF CONTENTS

Page

ACKNOWLEDGEMENTS ............................................................................................................ v

LIST OF TABLES .......................................................................................................................... x

LIST OF FIGURES ........................................................................................................................ xi

CHAPTER

1 INTRODUCTION AND LITERATURE REVIEW ..................................................... 1

2 DETRITAL STOICHIOMETRY AS A CRITICAL NEXUS FOR THE EFFECTS

OF STREAMWATER NUTRIENTS ON LEAF LITTER BREAKDOWN RATES

............................................................................................................................... 16

3 CONVERGENCE OF DETRITAL STOICHIOMETRY PREDICTS THRESHOLDS

OF NUTRIENT-STIMULATED BREAKDOWN IN STREAMS ............................ 53

4 NUTRIENTS AND TEMPERATURE ADDITIVELY INCREASE STREAM

MICROBIAL RESPIRATION ................................................................................... 93

5 NUTRIENTS ARE MORE IMPORTANT THAN DOC FOR INCREASING LEAF

LITTER DECOMPOSITION DESPITE THEIR COMBINED EFFECTS ON

MICROBIAL BIOMASS AND ACITIVITY ........................................................... 131

6 CONCLUSIONS ....................................................................................................... 165

APPENDICES

A Chapter 2: Additional path model results .................................................................. 177

B Chapter 2: Supplementary shredder biomass results ................................................. 183

ix

C Chapter 3: Linear models for nutrient enrichment effects on detrital stoichiometry 184

D Chapter 4: Streamwater nutrient concentrations and mean seasonal temperatures .. 190

E Chpater 4: Models for respiration on naturally occurring detritus ........................... 193

F Chapter 4: Models for temperature and nutrient effects on deployed detritus .......... 195

G Chapter 4: Comparison of respiration across different temperature ranges ............. 197

H Chapter 5: Stream mesocosm DOC, N, and P concentrations .................................. 200

x

LIST OF TABLES

Page

Table 2.1: Mean ambient and enriched nutrient concentrations during each litter breakdown

experiment for the five treatment reaches used in this study ............................................ 44

Table 2.2: Mean breakdown rates reported as decay coefficients of the negative exponential

model. ............................................................................................................................... 46

Table 2.3: Mean litter C:P and C:N ratios on d 70 and standard error for Acer rubrum (maple)

and Rhododendron maximum (rhododendron) .................................................................. 48

Table 3.1: Parameter estimates based on linear models for drivers of leaf litter and wood

stoichiometry. .................................................................................................................... 84

Table 4.1: Parameter estimates and 95% confidence intervals (95% CI) from the linear model for

respiration rates on naturally occurring detritus .............................................................. 122

Table 4.2: Parameter estimates and 95% confidence intervals (95% CI) for linear models relating

leaf litter and wood veneer respiration rates to temperature and nutrient enrichment .... 124

Table 5.1: Stream channel experimental design, and mean water chemistry during the stream

mesocosm experiment ..................................................................................................... 161

xi

LIST OF FIGURES

Page

Figure 2.1: Hypothesized path model describing how nutrients affect litter breakdown rates. .... 51

Figure 2.2: The best supported path models for PRE, YR1 and YR2 relating N (a) or P (b)

concentrations to drivers of litter breakdown rates. .......................................................... 52

Figure 3.1: A conceptual representation of how nutrients could affect microbially mediated

conditioning and detrital stoichiometry and the quality of the resource for shredders. .... 88

Figure 3.2: Leaf litter and wood veneer C:N and C:P under pretreatment and during YR1 and

YR2 of enrichment during early, middle and late stages of decay. .................................. 89

Figure 3.3: Breakdown rates of leaf litter and wood increased as a function of initial C:N and

C:P. ............................................................................................................................... 90

Figure 3.4: The contribution of shredder mediated breakdown in each year of the study for all

leaf litter types. .................................................................................................................. 91

Figure 3.5: Total breakdown rates as a function of middle-stage C:N or C:P ratios of all leaf litter

species used in this study .................................................................................................. 92

Figure 4.1: Temperature dependence of substrate-specific respiration rates for FBOM, leaf litter

and wood ......................................................................................................................... 127

Figure 4.2: Temperature dependence of respiration rates associated with leaf litter and wood

veneers deployed in our study streams for known periods of time ................................. 128

Figure 4.3: Fungal biomass and temperature for naturally occurring detritus and deployed

detritus ............................................................................................................................. 129

xii

Figure 4.4: Activation energy as a function of initial C:N and C:P for each of the detritus types

deployed for our study ..................................................................................................... 130

Figure 5.1: Mean litter decomposition rates for maple and rhododendron leaf litter in each DOC

treatment ......................................................................................................................... 161

Figure 5.2: Mean respiration rates for maple and rhododendron leaf litter or both leaf litter

species in each DOC treatment ....................................................................................... 162

Figure 5.3: Mean fungal biomass measured on day 35 of the experiment for maple and

rhododendron leaves by DOC treatment. ........................................................................ 163

Figure 5.4: Mean decomposition rates normalized by either fungal biomass or respiration rates

and respiration rates normalized by fungal biomass for each DOC treatment and N+P

additions. ......................................................................................................................... 164

1

CHAPTER 1

INTRODUCTION AND LITERATURE REVIEW

Nitrogen and phosphorus enrichment of aquatic ecosystems

Increasing concentrations of biologically important nutrients, such as nitrogen (N) and

phosphorus (P), pervasively impact aquatic ecosystems worldwide (Smith and Schindler 2009).

Elevated concentrations of N and P in aquatic ecosystems can be linked to human activities in

watersheds that produce increased N and P loads, such as fossil fuel combustion, application of

fertilizers to crops, and discharge of sewage (Brown and Froemke 2012). Nitrogen and P loading

is associated with increased primary production and autotrophic carbon (C) in some systems (i.e.,

algal biomass, Smith 1978, Peterson et al. 1993, Slavik et al. 2004). In contrast, nutrient

enrichment of heterotrophic, detritus-based ecosystems, such as most forested, headwater

streams, is expected to decrease, rather than increase, the availability of basal C resources such

as terrestrially derived organic matter (i.e., leaf litter and wood, hereafter: detritus). This different

effect of nutrient enrichment in detritus-based systems is likely due to increases in microbial

activity and biomass (Suberkropp et al. 2010) and increases in detrital quality for consumers

(Rosemond et al. 2010, Scott et al. 2013, Prater et al. 2015), which lead to reduced C storage and

altered C fluxes through detrital food webs (Benstead et al. 2009).

Much of what remains unresolved regarding nutrient enrichment effects on aquatic

ecosystems is related to the relative importance of N vs. P effects, especially in the case of

detrital C processing. Despite considerable and continuing debate (Schindler et al. 2008, Sterner

2008), mounting evidence suggests that N and P co-limit autotrophic C gains, pointing to

2

coupled effects of both nutrients on autotrophic C responses (Harpole et al. 2011). The same may

be true in detritus-based systems (Ferreira et al. 2015); however, the importance of N vs. P

effects in detritus-based systems remains poorly understood. This knowledge gap is crucial to

bridge because N and P inputs to detritus-based streams are often highly skewed and dependent

on dominant land-use practices, resulting in scenarios where N availability is much higher than P

and vice versa. For example, watersheds containing predominantly row-crop agriculture can

have high N loads relative to P due to application of N-rich fertilizers, while watersheds with

greater prevalence of urban land cover or livestock agriculture can have high P loads relative to

N due to inputs of P-rich animal waste or sewage (e.g., Downing and McCauley 1992, Arbuckle

and Downing 2001, Peñuelas et al. 2012). As a result, N and P availability in streams and rivers

likely diverge from the strict N and P requirements for organismal growth and functioning

(Sterner and Elser 2002), in contrast with matched nutrient availability and demand in oceans

driven by phytoplankton-mediated uptake and remobilization of N and P at a fixed ratio of 16:1

(Redfield 1958). Microbial decomposers associated with detritus in streams and rivers also

require specific amounts of N and P (fungal biomass N:P estimates 7:1-9:1, Grimmett et al.

2013, V. Gulis unpublished data), suggesting that ecosystem processes dependent on their

growth and activity (such as detrital C processing) could be affected by the relative supply of N

vs. P in streamwater. Studies that quantify the effects of divergent availability and demand for N

and P on the controls of detrital C processing and the fate of C in aquatic ecosystems are

currently lacking. Thus, the series of studies presented here were part of a multi-year,

experimental N and P enrichment of five headwater streams; the objective of the project was to

determine the relative effects of N vs. P on detritus-based ecosystems. Within the context of this

3

larger collaborative project, I specifically characterized the dual effects of N and P on the

controls of detrital processing in detritus-based streams.

Controls of detrital C loss rates

Most terrestrial primary production is not consumed by herbivores and enters detrital

pathways after senescence (Cebrian and Lartigue 2004), contributing to the structure and

function of receiving ecosystems. Such detrital C (e.g., leaf litter and wood) from terrestrial

sources serves several important roles in streams and rivers, including habitat, and fuel for food

webs (Wallace et al. 1997), and its availability coincides with seasonal pulses of leaf litter, and

stochastic inputs of wood (Webster et al. 1999). Once detritus enters a stream, it can either be

broken down, or transported downstream (Webster et al. 1999). The fate of detrital C in streams

and rivers is controlled by interacting abiotic and biotic factors, including physical abrasion,

streamwater nutrient concentrations, intrinsic detrital attributes (i.e., chemical recalcitrance,

physical toughness, and/or nutrient content) and subsequent microbial decomposer and

detritivore activity (Tank et al. 2010). Aquatic fungi are the predominant decomposers of coarse

detritus (Gessner and Chauvet 1994, Hieber and Gessner 2002, Tant et al. 2013), and their

colonization and maceration of detritus conditions these resources for subsequent consumption

by shredding detritivores (hereafter: shredders; Cummins et al. 1973). Shredders tend to

preferentially consume detritus colonized by fungi, emphasizing the importance of fungal

colonization of detritus for breakdown (Arsuffi and Suberkropp 1985). Importantly, intrinsic

characteristics of detritus, such as its nutrient content relative to C (i.e., C:nutrient

stoichiometry), can be strong predictors of fungal colonization and detrital breakdown rates,

4

where detritus with high C:N or C:P content generally has slower breakdown rates (e.g.,

Enriquez et al. 1993, Hladyz et al. 2009).

Nutrient enrichment effects on detrital C loss rates

Several studies have illustrated that nutrient enrichment increases rates of detrital C loss

rates (e.g., Elwood et al. 1981, Pearson and Connolly 2000, Greenwood et al. 2007, Ferreira et

al. 2015), likely due to the enhancement of biological processing. For instance, both N and P

enrichment have been shown to increase the activity and biomass of microbial decomposers,

particularly aquatic fungi (Suberkropp and Chauvet 1995, Rosemond et al. 2002, Ferreira et al.

2006a, Tant et al. 2015). This increased microbial activity can be related to increased detrital C

loss rates (Kominoski et al. 2015), but fungi may also contribute to increased breakdown rates by

incorporating streamwater nutrients more than detrital nutrients into their biomass, hence altering

bulk detrital nutrient content (Suberkropp 1995, Cheever et al. 2013). Similar to N and P effects

on microbial biomass and activity, both N and P can be immobilized on leaf litter when either is

available (Rosemond et al. 2010, Scott et al. 2013), which could affect shredder consumption of

detritus via reduced elemental imbalances between shredders and the resources they consume

(i.e., detrital C:nutrient content approaches shredder nutrient requirements, Cross et al. 2003,

Halvorson et al. 2015). Therefore, nutrient enrichment likely affects detrital C loss rates largely

by increased microbial activity, detrital C:nutrient content, and shredder activity, but the relative

importance of N vs. P on these controls of detrital C loss are unknown.

5

Nutrient-stimulated biological effects on detrital C loss: interactions with physical/chemical

predictors

There are important links between the effects of nutrient enrichment on both microbial-

and shredder-mediated increases to detrital C loss rates; however, few studies have fully tested

how microbial decomposers vs. shredders may affect breakdown rates in response to

experimental nutrient additions. In general, detrital C loss driven by microbial decomposers is

partly a function of microbial acquisition and conversion of detrital C to CO2 via respiration

(Gessner et al. 2007). On the other hand, shredders convert coarse detritus to fine particles (e.g.,

fine benthic organic matter [FBOM]), contributing to this C pool in streams (Wallace et al.

1982). Evidence from landscape-scale gradients of nutrient enrichment suggests that the presence

of intact shredder communities is important for increased breakdown rates to occur (Woodward

et al. 2012). Therefore, we sought to elucidate the importance of N and P for driving microbial

vs. shredder effects on leaf litter breakdown by excluding macroinvertebrate access to leaf litter

(Chapter 3), and by examining processes driven predominantly by microbial decomposers (i.e.,

microbial breakdown rates, respiration rates; Chapters 3, 4 and 5).

In addition to nutrient affects, we quantified controls on C processing including stream

discharge (Chapter 2), temperature (Chapters 2, 4) and availability of dissolved organic carbon

(DOC; Chapter 5). Stream discharge likely corresponds to the amount of physical abrasion and

fragmentation of detritus that occurs, which may interact with nutrient effects (e.g., Ferreira et al.

2006b). Further, both temperature and DOC are expected to increase in aquatic ecosystems due

to land use change (DOC, temperature; Kaushal et al. 2010, Stanley et al. 2012) or climate

change (temperature, Kaushal et al. 2010), and may occur simultaneously with elevated nutrient

6

availability. Therefore, it is important to consider how the combined effects of nutrients and

temperature, or nutrients and DOC may affect detrital C processing rates.

Experimental design and dissertation overview

To assess the importance of N vs. P for driving detrital C loss rates, we enriched five

headwater streams that were similar physically and chemically with both N and P for two years

at Coweeta Hydrologic Laboratory, Macon Co., North Carolina. Gradients of both N and P

concentrations were used, which corresponded to a gradient of N:P ratios. Targeted

concentrations of dissolved inorganic N (DIN) increased from ~ 80-650 µg/L and corresponded

to decreasing concentrations of soluble reactive P (SRP ~ 90-11 µg/L), to achieve target molar

N:P treatments of 2, 8, 16, 32, and 128 for the five streams. The nutrient additions reflected low-

to-moderate concentrations of N and P that are commonly observed in the Blue Ridge region of

the Southern Appalachians (Scott et al. 2002), and elsewhere in the U.S. (Alexander and Smith

2006) and Europe (Woodward et al. 2012). Several factors related to detrital C processing were

quantified, including these parameters: microbial respiration rates (for both seasonally collected

and deployed detritus), fungal biomass, detrital C:nutrient content (stoichiometry), shredder

biomass, and breakdown rates of five distinct detrital substrates. Data collection occurred prior to

nutrient enrichment in all five streams, and during the two years of enrichment. Details about the

experimental infrastructure and treatments are described in detail in Rosemond et al. (2015), as

well as Chapters 2, 3, and 4.

We used a fully factorial DOC × nutrient experiment in stream mesocosms to determine

the combined effects of elevated DOC and nutrients on the decomposition of maple and

rhododendron leaf litter. We elevated DOC and modified the lability of DOC available by adding

7

either dextrose or leaf leachate to the mesocosms alone, or combined with N and P additions.

Further details about the experimental infrastructure and design of this study can be found in

Chapter 5.

Chapter 2: Detrital stoichiometry as a critical nexus for the effects of streamwater nutrients on

leaf litter breakdown rates

The objectives of this study were to assess the effects of N vs. P on detrital breakdown

rates, and to characterize the importance of the key drivers of breakdown rates. Specifically, we

constructed a conceptual model with hypothesized N and P effects on leaf litter breakdown rates

related to the specific abiotic/biotic drivers of litter breakdown (i.e., temperature, discharge,

streamwater nutrients, detrital stoichiometry, shredder and fungal biomass). We tested this

conceptual model using path analysis, in order to discern the relative importance of N vs. P on

drivers of litter breakdown for both pretreatment (ambient), and nutrient-enriched conditions.

This test of our conceptual model for N and P effects on detrital breakdown provides a strong

framework to guide further tests of the underlying mechanisms that drive increased breakdown

rates due to nutrients in streams.

Chapter 3: Convergence of detrital stoichiometry predicts thresholds of nutrient-stimulated

breakdown in streams

This study builds on the observed positive effects of nutrient enrichment on detrital

nutrient content (Chapter 2), and explicitly links these effects to breakdown rates of several

detrital resources. The objective in this case was to use the observed convergence of initially

distinct detrital resources (litter from four tree species, and wood veneers with different initial

8

C:N and C:P content) to predict increases in detrital breakdown rates, and to isolate the effects of

N and P on breakdown rates with and without shredders. Specifically, we related the

experimental nutrient enrichment effects to convergent detrital resource quality (in terms of

C:nutrient content) using breakpoint regression, due to hypothesized threshold effects of detrital

stoichiometry that approached values matching shredder nutrient requirements (Halvorson et al.

2015). This analysis provides further insight into the importance of N vs. P on detrital

stoichiometry, and its effects on microbial vs. shredder-mediated detrital breakdown.

Chapter 4: Nutrients and temperature additively increase stream carbon respiration

Beyond characterizing the mechanisms that drive increased breakdown rates, Chapter 4

examines the importance of nutrients vs. temperature as drivers of microbial respiration rates

associated with detritus. Increased microbial respiration rates are linked to both increased

temperature and nutrients, but our ability to predict how both combine to affect respiration rates

is limited. Thus, we measured respiration rates on naturally occurring detritus (leaf litter, wood,

and fine benthic organic matter [FBOM]), and five deployed detrital substrates (four leaf litter

species and wood veneers) across our gradient of N and P concentrations, and seasonal gradients

of temperature. We related these respiration rates to temperature using metabolic theory, and

assessed whether the temperature dependence of respiration rates was modified by nutrient

enrichment. We predicted that nutrients would increase the temperature dependence of

respiration rates due to disproportionate effects of nutrients on respiration rates occurring at

warmer temperatures. Microbial respiration rates may be helpful in predicting breakdown rates

of coarse detritus at larger (i.e., patch or reach) scales, therefore it is important to understand how

both increasing temperature and nutrients can affect respiration rates.

9

Chapter 5: Nutrients are more important than DOC for increasing leaf litter decomposition

despite their combined effects on microbial biomass and activity

Finally, in Chapter 5, we explored the combined effects of dissolved organic carbon

(DOC) and nutrient enrichment on detrital C loss rates. Increasing DOC concentrations and

altered DOC quality (greater prevalence of labile vs. recalcitrant DOC) have been observed in

human-modified watersheds (Giling et al. 2014, Lu et al. 2014), in addition to elevated N and P

availability. Increased detrital C loss rates are generally attributed to alleviation of nutrient

limitation, but increased DOC may have effects on breakdown rates as well via complex

‘priming’ effects that could either increase detrital C processing (Guenet et al. 2010). We

predicted that added labile DOC would ‘prime’ decomposition of detrital C (i.e., faster

processing rates when DOC was added), and that nutrients would add to this effect by further

stimulating microbial activity, and priming. This study was a step towards understanding the

combined effects of increased nutrient and DOC concentrations through microbial pathways, and

provides an important basis for potential future work on aquatic priming effects and their

interaction with nutrient availability.

Literature Cited

Alexander, R. B., and R. A. Smith. 2006. Trends in the nutrient enrichment of U. S. rivers during

the late 20th century and their relation to changes in probable stream trophic conditions.

Limnology and Oceanography 51:639-654.

Arbuckle, K. E., and J.A. Downing. 2001. The influence of watershed land use on lake N:P in a

predominantly agricultural landscape. Limnology and Oceanography 46:970-975.

10

Arsuffi, T. L., and K. Suberkropp. 1985. Selective feeding by stream caddisfly (Trichoptera)

detritivores on leaves with fungal-colonized patches. Oikos 45:50-58.

Benstead, J. P., A. D. Rosemond, W. F. Cross, J. B. Wallace, S. L. Eggert, K. Suberkropp, V.

Gulis, J. L. Greenwood, and C. J. Tant. 2009. Nutrient enrichment alters storage and

fluxes of detritus in a headwater stream ecosystem. Ecology 90:2556-2566.

Brown, T. C., and P. Froemke. 2012. Nationwide Assessment of Nonpoint Source Threats to

Water Quality. BioScience 62:136-146.

Cheever, B. M., J. R. Webster, E. E. Bilger, and S. A. Thomas. 2013. The relative importance of

exogenous and substrate-derived nitrogen for microbial growth during leaf

decomposition. Ecology 94:1614-1625.

Cross, W. F., J. P. Benstead, A. D. Rosemond, and J. B. Wallace. 2003. Consumer-resource

stoichiometry in detritus-based streams. Ecology Letters 6:721-732.

Cummins, K. W., R. C. Petersen, F. O. Howard, J. C. Wuycheck, and V. I. Holt. 1973. The

utilization of leaf litter by stream detritivores. Ecology 54:336-345.

Downing, J. A., and E. McCauley. 1992. The nitrogen: phosphorus relationship in lakes.

Limnology and Oceanography 37:936-945.

Elwood, J. W., J. D. Newbold, A. F. Trimble, and R. W. Stark. 1981. The limiting role of

phosphorus in a woodland stream ecosystem: effects of P enrichment on leaf

decomposition and primary producers. Ecology 62:146-158.

Enriquez, S., C. M. Duarte, and K. Sand-Jensen. 1993. Patterns in decomposition rates among

photosynthetic organisms: the importance of detritus C: N: P content. Oecologia 94:457-

471.

11

Ferreira, V., V. Gulis, and M. A. S. Graça. 2006a. Whole-stream nitrate addition affects litter

decomposition and associated fungi but not invertebrates. Oecologia 149:718-729.

Ferreira, V., M. A. S. Graça, J. L. M. P. de Lima. 2006b. Role of physical fragmentation and

invertebrate activity in the breakdown rate of leaves. Archiv fur Hydrobiologie 165:493-

513.

Ferreira V., B. Castagneyrol, J. Koricheva, V. Gulis, E. Chauvet, and M. A. S. Graça. 2015. A

meta-analysis of the effects of nutrient enrichment on litter decomposition in streams.

Biological Reviews doi:10.111/brv.12125.

Gessner, M. O., and E. Chauvet. 1994. Importance of stream microfungi in controlling

breakdown rates of leaf litter. Ecology 75:1807-1817.

Gessner, M. O., V. Gulis, K. A. Kuehn, E. Chauvet, and K. Suberkropp. 2007. Fungal

decomposers of plant litter in aquatic ecosystems. Pages 301-324 in C. P. Kubicek, and I.

S. Druzhinina, editors. The Mycota. Springer. Berlin.

Giling, D. P., M. R. Grace, J. R. Thomson, R. Mac Nally, and R. M. Thompson. 2014. Effect of

native vegetation loss on stream ecosystem processes: dissolved organic matter

composition and export in agricultural landscapes. Ecosystems 17:82-95.

Greenwood, J. L., A. D. Rosemond, J. B. Wallace, W. F. Cross, and H. S. Weyers. 2007.

Nutrients stimulate leaf breakdown rates and detritivore biomass: bottom-up effects via

heterotrophic pathways. Oecologia 151:637-649.

Grimmett, I. J., K. N. Shipp, A. Macneil, and F. Bärlocher. 2013. Does the growth rate

hypothesis apply to aquatic hyphomycetes? Fungal Ecology 6:493-500.

Guenet, B., M. Danger, L. Abbadie, and G. Lacroix. 2010. Priming effect: bridging the gap

between terrestrial and aquatic ecology. Ecology 91:2850-2861.

12

Halvorson, H. M., J. T. Scott, A. J. Sanders, and M. A. Evans-White. 2015. A stream insect

detritivore violates common assumptions of threshold elemental ratio bioenergetics

models. Freshwater Science 34:508-515.

Harpole, W. S., J. T. Ngai, E. E. Cleland, E. W. Seabloom, E. T. Borer, M. E. S. Bracken, J. J.

Elser, D. S. Gruner, H. Hillebrand, J. B. Shurin, and J. E. Smith. 2011. Nutrient co-

limitation of primary producer communities: community co-limitation. Ecology Letters

14:852-862.

Hieber, M., and M. O. Gessner. 2002. Contribution of stream detritivores, fungi, and bacteria to

leaf breakdown based on biomass estimates. Ecology 83:1026-1038.

Hladyz, S., M. O. Gessner, P. S. Giller, J. Pozo, and G. Woodward. 2009. Resource quality and

stoichiometric constraints on stream ecosystem functioning. Freshwater Biology 54:957-

970.

Kaushal, S. S., G. E. Likens, N. A. Jaworski, M. L. Pace, A. M. Sides, D. Seekell, K. T. Belt, D.

H. Secor, and R. L. Wingate. 2010. Rising stream and river temperatures in the United

States. Frontiers in Ecology and the Environment 8:461-466.

Kominoski, J. S., A. D. Rosemond, J. P. Benstead, V. Gulis, J. C. Maerz, and D. W. P. Manning.

2015. Low-to-moderate nitrogen and phosphorus concentrations accelerate microbially

driven litter breakdown rates. Ecological Applications 25:856-865.

Lu, Y. H., J. E. Bauer, E. A. Canuel, R. M. Chambers, Y. Yamashita, R. Jaffé, and A. Barrett.

2014. Effects of land use on sources and ages of inorganic and organic carbon in

temperate headwater streams. Biogeochemistry 119:275-292.

Pearson, R. G., and N. M. Connolly. 2000. Nutrient enhancement, food quality and community

dynamics in a tropical rainforest stream. Freshwater Biology 43:31-42.

13

Peñuelas, J., J. Sardans, A. Rivas-Ubach, and I. A. Janssens. 2012. The human-induced

imbalance between C, N and P in Earth’s life system. Global Change Biology 18:3-6.

Peterson, B. J., L. Deegan, J. Helfrich, J. E. Hobbie, M. Hullar, B. Moller, T. E. Ford, A.

Hershey, A. Hiltner, G. Kipphut, M. A. Lock, D. M. Fiebig, V. McKinley, M. C. Miller,

J. R. Vestal, R. Ventullo, and G. Volk. 1993. Biological responses of a tundra river to

fertilization. Ecology 74:653-672.

Prater, C., E. J. Norman, and M. A. Evans-White. 2015. Relationships among nutrient

enrichment, detritus quality and quantity, and large-bodied shredding insect community

structure. Hydrobiologia 753:219-232.

Redfield, A. C. 1958. The biological control of chemical factors in the environment. American

Scientist 46:205-221.

Rosemond, A. D., C. M. Pringle, A. Ramírez, M. J. Paul, and J. L. Meyer. 2002. Landscape

variation in phosphorus concentration and effects on detritus-based tropical streams.

Limnology and Oceanography 47:278-289.

Rosemond, A. D., C. M. Swan, J. S. Kominoski, and S. E. Dye. 2010. Non-additive effects of

litter mixing are suppressed in a nutrient-enriched stream. Oikos 119:326-336.

Rosemond, A. D., J. P. Benstead, P. M. Bumpers, V. Gulis, J. S. Kominoski, D. W. P. Manning,

K. Suberkropp, and J. B. Wallace. 2015. Experimental nutrient additions accelerate

terrestrial carbon loss from stream ecosystems. Science 347:1142-1145.

Schindler, D. W., R. E. Hecky, D. L. Findlay, M. P. Stainton, B. R. Parker, M. J. Paterson, K. G.

Beaty, M. Lyng, and S. E. M. Kasian. 2008. Eutrophication of lakes cannot be controlled

by reducing nitrogen input: results of a 37-year whole-ecosystem experiment.

Proceedings of the National Academy of Sciences 105:11254-11258.

14

Scott, M. C., G. S. Helfman, M. E. McTammany, E. F. Benfield, and P. V. Bolstad. 2002.

Multiscale influences on physical and chemical stream conditions across Blue Ridge

landscapes. Journal of the American Water Resources Association 38:1379-1392.

Scott, E. E., C. Prater, E. Norman, B. C. Baker, M. Evans-White, and J. T. Scott. 2013. Leaf-

litter stoichiometry is affected by streamwater phosphorus concentrations and litter type.

Freshwater Science 32:753-761.

Slavik, K., B. J. Peterson, L. A. Deegan, W. B. Bowden, A. E. Hershey, and J. E. Hobbie. 2004.

Long-term responses of the Kuparuk River ecosystem to phosphorus fertilization.

Ecology 85:939-954.

Smith, V. H. 1978. Nutrient dependence of primary productivity in lakes. Limnology and

Oceanography 24:1051-1064.

Smith, V. H., and D. W. Schindler. 2009. Eutrophication science: where do we go from here?

Trends in Ecology & Evolution 24:201-207.

Stanley, E. H., S. M. Powers, N. R. Lottig, I. Buffam, and J. T. Crawford. 2012. Contemporary

changes in dissolved organic carbon (DOC) in human-dominated rivers: is there a role for

DOC management? Freshwater Biology 57:26-42.

Sterner, R.W., and Elser, J. J. 2002. Ecological Stoichiometry. Princeton University Press,

Princeton, New Jersey, USA.

Sterner, R. W. 2008. On the Phosphorus Limitation Paradigm for Lakes. International Review of

Hydrobiology 93:433-445.

Suberkropp, K. 1995. The influence of nutrients on fungal growth, productivity, and sporulation

during leaf breakdown in streams. Canadian Journal of Botany 73:S1361-S1369.

15

Suberkropp, K., and E. Chauvet. 1995. Regulation of leaf breakdown by fungi in streams:

influences of water chemistry. Ecology 76:1433-1445.

Suberkropp, K., V. Gulis, A. D. Rosemond, and J. P. Benstead. 2010. Ecosystem and

physiological scales of microbial responses to nutrients in a detritus-based stream: results

of a 5-year continuous enrichment. Limnology and Oceanography 55:149-160.

Tank, J. L., E. J. Rosi-Marshall, N. A. Griffiths, S. A. Entrekin, and M. L. Stephen. 2010. A

review of allochthonous organic matter dynamics and metabolism in streams. Journal of

the North American Benthological Society 29:118-146.

Tant, C. J., A. D. Rosemond, and M. R. First. 2013. Stream nutrient enrichment has a greater

effect on coarse than on fine benthic organic matter. Freshwater Science 32:1111-1121.

Tant, C. J., A. D. Rosemond, A. S. Mehring, K. A. Kuehn, and J. M. Davis. 2015. The role of

aquatic fungi in transformations of organic matter mediated by nutrients. Freshwater

Biology 60:1354-1363.

Wallace, J. B., J. R. Webster, and T. F. Cuffney. 1982. Stream detritus dynamics: regulation by

invertebrate consumers. Oecologia 53:197-200.

Wallace, J. B. 1997. Multiple Trophic Levels of a Forest Stream Linked to Terrestrial Litter

Inputs. Science 277:102-104.

Webster, J. R., E. F. Benfield, T. P. Ehrman, M. A. Schaeffer, J. L. Tank, J. J. Hutchens, and D.

J. D’Angelo. 1999. What happens to allochthonous material that falls into streams? A

synthesis of new and published information from Coweeta. Freshwater Biology 41:687-

705.

Woodward, G., et al. 2012. Continental-scale effects of nutrient pollution on stream ecosystem

functioning. Science 336:1438-1440.

16

CHAPTER 2

DETRITAL STOICHIOMETRY AS A CRITICAL NEXUS FOR THE EFFECTS OF

STREAMWATER NUTRIENTS ON LEAF LITTER BREAKDOWN RATES1

1David W. P. Manning, Amy D. Rosemond, John S. Kominoski, Vladislav Gulis, Jonathan P. Benstead, and John C. Maerz. 2015. Ecology 96:2214-2224. Reprinted here with permission of the publisher.

17

Abstract. Nitrogen (N) and phosphorus (P) concentrations are elevated in many freshwater

systems, stimulating breakdown rates of terrestrially-derived plant litter; however, the relative

importance of N and P in driving litter breakdown via microbial and detritivore processing are

not fully understood. Here, we determined breakdown rates of two litter species, Acer rubrum

(maple) and Rhododendron maximum (rhododendron), before (PRE) and during two years (YR1,

YR2) of experimental N and P additions to five streams, and quantified the relative importance

of hypothesized factors contributing to breakdown. Treatment streams received a gradient of P

additions (low to high soluble reactive phosphorus [SRP]; ca. 10-85 µg L-1) crossed with a

gradient of N additions (high to low dissolved inorganic nitrogen [DIN]; ca. 472-96 µg L-1) to

achieve target molar N:P ratios ranging from 128 to 2. Litter breakdown rates increased above

pre-treatment levels by an average of 1.1-2.2× for maple, and 2.7-4.9× for rhododendron in YR1

and YR2. We used path analysis to compare fungal biomass, shredder biomass, litter

stoichiometry (nutrient content as C:N or C:P), discharge, and streamwater temperature as

predictors of breakdown rates and compared models containing streamwater N, P or N+P and

litter C:N or C:P using model selection criteria. Litter breakdown rates were predicted equally

with either streamwater N or P (R2 = 0.57). In models with N or P, fungal biomass, litter

stoichiometry, discharge, and shredder biomass predicted breakdown rates; litter stoichiometry

and fungal biomass were most important for model fit. However, N and P effects may have

occurred via subtly different pathways. Litter N content increased with fungal biomass (N-driven

effects) and litter P content increased with streamwater P availability (P-driven effects),

presumably via P storage in fungal biomass. In either case, the effects of N and P through these

pathways were associated with higher shredder biomass and breakdown rates. Our results

18

suggest that N and P stimulate litter breakdown rates via mechanisms in which litter

stoichiometry is an important nexus for associated microbial and detritivore effects.

Introduction

Understanding biogeochemical cycles and the impacts of human activity on ecosystem

dynamics requires the consideration of interactions among multiple elements (Schlesinger et al.

2011). Nitrogen (N) and phosphorus (P) both limit autotrophic production (Elser et al. 2007), but

less is known about the relative importance of N and P for heterotrophic processes such as

breakdown of detrital organic matter (but see Woodward et al. 2012). Increased anthropogenic

mobilization of N and P often occurs in disproportionate amounts, driving the relative

availability of N or P in recipient ecosystems (e.g., atmospheric N deposition vs. P-rich livestock

waste; Arbuckle and Downing 2001). Thus, there is a need to understand the specific effects of N

and P on fundamental ecosystem processes such as detrital organic matter breakdown.

Processing of detrital carbon (C) in aquatic ecosystems is a function of interacting abiotic

and biotic factors, including temperature, physical abrasion, litter stoichiometry, microbial

conditioning, and detritivore biomass (Hieber and Gessner 2002, Hladyz et al. 2009). Under low-

nutrient conditions, litter species identity can be used to predict litter breakdown rates, as initial

nutrient content and other physical and chemical traits can affect microbial colonization and

consumption by detritivores (Petersen and Cummins 1974). Nutrient enrichment has been shown

to increase nutrient content of decomposing litter, thereby reducing the natural variation in litter

stoichiometry (i.e. C:nutrient ratios; C:N, C:P) between litter species (Rosemond et al. 2010).

This effect may relax consumer-resource constraints on detritivore growth and consumption of

litter to affect breakdown rates (Cross et al. 2003, Tant et al. 2013). Aquatic fungi play a

19

generally larger role in breakdown of coarse particulate organic matter such as leaf litter than

bacteria (Findlay et al. 2002, Tant et al. 2013). Fungi can affect nutrient content of conditioned

litter by incorporating both streamwater- and litter-derived nutrients into their biomass

(Suberkropp and Chauvet 1995, Cheever et al. 2013). The relative influence of streamwater N

and P on microbial and detritivore mediated processes and on the links among fungi, litter

stoichiometry, and detritivores that drive breakdown of C remains poorly understood.

Streamwater N and P may similarly affect the links between fungal and shredder

pathways of detrital C loss, and therefore either nutrient may limit the rate of litter breakdown

(e.g., Ferreira et al. 2014). For example, P may be critically important for the growth and

biomass accrual of fungi on litter, given that P-rich RNA is needed for rapid metabolism (i.e. the

growth rate hypothesis; Sterner and Elser 2002, Grimmett et al. 2013). Alternatively, N has been

linked to increased fungal biomass (Ferreira et al. 2006), and may be important for fungi to

produce N-rich enzymes to acquire C from polymers (Sinsabaugh et al. 2009). Increases in

fungal biomass may alter litter stoichiometry (both C:N and C:P) via immobilization of dissolved

nutrients to substantially increase litter nutrient content. This altered stoichiometry may affect

shredder consumption and litter breakdown rates (Cheever et al. 2013, Scott et al. 2013).

Shredder growth and consumption rates have been associated with litter N (Rosemond et al.

2010) or P content of detritus (Danger et al. 2013). Thus, increased litter breakdown rates may

occur when N or P are elevated alone or together through similar stimulatory effects on fungal

biomass and activity, increased litter nutrient content, and ensuing shredder activity (Ferreira et

al. 2014); but the relative contributions of N or P to these processes are unknown.

This study used crossed N and P streamwater concentration gradients in five headwater

streams to test the effects of N and P on litter breakdown rates and identify the mechanisms by

20

which they occurred. We used path analysis to test hypothesized causal links among streamwater

N and P concentrations, conditioned litter stoichiometry (C:N, C:P), fungal biomass, shredder

biomass, discharge, temperature, and litter breakdown rates; breakdown of litter is hypothesized

to occur through microbial processing (e.g., litter mass loss due to respiration, biomass

production, and spore production in the case of fungi) and shredder feeding (Fig. 1). Our

experimental design precluded testing for the isolated effects of N and P, but allowed us to

examine the relative strength of their effects on these pathways. We predicted that the effects of

dissolved N and P on stream detrital food webs would propagate through microbial pathways,

whereby increases in fungal biomass increase litter nutrient content and enhance shredder

biomass (Fig. 1). We tested path models with N, P, and N + P to assess the relative strength of

their singular or combined effects on litter breakdown. We also tested path models using

stoichiometry of conditioned litter, as either C:N or C:P, to evaluate the importance of litter N or

P content for explaining litter breakdown rates.

Methods

Study site and experimental nutrient additions

This study was conducted at the Coweeta Hydrologic Laboratory (CWT), a United States

Forest Service research station located in Macon County, North Carolina, USA. Coweeta is a

heavily forested 2185-ha basin with mixed hardwoods (maple, poplar and oak) that are common

in the Blue Ridge physiographic region of the southern Appalachian Mountains (Swank and

Crossley 1988). The basin contains a network of low-order streams that are heavily shaded year-

round by Rhododendron maximum. Seventy-meter reaches in five first-order streams within the

Dryman Fork catchment were identified for the nutrient manipulations used in this study (35°02ʹ′

21

N, 83°45ʹ′ W). These five streams were physically and chemically similar in terms of elevation

(ca. 1200 m a.s.l.), aspect (four out of five on E facing slopes, one NE), gradient, pH, and

temperature and were in close proximity (<0.5 km apart). Experimental additions of aqueous

21% ammonium nitrate and 85% phosphoric acid occurred continuously for two years (July

2011 through July 2013 following a year of pretreatment data collection, hereafter: PRE, YR1

and YR2). Solar-powered metering pumps (LMI Milton Roy, Ivyland, Pennsylvania, USA)

delivered concentrated nutrient solutions into gravity-fed irrigation lines according to a program

based on continuously measured discharge using a CR800 data-logger (Campbell Scientific,

Logan, Utah, USA) and a Nanolevel pressure transducer (Keller America, Newport News,

Virginia, USA). Each irrigation line had drip spouts placed approximately every 5 m throughout

the experimental reach to ensure sufficient mixing. Each stream was assigned a unique target

concentration of N (as dissolved inorganic nitrogen, DIN [nitrate + ammonium], including both

mean background and added N: 81, 244, 366, 488, and 650 µg L-1) that corresponded to a unique

decreasing concentration of P (as soluble reactive phosphorus, SRP: 90, 68, 51, 34, and 11 µg L-

1), which resulted in five molar ratios of dissolved N:P very close to target values (2, 8, 16, 32,

and 128, respectively). Therefore, N and P concentrations were elevated above background

concentrations in each stream (target N concentrations were between ~2× and ~12× background,

and target P concentrations were between ~5× and ~31× background) and reflected low-to-

moderate enrichment consistent with observed concentrations in streams experiencing land-use

change in the region (Scott et al. 2002).

Water samples for nitrate (as NO3–-N), ammonium (NH4

+-N), and SRP were collected

biweekly within each experimental reach (n = 2) and upstream of the nutrient dosing system (n

= 2), filtered in the field using 0.45-µm nitrocellulose membrane filters (Millipore, Billerica,

22

Massachusetts, USA) and frozen until analysis. Nitrate-N, NH4+-N, and SRP concentrations were

measured with an Alpkem Rapid Flow Analyzer 300 (DIN; Alpkem, College Station, Texas,

USA) at the University of Georgia Analytical Chemistry Laboratory (Athens, Georgia, USA) or

spectrophotometrically (SRP) using the ascorbic acid method (APHA 1998, Shimadzu UV-1700,

Japan).

Litter breakdown rates and stoichiometry

We measured breakdown rates of maple (Acer rubrum) and rhododendron

(Rhododendron maximum) litter from December to June (PRE) and from December to April

(YR1, YR2). Maple and rhododendron represent dominant riparian tree species at CWT, have

distinct initial nutrient content (maple C:N/C:P ca. 78/2645; rhododendron C:N/C:P ca.

145/7552; D. W. P. Manning and A. D. Rosemond, unpublished data) and have been used

extensively for studying litter breakdown rates in southern Appalachian streams (Webster et

al.1999, Kominoski et al. 2007). Litter packs were constructed using 5-mm plastic mesh pecan

bags (22×40 cm; Cady Bag, Inc., Pearson, Georgia, USA) to allow access by shredders and to

maintain known quantities of litter. Freshly abscised litter was collected during peak leaf-fall in

October 2010, 2011, and 2012, air-dried in the laboratory for several weeks, and weighed into

10±0.1 g packs. The litter bags were anchored in the five experimental reaches on 1 December

2010, 27 November 2011, and 29 November 2012 for PRE, YR1, and YR2, respectively. Within

each experimental reach, we delineated four 17.5-m sub-reaches where 7 arrays (one for each

sampling date) of the single-species litter bags were deployed for a total of 280 bags for each

year (5 streams × 4 sub-reaches × 7 sampling dates × 2 species). Five additional litter bags of

23

each species were taken to the sites, submerged in the stream, and immediately collected to

account for handling losses (Benfield 2006).

Incubated leaf litter was removed over time and processed for mass remaining and litter

stoichiometry. In the PRE year, we collected leaf packs on days 7, 14, 21, 70, 109, 160, and 187

from each sub-reach. During YR1 and YR2, we expected higher breakdown rates, in particular

for maple, so we used a shorter sampling schedule (maple = days 7, 14, 21, 34, 55, 63, 77;

rhododendron = days 7, 14, 21, 63, 110, 126, 143). On each sampling date, litter bags were

removed from the streams, placed into individual plastic bags, and transported to the laboratory

on ice. In the laboratory, litter was rinsed over nested 1-mm and 250-µm sieves to remove

sediments and macroinvertebrates, placed into paper bags and dried for a minimum of 24 h at

55°C. The entire sample was weighed to determine dry mass and ground using an 8000-D ball

mill (Spex SamplePrep, Metuchen, New Jersey, USA). A sub-sample was combusted at 500°C

for 4.5 h to determine ash-free dry mass (AFDM). We estimated conditioned litter C:N or C:P

content for litter material collected on day 70 (PRE) or day 63 (YR1, YR2). Conditioned litter C

and N content were determined using a Carlo Erba NA 1500 CHN Analyzer (Carlo Erba, Milan,

Italy). Phosphorus content of the conditioned litter was determined using the plant dry ash/acid

extraction method followed by spectrophotometric analysis using the ascorbic acid method

(Allen 1974; APHA 1998).

Fungal biomass

Fungal biomass was estimated by measuring ergosterol concentration associated with five

ca. 2×2 cm litter pieces sub-sampled from each litter bag early in the breakdown experiments

(day 14). We measured ergosterol concentrations early in the breakdown process because early

24

fungal colonization of litter is indicative of subsequent fungal community development and

effects on litter breakdown rates (e.g., Duarte et al. 2008, Sridhar et al. 2009). Briefly, lipids

were extracted from the freeze-dried, weighed litter pieces using liquid-to-liquid extraction

(Gulis and Suberkropp 2006), and ergosterol concentrations were determined by HPLC (LC-

10VP, Shimadzu, Columbia, Maryland, USA) equipped with a Kinetex C18 column

(Phenomenex, Torrance, California, USA) and a UV detector set at 282 nm. External ergosterol

standards (Acros Organics, Geel, Belgium) were used. Ergosterol concentrations were converted

to fungal biomass using a conversion factor of 5.5 µg mg-1 of mycelial dry mass (Gessner and

Chauvet 1993).

Macroinvertebrate biomass

We focused our macroinvertebrate sampling efforts for both maple and rhododendron

litter collected on day 70 (PRE) and day 63 (YR1 and YR2), such that we captured the time to

~50% mass loss for maple and ~15% mass loss for rhododendron under pretreatment conditions.

After rinsing the litter, the two size-classes of macroinvertebrates were removed from the nested

sieves and preserved separately in 70% ethanol. The macroinvertebrates in each sample were

sorted, identified to the lowest taxonomic unit (typically genus; Merritt et al. 2008), and

measured to the nearest millimeter. Biomass was then determined using previously established

length-mass regressions for CWT stream taxa (Benke et al. 1999, J. B. Wallace, unpublished

data). We estimated shredder biomass per gram of litter AFDM remaining in each corresponding

litter bag based on the classification of specific taxa as shredders (Merritt et al. 2008).

25

Data analyses

Breakdown rate, k, was estimated using a linear regression of the ln-transformed fraction

of AFDM remaining vs. time (negative exponential model; sensu Benfield 2006). Specifically,

the model is Mt = M0 × e-kt, where M0 is the initial litter mass, Mt is the litter mass on a given

sampling day, and t is time (number of days incubated in the stream). We estimated a specific k

value in four sub-reaches within each experimental stream, such that our total number of litter

breakdown rate estimates was 120 (4 sub-reaches × 5 streams × 2 species × 3 years). The

primary predictor variables used in this study were either ambient (PRE) or enriched (YR1,

YR2) DIN or SRP concentrations. For enriched values, we used calculated concentrations based

on experimental additions of N and P. Evidence of concentration-dependent nutrient uptake in

the treatment reaches indicated that concentration estimates based on the amounts of nutrients

actually s were better than measured streamwater concentrations to characterize the experimental

treatments (A. D. Rosemond, unpublished data). Enriched concentrations were determined based

on the quantity of N or P added to each stream, using records of concentrated nutrient solution

refills, measured ambient water nutrient concentrations, and total daily discharge.

The path model

We constructed a path model with hypothesized causal links based on previous studies of

how nutrients, other abiotic drivers, and biological factors are predicted to affect litter

breakdown rates (e.g., Hieber and Gessner 2002, Hladyz et al. 2009; Fig. 1). We used six

predictor variables for litter breakdown: temperature, discharge, streamwater nutrient

concentrations, fungal biomass, shredder biomass, and conditioned litter stoichiometry (C:N or

C:P); a link between fungal biomass and litter breakdown is included to imply fungal

26

contribution to C losses via respiration. We assessed model fit based on comparisons of the

implied model covariance structure and observed covariance structure using χ2 tests (Grace

2006). A path model was deemed to be consistent with the data when modeled covariance

structure and observed covariance structure were not statistically different (i.e. non-significant χ2

test). If assessment of the χ2 test suggested that a model was inconsistent with the data, we re-

evaluated model structure using one-degree of freedom χ2 criteria and inspection of residual

covariance matrices to test the improvement in model fit gained by adding a specific link to the

model (Grace et al. 2012). We removed links from the model to improve model parsimony in

cases where maintaining a specific link had negligible impact on overall model fit based on non-

significant parameter estimates.

Once we arrived at an acceptable model to predict litter breakdown rates, we compared

models with this underlying structure using N alone, P alone, or N and P combined as predictors.

We tested for the importance of stoichiometry of decomposing litter (C:N or C:P) in the same

manner, such that we had two sets of three models (i.e. N, P, and N+P for litter C:N and C:P,

respectively). These six models allowed us to test for the importance of N, P, and C:N vs. C:P for

predicting litter breakdown rates. We evaluated the support for each model based on Akaike’s

Information Criterion (AIC; Burnham and Anderson 2002). In addition to an overall model that

included results from PRE, YR1, and YR2, we analyzed models using a separate group

(hereafter: single-year) approach, to compare PRE, YR1, and YR2 separately. To contrast the

path coefficients before and after experimental nutrient enrichment, we compared model fit when

all path coefficients were allowed to differ to model fit when coefficients were held constant

among years using χ2-difference tests. For single-year modeling, we focused on addressing

27

differences in path coefficients using the underlying structure of the best-supported overall

models.

Parameters of each model are reported as standardized path coefficients to allow for

direct comparison of variables measured at different scales and are indicative of the weight of

each predictor variable for explaining variation in the response variables (unstandardized

coefficients are reported in Appendix A: Table A4). Standardized coefficients were obtained

through z-transformations such that means and variances of the variables are adjusted to zero and

one, respectively. Because path analysis is a structured set of linear regressions, basic

assumptions of linear regression apply; thus, we ln-transformed our predictor and response

variables to meet assumptions of normality and linearity. All analyses were conducted using the

statistical software R, version 3.0.1 (R Core Development Team 2013) and the package ‘lavaan’

(version 0.5-16; Rosseel 2012).

Results

Whole-stream nutrient additions

Experimental enrichment of the five study reaches resulted in elevation of DIN and SRP,

which generally reflected target concentrations (Table 1). Enriched DIN and SRP levels were on

average between 0.85-8× and 3-28× background (PRE) concentrations, respectively, during the

enrichment period. Mean temperature during each breakdown experiment differed, at most, by

2.6°C across streams and years (mean temperature for all streams and years = 7.1°C); within

each stream temperature changed <15% compared to pre-treatment (Table 1). Mean discharge

ranged from 4.1-20.0 L s-1 and changes in discharge ranged from 3-78% of pre-treatment

depending on stream and year (Table 1).

28

Litter breakdown rates

Across all five streams, maple and rhododendron breakdown rates were higher compared

to PRE in YR1 and YR2 (Table 2). Rhododendron breakdown rates were affected by nutrients to

a greater extent than maple, and were 3.1-6.4× higher in YR1 and 2.4-4.7× higher in YR2 than

PRE. Maple breakdown rates were 1.1-1.8× higher in YR1, and 1.1-2.7× higher in YR2 than

PRE. Two-year averages for increases in rhododendron breakdown rates tended to be highest in

the two lowest N:P treatments (N:P = 2 and 8; 4.5× and 4.9×, respectively), with decreasing

response to nutrients in higher N:P treatments (Table 2). Two-year averages for increases in

maple breakdown rates were highest when treatment N:P was 128 (2.2×), but breakdown rates

also increased when treatment N:P was < 16 (Table 2).

Litter stoichiometry

Maple and rhododendron C:N and C:P were reduced during YR1 and YR2 compared to

PRE (Table 3) for all treatments, with relatively greater differences in C:P for both species.

Rhododendron C:N and C:P decreased ~1.2-1.8× and 1.8-4.8× compared to PRE, which were

relatively greater changes than those of maple. Maple C:N and C:P decreased ~1.2-1.4× and 1.1-

2.5×, respectively (Table 3).

Path model: Nutrient effects in an overall model

We arrived at a general model structure that indicated that the primary influences of

nutrients on litter breakdown rate were propagated through effects on fungal biomass,

conditioned litter stoichiometry, and shredders (Fig. 2a,b). The final model structure was similar

to our original hypothesized model (Fig. 1), except for an added link between discharge and

29

shredders, and a pruned link between temperature and litter breakdown (Fig. 2a,b). Thus,

candidate models maintained this general model structure, and excluded temperature as a

predictor variable. We tested six candidate models containing N, P, N+P and C:N and C:P that

had 13 to 15 path coefficient estimates (Appendix A: Table A1). Of these six candidate models,

five were found to have good agreement between modeled and observed covariance matrices

based on χ2 tests (Appendix A: Table A1). These five models included the three models with

litter C:N (and streamwater N, P and N+P) and two models with litter C:P (and streamwater P

and N+P) (Appendix A: Table A1). The model with the best support based on AIC included N

and litter C:N (χ2 = 9.3, d.f. = 5, P = 0.10; Appendix A: Table A1). Although the model with the

most support based on AIC contained N and C:N, we also found support for a path model

containing P and C:P based on χ2 tests (χ2 = 0.7, d.f. = 5, P = 0.95; Appendix A: Table A1).

We tested the importance of specific parameters (stoichiometry, discharge, fungal and

shredder biomass) for the fit of the overall path model by fixing path coefficients to zero, and

then ranked the importance of each parameter based on ΔAIC when the full and reduced models

were compared. For both C:P/P and C:N/N models, removing any of the four parameters resulted

in significantly reduced model fit (χ2 difference test P < 0.05 in all cases; Appendix A: Table

A2). For the C:P/P model, conditioned litter stoichiometry (C:P) was the most important

parameter for model fit (ΔAIC = -47; Appendix A: Table A2), while for C:N/N, the most

important parameter was fungal biomass (ΔAIC = -271; Appendix A: Table A2), followed by

litter stoichiometry (C:N; ΔAIC = -34; Appendix A: Table A2).

30

Nitrogen effects on litter breakdown

Nitrogen concentrations affected litter breakdown rates through positive effects on fungal

biomass, which were linked to decreases in litter C:N and positive indirect effects on shredder

biomass. Overall, the C:N/N model explained 57% of the variation in litter breakdown rates, and

20%, 36%, and 36% of the variation in fungal biomass, shredders, and litter C:N, respectively

(Fig. 2a). Streamwater N positively affected fungal biomass, which was linked to reduced litter

C:N, that then positively affected litter breakdown rates through increased shredder biomass

(Fig. 2a). There was a strong link between litter C:N and litter breakdown (Fig. 2a) and

significant positive effects of fungi through C:N on litter breakdown (compound path = -0.60 ×

-0.48 = 0.3; P < 0.05). Fungal biomass, discharge, and shredder biomass had comparable

influence on litter breakdown rates, but litter C:N was 2.2-2.7× more important compared to

these three variables (Fig. 2a).

Phosphorus effects on litter breakdown

Similar to N effects on litter breakdown, P concentrations affected litter breakdown rates

through fungal biomass, litter stoichiometry, and shredder pathways. Overall, the C:P/P model

explained 57% of the variation in litter breakdown, and 34%, 39%, and 51% of the variation in

fungal biomass, shredders, and litter C:P, respectively. Streamwater P positively affected fungal

biomass, which was linked to reduced litter C:P, that then positively affected litter breakdown

rates through increased shredder biomass (Fig. 2b), although the strength of this path was lower

compared to that of the C:N/N model. In contrast to the C:N/N model, the C:P/P model included

a direct link between SRP and litter C:P, and there was a strong link between litter C:P and litter

breakdown (Fig. 2b), and significant positive effects of fungi on breakdown rates via C:P

31

(compound path = -0.29 × -0.52 = 0.15; P < 0.05). As with the C:N/N model, fungal biomass,

discharge, and shredder biomass had similar influence on litter breakdown rates, but litter C:P

was 2.1-2.9× more important compared to these three variables (Fig. 2b).

Path models: single-year models

The models above include data from PRE, YR1, and YR2 together. These models reflect

how similar variation in N and P concentrations – due to time or space – would affect litter

breakdown. Insights into the effects of N vs. P were also obtained by contrasting model structure

between PRE (no added nutrients) to YR1 or YR2. We analyzed the two best-supported models

for C:N/N and C:P/P (Appendix A: Table A1) with each year treated as a subset of the data. For

both C:N/N and C:P/P models, the model structure was consistent among years (C:P/P χ2 = 3.1,

11.6, and 2.1 for PRE, YR1, YR2, d.f. = 12; C:N/N χ2 = 5.4, 4.7, and 8.0 for PRE, YR1, and

YR2, d.f. = 15; all P > 0.05). For the C:P/P model, we found significant reduction in model fit

when path coefficients were held constant (χ2 difference test; P << 0.05). We found marginal

evidence for differences in model fit when path coefficients were held constant for the C:N/N

model (χ2 difference test; P = 0.07). Prior to enrichment for both the C:N/N and C:P/P models,

conditioned litter C:N or C:P was the central predictor of litter breakdown rates and shredder

biomass, which in this case was solely determined by litter species identity, not streamwater

nutrient concentrations (Appendix A: Table A3). During the enrichment years for both C:P/P and

C:N/N, fungal biomass, shredders and discharge became stronger predictors of litter breakdown

rates (Appendix A: Table A3). Conditioned litter C:P, and to some extent C:N, were weaker

predictors of litter breakdown rates in YR1 and YR2 compared to PRE, but remained a nexus of

the paths linking fungal biomass, shredders and breakdown rates (Appendix A: Table A3). The

32

amount of variation in litter breakdown rates explained by the single-year models differed from

year to year, although in each case the models explained >30% of the variation in litter

breakdown.

Discussion

Our study showed that streamwater N and P affected litter breakdown through

stimulation of fungal biomass and changes in litter stoichiometry, which were associated with

higher shredder biomass and litter breakdown rates. These effects occurred via multiple

pathways and included a collection of interactions with litter stoichiometry at their center. Path

analysis indicated that the strength of N and P as predictors of these C loss pathways was similar.

Our study adds to evidence that N and P loading accelerates detrital C loss from ecosystems,

thereby reducing standing stocks of an important energy source (Benstead et al. 2009,

Suberkropp et al. 2010, Woodward et al. 2012), and reveals some of the fundamental

mechanisms by which these effects occur.

Nitrogen and phosphorus effects on litter breakdown pathways

Previous studies have revealed that a key effect of nutrient enrichment of detritus-based

systems is increased detrital quality for consumers, and our results emphasize the central

importance of this effect for predicting litter breakdown rates (Cross et al. 2003, Rosemond et al.

2010, Scott et al. 2013). Based on the overall models, litter breakdown in streams could largely

be predicted using conditioned litter stoichiometry (using either C:N or C:P) across gradients of

low-to-moderate nutrient enrichment, given that large ranges in N and P availability will be

reflected in corresponding gradients of litter C:N and C:P. By examining single-year path

33

models, we were able to ascertain that the strongest driver of litter breakdown before nutrient

enrichment was conditioned litter stoichiometry (C:N and/or C:P), owing to the large range in

litter C:N and C:P content driven by species differences and associated microbial activity.

During YR1 and YR2, we observed substantial decreases in conditioned maple and

rhododendron litter C:N and C:P, by as much as 1.8× for rhododendron C:N and 4.8× for

rhododendron C:P. As a result, litter species differences in terms of C:N and C:P were weaker

predictors of litter breakdown rates during YR1 and YR2.

The streamwater nutrient-mediated convergence of C:P content of different litter species

facilitated by microbial pathways is likely an important determinant of shredder biomass and

activity, because of reduced consumer-resource stoichiometric imbalances (Cross et al. 2003).

Consumer-resource imbalances are typically determined using threshold elemental ratios (TERs;

Sterner and Elser 2002); in this case, the C:P or C:N threshold at which growth limitation by

either element is minimized (e.g., Frost et al. 2006, Danger et al. 2013). The results of this study

support TER predictions, as we observed the highest shredder biomass in litter bags containing

litter with C:N and C:P content that approached or matched reported stream shredder TERs for

C:N and C:P (Frost et al. 2006, Tant et al. 2013) (Appendix B: Fig. B1). The effects we observed

based on fungal biomass measured at early stages of decay support the idea that the initial (ca.

two week) fungal colonization of litter is an important predictor of litter stoichiometry at later

stages of decay, shredder colonization, and breakdown rates (Duarte et al. 2008, Sridhar et al.

2009).

Path analysis showed that N and P had similar effects on litter breakdown via both fungal

biomass and litter stoichiometry, but the similar consequences of N and P on breakdown rates

appear to be driven by subtly different mechanisms. The key difference between the overall

34

C:N/N and C:P/P models was the inclusion of an apparent link between streamwater P and litter

C:P in the overall C:P/P model. In contrast, litter C:N was not predicted by streamwater N, but

was strongly predicted by fungal biomass. The differences in the models imply two alternative

mechanisms driving the effects of N and P on litter breakdown. First, the absence of direct

effects of streamwater N on litter C:N suggests that reductions in litter C:N are driven indirectly

by positive N effects on fungal biomass. This result is consistent with previous studies that

showed increased fungal biomass and increased litter N content due to elevated streamwater N

(e.g., Ferreira et al. 2006, Rosemond et al. 2010). Microcosm studies complementary to this

study also demonstrated that fungal growth rates were more strongly related to N concentrations

compared to P, indicating that N may be more important for fungal biomass accrual on litter (V.

Gulis, unpublished data). Second, the apparent direct effect of streamwater P on litter C:P in the

overall model suggests that litter C:P and fungal biomass may be decoupled, presumably because

fungi may exhibit flexible cellular C:P via P storage (e.g., as polyphosphate granules; Beever and

Burns 1980, V. Gulis, unpublished data). However, we cannot rule out increased litter P due to

abiotic sorption, microbial community shifts (e.g., Gulis and Suberkropp 2004), or the effects of

bacteria (but see Gulis and Suberkropp 2003, Tant et al. 2013).

Nutrient enrichment resulted in increases in litter breakdown rates via pathways that were

driven by both microorganisms and shredders. Our findings illustrate that losses due to shredder

feeding were stimulated by initial fungal colonization and subsequent changes in litter

stoichiometry; thus it is difficult to adequately partition contributions by either microorganisms

or shredders. Litter mass loss driven by microorganisms includes multiple mechanisms:

production of microbial biomass, respiration, production of exoenzymes and in the case of fungi,

production of spores. We were not able to measure all of these microbial-driven C loss pathways,

35

which together may result in high C loss particularly in the early stages of litter breakdown, such

that less litter C is subsequently available to shredders (Tant et al. 2015). However, comparing a

primary measure of microbial driven C loss—respiration—to shredder driven C loss illustrates

that losses directly attributed to microorganisms alone can be smaller than the effects of

microorganisms and shredders combined. Specifically, we found estimated mass (mg C d-1) of

maple and rhododendron litter respired by microorganisms or consumed by shredders increased

by 1.7× and 9.4× under nutrient enriched conditions, respectively. Our path analyses are

consistent with this contrast illustrating the important interactions between microorganisms and

shredders in driving litter breakdown rates, which resulted in greater C losses compared to the

effect of one microbially-driven pathway alone.

Overall effects of gradients of N and P on litter breakdown

Increased litter breakdown rates across the experimental gradient of N:P were likely

because of similar effects of N and P on fungi, litter stoichiometry, and eventually shredders,

demonstrating that rapid C loss from detritus-based aquatic ecosystems could occur in situations

where either N or P is elevated relative to the other nutrient. Our experimental design included

treatments with relatively low levels of added P relative to N and vice versa (e.g., ca. 430 µg N

L-1 and 9 µg P L-1 vs. 100 µg N L-1 and 80 µg P L-1), suggesting that large changes in breakdown

can occur with elevated concentrations of one nutrient and minor alleviation of nutrient

limitation by the other. For this reason, the ratio of nutrients was found to be a poor predictor of

litter breakdown, as shown by stronger support for models containing N and P separately

compared to N and P combined, and poor agreement between observed and modeled covariance

36

matrices when N:P was used as a predictor (D. W. P. Manning and A. D. Rosemond,

unpublished data).

Species-specific differences in initial litter nutrient ratios may have been important in the

context of differential responses to N vs. P enrichment. Breakdown rates for both litter species

were elevated across all nutrient treatments, but generally the highest breakdown rates for

rhododendron occurred when streamwater P concentrations were greatest and the highest

breakdown rates for maple occurred when streamwater N concentrations were greatest.

Deficiencies in litter nutrient content may help explain these patterns. Rhododendron litter is

much lower in P content than maple and thus colonizing microorganisms require P from the

water column, and respond most when it is available. Rhododendron litter gained much more P

in low vs. high N:P treatments (~4× vs. ~2× increase in P content compared to PRE in N:P = 2,

128, respectively). Maple litter may have had adequate P availability for a stronger response to

streamwater N in the high N:P treatment. Specifically, maple litter gained similar P content in

both low and high N:P treatments (~2× vs. ~1.5× increase in P content compared to PRE in N:P

= 2, 128, respectively). Thus, because rhododendron litter was initially more P-deficient,

differential changes to litter P content created a more defined gradient in litter P content

compared to maple and potentially limited the increases in rhododendron breakdown rate where

streamwater N:P treatments were high (128) and litter P gains were low.

Our results show that low-to-moderate enrichment of aquatic ecosystems with gradients

of N and P concentrations caused substantial acceleration of C loss, and that streamwater N and

P and associated effects on litter C:N and C:P had similar magnitude effects on breakdown rates

via microbial and detritivore pathways. We propose that dissolved N and P modulate litter

breakdown rates through effects on fungal biomass and litter C:N (N-driven effects), as well as

37

effects on litter C:P owing to abiotic or biotic P immobilization on detritus (P-driven effects).

The N and P effects on litter stoichiometry appear to be important for shredder pathways,

because litter C:N and C:P can be reduced to levels that approach shredder nutrient demand

(Frost et al. 2006). The N and P concentrations we used for this study are comparable or lower

than those observed in streams experiencing moderate land-use change across the southern

Appalachians (Scott et al. 2002), and are in the lower range of continental nutrient gradients in

the US and Europe (Alexander and Smith 2006, Woodward et al. 2012). Mechanisms for

accelerated litter breakdown described in this study likely occur in many systems with similarly

elevated nutrient concentrations. Our results imply that elevated N and P throughout river

networks could lead to increased litter breakdown rates, reduced C retention, and altered delivery

of C to downstream ecosystems (Cole et al. 2007, Benstead et al. 2009, Woodward et al. 2012).

Acknowledgements

This work was supported by NSF (DEB-0918894 to ADR and JCM, DEB-0918904 to

JPB, and DEB-0919054 to VG). This research leveraged logistical support from the CWT LTER

Program at the University of Georgia, which is supported by NSF award DEB-0823293 from the

Long Term Ecological Research Program (JCM co-PI). Rob Case, Daniel Hutcheson, and Kevin

Simpson of YSI Integrated Systems and Services implemented the infrastructure for the nutrient-

dosing system. Aqueous ammonium nitrate was provided by The Andersons, Inc. through David

Plank. We are grateful for data collection and maintenance of the experimental dosing system by

Jason Coombs and Katie Norris. Christian Barrett, Phillip Bumpers, Jason Coombs, John Davis,

Hannah Dolan, Kait Farrell, Tom Maddox, Chelsea Norman, Katie Norris, and James Wood

38

helped with fieldwork or in the laboratory. This manuscript was improved by helpful comments

from two anonymous reviewers, the Rosemond lab group, and Chao Song.

Literature Cited

Alexander, R. B., and R. A. Smith. 2006. Trends in the nutrient enrichment of U. S. rivers during

the late 20th century and their relation to changes in probable stream trophic conditions.

Limnology and Oceanography 51:639-654.

Allen, S. E., et al. 1974. Chemical analysis of ecological materials. Wiley, New York, New

York, USA.

American Public Health Association (APHA). 1998. Standard methods for the

examination of water and wastewater. 20th edition. American Public Health

Association, Washington, D.C.

Arbuckle, K. E., and J.A. Downing. 2001. The influence of watershed land use on lake N:P in a

predominantly agricultural landscape. Limnology and Oceanography 46: 970-975.

Beever, R. E., and D. J. W. Burns. 1980. Phosphorus uptake, storage and utilization by fungi.

Advances in Botanical Research. 8: 127-219.

Benfield, E. F. 2006. Decomposition of leaf material. Pages 711-720 in F.R. Hauer and G.A.

Lamberti, editors. Methods in Stream Ecology. Second Edition. Academic Press.

Benke, A. C., A. D. Huryn, L. A. Smock, and J. B. Wallace. 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.

39

Benstead, J. P., A. D. Rosemond, W. F. Cross, J. B. Wallace, S. L. Eggert, K. Suberkropp, V.

Gulis, J. L. Greenwood, and C. J. Tant. 2009. Nutrient enrichment alters storage and

fluxes of detritus in a headwater stream ecosystem. Ecology 90:2556-2566.

Burnham, K. P., and D. R. Anderson. 2002. Model selection and multi-model inference: A

practical information-theoretic approach. Second Edition. Springer, New York, USA.

Cheever, B. M., J. R. Webster, E. E. Bilger, and S. A. Thomas. 2013. The relative importance of

exogenous and substrate-derived nitrogen for microbial growth during leaf

decomposition. Ecology 94:1614-1625.

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

Duarte, P. Kortelainen, J. A. Downing, J. J. Middelburg, and J. Melack. 2007. Plumbing

the global carbon cycle: Integrating inland waters into the terrestrial carbon budget.

Ecosystems 10:172-185.

Cross, W. F., J. P. Benstead, A. D. Rosemond, and J. B. Wallace. 2003. Consumer-resource

stoichiometry in detritus-based streams. Ecology Letters 6:721-732.

Danger, M., J. Arce Funck, S. Devin, J. Heberle, and V. Felten. 2013. Phosphorus content in

detritus controls life-history traits of a detritivore. Functional Ecology 27:807-815.

Duarte, S., C. Pascoal, and F. Cássio. 2008. High Diversity of fungi may mitigate the impact of

pollution on plant litter decomposition in streams. Microbial Ecology 56:688-695.

Elser, J. J., M. E. S. Bracken, E. E. Cleland, D. S. Gruner, W. S. Harpole, H. Hillebrand, J. T.

Ngai, E. W. Seabloom, J. B. Shurin, and J. E. Smith. 2007. Global analysis of nitrogen

and phosphorus limitation of primary producers in freshwater marine and terrestrial

ecosystems. Ecology Letters 10:1135-1142.

40

Ferreira, V., V. Gulis, and M. A. S. Graça. 2006. Whole-stream nitrate addition affects litter

decomposition and associated fungi but not invertebrates. Oecologia 149:718-729.

Ferreira, V., B. Castagneyrol, J. Koricheva, V. Gulis, E. Chauvet, and M. A. S. Graça. 2014. A

meta-analysis of the effects of nutrient enrichment on litter decomposition in streams.

Biological Reviews DOI: 10.1111/brv.12125

Findlay, S., J. Tank, S. Dye, H. M. Valett, P. J. Mulholland, W. H. McDowell, S. L. Johnson, S.

K. Hamilton, J. Edmonds, W. K. Dodds, W. B. Bowden. 2002. A Cross-system

comparison of bacterial and fungal biomass in detritus pools of headwater streams.

Microbial Ecology 43: 55-66.

Frost, P. C., J. P. Benstead, W. F. Cross, H. Hillebrand, J. H. Larson, M. A. Xenopoulos, and T.

Yoshida. 2006. Threshold elemental ratios of carbon and phosphorus in aquatic

consumers. Ecology Letters 9:774-779.

Gessner, M. O., and E. Chauvet. 1993. Ergosterol-to-biomass conversion factors for aquatic

hyphomycetes. Applied and Environmental Microbiology 59:502-07.

Grace, J. B. 2006. Structural equation modeling and natural systems. Cambridge University

Press, Cambridge, UK.

Grace, J. B., T. M. Anderson, H. Olff, and S. M Scheiner. 2012. On the specification of

structural equation models for ecological systems. Ecological Monographs 80:67-87.

Grimmett, I. J., K. N. Shipp, A. Macneil, and F. Bärlocher. 2013. Does the growth rate

hypothesis apply to aquatic hyphomycetes? Fungal Ecology 6:493-500.

Gulis, V., and K. Suberkropp. 2003. Interactions between stream fungi and bacteria associated

with decomposing leaf litter at different levels of nutrient availability. Aquatic Microbial

Ecology 30:149–157.

41

Gulis, V., and K. Suberkropp. 2004. Effects of whole-stream nutrient enrichment on the

concentration and abundance of aquatic hyphomycete conidia in transport. Mycologia

96:57-65.

Gulis, V., and K. Suberkropp. 2006. Fungi: Biomass, production, and sporulation of aquatic

hyphomycetes. Pages 311-326 in F.R. Hauer and G.A. Lamberti, editors. Methods in

Stream Ecology. Second Edition. Academic Press.

Hieber, M., and M. O. Gessner. 2002. Contribution of stream detritivores, fungi, and bacteria to

leaf breakdown based on biomass estimates. Ecology 83:1026-1038.

Hladyz, S., M. O. Gessner, P. S. Giller, J. Pozo, and G. Woodward. 2009. Resource quality and

stoichiometric constraints on stream ecosystem functioning. Freshwater Biology 54:957-

970.

Kominoski, J. S., C. M. Pringle, B. A. Ball, M. A. Bradford, D. C. Coleman, D. B. Hall, and M.

D. Hunter. 2007. Nonadditive effects of leaf litter species diversity on breakdown

dynamics in a detritus-based stream. Ecology 88:1167-1176.

Merritt, R. W., K. W. Cummins and M. B. Berg. 2008. An introduction to the aquatic insects of

North America. Fourth edition. Kendall/Hunt.

Petersen, R. C., and K. W. Cummins. 1974. Leaf processing in a woodland stream. Freshwater

Biology 4:343-368.

R Development Core Team. 2013. R: a language and environment for statistical computing. R

Foundation for Statistical Computing, Vienna, Austria.

Rosseel, Y. 2012. Lavaan: an R package for structural equation modeling. Journal of Statistical

Software. 48:1-36.

42

Rosemond, A. D., C. M. Swan, J. S. Kominoski, and S. E. Dye. 2010. Non-additive effects of

litter mixing are suppressed in a nutrient-enriched stream. Oikos 119:326-336.

Scott, M. C., G. S. Helfman, M. E. McTammany, E. F. Benfield, and P. V. Bolstad. 2002.

Multiscale influences on physical and chemical stream conditions across Blue Ridge

landscapes. Journal of the American Water Resources Association 38:1379-1392.

Scott, E. E., C. Prater, E. Norman, B. C. Baker, M. Evans-White, and J. T. Scott. 2013. Leaf-

litter stoichiometry is affected by streamwater phosphorus concentrations and litter type.

Freshwater Science 32:753-761.

Schlesinger, W. H., J. J. Cole, A. C. Finzi, and E. A. Holland. 2011. Introduction to coupled

biogeochemical cycles. Frontiers in Ecology and the Environment 9:5-8

Sinsabaugh, R.L., B. H. Hill and J. J. Folstad-Shah. 2009. Ecoenzymatic stoichiometry of

microbial nutrient acquisition in soils and sediment. Nature 462:795-798.

Sridhar, K. R., S. Duarte, F. Cássio, and C. Pascoal. 2009. The role of early fungal colonizers in

leaf-litter decomposition in Portuguese streams impacted by agricultural runoff.

International Review of Hydrobiology 94:399-409.

Sterner, R.W., and Elser, J. J. 2002. Ecological Stoichiometry. Princeton University Press,

Princeton, New Jersey, USA.

Suberkropp, K., and E. Chauvet. 1995. Regulation of leaf breakdown by fungi in streams:

Influences of water chemistry. Ecology 76:1433-1455.

Suberkropp, K., V. Gulis, A. D. Rosemond, and J. P. Benstead. 2010. Ecosystem and

physiological scales of microbial responses to nutrients in a detritus-based stream:

Results of a 5-year continuous enrichment. Limnology and Oceanography 55:149-160.

43

Swank, W. T., and D. A. Crossley. 1988. Forest hydrology and ecology at Coweeta. Springer-

Verlag. New York, New York, USA.

Tant, C. J., A. D. Rosemond, and M. R. First. 2013. Stream nutrient enrichment has a greater

effect on coarse than on fine benthic organic matter. Freshwater Science 32:1111-1121.

Tant, C. J., A. D. Rosemond, A. M. Helton, and M. R. First. In press. Nutrient enrichment alters

the magnitude and timing of fungal, bacterial, and detritivore contributions to litter

breakdown. Freshwater Science.

Webster, J. R., E. F. Benfield, T. P. Ehrman, M. A. Schaeffer, J. L. Tank, J. J. Hutchens, and D.

J. D’angelo. 1999. What happens to allochthonous material that falls into streams? A

synthesis of new and published information from Coweeta. Freshwater Biology 41:687-

705.

Woodward, G., et al. 2012. Continental-scale effects of nutrient pollution on stream ecosystem

functioning. Science 336:1438-1440.

Supplementary Material

Appendix A. Additional path model results, and unstandardized path coefficients for overall

models.

Appendix B. Supplementary results: Shredder biomass in maple and rhododendron litter bags

during PRE, YR1 and YR2 as a function of litter C:N and C:P.

44

Table 2.1. Mean (±SE) ambient (PRE) and enriched (YR1, YR2) nutrient concentrations (µg L-1) during each litter breakdown

experiment for the five treatment reaches used in this study (n = 9, 11, 23, respectively). Nutrient concentrations for PRE were

measured ambient concentrations; YR1 and YR2 concentrations are based on the amounts of DIN or SRP added to each stream

estimated using records of total daily discharge, concentrated nutrient solution refills and background nutrient concentrations. Also

reported are the mean, maximum and minimum daily discharge (L s-1) observed for each treatment reach, in addition to the mean

(±SE) daily temperature (°C) recorded during each litter breakdown experiment (PRE, YR1, YR2).

   

Nutrients (µg L-1)

Discharge (L s-1)

 

Temperature (°C)

Target N:P Year N:P DIN (±SE) SRP (±SE)

Mean Max Min

 

Mean (±SE)

2 PRE 12.5 17.0 (2.0) 3.0 (0.0)

6.3 16.4 2.2

 

7.09 (0.29)

 

YR1 3.0 120.5 (15.5) 90.1 (6.5)

5.2 16.5 1.7

 

7.79 (0.19)

 

YR2 2.6 80.4 (7.9) 69.4 (6.5)

8.3 34.6 1.6

 

6.48 (0.17)

    8 PRE 127.6 173.0 (10.0) 3.0 (0.3)

21.9 43.2 13.2

 

7.66 (0.27)

 

YR1 14.3 302.8 (26.2) 46.9 (4.1)

18.1 75.7 6.1

 

8.28 (0.15)

 

YR2 8.6 149.1(10.7) 38.6 (2.7)

17.9 74.5 8.3

 

7.35 (0.13)

    16 PRE 27.1 49.0 (8.0) 4.0 (1.0)

9.6 25.2 3.1

 

6.69 (0.26)

45

 

YR1 18.0 429.5 (51.2) 52.8 (7.3)

5.7 26.0 3.0

 

7.22 (0.18)

 

YR2 16.0 409.1 (85.3) 56.7 (11.9)

5.7 8.9 2.1

 

6.32 (0.16)

    32 PRE 125.1 238.0 (22.0) 4.0 (0.4)

12.0 23.0 6.3

 

7.06 (0.27)

 

YR1 42.7 362.8 (26.5) 18.8 (1.9)

6.1 16.0 3.8

 

8.00 (0.17)

 

YR2 30.6 388.1 (12.0) 28.1 (1.2)

7.5 17.8 3.6

 

6.98 (0.16)

    128 PRE 57.5 78.0 (9.0) 3.0 (0.3)

18.7 118.4 6.7

 

6.38 (0.29)

 

YR1 103.3 366.9 (43.1) 7.9 (1.0)

9.8 45.2 2.1

 

6.95 (0.20)

 

YR2 105.6 494.1 (32.6) 10.4 (0.5)

10.7 44.2 0.1

 

5.72 (0.17)

46

Table 2.2 Mean breakdown rates (±SE) reported as decay coefficients (k, d-1) of the negative

exponential model. Also reported are the YR1/PRE and YR2/PRE ratios (and their means) that

indicate the multiplicative increase in breakdown rate between PRE and enrichment years (i.e.,

YRx/PRE = 2 indicates an increase in k by 2×).

   

Maple k

 

Rhododendron k

Target N:P Year mean ±SE YRx/PRE

 

mean ±SE YRx/PRE

2 PRE 0.0106 0.004

 

0.0019 0.000

YR1 0.0115 0.004 1.09

 

0.0099 0.003 5.27

YR2 0.0215 0.005 2.04

 

0.0069 0.003 3.70

 

mean: 1.56

   

mean: 4.48

      8 PRE 0.0133 0.004

 

0.0047 0.001

YR1 0.0207 0.002 1.56

 

0.0300 0.001 6.36

YR2 0.0252 0.000 1.90

 

0.0159 0.001 3.37

mean: 1.73

 

mean: 4.86

  16 PRE 0.0096 0.001

 

0.0020 0.000

YR1 0.0124 0.002 1.30

 

0.0083 0.002 4.06

YR2 0.0191 0.003 2.00

 

0.0095 0.003 4.70

mean: 1.65

 

mean: 4.38

  32 PRE 0.0152 0.003

 

0.0039 0.001

YR1 0.0177 0.001 1.16

 

0.0210 0.006 5.35

YR2 0.0166 0.002 1.09

 

0.0093 0.001 2.37

47

mean: 1.12

 

mean: 3.86

  128 PRE 0.0074 0.001

 

0.0035 0.001

YR1 0.0135 0.002 1.83

 

0.0109 0.002 3.08

YR2 0.0195 0.003 2.65

 

0.0080 0.001 2.26

     

mean: 2.24

   

mean: 2.67

48

Table 2.3. Mean litter C:P and C:N ratios on d 70 and standard error for Acer rubrum (maple)

and Rhododendron maximum (rhododendron) leaves during PRE, YR1 and YR2. Also reported

are the YR1/PRE and YR2/PRE ratios, indicating the magnitude of change in C:P or C:N

compared to PRE.

Maple

litter C:P

litter C:N

Target N:P Year mean ±SE YRx/PRE

mean ±SE YRx/PRE

2 PRE 2746 172

55 3

YR1 1102 33 0.40

40 2 0.72

YR2 1448 255 0.53

42 3 0.75

8 PRE 2254 244

47 3

YR1 986 87 0.44

38 5 0.81

YR2 1025 119 0.45

36 2 0.75

16 PRE 2124 201

51 3

YR1 1324 175 0.62

44 6 0.85

YR2 1170 106 0.55

37 2 0.73

32 PRE 2107 211

51 3

YR1 1175 163 0.56

34 4 0.67

YR2 1825 510 0.87

35 1 0.69

128 PRE 2331 123

52 5

YR1 1234 180 0.53

44 2 0.84

YR2 2062 273 0.88

40 2 0.76

                 

49

   

Rhododendron

   

litter C:P

litter C:N

Target N:P Year mean ±SE YRx/PRE

mean ±SE YRx/PRE

2 PRE 6223 683

112 1

YR1 1312 87 0.21

66 4 0.58

YR2 1873 151 0.30

64 4 0.57

8 PRE 5827 614

73 23

YR1 1294 109 0.22

59 2 0.81

YR2 2066 253 0.35

59 3 0.81

16 PRE 5023 217

103 3

YR1 1886 141 0.38

64 3 0.63

YR2 2442 198 0.49

63 2 0.62

32 PRE 4430 384

101 6

YR1 1626 119 0.37

60 3 0.59

YR2 1425 114 0.32

63 4 0.63

128 PRE 5026 72

113 n.a.

YR1 2875 90 0.57

69 1 0.61

YR2 2293 645 0.46

64 2 0.57

50

Figure Legends

Fig. 2.1. Hypothesized path model describing how nutrients affect litter breakdown rates. Arrows

indicate hypothesized causal links between variables, with the direction of the effect denoted by

a (+) or (-) symbol. The structured set of linear equations that correspond to each response

variable can be described based on the links associated with each variable (e.g., Shredders ~

Fungal biomass + Leaf C:N/C:P, Leaf C:N/C:P ~ Fungal biomass, etc.). We hypothesized that

aquatic fungi play a central role in mediating the effects of nutrients on leaf breakdown due to

their direct positive effects on shredders, and positive indirect effects on shredders due to

increased microbially mediated litter nutrient content.

Fig. 2.2a,b. The best supported models for PRE, YR1 and YR2 relating N (a) or P (b)

concentrations to drivers of litter breakdown rates. Standardized path coefficients are reported,

and the sign of the coefficient indicates the direction of the correlation between variables. The

models explained 57% of the variation in litter breakdown rates. Weights of the arrows

correspond to path coefficients adjusted based on standard deviations, with strength of the

correlations indicated by arrow width. Small, medium, large, and extra-large arrows denote

adjusted coefficients <0.15, <0.30, <0.45, and >0.45, respectively. Dashed arrows indicate non-

significant path coefficients.

51

Fig. 2.1.

Shredders'

Nutrients'

Fungi'

Li0er'C:N/C:P'

Abio8c'Factors'

;'

+

"#

+#

"

Temperature' Discharge'

+#

Li#er&breakdown&

"#

+#

+

+#

52

Fig. 2.2a,b.

a. b.

Li#er&Breakdown&

Shredder&Biomass&

DIN&

Discharge&

Li3er&C:N&

Fungal&Biomass&:0.48&

0.21&

0.45&

:0.34&

:0.60&

0.23&

R2&=&0.57&

R2&=&0.36&

R2&=&0.36&

R2&=&0.20&0.18&

0.22&

:0.30&Li#er&Breakdown&

Shredder&Biomass&

SRP&

Discharge&

Li3er&C:P&

Fungal&Biomass&:0.52&

0.25&

0.59&

:0.43&

:0.50&

:0.29&

0.18&

R2&=&0.57&

R2&=&0.39&

R2&=&0.51&

R2&=&0.34&0.18&

0.20&

:0.21&

53

CHAPTER 3

CONVERGENCE OF DETRITAL STOICHIOMETRY PREDICTS THRESHOLDS OF

NUTRIENT-STIMULATED BREAKDOWN IN STREAMS2

2David W. P. Manning, Amy D. Rosemond, Vladislav Gulis, Jonathan P. Benstead, John S. Kominoski, and John C. Maerz. Submitted to Ecological Applications.

54

Abstract. Nutrient enrichment of detritus-based aquatic ecosystems increases detrital resource

quality for consumers and stimulates breakdown rates of particulate organic carbon (C). The

relative importance of dissolved inorganic nitrogen (N) vs. phosphorus (P) for detrital quality

and their effects on microbial- vs. detritivore-mediated detrital breakdown are poorly understood.

Here, we tested effects of experimental N and P additions on detrital stoichiometry (C:N, C:P)

and total and microbial breakdown (i.e., with and without detritivorous shredders, respectively)

of five substrates (four leaf litter species and wood) that differed in initial C:nutrient content. We

enriched five headwater streams continuously for two years at different relative availabilities of

N and P and compared breakdown rates and detrital stoichiometry to pretreatment conditions.

Breakdown rates increased with nutrient enrichment and were predicted by altered detrital

stoichiometry. Streamwater N and P, fungal biomass, and their interactions affected conditioned

detrital stoichiometry. Streamwater N and P both had significant effects on detrital C:N, while

streamwater P had stronger effects on detrital C:P. Nutrient addition and microbial effects

reduced C:N by 70% and C:P by 83% on average after conditioning, compared to only 26% for

C:N and 10% for C:P under pretreatment conditions; substrates with highest initial C:nutrient

content changed the most. Detrital stoichiometry was reduced and homogenized by nutrient

enrichment. Values of detrital nutrient content approached detritivore nutritional requirements,

and corresponded to greater consumptive effects of detritivores on litter breakdown with nutrient

enrichment. We used breakpoint regression to estimate values of detrital stoichiometry that can

be used to indicate elevated breakdown rates. Breakpoint ratios for total breakdown were 41

(C:N) and 1518 (C:P), coinciding with total breakdown rates that were ~1.9× higher when C:N

or C:P fell below these breakpoints. Microbial and shredder-mediated breakdown rates both

increased when C:N and C:P were reduced, suggesting that detrital stoichiometry is useful for

55

predicting litter breakdown dominated by either microbial or shredder activity. Our results show

strong effects of nutrient enrichment on detrital stoichiometry and offer a robust link between a

potential holistic metric of nutrient loading (decreased and homogenized detrital stoichiometry)

and increased C loss from stream ecosystems.

Introduction

Nutrient pollution from nonpoint sources affects aquatic ecosystems worldwide

(Carpenter et al. 1998). Among aquatic ecosystems, stream ecosystems are particularly

vulnerable to the effects of nutrient pollution from nonpoint sources due to their connection to

the surrounding landscape via groundwater and/or surface water runoff (Mulholland 1992,

Sudduth et al. 2013). As a result of human activities such as urbanization and agriculture,

nutrient concentrations (i.e., nitrogen [N] and phosphorus [P]) in streams have increased

dramatically in recent decades (Alexander and Smith 2006), contributing to degradation of water

quality (Brown and Froemke 2012). Increased availability of nutrients can have fundamental

effects on ecosystem-level processes, such as macroinvertebrate production (Cross et al. 2006)

and organic matter breakdown (Ferreira et al. 2015). However, nutrient effects on detritus-based

pathways are typically outside of the scope of nutrient pollution management practices, which

tend to rely on metrics related to primary production (e.g., Evans-White et al. 2013).

Particulate organic carbon (C) is a crucial energy base for stream ecosystems (Wallace et

al. 1997, Hall et al. 2000), and nutrient loading from watershed sources can alter its availability

(Kominoski and Rosemond 2012). Recent evidence suggests that increased nutrients can rapidly

deplete terrestrially derived C (hereafter, detritus; Rosemond et al. 2015), but development of

robust relationships between added nutrients and the mechanisms driving detrital breakdown are

56

lacking. In contrast, relationships between nutrients and production of algal C or stream

periphyton community structure are more established (e.g., Elser et al. 2007, Taylor et al. 2014).

A potential indicator of nutrient pollution effects in stream ecosystems is the rate of detrital

breakdown (Gessner and Chauvet 2002). However, stressors that occur in parallel with excessive

nutrient loading (e.g., toxic pollutants) may reduce a key trophic guild, detritivores (e.g.,

shredding macroinvertebrates, hereafter shredders), causing inconsistent effects of nutrient

concentrations on detrital breakdown, and impeding its use as an indicator of ecosystem health in

streams affected by multiple stressors (Woodward et al. 2012). Stronger mechanistic links

between consistent nutrient-mediated changes to ecosystem structure and function are needed to

adequately assess integrity of detritus-based aquatic ecosystems (Palmer and Febria 2012).

Detrital breakdown in streams is driven by interacting abiotic and biotic factors,

including abrasion, microbial conditioning and consumption by shredders (Tank et al. 2010). In

terrestrial and aquatic systems, detrital breakdown rates can be partially explained by initial

detrital quality (i.e., C:N, or C:P), with slower rates of decomposition for detritus that is nutrient-

poor (i.e., high C:N, C:P; Cornwell et al. 2008, Hladyz et al. 2009). Nutrient enrichment can

increase the nutrient content of detritus (Rosemond et al. 2010, Scott et al. 2013, Prater et al.

2015), potentially homogenizing formerly diverse detrital resources in terms of elemental

stoichiometry (i.e., C:N or C:P). These effects occur through microbial colonization and biomass

accrual on detritus and associated immobilization of dissolved inorganic nutrients (Suberkropp

1995, Cheever et al. 2012, Tant et al. 2013, Mehring et al. 2015), such that wide-ranging detrital

C:nutrient content becomes increasingly homogenous via microbial conditioning (e.g., Wickings

et al. 2012; Fig. 1).

57

Reduction and homogenization of detrital C:nutrient content driven by microbial

decomposers in streams (especially aquatic hyphomycetes; Findlay et al. 2002) can affect the

activity of shredders. Specifically, microbial enhancement of detrital nutrient content can cause

detrital C:N or C:P to more closely match the shredder threshold elemental ratio (TER), or the

point at which the consumer switches from nutrient limitation to C limitation (Sterner and Elser

2002). The implication of such reduced imbalances between detritus and detritivore TERs is

increased consumption (Cornut et al. 2015), and/or growth by individuals (Cornut et al. 2015,

Fuller et al. 2015, Halvorson et al. 2015), and potentially increased reproduction and survival at

the population level (e.g., Danger et al. 2013). Combined, these effects of increased nutrient

content on shredder growth, consumption and survival could affect the rate at which detritus is

processed in streams. However, the association between nutrient-stimulated changes to detrital

C:N and C:P and breakdown rates remains poorly characterized, particularly with respect to

differential effects of streamwater N vs. P on detrital C:nutrient content and subsequent

microbial vs. shredder-mediated effects on breakdown rates.

Our objective was to assess if detrital C:nutrient stoichiometry affected by nutrient

pollution could be linked to increased microbial and shredder-mediated detrital breakdown rates.

We explored these relationships using multi-year, whole-ecosystem enrichments of five

headwater streams with a crossed gradient of N and P concentrations. We used a set of five

substrates (four leaf litter species and wood veneers) that spanned a wide range of initial C:N and

C:P ratios. We predicted that nutrient enrichment would enhance microbial conditioning and

stoichiometric homogenization of detritus (Fig. 1), that P would be more important for

decreasing C:P, and that N would be more important for reducing C:N. We also predicted that

microbial effects on detrital stoichiometry would produce C:nutrient content that approached

58

estimated shredder TERs and thereby facilitate shredder-mediated breakdown (Frost et al. 2006,

Tant et al. 2013). If shredder activity is stimulated by such convergence of detrital stoichiometry

toward shredder TERs, we would also expect breakdown rates to exhibit threshold responses at

or around detrital C:N and/or C:P ratios that matched detritivore TERs.

Methods

Site description and experimental design

Our study took place at Coweeta Hydrologic Laboratory (CWT), a USDA Forest Service

research station and Long Term Ecological Research (LTER) site located in the southern

Appalachian Mountains in Macon County, North Carolina, USA (see Swank and Crossley 1988).

We selected five 70-m reaches in first-order streams within the Dryman Fork catchment to

receive continuous dosing of N and P for two years. The streams were similar physically and

chemically, and contained similar abundance and biomass of shredders prior to nutrient additions

(based on analysis of similarity [ANOSIM]) of shredder community data from litterbags [see

below]; ANOSIM R = -0.005, 0.004, respectively; P > 0.05 in both cases). Following a year of

pretreatment data collection, we began dosing the entire length of each 70-m reach on 11 July

2011 with solutions of ammonium nitrate (NH4NO3) and phosphoric acid (H3PO4) using solar-

powered metering pumps (LMI Milton Roy, Ivyland Pennsylvania, USA) connected to gravity-

fed irrigation lines supplied with streamwater. The dosing system in each stream was programed

to be proportional to discharge measured continuously using pressure transducers (Keller

America, Newport News, Virginia, USA) and CR800 dataloggers (Campbell Scientific, Logan,

Utah, USA). Dripper spouts were placed ~5 m apart along the irrigation line to ensure adequate

mixing along each 70-m reach.

59

Each stream reach received a unique concentrated solution of N and P to target five

increasing concentrations of N (added + background = 81, 244, 365, 488, 650 µg/L as dissolved

inorganic nitrogen [DIN]) and corresponding decreasing concentrations of P (added +

background = 90, 68, 51, 33, and 11 µg/L as soluble reactive phosphorus [SRP]), resulting in a

unique target N:P ratio for each stream (2, 8, 16, 32 and 128 respectively). Multiple streamwater

samples (n = 4; collected and analyzed as described below) were taken every ~15 m along each

of the 70-m reaches on days 1, 4, 7, 14, 23, 29 and 34 of enrichment to confirm adequate mixing

of added nutrients. After day 34 of enrichment, streamwater was collected above (n = 1) and

below (n = 3, at 10, 17 and 70-m) the nutrient dosing system biweekly, filtered in the field (0.45-

µm nitrocellulose membrane filters; Millipore), frozen and analyzed for DIN (NH4-N + NO3-N)

and SRP concentrations within 28 d (Alpkem Rapid Flow Analyzer 300 for DIN,

spectrophotometric method with Shimadzu UV-1700 for SRP). Two-year average measured DIN

concentrations during enrichment (83, 198, 330, 363, and 309 µg/L; all values respective to

treatment targets above) were close to but lower than target concentrations. Two-year average

measured SRP concentrations (49, 55, 36, 22, and 7 µg/L) were also close to but lower than

target concentrations. Measured concentrations reflected the effects of uptake, so we also

calculated the actual amounts of DIN and SRP that were added to the streams; the latter values

were used for the analyses presented here. These enriched concentrations were determined using

the quantity of nutrients added to each stream based on detailed nutrient solution refill records,

total daily discharge, and measured background concentrations from samples collected above the

nutrient dosing system. Further details about the experimental design, infrastructure, and stream

physicochemical characteristics can be found in Rosemond et al. (2015) and Manning et al.

(2015).

60

Detrital breakdown rates

We determined breakdown rates of red maple (Acer rubrum L.), tulip poplar

(Liriodendron tulipifera L.), chestnut oak (Quercus prinus L.), rhododendron, (Rhododendron

maximum L.), and white oak wood veneers (Quercus alba L.) prior to nutrient enrichment (PRE),

and during two consecutive years of nutrient addition (YR1, YR2). We collected freshly abscised

leaf litter of each type during peak leaf fall (October 2010, 2011 and 2012). Collected litter was

air-dried for several weeks in the laboratory. Air-dried litter was then weighed into 10±0.1 g

litter packs and placed into 5-mm plastic mesh bags to allow macroinvertebrate access (22 × 40

cm, Cady Bag Inc., Pearson, Georgia, USA). To prevent macroinvertebrate access to litter, we

placed additional 0.5–1-g litter packs into 0.5-mm mesh bags (20-cm right triangles, Industrial

Netting Inc., Minneapolis, Minnesota, USA) within corresponding coarse-mesh litterbags. Initial

mass of the litter in coarse- or fine-mesh litterbags was determined to the nearest 0.01 g. Wood

veneers were cut into ~2.5 × 15-cm pieces and weighed to the nearest 0.01 g. Three veneers were

fastened to 12 × 17-cm nylon gutter mesh rafts with cable ties. Seven sets of litterbags of each

species and two wood veneer rafts were anchored into four 17.5-m sub-reaches within the 70-m

experimental reach in each of the five streams (n = 7 sampling dates × 4 sub-reaches × 5 streams

= 140 per year for leaf litter types; n = 3 sampling dates × 4 sub-reaches × 2 rafts × 5 streams =

120 per year for wood veneers). Thus, the number of substrates sampled was 140 litterbags × 4

litter species + 120 wood veneers per year for a total of 2040 substrates used during the three

years of the study.

After incubation in the stream began (day 0 = 1 December 2010, 27 November 2011, and

29 November 2012 for PRE, YR1, and YR2, respectively), litterbags were collected on seven

dates (PRE = day 7, 14, 21, 70, 109, 160, 187; YR1, YR2 = day 7, 14, 21, 34, 55, 63, 77 [maple,

61

poplar] or day 7, 14, 21, 63, 110, 126, 143 [oak, rhododendron]), placed into individual plastic

bags, and transported to the laboratory on ice. Wood veneers were placed in the streams at the

same time as litterbags and collected on day 21, 109 and 160 (PRE) and day 21, 109 and 143

(YR1, YR2). For both leaf litter and wood veneers, sampling dates differed in PRE vs. YR1 and

YR2 because of accelerated breakdown during nutrient enrichment. Within 24 h, the litter was

removed from the coarse-mesh bags, rinsed over nested 1-mm and 0.25-mm sieves, subsampled

for microbiological analyses (see below) and then dried for 24 h at 55°C. Litter in the fine-mesh

bags was rinsed over the nested sieves and dried while still inside the mesh bag to minimize loss

of litter material. Once dry, the litter from coarse-mesh bags was weighed to determine dry mass

remaining, and then ground with a ball mill (Spex Certiprep 8000D, Metuchen, New Jersey,

USA). A 1–2-g subsample was weighed and combusted for 4.5 h at 500°C to determine ash-free

dry mass (AFDM). The entire sample from each fine-mesh bag was weighed to determine dry

mass remaining and combusted in a similar manner. After subsampling for fungal biomass, wood

veneers were also dried for 24 h at 55°C, ground using a ball mill, and ~0.5 g subsamples

combusted to determine AFDM.

Detrital stoichiometry

We measured C:N and C:P at early, middle and late stages of leaf litter decay in an effort

to target litter with different levels of microbial conditioning. Sampling schedules for middle and

late stages of decay differed depending on litter species and year due to faster processing of

detritus under nutrient enrichment (PRE = day 14, 70, 160; YR1 and YR2 = day 14, 34, 77

[maple, poplar]; day 14, 63, 126 [oak, rhododendron]). Wood veneers were sampled for

stoichiometry at middle stages of decay (day 109) for all years. The majority of the analyses

62

presented here uses only detrital stoichiometry data from middle stages of decay to assess the

effect of nutrient concentrations on stoichiometric homogenization of detrital C:N and C:P after

substantial microbial conditioning. A 2–4-mg subsample of the dried and ground litter or wood

was used to determine conditioned litter %C and %N with a Carlo Erba 1500 CHN analyzer

(Milan, Italy). Phosphorus content of conditioned leaf litter and wood was determined using the

plant dry ash/acid extraction method (Allen 1974) followed by spectrophotometric analysis

(Shimadzu UV-1700, Japan) of the extracted solution using the ascorbic acid method (APHA

1998).

Fungal biomass

We analyzed ergosterol concentrations to estimate fungal biomass associated with leaf

litter and wood veneers during middle stages of decay for leaf litter (day 34 [YR1, YR2 for

maple, poplar], 63 [YR1, YR2 for oak, rhododendron], or 70 [all leaf litter during PRE], and day

109 [wood veneers]) that corresponded with the samples used for detrital stoichiometry (as

described above). We subsampled and froze ~2 × 2 cm pieces from rinsed leaf litter or wood

until analysis. Lipids were extracted from freeze-dried, weighed leaf litter and wood pieces using

liquid-to-liquid extraction. Ergosterol concentrations were determined by HPLC (LC-10VP,

Shimadzu, Columbia, Maryland, USA) equipped with a Kinetex C18 column (Phenomenex,

Torrance, California, USA) and a UV detector set at 282 nm (Gulis and Suberkropp 2006). We

used external ergosterol standards (Acros Organics, Geel, Belgium), and ergosterol

concentrations were converted to fungal biomass using a standard conversion factor of 5.5 µg of

ergosterol per mg of fungal dry mass (Gessner and Chauvet 1993).

63

Data analysis

Total leaf litter and wood breakdown rates (i.e., microbial + shredder = ktotal) were

estimated based on the percent litter or wood mass remaining over time according to the

negative-exponential model (Petersen and Cummins 1974):

𝑚! = 𝑚!𝑒!!",

where mt is the detrital mass remaining at time t, m0 is the initial litter or wood mass, and k is the

breakdown rate. We determined a unique decay rate k for each leaf litter type and wood veneers

from each sub-reach within the 70-m experimental reach (n = 4 sub-reaches × 5 streams × 5

detritus types × 3 years = 300). Breakdown rates of litter in fine-mesh bags were determined in

the same way, and we considered these estimates to reflect microbial breakdown rates (kmicrobe).

We determined shredder contributions to breakdown rates by subtracting kmicrobe from total

breakdown rates (i.e., ktotal – kmicrobe = kshredder, Woodward et al. 2012).

All analyses were conducted using the statistical software R v. 3.0.1 (R Core

Development Team 2013). Response variables were ln-transformed when appropriate to meet

assumptions of normality or linearity. We used a linear model with categorical predictor

variables (year, detritus type were the predictors) to test for overall nutrient enrichment effects

on detrital C:N and C:P, by assessing differences in means (intercepts in this case) between years

and detritus types (i.e., analysis of variance [ANOVA]). We then used a linear model with

continuous predictors to assess the effects of N and P concentrations and fungal biomass on

detrital C:N and C:P, based on the assumption that N and P effects would be additive, and fungal

biomass effects on detrital C:N and C:P would interact with nutrient concentrations (e.g.,

Kominoski et al. 2015, Manning et al. 2015). This model also included a categorical predictor for

detritus type, such that we tested for differences in mean C:N or C:P among the five detritus

64

types (i.e., differences in the intercept for the detritus). Therefore, our model predicting detrital

C:N and C:P included N and P concentration, fungal biomass, their interactions and detritus type.

We standardized predictor variables using z-scores to compare predictor variables measured at

different scales, and to aid the interpretation of interactions between continuous predictors (i.e.,

N and P concentrations and fungal biomass; Gelman and Hill 2007). Slope coefficients for this

model can be interpreted as the expected change in detrital C:N or C:P for an increase in N, P or

fungal biomass concentration by one standard deviation. Because we centered our data, main

effects in the models can be interpreted as the predicted effect of a given parameter on detrital

C:N or C:P for detritus with mean fungal biomass, or in streamwater with mean N or P

concentration.

We analyzed the contributions of microbial and shredder-mediated breakdown rates (i.e.,

kmicrobe vs. ktotal - kmicrobe = kshredder; wood k was excluded for this analysis) with respect to nutrient

enrichment (i.e., by year) and detritus type. The relative contribution of microorganisms vs.

shredders can be determined based on the ratio of kshredder/kmicrobe, where a ratio < 1 indicates

greater microbial contributions in this case (modified from Gessner and Chauvet 2002, which

used ktotal/kmicrobe). We tested for differences in contributions of microbial vs. shredder-mediated

breakdown using ANOVA, with year as the main predictor of ln-transformed kshredder/kmicrobe.

Finally, we used breakpoint regression to pinpoint possible thresholds in the relationship

between leaf litter breakdown and detrital C:N or C:P (Dodds et al. 2010). Breakpoint regression

allows for the estimation of piece-wise linear relationships, and can be used to pinpoint the value

where the linear relationship between two variables changes (Muggeo 2003). We specifically

used these breakpoint models to identify the point at which the relationship between C:N or C:P

and leaf litter breakdown changed significantly. We then compared the breakpoint C:N and C:P

65

values identified by the models to previously reported shredder TERs (e.g., Tant et al. 2013).

Specifically, we compared breakpoints ad hoc to the values reported in Tant et al. (2013), which

was specific to a dominant shredder in our system (larvae of a caddisfly in the genus

Pycnopsyche, TER C:N = 27, C:P = 1992), and are also in the range of reported values for

detritivores from Frost et al. (2006; mean TER CP: = 1187, ± one standard deviation = 493 to

2440 for detritivores) and Halvorson et al. (2015; Pycnopsyche lepida TER C:P = 1620). We

used the package ‘segmented’ in R (Muggeo 2003) to analyze our data for possible breakpoints

that could correspond to shredder TERs.

Results

Nutrient enrichment effects on detrital stoichiometry

Decreases in detrital C:N and C:P after a period of microbial conditioning were enhanced

by nutrient enrichment (Appendix A: Table A1, Fig. 2). In the pretreatment year, conditioning

resulted in relatively small changes in stoichiometry. Specifically, in the pretreatment year, leaf

litter C:N and C:P were reduced 30% and 24%, respectively, and wood C:N and C:P were

reduced 11% and increased 47%, respectively, after a period of colonization (Appendix A: Table

a1). In contrast, leaf litter C:N was reduced by ~50%, and leaf litter C:P was reduced by ~60%

after conditioning under nutrient-enrichment (YR1 and YR2). Wood exhibited the sharpest

reductions in C:N and C:P (compared to initial ratios) under nutrient-enriched conditions, and

was reduced by >70% for both C:N and C:P (Appendix A: Table A1).

Greater reductions in C:N and C:P during the period of microbial conditioning also

resulted in pronounced differences in detrital stoichiometry between the pretreatment and

nutrient-enriched years for similar stages of decay. Conditioning under nutrient enrichment led to

66

27% and 29% reductions in mean detrital C:N for YR1 and YR2 compared to pretreatment

values across leaf litter species (Appendix A: Table A1). For wood, nutrient enrichment led to

66% and 73% reductions in mean detrital C:N for YR1 and YR2 compared to pretreatment.

Significantly greater reductions in mean detrital C:N were observed for poplar, rhododendron,

and wood veneers compared to red maple leaf litter in YR1 and YR2 (Appendix A: Table A1).

Detrital C:P was reduced to a greater degree than C:N during YR1 and YR2 (Appendix

A: Table A1, Fig. 2). Conditioning under nutrient enrichment led to 52% and 48% reductions in

mean C:P for YR1 and YR2 compared to pretreatment values across leaf litter species (Appendix

A: Table A1). Nutrient enrichment led to 80% and 93% reductions in wood C:P compared to

pretreatment. Significantly greater reductions in mean detrital C:P were observed for wood

veneers and rhododendron leaf litter compared to red maple leaf litter, but not other leaf litter

species in YR1 and YR2 (Appendix A: Table A1).

Microbial and streamwater nutrient effects on stoichiometric homogenization of detritus

The effects of N and P concentrations, fungal biomass, their interactions, and detritus

type explained 54% and 58% of the variation in conditioned detrital C:N and C:P, respectively,

based on adjusted R2 values from each model. The effects of fungal biomass on detrital C:N and

C:P were dependent on N and P availability (Table 1); specifically, there was a positive

association between added N and P concentrations, fungal biomass, and detrital stoichiometry,

corresponding to stronger negative effects of fungal biomass on detrital C:N and C:P with

increased N and P concentrations. The combined effects of N and P on detrital C:N and C:P

ratios were comparable to the effects of fungal biomass alone based on the scaled coefficients in

67

the models. That is, the N and P effects on detritus with mean fungal biomass were comparable

to fungal biomass effects given mean N and P concentrations (Table 1).

The C:N stoichiometry of leaf litter and wood was driven by fungal biomass, streamwater

N, streamwater P and the interaction between fungal biomass and streamwater N. Based on

comparisons among the scaled coefficients, the strength of the effect of streamwater N on detrital

C:N was weaker than the effect of streamwater P on detrital C:N for detritus with mean fungal

biomass (Table 1). In general, the most N-poor substrates (i.e., rhododendron, wood) tended to

gain the most N and showed significant decreases in C:N compared to relatively nutrient-rich

detritus (e.g., red maple, Appendix A: Table A2), although poplar also showed decreased C:N

compared to maple. As a result, differences between rhododendron and wood C:N vs. maple,

poplar and oak C:N were reduced under nutrient-enriched conditions (Appendix A: Table A3).

For example, mean wood C:N was ~2.9× greater than maple C:N under pretreatment conditions,

whereas mean wood C:N was essentially equivalent to maple C:N (~1.0×) under nutrient-

enriched conditions (Appendix A: Table A3). Likewise, the range of C:N values was reduced by

65% under nutrient-enriched conditions; C:N values of all five detritus types spanned a range

between 27-133 compared to 37-337 under pretreatment conditions. Across our N:P treatments,

we found that the greatest reductions in detrital C:N occurred when dissolved N:P ≥ 32, but the

differences between reduced C:N values among streams were small (mean C:N = 48 vs. C:N =

43 for treatment N:P = 2 and 128, respectively).

The C:P stoichiometry of leaf litter and wood was also driven by fungal biomass,

streamwater P, and the interaction between fungal biomass and streamwater N and P (Table 1).

The strongest driver of conditioned detrital C:P ratios was fungal biomass, followed by nearly

equal control by streamwater P (Table 1). The most P-poor substrates (i.e., rhododendron, wood)

68

gained the most P, leading to sharp decreases in C:P compared to more nutrient-rich substrates

(e.g., maple, Appendix A: Table A2). As a result, differences between rhododendron and wood

C:P vs. maple, poplar and oak C:P were reduced under nutrient-enriched conditions. For

instance, mean wood C:P was ~6.5× greater than mean maple C:P under pretreatment conditions,

whereas wood C:P was only ~1.5× higher than maple C:P under nutrient-enriched conditions,

(Appendix A: Table A3). Likewise, the range of C:P values was reduced by 60% under nutrient-

enriched conditions; C:P values of all five detritus types spanned a range between 669-19,956

compared to 1042-49,496 under pretreatment conditions. Across our N:P treatments, we found

that the differences in litter stoichiometry values among streams were larger for C:P compared to

C:N (mean C:P =1205 vs. C:P = 2705 for treatment N:P = 2 and 128, respectively).

Effects of nutrient enrichment and initial stoichiometry on detrital breakdown rates

Reduced and homogenized detrital stoichiometry corresponded to increased detrital

breakdown rates, especially for ktotal. Across litter types and years, total breakdown rates were

~2.5× pretreatment rates, whereas microbial breakdown rates in YR1 and YR2 were only ~1.4×

pretreatment rates (Appendix A: Table A4). Similar to reductions in detrital C:nutrient ratios,

increases in total breakdown rates were more pronounced for recalcitrant leaf litter species:

average ktotal for oak and rhododendron was 2.4 and 3.9× higher during nutrient enrichment,

compared to 1.8 and 2.3× higher breakdown rates for maple and poplar, respectively (Appendix

A: Table A4). We found a positive, linear relationship between the magnitude of the increase in

breakdown rate under nutrient enrichment (mean YR1 and YR2 ktotal / PRE ktotal; Fig. 3a,b,

Appendix A: Table A4) and the initial C:N or C:P of the detritus, such that the most pronounced

69

increases in breakdown rates relative to PRE were for the most nutrient-poor substrates (i.e.,

highest C:N, C:P, Fig. 3a,b).

Effects of shredders on detrital breakdown rates

We found that shredder contributions increased more than microbial contributions to

breakdown in YR1 and YR2 based on the ln-transformed ratio, kshredder/kmicrobe (Fig. 4a-d). We

tested for differences in shredder contributions between years for each detritus type because we

found no evidence for interactions between year and detritus type, or differences among streams

in terms of the contribution of shredders vs. microorganisms to breakdown rates. We found

significantly increased shredder contributions to breakdown for all leaf litter types in all years,

with the exception of rhododendron in YR2 (Tukey’s HSD, all P < 0.05, Fig. 4a-d). Maple,

poplar and oak switched from greater relative microbial vs. shredder contribution to breakdown

under pretreatment conditions to greater relative shredder contribution in YR1 and YR2 (i.e., ln

kshredder/kmicrobe > 0, Fig. 4a-c). For rhododendron, shredder contributions were greater than for

other litter types under pretreatment conditions; shredder contributions increased in YR1, but not

in YR2, compared to pretreatment conditions (Fig. 4d).

Identifying stoichiometrically explicit breakpoints for detrital breakdown rates

Breakpoints were found in the relationship between total breakdown rates and C:N

stoichiometry (Fig. 5a; P < 0.05, Adj-R2 = 0.25) of decaying leaf litter. The identified breakpoint

for litter C:N and total litter breakdown was 41 (±2), which represents a 56% decrease from

initial C:N averaged across leaf litter species. Total breakdown rates differed above and below

breakpoints for C:N. Total breakdown rates were 1.8× higher when C:N was below the

70

breakpoint (Fig. 5a). There were also significant breakpoints in the relationship between total

litter breakdown rates and litter C:P stoichiometry (Fig. 5b; P < 0.05, Adj-R2 = 0.31). The

breakpoint for C:P and total litter breakdown was 1518 (±167), which represents a 65% decrease

from initial C:P averaged across leaf litter types. As with C:N, total breakdown rates differed

above and below breakpoints for C:P. Total breakdown rates were 1.9× higher when C:P was

below the breakpoint (Fig. 5b). These breakpoint C:N and C:P ratios for total breakdown rates

differed considerably from mean leaf litter C:N (63) and C:P (3167) values for pretreatment

conditions, while mean C:N (46) and C:P (1453) during enriched conditions were more similar

to breakpoint C:N and C:P for total breakdown rates. Mean C:N and C:P of leaf litter as well as

the breakpoint ratios were comparable to TERs for C:N (27) and C:P (1992) of an important

shredder taxon in our study system (caddisfly larvae in the genus Pycnopsyche, Tant et al. 2013),

and to the range of TER C:Ps reported for detritivores generally (mean TER CP: = 1187, ± one

standard deviation = 493 to 2440; Frost et al. 2006).

Discussion

Linking resource structure and ecosystem function is a crucial step toward developing

robust metrics of ecosystem integrity (Gessner and Chauvet 2002, Palmer and Febria 2012),

particularly with regard to detecting and predicting responses to widespread ecosystem stressors

such as nutrient pollution. Increases in streamwater nutrient concentrations are known to increase

litter and wood nutrient content (Stelzer et al. 2003, Gulis et al. 2004); here, we explicitly linked

these stoichiometric changes to causative experimental gradients of N and P and resulting

ecosystem function (i.e., increased rates of detrital breakdown).

71

Our study targeted low-to-moderate nutrient enrichment common in human-modified

landscapes (Scott et al. 2002, Alexander and Smith 2006), but these concentrations were

sufficient to induce decreased detrital stoichiometry and increased breakdown rates. Thus, we

suggest standardized, nutrient-poor detrital substrates (e.g., wood veneers) could be used to

indicate nutrient pollution and its effects on an ecosystem function. For instance, we found that

after ~100 days of conditioning, wood veneers had C:N and C:P that was 73% and 80% lower

than initial under nutrient-enriched conditions, compared to only 10% lower than initial for C:N

and no difference for C:P under pretreatment conditions. Thus our data suggest that increased

rates of ecosystem-scale C loss may be expected if detrital C:N and C:P are found to be

substantially reduced (e.g., >70%) compared to corresponding initial C:N and C:P. These

findings offer a potentially significant and useful metric of nutrient-mediated changes to a

critical basal resource that can be used to predict where and when accelerated C loss might occur

in stream ecosystems in response to nutrient pollution. More studies at the level of stream

mesocosms and/or whole streams are needed to further develop the predicted linkage we make

here between changes in nutrient content of detritus and detrital loss rates.

Streamwater nutrient and microbial effects on detrital stoichiometry

Nutrient loading from distinct land use can result in skewed N and P availability in

streams, thus, the concentrations and ratios we used mimicked such patterns of N and P loading

(e.g., high N:P from N-rich fertilizers vs. low N:P from sewage effluent; Arbuckle and Downing

2001, Peñuelas et al. 2012). Detrital C:N and C:P were reduced across our N:P treatments, but

there were differences in this response when comparing detrital C:N and C:P. For example, our

findings suggest that C:N can be affected by increases in either N or P concentrations, as

72

evidenced by similar C:N values across our treatment N:P ratios. In contrast, P effects were

stronger than N effects on C:P, indicating that changes in detrital C:P are more dependent on the

availability of P when N is also available. These different responses of C:N and C:P to

streamwater N compared to P could be important in the context of detritivore nutrient limitation

when N and P loading are skewed. Several studies have observed both N- and P-limitation of

decomposers (Rosemond et al. 2002, Ferreira et al. 2006) and detritivores (Danger et al. 2013,

Frainer et al. 2015). Thus, our findings suggest that N-limitation via detrital resources could be

alleviated across elevated N or P concentrations, while P-limitation could be alleviated more

when N:P loading is skewed toward lower N:P ratios.

Increased microbial biomass has been implicated as a driver of increased detrital nutrient

content (Gulis and Suberkropp 2003, Tant et al. 2013). Our study found that streamwater nutrient

availability and fungal biomass together control detrital C:nutrient stoichiometry. Specifically,

the effects of streamwater N and P on detrital C:N and C:P were dependent on fungal biomass,

but streamwater P concentration effects were stronger than the interaction of streamwater P with

fungal biomass. These results suggest that while DIN effects on detrital C:N are particularly

dependent on fungal biomass accrual, streamwater P appears to control detrital C:P to a greater

degree, as fungi can likely store P without considerable increases in biomass when streamwater P

availability is elevated (Gulis et al., unpublished data). Aquatic fungi typically contribute >95%

of microbial biomass on coarse detritus such as leaf litter and wood (Gessner et al. 2007), and

thus were likely the major drivers of stoichiometric changes of detritus in this study. Detrital

stoichiometry could be affected in a similar way in other systems where fungal decomposers are

predominant (e.g., soils/terrestrial detritus; Barantal et al. 2014). However, we cannot rule out

the potential effects of bacteria (but see Tant et al. 2013), abiotic sorption (e.g., Mehring et al.

73

2015), or differential allocation of resources to fungal sporulation in response to N vs. P

availability in affecting detrital stoichiometry.

Shredder-driven effects on leaf litter breakdown rates

We found that relatively recalcitrant detritus showed greater responses to streamwater

nutrients in terms of reduced C:nutrient content and that shredders contributed more to the

breakdown of all four leaf litter species under nutrient enrichment. These findings combined with

our breakpoint ratios for increased total breakdown rates suggest that reduced C:N and C:P of

detritus is an important driver of increased breakdown rates, particularly when nutrient-poor

resources approach shredder nutrient requirements. Thus, the effects of nutrient enrichment on

detrital breakdown could be amplified in systems characterized by intact shredder communities

and nutrient-poor detritus. Although the presence of shredders may increase nutrient-enrichment

effects on detrital breakdown, our analysis demonstrated that microbial breakdown rates also

increased. These findings imply that, across diverse systems with and without detritivores,

nutrient enrichment will predictably increase detrital C loss rates.

Breakpoints for detrital breakdown

Shredder growth, reproduction and survival are related to the nutrient content of basal

resources consumed, particularly if the food resource meets shredder nutrient demands (i.e.,

approaches the TER; Danger et al. 2013, Halvorson et al. 2015). Our findings support the

existence of a causative link between 1) nutrient-mediated microbial processing that drives

detrital C:N or C:P toward shredder TERs and 2) subsequent stimulation of shredder activity

leading to increased litter breakdown rates. As a result of effects of streamwater N and P and

74

fungal biomass, greater than 78% of conditioned detrital C:P was found to be below the

estimated C:P TER for a dominant shredder in our study system (larvae of the caddisfly

Pycnopsyche; C:P TER = 1992; Tant et al. 2013) under enriched conditions, compared to 17%

falling below this threshold during pretreatment. Other detritivores can exhibit a range of C:P

TERs depending on diet or taxon-specific physiology (e.g., larvae of the cranefly Tipula C:P

TER = 1000-2500, Fuller et al. 2015) and other reported values of shredder TERs tend to fall in

the range of reduced detrital stoichiometry found in this study (Frost et al. 2006). However, it is

known that some shredders selectively consume and assimilate patches of detritus that contain

greater amounts of fungal biomass and nutrients (Arsuffi and Suberkropp 1985, Dodds et al.

2014), complicating the inferences made here based on bulk detrital resources. Nevertheless, we

show that nutrient enrichment did increase the prevalence of low C:nutrient detritus that could be

consumed by shredders, and that these patterns likely correspond to the breakpoints we have

identified in the relationship between total litter breakdown rates and detrital C:N and C:P

stoichiometry.

Breakpoints in litter breakdown rates were also observed when detritivores were

excluded from litter using fine-mesh litterbags (data not shown). In this case, large changes in

detrital stoichiometry indicate increased colonization of detritus by microorganisms.

Consequently, changes in detrital stoichiometry may indicate increased detrital loss from either

decomposers and/or detritivores and thus be applicable to a range of stream types and conditions.

Thus, we expect that in areas where shredder abundance is low due to biogeographic factors

(Boyero et al. 2011) or stressors (Griffiths et al. 2009, Woodward et al. 2012), microbially driven

breakdown will be more important and detrital loss rates would be increased with nutrient

enrichment unless decomposers are suppressed by associated contaminants.

75

Comparing nutrient-induced changes to detrital vs. biofilm stoichiometry

We found that changes in detrital stoichiometry were induced at low-to-moderate nutrient

concentrations. The changes in detrital stoichiometry we observed were similar to changes in

biofilm stoichiometry across gradients of N and P concentrations (Taylor et al. 2014).

Specifically, Taylor et al. (2014) found that biofilm C:N decreased by 26% and biofilm C:P

decreased by 38% when total P increased from 10 to 20 µg/L. When comparing our data to this

range in P concentrations, we found similar responses: detrital C:N decreased by 10% and

detrital C:P decreased by 38%. These findings indicate that changes in basal resource

stoichiometry can occur due to effects on either autotrophic (e.g., biofilm) or heterotrophic

microbial communities and that they may respond to nutrient gradients similarly. Further, as a

result of our experimental nutrient additions, detrital stoichiometry that formerly spanned large

ranges also became more similar and approached TERs for certain detritivores. A similar pattern

occurred in the Taylor et al. (2014) study, where biofilms were more variable in terms of nutrient

content at low nutrient concentrations and more homogeneous at high nutrient concentrations.

This similar pattern suggests that the diversity in nutrient content of detritus or biofilms can be

reduced by nutrient enrichment, which may affect detritivore and biofilm consumer communities

(LeRoy and Marks 2006, Taylor et al. 2014). Given that diverse detritus or biofilm nutrient

content typically supports diverse consumer assemblages, nutrient-induced reduction and

homogenization of basal resource stoichiometry could potentially lead to low diversity of food

quality and associated reduced diversity of consumers (e.g., Evans-White et al. 2009). Thus,

nutrient enrichment may diminish stream consumer biodiversity related to either heterotrophic or

autotrophic food web pathways in some cases.

76

We identified relevant breakpoints for a key ecosystem function in response to microbial

homogenization of a key detrital resource trait under nutrient enrichment, a pattern similar to

observations made for stream biofilms. Thus, the response of detrital or autotrophic

stoichiometry could be an important tool for predicting ecosystem-scale consequences of nutrient

enrichment. Such assessments of changes in basal resource structure (i.e., reduced and

homogenized detrital stoichiometry associated with stream carbon loss) could be useful if

incorporated into management strategies for mitigating the effects of nutrient pollution on the

functioning of stream ecosystems.

Acknowledgements

We are grateful for the maintenance and sampling of the five study streams by Jason

Coombs and Katie Norris. Phillip Bumpers, Jason Coombs, Katie Norris, Kait Farrell, James

Wood, Tom Maddox, and Emmy Deng helped in the laboratory or in the field. This study

received support from the NSF (DEB-0918894 to ADR and JCM, DEB-0918904 to JPB, and

DEB-0919054 to VG). This study also leveraged logistical support from the CWT LTER

Program at the University of Georgia, which is supported by NSF award DEB-0823293 from the

Long Term Ecological Research Program (JCM co-PI). Rob Case, Daniel Hutcheson, and Kevin

Simpson of YSI Integrated Systems and Services constructed the infrastructure for the nutrient-

dosing system. Aqueous ammonium nitrate was provided by The Andersons, Inc. through David

Plank. We thank Phillip Bumpers, Alan Covich, Chao Song, Nina Wurzburger and two

reviewers for helpful comments that improved this manuscript.

77

Literature Cited

Alexander, R. B., and R. A. Smith. 2006. Trends in the nutrient enrichment of U. S. rivers during

the late 20th century and their relation to changes in probable stream trophic conditions.

Limnology and Oceanography 51:639-654.

Allen, S. E., et al. 1974. Chemical analysis of ecological materials. Wiley, New York, New

York, USA.

American Public Health Association (APHA). 1998. Standard Methods for the

Examination of Water and Wastewater. 20th edition. American Public Health

Association, Washington, D.C.

Arbuckle, K. E., and J.A. Downing. 2001. The influence of watershed land use on lake N:P in a

predominantly agricultural landscape. Limnology and Oceanography 46: 970-975.

Arsuffi, T. L., and K. Suberkropp. 1985. Selective feeding by stream caddisfly (Trichoptera)

detritivores on leaves with fungal-colonized patches. Oikos 45:50-58.

Barantal, S., H. Schimann, N. Fromin, and S. Hattenschwiler. 2014. C, N and P fertilization in an

Amazonian rainforest supports stoichiometric dissimilarity as a driver of litter diversity

effects on decomposition. Proceedings of the Royal Society B: Biological Sciences

281:20141682-20141682.

Brown, T. C., and P. Froemke 2012. Nationwide assessment of nonpoint source threats to water

quality. Bioscience 62:136-146.

Boyero, L., et al. 2011. Global distribution of a key trophic guild contrasts with common

latitudinal diversity patterns. Ecology 92:1839-1848.

78

Carpenter, S. R., N. F. Caraco, D. L. Correll, R. W. Howarth, A. N. Sharpley, and V. H. Smith.

1998. Nonpoint pollution of surface waters with phosphorus and nitrogen. Ecological

Applications 8:559-568.

Cheever B. M., E.B. Kratzer, J. R. Webster. 2012. Immobilization and mineralization of N and P

by heterotrophic microbes during leaf decomposition. Freshwater Science 31:133-147

Cross, W. F., J. B. Wallace, A. D. Rosemond, and S. L. Eggert. 2006. Whole-system nutrient enrichment

increases secondary production in a detritus-based ecosystem. Ecology 87:1556-1565.

Cornwell, W. K., et al. 2008. Plant species traits are the predominant control on litter

decomposition rates within biomes worldwide. Ecology Letters 11: 1065-1071.

Danger, M., J. Arce Funck, S. Devin, J. Heberle, and V. Felten. 2013. Phosphorus content in

detritus controls life-history traits of a detritivore. Functional Ecology 27:807-815.

Dodds, W. K., W. H. Clements, K. Gido, R. H. Hilderbrand, and R. S. King. 2010. Thresholds,

breakpoints, and nonlinearity in freshwaters as related to management. Journal of the

North American Benthological Society 29:988-997.

Dodds, W. K., et al. 2014. You are not always what we think you eat: selective assimilation

across multiple whole-stream isotopic tracer studies. Ecology 95:2757-2767.

Elser, J. J., M. E. S. Bracken, E. E. Cleland, D. S. Gruner, W. S. Harpole, H. Hillebrand, J. T.

Ngai, E. W. Seabloom, J. B. Shurin, and J. E. Smith. 2007. Global analysis of nitrogen

and phosphorus limitation of primary producers in freshwater marine and terrestrial

ecosystems. Ecology Letters 10:1135-1142.

Evans-White, M. A., W. K. Dodds, D. G. Huggins, and D. S. Baker. 2009. Thresholds in

macroinvertebrate biodiversity and stoichiometry across water-quality gradients in

Central Plains (USA) streams. Journal of the North American Benthological Society

28:855-868.

79

Evans-White, M. A., B. E. Haggard, and J. T. Scott. 2013. A review of nutrient criteria

development in the United States. Journal of Environmental Quality 42:1002-1014.

Ferreira V., V. Gulis, and M. A. S. Graça. 2006. Whole stream nitrate addition affects litter

decomposition and associated fungi but not invertebrates. Oecologia 149:718-729.

Ferreira V., B. Castagneyrol, J. Koricheva, V. Gulis, E. Chauvet, and M.A.S. Graça. 2015. A

meta-analysis of the effects of nutrient enrichment on litter decomposition in streams.

Biological Reviews 90:669-688.

Findlay, S., J. Tank, S. Dye, H. M. Valett, P. J. Mulholland, W. H. McDowell, S. L. Johnson, S.

K. Hamilton, J. Edmonds, W. K. Dodds, W. B. Bowden. 2002. A cross-system

comparison of bacterial and fungal biomass in detritus pools of headwater streams.

Microbial Ecology 43:55-66.

Frainer A., J Jabiol, M. O. Gessner, A. Bruder, E. Chauvet, and B. G. McKie. 2015.

Stoichiometric imbalances between detritus and detritivores are related to shifts in

ecosystem functioning. Oikos 000:001-011

Frost, P. C., J. P. Benstead, W. F. Cross, H. Hillebrand, J. H. Larson, M. A. Xenopoulos, and T.

Yoshida. 2006. Threshold elemental ratios of carbon and phosphorus in aquatic

consumers. Ecology Letters 9:774-779.

Fuller, C. L., M. A. Evans-White, and S. A. Entrekin. 2015. Growth and stoichiometry of a

common aquatic detritivore respond to changes in resource stoichiometry. Oecologia

177:837-848.

Gelman, A, and J. Hill. 2007. Data analysis using regression and multilevel/hierarchical models.

Cambridge University Press, Cambridge, UK.

80

Gessner, M. O., and E. Chauvet. 1993. Ergosterol-to-biomass conversion factors for aquatic

hyphomycetes. Applied and Environmental Microbiology 59:502-507.

Gessner, M. O., and E. Chauvet. 2002. A case for using litter breakdown to assess functional

stream integrity. Ecological Applications 12:498-510.

Gessner, M. O., V. Gulis, K. A. Kuehn, E. Chauvet, and K. Suberkropp. 2007. Fungal

decomposers of plant litter in aquatic ecosystems. Pages 301-324 in C. P. Kubicek, and I.

S. Druzhinina, editors. The Mycota. Springer. Berlin.

Gulis, V., and K. Suberkropp. 2003. Leaf litter decomposition and microbial activity in nutrient-

enriched and unaltered reaches of a headwater stream. Freshwater Biology 48:123-134.

Gulis, V., A. D. Rosemond, K. Suberkropp, H. S. Weyers, and J. P. Benstead. 2004. Effects of

nutrient enrichment on the decomposition of wood and associated microbial activity in

streams. Freshwater Biology 49:1437-1447.

Gulis, V., and K. Suberkropp. 2006. Fungi: Biomass, production, and sporulation of aquatic

hyphomycetes. Pages 311-326 in F.R. Hauer and G.A. Lamberti, editors. Methods in

Stream Ecology. Second Edition. Academic Press.

Hall, R. O., J. B. Wallace, and S. L. Eggert. 2000. Organic matter flow in stream food webs with

reduced detrital resource base. Ecology 81:3445-3463.

Halvorson, H. M., J. T. Scott, A. J. Sanders, and M. A. Evans-White. 2015. A stream insect

detritivore violates common assumptions of threshold elemental ratio bioenergetics

models. Freshwater Science 34:508–515.

Hladyz, S., M. O. Gessner, P. S. Giller, J. Pozo, and G. Woodward. 2009. Resource quality and

stoichiometric constraints on stream ecosystem functioning. Freshwater Biology 54:957–

970.

81

Kominoski, J. S., and A. D. Rosemond. 2012. Conservation from the bottom up: forecasting

effects of global change on dynamics of organic matter and management needs for river

networks. Freshwater Science 31:51-68.

Kominoski, J. S., A. D. Rosemond, J. P. Benstead, V. Gulis, J. C. Maerz, and D. W. P. Manning.

2015. Low-to-moderate nitrogen and phosphorus concentrations accelerate microbially

driven litter breakdown rates. Ecological Applications 25:856-865.

LeRoy, C. J., and J. C. Marks. 2006. Litter quality, stream characteristics and litter diversity

influence decomposition rates and macroinvertebrates. Freshwater Biology 51: 605-617.

Mehring, A. S., K. A. Kuehn, A. Thompson, C. M. Pringle, A. D. Rosemond, M. R. First, R. R.

Lowrance, and G. Vellidis. 2015. Leaf litter nutrient uptake in an intermittent blackwater

river: influence of tree species and associated biotic and abiotic drivers. Functional

Ecology doi:10.1111/1365-2435.12399.

Manning, D. W. P., A. D. Rosemond, J. S. Kominoski, V. Gulis, J. P. Benstead, and J. C. Maerz.

2015. Detrital stoichiometry as a critical nexus for the effects of streamwater nutrients on

leaf litter breakdown rates. Ecology 96:2214-2224.

Muggeo, V. M. 2003. Estimating regression models with unknown break-points. Statistics in

Medicine 22:3055-3071.

Mulholland P. J. 1992. Regulation of nutrient concentrations in a temperate forest stream: Roles

of upland, riparian, and instream processes. Limnology and Oceanography 37:1512-1526.

Palmer, M. A., and C. M. Febria. 2012. The heartbeat of ecosystems. Science 336:1393-1394.

Prater, C., E. J. Norman, and M. A. Evans-White. 2015. Relationships among nutrient

enrichment, detritus quality and quantity, and large-bodied shredding insect community

structure. Hydrobiologia: 753:219-232.

82

Peñuelas, J., J. Sardans, A. Rivas-Ubach, and I. A. Janssens. 2012. The human-induced

imbalance between C, N and P in Earth’s life system. Global Change Biology 18:3-6.

Petersen, R. C., and K. W. Cummins. 1974. Leaf processing in a woodland stream. Freshwater

Biology 4:343-368.

R Development Core Team. 2013. R: a language and environment for statistical computing. R

Foundation for Statistical Computing, Vienna, Austria.

Rosemond, A. D., C. M. Pringle, A. Ramírez, M. J. Paul, and J. L. Meyer. 2002. Landscape

variation in phosphorus concentration and effects on detritus-baed tropical streams.

Limnology and Oceanography 47:278-289.

Rosemond, A. D., C. M. Swan, J. S. Kominoski, and S. E. Dye. 2010. Non-additive effects of

litter mixing are suppressed in a nutrient-enriched stream. Oikos 119:326-336.

Rosemond, A. D., J. P. Benstead, P. M. Bumpers, V. Gulis, J. S. Kominoski, D. W. P. Manning,

K. Suberkropp, and J. B. Wallace. 2015. Experimental nutrient additions accelerate

terrestrial carbon loss from stream ecosystems. Science 347:1142-1145.

Scott, M. C., G. S. Helfman, M. E. McTammany, E. F. Benfield, and P. V. Bolstad. 2002.

Multiscale influences on physical and chemical stream conditions across Blue Ridge

landscapes. Journal of the American Water Resources Association 38:1379-1392.

Scott, E. E., C. Prater, E. Norman, B. C. Baker, M. Evans-White, and J. T. Scott. 2013. Leaf-

litter stoichiometry is affected by streamwater phosphorus concentrations and litter type.

Freshwater Science 32:753-761.

Suberkropp, K. 1995. The influence of nutrients on fungal growth, productivity, and sporulation during

leaf breakdown in streams. Canadian Journal of Botany 73:S1361-S1369.

Sudduth E. B., S. S. Perakis, and E. S. Bernhardt. 2013. Nitrate in watersheds: Straight from soils to

streams? Journal of Geophysical Research: Biogeosciences 118:291-302.

83

Swank, W. T., and D. A. Crossley. 1988. Forest hydrology and ecology at Coweeta. Springer-

Verlag. New York, New York, USA.

Tank, J. L., E. J. Rosi-Marshall, N. A. Griffiths, S. A. Entrekin, and M. L. Stephen. 2010. A

review of allochthonous organic matter dynamics and metabolism in streams. Journal of

the North American Benthological Society 29:118-146.

Tant, C. J., A. D. Rosemond, and M. R. First. 2013. Stream nutrient enrichment has a greater

effect on coarse than on fine benthic organic matter. Freshwater Science 32:1111-1121.

Taylor, J. M., R. S. King, A. A. Pease, and K. O. Winemiller. 2014. Nonlinear response of

stream ecosystem structure to low-level phosphorus enrichment. Freshwater Biology

59:969-984.

Wallace, J. B., S. L. Eggert, J. L. Meyer, and J. R. Webster. 1997. Multiple trophic levels of a

forest stream linked to terrestrial litter inputs. Science 277:102-104.

Wickings, K., A. S. Grandy, S. C. Reed, and C. C. Cleveland. 2012. The origin of litter chemical

complexity during decomposition. Ecology Letters 15:1180-1188.

Woodward, G., et al. 2012. Continental-scale effects of nutrient pollution on stream ecosystem

functioning. Science 336:1438-1440.

Supplementary Material

Appendix C: Tables of summarized detrital stoichiometry, linear model results for nutrient

enrichment effects on detrital stoichiometry, difference matrices comparing all detrital

stoichiometry, and increases in breakdown rates.

84

Table 3.1. Parameter estimates based on linear models for drivers of leaf litter and wood

stoichiometry. Intercepts correspond to mean ln-transformed maple (M), poplar (P), oak (O),

rhododendron (R), and wood (W) C:N or C:P, and standardized slopes correspond to effects of

added N (N conc.) or P (P conc.) concentrations on detrital stoichiometry, and their interactions

with fungal biomass (i.e., N*Fungi and P*Fungi). These predictors explained 54% and 58% of

the variation in detrital C:N and C:P, respectively. All slopes are based on relationships between

scaled predictors (using z-scores) and ln-transformed response variables (i.e., middle-stage C:N,

C:P). Standardized slopes are useful for direct comparisons of parameters measured at different

scales and for interpreting interactions between continuous variables (Gelman and Hill 2007).

Bold text corresponds to significant differences (P < 0.05) between intercepts for different

detritus types in comparison to maple, or significant slope estimates (i.e., slope estimates are for

fungi, N conc., P conc., and their interactions)

C:N model Estimate SE

C:P model Estimate SE

Intercepts

M 3.789 0.040

M 7.276 0.068

P -0.202 0.054

P -0.199 0.091

O 0.071 0.056

O 0.147 0.094

R 0.408 0.057

R 0.400 0.098

W 0.243 0.054

W 0.385 0.092

Slopes

Fungi -0.123 0.021

Fungi -0.263 0.035

N conc. -0.049 0.020

N conc. -0.043 0.034

P conc. -0.081 0.020

P conc. -0.238 0.034

85

N*Fungi 0.078 0.021

N*Fungi 0.159 0.035

P*Fungi 0.070 0.021

P*Fungi 0.166 0.036

86

Figure legends

Fig. 3.1. A conceptual representation of how nutrients could affect microbially mediated

conditioning and detrital stoichiometry (modified from Wickings et al. 2012) and the quality of

the resource for shredders. We show reference (left panel) and nutrient-enriched conditions (right

panel), where the leaf litter species used in this study (poplar, maple, oak and rhododendron)

converge to a similar C:nutrient stoichiometry, and approach shredder threshold elemental ratios

(TERs) under nutrient enrichment. We expected that detrital breakdown rates would be increased

under nutrient enrichment and could be predicted based on litter C:N and C:P approaching

optimal C:nutrient content after nutrient-enhanced microbial conditioning.

Fig. 3.2. Mean C:N (top row) and C:P (bottom row) for leaf litter collected during pretreatment

(open circles) or nutrient-enriched conditions (YR1 and YR2; gray and black circles,

respectively). Error bars indicate standard error. Early (d 14), middle (d 70 [PRE], 34, or 62) and

late (d 160 [PRE], 77, or 143) C:N and C:P are shown depending on the given leaf litter type and

year. Conditioned C:N or C:P were only measured on wood veneers on d 109 in all years, and

are compared among years using boxplots (rightmost graph on top and bottom rows of the

figure; note different y-axis scales). All ratios are molar.

Fig. 3.3a,b. Breakdown rates of leaf litter and wood increased as a function of initial C:N (a) and

C:P (b) (solid lines indicate linear relationships between breakdown rate response ratios and

intial C:N [a] or C:P [b]); P < 0.05 in both cases, R2 = 0.86, 0.90, respectively). The magnitude

of the increase in breakdown rates was calculated as the ratio of the average YR1 and YR2

(ENR) ktotal / PRE ktotal (i.e., the response ratio). Letters correspond to each detritus type: maple

(M), poplar (P), oak (O), rhododendron (R), and wood (W).

87

Fig. 3.4a-d. The contribution of shredder-mediated breakdown increased with nutrient

enrichment for all leaf litter types, with the exception of rhododendron in YR2. We tested for

differences in shredder contributions between years for each detritus type because the interaction

between year and detritus type was not significant. Shown are mean ln-transformed

kshredder/kmicrobe for each leaf litter type (maple [a], poplar [b], oak [c], and rhododendron [d]) and

year (±SE). The dashed horizontal line at kshredder/kmicrobe = 0 corresponds to the point at which

contributions of shredders and microorganisms are equivalent (i.e., ln (1) = 0). Differing letters

between years denote significant differences in ln(kshredder/kmicrobe) between years based on

ANOVA and Tukey’s HSD post hoc tests at P < 0.05.

Fig. 3.5a,b. Total breakdown rates as a function of middle-stage C:N (a) or C:P (b) ratios of all

leaf litter species used in this study (wood veneers not included). Vertical dashed lines depict

breakpoints for these relationships. Identified breakpoints were 41 and 1519 for C:N and C:P,

respectively. Open, closed gray, and closed black circles represent PRE, YR1, and YR2,

respectively.

88

Fig. 3.1.

Low$$$C:N$C:P$$$$High$$

Nutrient(enriched,,Reference,

Low$ $$C:N$$C:P $ $$High$

$!$C$limited$ Nutrient$limited$"$ !$C$limited$ Nutrient$limited$"$

Consumer)TER)Consumer)TER)

Microbial$condi9oning$

Microbial$condi9oning$

David W. P. Manning et al.

89

Fig. 3.2.

2040

6080

100

140

Maple

Days in stream

Litte

r C:N

0 50 100 150

PREYR1YR2

2040

6080

100

140

Poplar

Days in stream

Litte

r C:N

0 50 100 150

PREYR1YR2

2040

6080

100

140

Oak

Days in stream

Litte

r C:N

0 50 100 150

PREYR1YR2

2040

6080

100

140

Rhododendron

Days in stream

Litte

r C:N

0 50 100 150

PREYR1YR2

PRE YR1 YR2

50100

200

Wood veneers

Year

d 10

9 C

:N

2000

4000

6000

8000

Maple

Days in stream

Litte

r C:P

0 50 100 150

PREYR1YR2

2000

4000

6000

8000

Poplar

Days in stream

Litte

r C:P

0 50 100 150

PREYR1YR2

2000

4000

6000

8000

Oak

Days in stream

Litte

r C:P

0 50 100 150

PREYR1YR2

2000

4000

6000

8000

Rhododendron

Days in streamLi

tter C

:P

0 50 100 150

PREYR1YR2

PRE YR1 YR2

1000

5000

20000

Wood veneers

Year

d 10

9 C

:P

David W. P. Manning et al.

90

Fig. 3.3a,b.

M

P O

R

W

60 100 140 180

23

45

6

Initial C:N

(ktotal (ENR)k total (PRE))

ABre

akdo

wn

Rat

e R

espo

nse

Rat

io

M

P O

R

W

2000 6000 10000

23

45

6

Initial C:P

B

David W. P. Manning et al.

91

Fig. 3.4a-d.

-0.5

0.51.01.52.0

Maple

Year

(ln kshredderk m

icrobe)

PRE YR1 YR2

Shr

edde

r vs.

mic

robi

al c

ontri

butio

n

a b bc

A -0.5

0.51.01.52.0

Poplar

Year(ln

kshredderk m

icrobe)

PRE YR1 YR2

Shr

edde

r vs.

mic

robi

al c

ontri

butio

n

a b bc

B

-0.5

0.51.01.52.0

Oak

Year

(ln kshredderk m

icrobe)

PRE YR1 YR2

Shr

edde

r vs.

mic

robi

al c

ontri

butio

n

a b bc

C -0.5

0.51.01.52.0Rhododendron

Year

(ln kshredderk m

icrobe)

PRE YR1 YR2

Shr

edde

r vs.

mic

robi

al c

ontri

butio

n a bc ac

D

David W. P. Manning et al.

92

Fig. 3.5a,b.

40 60 80 100

0.00

0.02

0.04

0.06

Litter C:N

Leaf

litter breakdown rate

(ktotal d−1)

A

2000 4000 6000 80000.00

0.02

0.04

0.06

Litter C:P

B

93

CHAPTER 4

NUTRIENTS AND TEMPERATURE ADDITIVELY INCREASE STREAM MICROBIAL

RESPIRATION1

1 David W. P. Manning, Amy D. Rosemond, Jonathan P. Benstead, Vladislav Gulis, John S. Kominoski and John C. Maerz. To be submitted to Limnology and Oceanography.

94

Abstract. Nutrient enrichment and rising temperatures are expected to stimulate metabolic

process rates such as respiration of detrital carbon (C). However, few studies have examined how

the temperature dependence of respiration may be altered by nutrient enrichment in aquatic

ecosystems. Here, we measured respiration rates associated with naturally occurring coarse and

fine particulate detrital C (leaf litter, wood and fine benthic organic matter [FBOM]), and

deployed leaf litter and wood across seasonal temperature gradients in response to experimental

nutrient additions to five streams. We assessed the temperature dependence of both AFDM-

specific and fungal biomass-specific respiration rates using metabolic theory. Respiration rates

increased with temperature and exhibited activation energies (E) that were equivalent for all

three naturally occurring substrates, and were below predicted values (E = 0.43 eV). Activation

energies for deployed leaf litter and wood were higher than naturally occurring detritus.

Activation energy for deployed detritus increased with initial C:nutrient content (E = 0.60-2.28

eV). Nutrient enrichment had no effect on the temperature dependence of respiration for

naturally occurring leaf litter, wood or FBOM, but did increase respiration rates on average

across the seasonal temperature gradient for leaf litter and wood by 1.31× and 1.37×,

respectively. Increases in respiration corresponded to stimulation of fungal biomass on these

substrates across the entire seasonal temperature gradient. Temperature and nutrient effects were

additive, implying that stream microbial respiration could increase up to 1.59× with a 4°C

increase in stream temperature (1.25×) and moderately increased nutrient concentrations

(~1.34×). Temperature dependence of deployed detritus was modified under nutrient enrichment,

likely due to higher fungal biomass at early stages of decay. Specifically, fungal colonization and

respiration at early stages of decay increased the most in response to nutrients despite colder

streamwater temperatures. This effect corresponded to decreased temperature dependence of

95

AFDM-specific respiration for deployed detritus in nutrient-enriched conditions. Our data

suggest that nutrient enrichment is unlikely to change the temperature dependence of stream

microbial respiration once microbial communities are established on detritus (i.e., for naturally

occurring detritus), while nutrients will stimulate microbial biomass and respiration the most at

early stages of decay.

Introduction

Nutrient availability and temperature are two key drivers of energy and material

processing in ecosystems, and both are increasing worldwide (IPCC, 2007, Peñuelas et al. 2012).

In aquatic ecosystems, water temperatures are predicted to track rising temperatures driven by

global climate change, in addition to altered thermal regimes driven by land use change such as

urbanization or deforestation (e.g., Kaushal et al. 2010, Ferreira et al. 2014). Along with

increased temperatures, nutrient availability in aquatic ecosystems has increased as a result of

widespread anthropogenic nutrient inputs (Alexander and Smith 2006, Woodward et al. 2012).

These two important global change drivers of ecosystem processes have been studied separately

to a greater degree than their combined effects. Thus, predicting how both nutrients and

temperature will affect ecosystem functions remains largely unresolved (Cross et al. 2015).

Among aquatic ecosystems, streams and rivers are increasingly recognized as a

significant component of the global carbon cycle, particularly through transformations of

terrestrially derived organic matter to CO2 via respiration (Cole et al. 2007, Butman and

Raymond 2011, Hotchkiss et al. 2015). Processing of detritus such as leaf litter, wood and fine

particles is a predominant energy pathway in streams and rivers, and fungal more than bacterial

decomposers typically control processing and respiration of coarse detrital substrates to CO2

96

(Hieber and Gessner 2002, Findlay et al. 2002, Tant et al. 2013). Increased temperatures are

expected to increase rates of respiration in predictable ways, based on well-described

temperature-respiration models founded on first principles (e.g., Arrhenius, 1889, metabolic

theory of ecology [MTE]; Gillooly et al. 2001, Yvon-Durocher et al. 2012). Likewise, increased

nutrient availability can also increase microbial respiration of detrital C by stimulating microbial

biomass and activity on these substrates (Suberkropp et al. 2010, Tant et al. 2013, Kominoski et

al. 2015). Increased processing and loss rates of detrital C via respiration could therefore occur

when streams are warmer due to climate change, and when nutrients are elevated from

anthropogenic sources, thereby altering the role of streams and rivers in global C budgets

(Aufdenkampe et al. 2011).

Detrital organic matter respiration is controlled by several factors including, temperature,

streamwater nutrient availability, microbial (i.e., fungal) biomass (Gulis and Suberkropp 2003,

Tant et al. 2013, Cheever et al. 2013), and substrate nutrient content. Among these, both

increased streamwater nutrients and temperature are expected to stimulate microbial biomass and

hence respiration rates associated with detritus. For example, increased streamwater nutrients are

particularly important for increasing fungal biomass and respiration rates of detrital C in streams

(Suberkropp et al. 2010, Kominoski et al. 2015). Likewise, experimentally elevated stream

temperatures stimulate fungal biomass accrual on leaf litter, with subsequent effects on

respiration rates (Ferreira and Chauvet 2011a). Combined, these two drivers may increase fungal

biomass relatively more at high temperatures and nutrient availability, i.e., respiration rates could

be increased more than would be predicted if either driver is altered alone (e.g., Ferreira and

Chauvet 2011b). However, few studies have examined the effects of nutrient enrichment across

seasonal temperature gradients to determine how increased nutrient availability may modulate

97

the temperature dependence of microbial respiration rates on detrital C (but see Ferreira et al.

2011b, Welter et al. 2015).

In addition to extrinsic factors such as temperature and streamwater nutrients, studies in

terrestrial and aquatic systems suggest that intrinsic substrate characteristics such as C quality

also play an important role in determining the temperature dependence of detrital C processing

(e.g., Fierer et al. 2005, Jankowski et al. 2014). This hypothesis stems from the expectation that

more recalcitrant, complex C compounds require more steps to complete conversion to CO2, and

therefore require greater activation energy (e.g., Bosatta and Ågren, 1999). Despite this apparent

C-quality vs. temperature dependence trend, few studies have examined how nutrient enrichment

could modify the temperature dependence of microbial respiration rates for diverse substrates

with differing initial C quality (defined here as initial C:nutrient content).

Our objective for this study was to examine the temperature dependence of microbial

respiration rates associated with detrital C under ambient and nutrient-enriched conditions. We

predicted that nutrient enrichment would stimulate microbial biomass (especially fungal

biomass) and activity on detritus, with relatively greater increases for coarse detritus vs. fine

particles (Stelzer et al. 2003, Tant et al., 2013). We also predicted that the temperature

dependence of respiration would be greater for relatively nutrient-poor detrital substrates (i.e.,

higher C:N, C:P, Jankowski et al. 2014). Finally, we hypothesized that the effects of nutrient

availability would be dependent on temperature, with stronger, positive effects of nutrients at

higher temperatures (e.g., Ferreira and Chauvet, 2011b). We tested these predictions using multi-

year experimental nutrient (nitrogen [N] and phosphorus [P]) additions to five streams. To assess

the effect of nutrient enrichment on respiration of detrital organic matter, we measured

respiration rates associated with conditioned, naturally occurring coarse (leaf litter, wood) and

98

fine benthic organic matter, as well as leaf litter and wood that was deployed for known amounts

of time in our study streams. We examined the response of the respiration rates before and

during two years of nutrient enrichment, using seasonal gradients in temperature. To assess the

temperature dependence of respiration we used the MTE (i.e., Van’t Hoff-Arrhenius

relationship) to estimate the activation energy (E) of respiration under pretreatment and nutrient-

enriched conditions.

Methods

Site description

Our study was conducted in five first-order streams at the Coweeta Hydrologic

Laboratory (CWT), a USDA Forest service and Long-term Ecological Research site in Macon

Co., North Carolina, USA. The CWT basin is characterized by mature hardwood forest, and

contains several low-order streams that are heavily shaded year-round by Rhododendron

maximum. Detailed descriptions of the study site can be found in Swank and Crossley (1988).

We identified five unnamed first-order reaches of the Dryman Fork basin at CWT

(35°03’35” N 83°25’48” W) for the nutrient additions used in this study. All five 70-m reaches

had similar chemical and physical characteristics, and were characterized by low ambient

nutrient concentrations (<0.2 mg/L dissolved inorganic nitrogen [DIN] and <0.005 mg/L soluble

reactive phosphorus [SRP], Appendix C: Table C1). Detailed methods for nutrient additions can

be found in Rosemond et al. (2015) and Manning et al. (2015), and are presented here in brief.

Concentrated nutrient solutions (NH4NO3 and H3PO4) were added to the five streams using

solar-powered metering pumps (LMI Milton Roy) that delivered the nutrients to stream-fed

irrigation lines continuously for two years (YR1, YR2) following a year of pre-treatment data

99

collection (PRE). Nutrients were delivered to the irrigation lines proportional to continuously

measured discharge using pressure transducers (Keller America) and CR800 dataloggers

(Campbell Scientific). We targeted concentrations that reflected low-to-moderate elevated

concentrations of N and P that were common in the region (Scott et al. 2002). Target

concentrations of N and P encompassed an increasing range of DIN (~80-650 µg/L) that also

corresponded to decreasing concentrations of SRP (~90-11 µg/L), and target N:P from 2-128.

Naturally occurring detritus

We collected three types of organic matter on a quarterly basis (July 2010-July 2013;

summer, autumn, winter and spring) for one year before nutrients were added (PRE) and during

two consecutive years of nutrient enrichment (YR1, YR2) for analysis of microbial respiration

rates. Leaf litter, wood (small sticks <2 cm in diameter), and fine benthic organic matter (FBOM)

were collected from four randomly selected transects in each of the 70-m treatment reaches. We

collected five submerged leaves from each transect without regard to litter type, such that the

composite sample of the five leaves reflected the relative abundance of available leaf litter at the

time of sampling. We used a similar method for sampling wood, where five small, submerged

sticks were collected and reduced in size with pipe cutters as needed. We collected surface-layer

FBOM from obvious depositional areas in the same transect as leaf litter and wood. All samples

were placed in whirl-pak bags with streamwater, and transported to the laboratory on ice until

microbiological analyses were conducted (see below).

100

Deployed leaf litter and wood veneers

We deployed known amounts of leaf litter and wood and measured respiration rates in

response to seasonal gradients in temperature and experimentally elevated nutrient

concentrations. We used four leaf litter types from tree species common in the CWT basin (Acer

rubrum L. [maple], Liriodendron Tulipifera L. [poplar], Quercus prinus L. [oak], and

Rhododendron maximum L. [rhododendron]) and wood veneers (Quercus alba L.). Leaf litter

was deployed in 10-g packs enclosed in nylon-mesh bags (Cady Bag Inc., Pearson, Georgia,

USA). We deployed the litterbags in four 17.5-m sub-reaches within each 70-m treatment reach

(n = 4 litterbags × 4 litter species × 5 streams × 7 collection dates = 560 for each year). Wood

veneers were cut into 2.5 × 20 cm strips and fastened to nylon gutter mesh rafts. Litterbags and

wood veneers were collected periodically after incubation in the stream to analyze litter

breakdown rates (Manning et al., 2015, D. W. P. Manning unpublished data), and microbial

respiration rates. We measured microbial respiration rates associated with the deployed leaf litter

and wood at early, middle and late stages of decay, with shortened sampling schedules during

nutrient enrichment due to faster processing of leaf litter and wood (all leaf litter PRE = days 14,

70 160; YR1, YR2 = 14, 34, 63 [maple, poplar], 14, 63, 126 [oak, rhododendron]; wood veneers

PRE = days 21, 109, 160, YR1, YR2 = days 21, 109, 143). We determined initial (i.e., day 0)

leaf litter and wood veneer %C and %N content using a Carlo Erba CHN analyser (Milan, Italy),

and initial %P was determined spectrophotometrically with the ascorbic acid method (APHA

1998) using a UV-1700 spectrophotometer (Shimadzu, Japan) after acid/dry-ash extraction of

phosphorus (Allen 1974).

101

Microbial respiration rates

Microbial respiration rates associated with leaf litter, wood, and FBOM were measured

using methods outlined by Gulis and Suberkropp (2003), Gulis et al. (2004), and Tant et al.

(2013) within 24 h after collection. Briefly, microbial respiration rates were determined as the

rate of oxygen (O2) uptake associated with the leaf litter, wood, or FBOM normalized per gram

ash-free dry mass (AFDM) of the sample (mg O2/g AFDM/hr). We placed two 1-cm diameter

leaf discs from each of the five leaves in the sample into 30 mL of streamwater in glass

respiration chambers. We then measured oxygen concentrations and temperature periodically

over a 30-minute interval using YSI 5100 Dissolved Oxygen Meters (YSI Inc., Yellow Springs,

Ohio, USA) in a walk-in incubator set to stream temperatures measured at the time of sample

collection. Respiration rates were computed based on the slope of the decline in dissolved

oxygen concentrations over time. Microbial respiration rates associated with five wood discs cut

from the small sticks were measured in the same way. Respiration rates of FBOM were

determined using longer (~2 hour) incubations of 100 mL subsamples of continuously agitated

FBOM and streamwater in 150 mL glass bottles, and were computed using the difference in

dissolved oxygen before and after the two-hour incubation period. After respiration rates were

measured on leaf litter and wood, the samples were removed from the chambers, dried for 24h at

55 °C, weighed to the nearest 0.001 g, combusted for 4.5 h at 550 °C and reweighed to determine

AFDM. We determined AFDM in the same manner for freeze-dried FBOM samples.

We measured respiration rates associated with deployed leaf litter and wood in the same

manner as for naturally occurring detritus, by measuring dissolved O2 over a 30-minute interval

within 24-48 h of collection. In the case of leaf litter, all litter was removed from the litterbags,

and rinsed over nested sieves. We then subsampled ten ~2 × 2 cm pieces from separate leaves or

102

wood veneers and incubated this material in the same manner as the leaf litter or wood discs as

described above. After the leaf litter or wood veneers were incubated, each sample was dried and

combusted as above to determine AFDM.

Fungal biomass

We measured fungal biomass associated with naturally occurring leaf litter and wood, as

well as deployed leaf litter and wood veneers. We did not measure fungal biomass associated

with FBOM, due to relatively greater importance of bacteria on this substrate (Findlay et al.

2002). We estimated fungal biomass by quantifying ergosterol concentrations associated with the

detritus and converting to fungal biomass using standard conversions (5.5 µg of ergosterol per

mg fungal dry mass; Gessner and Chauvet 1993). Ergosterol was extracted from the 5 leaf discs,

or 5 wood discs preserved in the field in methanol (for naturally occurring detritus) or from ~2 ×

2 cm freeze-dried, weighed leaf litter or wood veneer pieces (for deployed detritus) using HPLC

((LC-10VP, Shimadzu, Columbia, Maryland, USA). We used a Kinetex C18 column

(Phenomenex, Torrance, California, USA) and a UV detector set at 282 nm. We used external

ergosterol standards (Acros Organics, Geel, Belgium). Further details about this method can be

found in Gulis and Suberkropp (2006).

Statistical Analyses

We used the linearized form of the Arrhenius equation (Arrhenius 1889, Perkins et al.

2012) to estimate the temperature dependence of respiration associated with naturally occurring

and deployed detritus:

ln𝑅 𝑇 = −𝐸 !!"− !

!"!+ ln[𝑅 𝑇! ]  ,

103

where R(T) is the respiration rate at absolute temperature T, E is the activation energy, and k is

the Boltzmann constant (8.617 × 10-5 eV/K, 1 eV = 1 × 10-19 J). We centered our data using the

approximate mean annual temperature for the study in the five streams using the reciprocal of

absolute temperature at 10 °C (i.e., 1/kTc ), such that the intercept of the linear equation describes

the average respiration rate at this temperature. The slope of this model is considered the

activation energy (E) of respiration for a given substrate and year (Perkins et al. 2012, Yvon-

Durocher et al. 2012,). We used the predictor ‘year’ (PRE = ambient nutrients, YR1 and YR2 =

elevated nutrients) as a surrogate for nutrient effects in our models. We tested for the effects of

nutrient enrichment and substrate type on the temperature dependence of respiration rates by

estimating E before (PRE) and during nutrient enrichment (YR1, YR2) with this linear model for

each substrate, including FBOM, leaf litter and wood that was naturally occurring, and leaf litter

and wood that was deployed for known amounts of time in our study streams. In addition to

respiration rates corrected for AFDM of the sample, we also modeled respiration rates corrected

for fungal biomass (mg) in the same way to account for fungal biomass effects on respiration

rates that were not related to temperature. All analyses were conducted using the statistical

software R v. 3.0.1 (R Development Core Team, 2013).

Results

Nutrient treatments and stream temperatures

Streams were similar in terms of temperature regimes. Temperatures spanned a gradient

of ~12.3°C (minimum temperature = 4.6°C, maximum = 16.9°C) for the entire study period. Our

study streams had relatively consistent temperatures within each season; the largest differences

in temperature among streams were found during summer (range = 2.3°C), and smallest

104

differences were found in autumn (range = 0.5°C) (Appendix D: Table D1). We found smaller

temperature ranges corresponding to deployed detritus sampling schedules (from winter to spring

in all streams, range = 10.9°C, Appendix D: Table D2), compared to the entire year in our study

streams (range = 14.4°C).

Nutrient concentrations measured biweekly during the study generally reflected target

concentrations (Appendix D: Table D3, also see Manning et al. 2015, Rosemond et al. 2015),

and as a result, N and P concentrations differed among streams. Both N and P were added

continuously for two years, and concentrations that were measured in the streams spanned a

gradient of 66.3-510.6 µg/L (for DIN) and 6.2-78.4 µg/L (for SRP). Despite these gradients in N

and P concentrations, we found little evidence for concentration-dependent effects for either N or

P on respiration rates (data not shown); therefore, we used nutrient-enriched conditions (i.e.,

YR1, YR2) in comparison to pretreatment (PRE) for our analyses.

Naturally occurring detritus respiration rates

Respiration rates on naturally occurring substrates showed a positive relationship with

temperature (i.e., negatively related to inverse temperature, 1/kT). Activation energies were

statistically indistinguishable among the two coarse detrital substrates, and FBOM (P > 0.05,

Table 1, Fig. 1a-c); all three substrates had average activation energy of 0.43 eV. The 95%

confidence intervals for activation energies included the predicted E value of 0.65 eV (95% CI =

0.19-0.68; Table 1, Gillooly et al. 2001). Nutrient enrichment stimulated AFDM-specific rates of

respiration for leaf litter and wood, but not FBOM, based on significantly higher intercepts for

leaf litter and wood respiration vs. standardized temperature in the enrichment years (Table 1,

Fig. 1a-c). Based on these differences in intercepts, respiration rates increased by 1.31× in YR1

105

(YR2 not significant) for leaf litter, and 1.24-1.50× in YR1 and YR2 for wood, respectively (P <

0.05, Table 1, Fig. 1a-c). The temperature dependence (i.e., activation energy) of respiration

rates on naturally occurring leaf litter, wood, and FBOM were unchanged by nutrient enrichment

(Table 1, Fig. 1a-c).

Deployed leaf litter and wood respiration rates

Respiration rates associated with litterbags and wood veneers that were deployed in our

study streams were also positively related to temperature. In contrast to naturally occurring

detritus, deployed detritus showed greater temperature dependence (i.e., higher activation

energies), with average E values for all five substrates ~1.15 eV. Activation energy was closer to

the MTE predicted range of activation energy (0.6-0.7 eV) for maple, poplar and oak leaf litter,

but rhododendron and wood had activation energies 2.1× and 3.4× higher than maple,

respectively (Appendix E). Wood veneers exhibited the highest estimated activation energy (2.28

eV) of all detritus types sampled in this study (Table 2, Appendix E). Unlike naturally occurring

detritus, we were unable to detect nutrient effects on either respiration rates, or temperature

dependence when each detritus type was considered separately (Appendix E). When all leaf litter

types were considered together and compared to wood veneers, we found there was a significant

effect of nutrient enrichment on the temperature dependence of respiration on deployed leaf litter

(Table 2, Fig. 2a,b). Specifically, nutrient enrichment reduced the temperature dependence of

respiration rates on average for the four leaf litter species in YR1 only (YR2 not significant,

Table 2, Fig. 2a). In contrast, nutrient enrichment was associated with stronger temperature

dependence of respiration rates for wood veneers compared to leaf litter in YR1; YR2 was not

significantly different than PRE. Wood veneers also generally had higher respiration rates during

106

YR1 and YR2 than PRE (~2.8 and 3.0× higher than pretreatment in YR1 and YR2, respectively;

Table 2, Fig. 2b).

Fungal biomass-specific respiration rates

Fungal biomass-specific respiration rates (i.e., respiration rate/mg fungal biomass) for

naturally occurring leaf litter and wood (FBOM fungal biomass was not measured), were

positively related to temperature, and had activation energies that were similar to observed

AFDM-specific rates (fungal biomass-specific E =0.47 eV; Appendix F: Table F1). However,

unlike AFDM-specific respiration rates, nutrient enrichment generally did not increase fungal

biomass-specific respiration rates on naturally occurring leaf litter and wood, except for a

marginal increase in the first year of enrichment for leaf litter only (P = 0.065; Appendix F:

Table F1), implying that the nutrient effect was through increased fungal biomass, particularly in

the second year of nutrient enrichment.

Patterns of fungal biomass-specific respiration rates for deployed detritus across seasonal

temperature gradients were different than AFDM-specific respiration rates for these substrates

(Fig. 2a-d). Fungal biomass-specific respiration rates were generally higher at early stages of

decay during pretreatment for leaf litter, and fungal biomass-specific respiration tended to

decrease with increasing streamwater temperatures in both pretreatment and the first year of

nutrient-enriched conditions. Nutrient enrichment in YR2 significantly changed the relationship

between fungal biomass-specific respiration and temperature for leaf litter from decreasing, to

increasing (Appendix F: Table F2). There was no detectable effect of nutrient enrichment on the

relationship between fungal biomass-specific respiration rates and temperature for wood veneers

(Appendix F: Table F2). Mean fungal biomass-specific respiration rates on wood veneers were

107

generally higher than leaf litter. Fungal biomass-specific respiration rates for wood veneers were

significantly lower under nutrient-enriched conditions in YR2 only (Appendix F: Table F2).

Fungal biomass on naturally occurring and deployed detritus

Nutrient enrichment increased fungal biomass on naturally occurring detritus (leaf litter

and wood only) in YR2 (1.23× higher, P < 0.05), but was not significantly higher in YR1

compared to PRE (P = 0.42), consistent with patterns of fungal biomass-specific respiration.

Fungal biomass on naturally occurring detritus was unrelated to temperature, and this

relationship was unchanged by nutrient enrichment (P = 0.32, Fig. 3a).

Nutrient enrichment increased fungal biomass on deployed detritus by a factor of 1.67×

in both years (both P < 0.05). In contrast to naturally occurring detritus, fungal biomass

increased significantly as a function of temperature in pretreatment (Fig. 3b). The relationship

between fungal biomass and temperature became less pronounced under nutrient enrichment

(Fig. 3b). Specifically, there was a significant interaction between nutrient enrichment (year) and

temperature. In this case, we found fungal biomass was 4.9× higher at colder temperatures under

nutrient-enriched conditions compared to pretreatment, whereas fungal biomass increased to a

lesser degree when stream temperatures were warmer (fungal biomass was 2.6 and 2.7×

pretreatment at middle and late stages of decay, respectively). This pattern of relatively greater

increases in fungal biomass when stream temperature was colder corresponded to slopes between

fungal biomass and inverse temperature that were significantly greater under nutrient-enriched

conditions compared to pretreatment (P < 0.05, Fig. 3b).

108

Detrital C:nutrient content and temperature dependence

We compared the slopes (activation energy, E) for naturally occurring and deployed

detritus for each detritus type to test for differential responses to temperature based on substrate

C:nutrient content under pretreatment conditions. Naturally occurring wood showed the highest

C:nutrient content for both C:N and C:P compared to leaf litter and FBOM. Wood substrates had

C:N values of 155 on average in pretreatment, which was 2.1× and 7.3× higher than leaf litter

(mean C:N = 74) and FBOM (mean C:N = 21), respectively. Wood C:P was 11,147 on average

during pretreatment, which was 2.7× and 27.9×, higher than leaf litter (mean C:P = 4060) and

FBOM (mean C:P = 399), respectively. For naturally occurring wood, leaf litter, and FBOM we

found that the differences in detrital C:nutrient content were unrelated to the temperature

dependence of respiration for all three substrates (Table 1, Fig 1a-c).

Deployed wood showed the highest initial C:nutrient for both C:N and C:P compared to

leaf litter. In general, initial C:N values ranged between 66-167, and initial C:P ranged between

2077-10215 (Fig. 4a,b). For deployed detritus, we found significantly greater temperature

dependence for detritus with higher initial C:nutrient content (Fig. 4a,b). The highest activation

energy and initial C:nutrient content was for wood veneers (2.28 eV), followed by rhododendron

(1.4 eV), tulip poplar (0.82 eV), maple (0.67) and oak (0.60). Tulip poplar and oak had activation

energies that were statistically similar to maple (all P > 0.05, Fig. 4a,b), while rhododendron was

marginally higher than maple (P = 0.08), and wood was significantly higher than maple (P <

0.05; Fig. 4a,b).

Temperature dependence of respiration on deployed substrates was also related to initial

fungal biomass (d 14 [leaf litter], d21 [wood veneers) during pretreatment (Fig. 4c). In this case,

substrates with lower fungal biomass tended to exhibit higher activation energies (Fig. 4c).

109

Likewise, fungal biomass at early stages showed marginal declines with higher initial C:N (data

not shown, P = 0.07), and early fungal biomass significantly decreased as a function of initial

C:P (data not shown, P < 0.04).

Discussion

Respiration rates increased as a function of elevated nutrients and temperature, and the

combined effects of both drivers were additive. Our models describing the relationship between

temperature and respiration suggest that respiration rates associated with naturally occurring

coarse detritus will increase by ~6.4%, on average (95% confidence interval: 2.7-10.2%), per

1°C increase in stream temperatures due to climate change or thermal pollution from land-use

change. Nutrient enrichment also increased respiration rates on naturally occurring coarse

detritus, especially wood, and this effect was consistent across our seasonal temperature gradient.

As a result, predicted increases in respiration based on temperature alone were found to be

independent of increases in respiration rates during nutrient enrichment, which were up to 1.37×

higher than pretreatment rates. Thus, our models suggest that on naturally occurring detritus,

respiration rates could increase by ~1.59× for a 4°C increase in streamwater temperatures and

moderate increases in nutrient availability. This pattern of additive effects of temperature and

nutrients contrasts with hypothesized and previously observed synergistic effects of elevated

temperature and nutrients on respiration rates (Ferreira and Chauvet 2011b). Our estimates of

activation energies for naturally occurring detritus were slightly different than predictions of

MTE, with mean activation energies that were lower than those reported for cellular respiration

(e.g., Gillooly et al. 2001), or short-term respiration rates in rivers (Yvon-Durocher et al. 2012).

110

Taken together, these results imply that respiration rates associated with detrital C in streams

may be less sensitive to temperature changes than previously predicted.

Our findings suggest that it is important to consider how fungi drive differences in

temperature dependence of respiration when nutrients are elevated. Fungal biomass-specific

respiration rates were generally unrelated to temperature (deployed leaf litter), or negatively

related to temperature (deployed wood) under pretreatment or nutrient-enriched conditions,

implying that fungi that initially colonize detritus respire as much, or more per unit biomass

compared to later stages of decay. Further, we found that fungal biomass on deployed detritus

generally increased with temperature under pretreatment conditions. Nutrient enrichment had

stronger effects on fungal biomass at early stages of decay, which weakened the relationship

between fungal biomass and temperature. These findings for deployed detritus are informative in

the context of temperature and nutrient effects on detritus as colonization by fungi progresses.

Typically, successional patterns of fungal colonization of leaf litter and wood follow distinct

stages, with peak biomass, fungal community diversity and respiration rates found at middle

stages of decay (e.g., Gessner et al. 1993, Gessner et al. 2007). We found that this pattern was

generally true during pretreatment for fungal biomass and respiration rates, while biomass and

respiration rates responded nearly 2-fold more at earlier stages of decay compared to middle or

late stages when nutrients were elevated. Thus, alleviation from nutrient limitation at early stages

of decay appears to increase fungal biomass and respiration rates to a greater degree compared to

middle or late stages of decay, thereby obscuring temperature-driven effects on respiration rates.

Comparing the temperature dependence of respiration for naturally occurring detritus and

deployed detritus reveals different responses to temperature and nutrients. Specifically, we

detected little evidence for nutrient effects on temperature dependence of naturally occurring

111

detritus vs. some evidence for reduced temperature dependence of respiration for deployed

detritus. Several factors may have contributed to this difference between deployed and naturally

occurring detritus. For example, we used truncated sampling schedules for deployed detritus

during nutrient enrichment that resulted in smaller temperature ranges for our study in YR1 and

YR2 that make year-to-year comparisons difficult. In addition, we deployed detritus in winter,

which also meant that rising stream temperatures coincided with the progression of fungal

biomass accrual on deployed detritus. These consequences of our study design necessitate

cautious interpretation of the temperature dependence of respiration for deployed detritus, and

arguably give greater weight to inferences regarding temperature dependence of respiration for

naturally occurring detritus that involved more comparable temperature gradients and microbial

communities. Nonetheless, the relationship between temperature and respiration rates for

deployed detritus could still be useful for determining the temperature dependence of respiration

rates, but may be more aptly described in terms of the ‘apparent’ activation energy, as dictated

by the rate of fungal biomass accrual through time (e.g., Yvon-Durocher et al. 2012, Cross et al.

2015, Welter et al. 2015). Therefore, our results demonstrate that nutrient-stimulated fungal

biomass accrual, particularly during low temperatures at early stages of decay, could potentially

reduce the apparent activation energy of leaf litter respiration.

Detrital quality and temperature dependence

Our study shows that nutrients and temperature increase respiration rates associated with

diverse detritus. Several studies have highlighted the importance of considering substrate C

quality (here, C:nutrient content) for determining the response of respiration rates to nutrients

(Stelzer et al. 2003, Ferrieria et al. 2015) and temperature (Fierer et al. 2005, Jankowski et al.

112

2014). For naturally occurring detritus, we found that leaf litter and wood responded comparably

to temperature, despite distinct C:nutrient content of these two detritus types. For example,

naturally occurring wood had mean C:N that was ~2× higher than collected leaf litter under

pretreatment conditions, but this difference in detrital C:N did not translate to differences

between leaf litter and wood in terms of temperature dependence. This finding is counter to the

hypothesis that more recalcitrant detritus would show greater temperature dependence, but again,

patterns of fungal colonization may help explain this trend. Wood generally has greater residence

time in streams compared to leaf litter (Webster et al. 1999), which suggests that the fungal

communities on the wood we sampled may have had more time to develop, and therefore may

exhibit similar temperature dependence as more recently colonized leaf litter. Consistent with

this, we found that fungal biomass was generally unrelated to temperature on naturally occurring

substrates. This trend suggests that once similar levels of fungal biomass are reached, respiration

rates will consistently respond to increasing temperatures regardless of differences in initial

substrate C:nutrient content.

For deployed detritus, we used four leaf litter species and wood, which spanned a wide

range of initial C:nutrient content. In contrast to naturally occurring leaf litter and wood, we

found larger differences in apparent activation energies corresponding to differences in initial

C:nutrient content when leaf litter species and wood were considered individually under

pretreatment conditions. These data are more consistent with the hypothesis that recalcitrant

detritus (in terms of C:nutrient content) is associated with greater activation energies (Jankowski

et al. 2014), and suggest that initial C:N or C:P content could be used to predict temperature

dependence of respiration rates on diverse carbon resources to some extent. However, similar to

nutrient effects on temperature dependence, we cannot fully partition the effects of exposure time

113

vs. temperature on these activation energies, which are correlated in our study. It is possible that

higher activation energy of wood veneers, or rhododendron is also related to the lower initital

fungal colonization due to high C:nutrient content, rather than responses to increasing

temperature per se. For example, we found that lower fungal biomass at early stages of decay on

leaf litter or wood veneers corresponded to higher activation energies. Greater temperature

dependence of wood veneers and rhododendron could be a function of slower fungal biomass

colonization stemming from low nutrient availability in these substrates. Therefore, the

differences in activation energy we observed related to C:nutrient content likely reflect the

apparent activation energy for each substrate as dictated by differences in initial fungal biomass

accrual rates.

Conclusions

Our study aimed to assess the effects of nutrients across seasonal gradients of

temperature. However, the sampling design we used for deployed substrates sometimes

prevented us from using comparable temperature gradients across different years of the study.

Thus, we also compared temperature dependence of respiration on deployed substrates using

smaller, but comparable temperature ranges (i.e., by excluding late stages of decay where

greatest differences in temperature among years occurred and across the interquartile range of

temperatures observed in this study). The results of this analysis suggest that across more

comparable temperature gradients (interquartile range), temperature may have played less of a

role in determining respiration rates on deployed substrates than might be implied by examining

the entire range of temperature, or the range observed for early and middle stages of decay (i.e.

excluding late stages; Appendix G). Thus, respiration rates associated with leaf litter or wood

114

veneers deployed for similar incubation times (e.g., 14-28 d) and measured monthly throughout

the year could be more informative for building temperature-respiration relationships in future

studies.

Determining the dual effects of two dominant anthropogenic global change drivers,

nutrient enrichment and rising temperatures, on detrital C processing is critical in aquatic

ecosystems, given their role in the global C cycle. Our study shows that across seasonal

temperature gradients, nutrient enrichment consistently increased respiration rates of naturally

occurring detrital C, which was likely driven by stimulation of fungal biomass via relaxation of

nutrient limitation. Respiration rates responded more to our moderate nutrient enrichment

compared to predicted increases in respiration rates with the higher temperatures due to climate

change. The temperature dependence of respiration was unchanged by nutrient enrichment for

naturally occurring detritus, and was slightly lower than predictions based on MTE (Gillooly et

al. 2001). In contrast, the temperature dependence of respiration for each deployed detrital

substrate was predicted by C:nutrient content, with no detectable effects of nutrient enrichment

on either respiration rates or their dependence on temperature. The average temperature

dependence of all four leaf litter species vs. wood veneers was weaker with nutrient enrichment,

likely due to increased fungal biomass on detritus at early, low-temperature stages of decay.

These findings suggest that nutrient enrichment could affect the timing of fungal biomass accrual

on detritus, with subsequent effects on the apparent activation energy of respiration rates. Thus,

our findings imply that for coarse detrital C, respiration rates will increase additively with

elevated temperature and nutrients once nutrient-stimulated fungal colonization of detritus has

reached comparable levels.

115

Acknowledgements

We thank Jason Coombs and Katie Norris for maintenance and sampling of the study

streams. Phillip Bumpers, Jason Coombs, Emmy Deng, Jenna Martin, Tom Maddox, and Katie

Norris helped in the laboratory or in the field. This study received support from the NSF (DEB-

0918894 to ADR and JCM, DEB-0918904 to JPB, and DEB-0919054 to VG). This study also

leveraged logistical support from the CWT LTER Program at the University of Georgia, which is

supported by NSF award DEB-0823293 from the Long Term Ecological Research Program

(JCM co-PI). Rob Case, Daniel Hutcheson, and Kevin Simpson of YSI Integrated Systems and

Services constructed the infrastructure for the nutrient-dosing system. Aqueous ammonium

nitrate was provided by The Andersons, Inc. through David Plank. We thank Alan Covich and

Nina Wurzburger for helpful comments on earlier versions of this manuscript.

Literature Cited

Allen, S. E., et al. 1974. Chemical analysis of ecological materials. Wiley, New York, New

York, USA.

Alexander R. B., and R. A. Smith. 2006. Trends in nutrient enrichment of U.S. rivers during the

late 20th century and their relation to changes in probable stream trophic conditions.

Limnology and Oceanography 51:639-654.

American Public Health Association (APHA). 1998. Standard Methods for the

Examination of Water and Wastewater. 20th edition. American Public Health

Association, Washington, D.C.

Arrhenius S. 1889. Über die reaktionsgeschwindigkeit bei der inversion von rohrzucker durch

säuren. Zeitschrift für Physikalische Chemie. 4:266-248.

116

Aufdenkampe, A. K., E. Mayorga, P. A. Raymond, J. M. Melack, S. C. Doney, S. R. Alin, R. E.

Aalto, and K. Yoo. 2011. Riverine coupling of biogeochemical cycles between land,

oceans, and atmosphere. Frontiers in Ecology and the Environment 9:53-60.

Bosatta E., and G. I. Ågren. 1999. Soil organic matter quality interpreted thermodynamically.

Soil Biology and Biochemistry 31:1889-1891.

Butman D., and P. A. Raymond. 2011. Significant efflux of carbon dioxide from streams and

rivers in the United States. Nature Geoscience 4:839-842.

Cheever B. M., J. R. Webster, E. E. Bilger, and S. A. Thomas. 2013. The relative importance of

exogenous and substrate-derived nitrogen for microbial growth during leaf

decomposition. Ecology 94:1614-1625.

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

Duarte, P. Kortelainen, J. A. Downing, J. J. Middelburg, and J. Melack. 2007. Plumbing

the global carbon cycle: Integrating inland waters into the terrestrial carbon budget.

Ecosystems 10:172-185.

Cross W. F., J. M. Hood, J. P. Benstead, A. D. Huryn, and D. Nelson. 2015. Interactions between

temperature and nutrients across levels of ecological organization. Global Change

Biology 21:1025-1040.

Dang C.K., M. Schindler, E. Chauvet, and M. O. Gessner. 2009. Temperature oscillation coupled

with fungal community shifts can modulate warming effects on litter decomposition.

Ecology 90:122-131.

Findlay, S., J. Tank, S. Dye, H. M. Valett, P. J. Mulholland, W. H. McDowell, S. L. Johnson, S.

K. Hamilton, J. Edmonds, W. K. Dodds, and W. B. Bowden. 2002. A cross-system

117

comparison of bacterial and fungal biomass in detritus pools of headwater streams.

Microbial Ecology 43:55-66.

Ferreira, V., and E. Chauvet. 2011. Future increase in temperature more than decrease in litter

quality can affect microbial litter decomposition in streams. Oecologia 167:279-291.

Ferreira, V., and E. Chauvet. 2011. Synergistic effects of water temperature and dissolved

nutrients on litter decomposition and associated fungi. Global Change Biology 17:551-

564.

Ferreira V., V. Gulis, C. Pascoal, and M. A. S. Graça. 2014. Stream pollution and fungi. Pages

389-412 in G. Jones, K. D. Hyde and K-L Pang editors. Freshwater Fungi and Fungal-

like Organisms.De Gruyter, Berlin, Germany.

Ferreira V., B. Castagneyrol, J. Koricheva, V. Gulis, E. Chauvet, and M. A. S. Graça. 2015. A

meta-analysis of the effects of nutrient enrichment on litter decomposition in streams:

Nutrient enrichment and litter decomposition. Biological Reviews 90:669-688.

Fierer N., J. M. Craine, K. McLauchlan, and J. P. Schimel. 2005. Litter quality and the

temperature sensitivity of decomposition. Ecology: 86:320-326.

Gulis V., and K. Suberkropp. 2003. Leaf litter decomposition and microbial activity in nutrient-

enriched and unaltered reaches of a headwater stream. Freshwater biology 48:123-134.

Gulis V., A. D. Rosemond, K. Suberkropp, H. S. Weyers, and J. P. Benstead. 2004. Effects of

nutrient enrichment on the decomposition of wood and associated microbial activity in

streams. Freshwater Biology 49:1437-1447.

Gulis, V., and K. Suberkropp. 2006. Fungi: Biomass, production, and sporulation of aquatic

hyphomycetes. Pages 311-326 in F.R. Hauer and G.A. Lamberti, editors. Methods in

Stream Ecology. Second Edition. Academic Press.

118

Gillooly J. F., J. H. Brown, G. B. West, V. M. Savage, and E. L. Charnov. 2001. Effects of Size

and Temperature on Metabolic Rate. Science 293:2248-2251.

Gessner M. O., and E. Chauvet. 1993. Ergosterol-to-biomass conversion factors for aquatic

hyphomycetes. Applied and environmental microbiology 59:502-507.

Gessner M. O., M. Thomas, A. M. Jean-Louis, and E. Chauvet. 1993. Stable successional

patterns of aquatic hyphomycetes on leaves decaying in a summer cool stream.

Mycological Research 97:163-172.

Gessner, M. O., V. Gulis, K. A. Kuehn, E. Chauvet, and K. Suberkropp. 2007. Fungal

decomposers of plant litter in aquatic ecosystems. Pages 301-324 in C. P. Kubicek, and I.

S. Druzhinina, editors. The Mycota. Springer. Berlin.

Hieber M., and M. O. Gessner. 2002. Contribution of stream detrivores, fungi, and bacteria to

leaf breakdown based on biomass estimates. Ecology 83:1026-1038.

IPCC. 2007. Climate Change 2007: The Physical Sciences Basis. Pages 7-22 in Contribution of

Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on

Climate Change. M. L. Parry, O. F. Canziani, J. P. Palutikof, P. H. van der Linden and C.

E. Hanson, editors. Cambridge University Press, Cambridge, UK.

Jankowski K., D. E. Schindler, and P. J. Lisi. 2014. Temperature sensitivity of community

respiration rates in streams is associated with watershed geomorphic features. Ecology

95:2707-2714.

Kaushal, S. S., G. E. Likens, N. A. Jaworski, M. L. Pace, A. M. Sides, D. Seekell, K. T. Belt, D.

H. Secor, and R. L. Wingate. 2010. Rising stream and river temperatures in the United

States. Frontiers in Ecology and the Environment 8:461-466.

119

Kominoski J. S., A. D. Rosemond, J. P. Benstead, V. Gulis, J. C. Maerz, and D. W. P. Manning

2015. Low-to-moderate nitrogen and phosphorus concentrations accelerate microbially

driven litter breakdown rates. Ecological Applications 25:856-865.

Perkins, D. M., G. Yvon-Durocher, B. O. L. Demars, J. Reiss, D. E. Pichler, N. Friberg, M.

Trimmer, and G. Woodward. 2012. Consistent temperature dependence of respiration

across ecosystems contrasting in thermal history. Global Change Biology 18:1300-1311.

R Development Core Team. 2013. R: a language and environment for statistical computing. R

Foundation for Statistical Computing, Vienna, Austria.

Rosemond, A. D., J. P. Benstead, P. M. Bumpers, V. Gulis, J. S. Kominoski, D. W. P. Manning,

K. Suberkropp, and J. B. Wallace. 2015. Experimental nutrient additions accelerate

terrestrial carbon loss from stream ecosystems. Science 347:1142-1145.

Scott, M. C., G. S. Helfman, M. E. McTammany, E. F. Benfield, and P. V. Bolstad. 2002.

Multiscale influences on physical and chemical stream conditions across Blue Ridge

landscapes. Journal of the American Water Resources Association 38:1379-1392.

Sinsabaugh R. L., and J. J. Follstad Shah. 2012. Ecoenzymatic Stoichiometry and Ecological

Theory. Annual Review of Ecology, Evolution, and Systematics, 43:313-343.

Sinsabaugh R. L., J. J. F. Shah, S. G. Findlay, K. A. Kuehn, and D. L. Moorhead. 2015. Scaling

microbial biomass, metabolism and resource supply. Biogeochemistry 122:175-190.

Suberkropp K., V. Gulis, A. D. Rosemond, J. P. Benstead. 2010. Ecosystem and physiological

scales of microbial responses to nutrients in a detritus-based stream: Results of a 5-year

continuous enrichment. Limnology and Oceanography 55:149-160.

120

Stelzer R. S., J. Heffernan J, and G. E. Likens. 2003. The influence of dissolved nutrients and

particulate organic matter quality on microbial respiration and biomass in a forest stream.

Freshwater Biology, 48:1925-1937.

Swank W. T., and D. A. Crossley. 1988. Forest hydrology and ecology at Coweeta. Springer-

Verlag. New York, New York, USA.

Tant C. J., A. D. Rosemond AD, and M. R. First. 2013. Stream nutrient enrichment has a greater

effect on coarse than on fine benthic organic matter. Freshwater Science 32:1111-1121.

Yvon-Durocher, G., J. M. Caffrey, A. Cescatti, M. Dossena, P. del Giorgio, J. M. Gasol, J. M.

Montoya, J. Pumpanen, P. A. Staehr, M. Trimmer, G. Woodward, and A. P. Allen. 2012.

Reconciling the temperature dependence of respiration across timescales and ecosystem

types. Nature 487:472–476.

Webster J. R., E. F. Benfield, T. P. Ehrman, M. A. Schaeffer, J. L. Tank, J. J. Hutchens, D. J.

D’Angelo. 1999. What happens to allochthonous material that falls into streams? A

synthesis of new and published information from Coweeta. Freshwater Biology 41:687-

705.

Welter J. R., J. P. Benstead, W. F. Cross, J. M. Hood, A. D. Huryn, P. W. Johnson, and T. J.

Williamson. 2015. Does N2 fixation amplify the temperature dependence of ecosystem

metabolism? Ecology 96:603-610.

Woodward, G., et al. 2012. Continental-scale effects of nutrient pollution on stream ecosystem

functioning. Science 336:1438-1440.

Supplementary Material

Appendix D. Stream nutrient treatments, measured nutrient concentrations and mean seasonal

temperatures for each study stream during the experiment.

121

Appendix E. Parameter estimates for linear models describing deployed litterbag and wood

veneer respiration rates.

Appendix F. Linear models describing temperature and nutrient enrichment effects on fungal

biomass-specific respiration rates.

Appendix G. Comparison of respiration on deployed subtrates when considering all data, early

and middle respiration only, and data within the interquartile range of all available data.

122

Tables

Table 4.1. Parameter estimates and 95% confidence intervals (95% CI) from the linear model for

respiration rates on naturally occurring detritus (Adj-R2 = 0.60). Intercepts correspond to the

mean respiration rate (ln R[T], originally in mg O2/g AFDM/hr) expected at 10°C for each

substrate (FBOM, leaf litter and wood) and for each substrate during YR1 and YR2 compared to

PRE. Slope estimates for this model are equivalent to the activation energy (E [eV, 1 eV = 1.6 ×

10-19 J]) of respiration for each substrate; interactions between substrate and year test for

differences in E between PRE and enrichment years (YR1, YR2). Significant differences

between intercepts or slopes (P < 0.05) are emphasized with bold text, and can be interpreted as

the mean difference between a given year and substrate in comparison to FBOM (for substrates)

or PRE (for year).

Parameter Estimate 95% CI

Intercepts

FBOM -2.825 (-2.97, -2.68)

Leaf litter 0.606 (0.40, 0.81)

Wood -1.228 (-1.43, -1.03)

FBOM * YR1 -0.177 (-0.39, 0.04)

FBOM * YR2 0.016 (-0.20, 0.23)

YR1* Leaf litter 0.478 (0.17, 0.78)

YR2 *Leaf litter 0.222 (-0.08, 0.53)

YR1 * Wood 0.394 (0.09, 0.69)

YR2 * Wood 0.387 (0.08, 0.69)

123

Slopes (E)

Tc * FBOM -0.432 (-0.68 -0.19)

Tc * Leaf litter 0.133 (-0.20 0.46)

Tc * Wood 0.009 (-0.32 0.34)

Tc * FBOM * YR1 -0.222 (-0.58 0.13)

Tc * FBOM * YR2 0.146 (-0.33 0.62)

Tc * Leaf litter * YR1 -0.108 (-0.61 0.40)

Tc * Leaf litter * YR2 0.004 (-0.65 0.66)

Tc * Wood * YR2 0.115 (-0.39 0.62)

Tc * Wood * YR2 0.165 (-0.49 0.82)

124

Table 4.2. Parameter estimates (95% confidence intervals [95% CI]) for linear models relating

leaf litter and wood veneer respiration rates to temperature and nutrient enrichment (i.e., year).

Intercepts are equivalent to the mean respiration rate (ln R[T], originally in mg O2/g AFDM/hr)

at 10°C for a given substrate and year. Slopes (all parameters including the term Tc) are

interpreted as the apparent activation energy of respiration (E in eV [1 eV = 1.6 × 10-19 J]).

Confidence intervals that did not include zero are emphasized with bold text (i.e., P < 0.05), and

can be interpreted as the mean difference in respiration rate, or E between a given year and

substrate in comparison to leaf litter (for substrates) or PRE (for year).

Parameter Estimate 95% CI

Intercepts

Litter -2.080 (-2.247, -1.913)

Wood -0.871 (-1.175, -0.567)

YR1 * Litter -0.242 (-0.474, -0.010)

YR2 * Litter 0.147 (-0.249, 0.542)

YR1 * Wood 1.030 (0.611, 1.448)

YR2 * Wood 1.093 (0.419, 1.768)

Slopes (E)

Tc * Litter -0.858 (-1.166, -0.551)

Tc * Wood -1.418 (-1.980, -0.855)

Tc * YR1 * Litter 1.175 (0.712, 1.639)

Tc * YR2 * Litter 0.325 (-0.493, 1.144)

Tc * YR1 * Wood -1.798 (-2.743, -0.853)

Tc * YR2 * Wood 0.335 (-1.343, 2.012)

125

Figures

Fig. 4.1a-c. Temperature dependence of substrate-specific respiration rates (ln R[T]) for FBOM

(a), leaf litter (b) and wood (c) (Adj-R2 = 0.60). Standardized temperature (1/kT – 1/kTc, k =

Boltzmann constant 8.617 * 10-5 eV/K [1 eV = 1 * 10-19 J]) was centered at 10°C (mean

temperature for the duration of the study period in all five study streams). Open circles

correspond to pretreatment conditions, filled gray and black circles correspond to YR1 and YR2

of enrichment, respectively. Dashed black lines correspond to pretreatment slopes, and thicker

gray or black lines correspond to the slopes for YR1 and YR2. Activation energies (E) were

similar all three substrate sampled in this study in all three years, (E = 0.43 eV), but intercepts

(i.e., mean respiration rates) were significantly higher in YR1 for leaf litter and YR1 and YR2

for wood (all P < 0.05). Relationships that were not statistically different compared to PRE are

not shown for visual clarity (e.g., FBOM respiration vs. inverse temperature in YR1, YR2).

Fig. 4.2a-d. Temperature dependence of respiration rates associated with leaf litter (a) and wood

veneers (b) deployed in our study streams for known periods of time (14-160 days) (Adj-R2 =

0.27). Also depicted are the fungal biomass-specific respiration rates for leaf litter (c) and wood

veneers (d). Temperatures were centered at 10°C (approximate mean temperature for the

duration of the study period in all five study streams). Open circles correspond to PRE

respiration rates; filled gray and black circles denote YR1 and YR2, respectively. Dashed black

lines denote slopes (E [eV]) for pretreatment; thick gray and black lines denote slopes for YR1

and YR2, respectively.

Fig. 4.3a,b. Fungal biomass (mg/g AFDM) was unrelated to temperature (1/kT-1/kTc) for

naturally occurring detritus, and decreased as a function of temperature for deployed detritus

(Adj-R2 = 0.32 for deployed detritus). Overall fungal biomass increased during nutrient

126

enrichment for both naturally occurring detritus (1.23× in YR2 only, P < 0.05), and deployed

detritus (1.67× in both YR1 and YR2, P < 0.05) For deployed detritus under nutrient-enriched

conditions (both YR1 and YR2), the relationship between fungal biomass and temperature

became less negative (i.e., approached zero; P < 0.05), corresponding to greater response of

fungal biomass to nutrients at early stages of decay and colder stream temperatures. Open circles

denote pretreatment values, and the thin solid black line denotes the slope between fungal

biomass and inverse temperature for pretreatment. Closed gray and black circles correspond to

YR1 and YR2, respectively. The relationship between inverse temperature and fungal biomass

during nutrient enrichment is shown with a solid gray line (YR1) or solid thick black line (YR2).

Fig. 4.4a-c. Activation energy as a function of initial C:N (a) and C:P (b) and early fungal

biomass ([c]; d 14 [leaf litter] or d 21 [wood veneers]) for each of the detritus types deployed for

our study (maple [A], tulip poplar [L], oak [Q], rhododendron [R] and wood veneers [W]). Wood

activation energy was significantly greater than activation energy for maple (P < 0.05), and

rhododendron was marginally higher than maple (P = 0.08). Early fungal biomass on leaf litter

and wood veneers during pretreatment tended to decrease as a function of initial C:N and C:P

(data not shown, P = 0.07, 0.04, respectively).

127

Fig. 4.1a-c.

-1.0 -0.5 0.0 0.5 1.0

-6-5

-4-3

-2-1

0

FBOM

1 kT − 1 kT c

ln R(T)

PREYR1YR2

(a)

-1.0 -0.5 0.0 0.5 1.0

-6-5

-4-3

-2-1

0

Leaves

1 kT − 1 kT cln

R(T)

PREYR1YR2

(b)

-1.0 -0.5 0.0 0.5 1.0

-6-5

-4-3

-2-1

0

Wood

1 kT − 1 kT c

ln R(T)

PREYR1YR2

(c)

128

Fig. 4.2a-d.

-0.4 -0.2 0.0 0.2 0.4 0.6 0.8

-8-6

-4-2

0

1 kT − 1 kT c

ln R(T)

(a)PREYR1YR2 (ns)

-0.5 0.0 0.5

-8-6

-4-2

0

1 kT − 1 kT c

ln R(T)

(b)PREYR1YR2

-0.4 -0.2 0.0 0.2 0.4 0.6 0.8

-12

-10

-8-6

-4-2

0

1 kT − 1 kT c

Fung

al b

iom

ass-

spec

ific

resp

iratio

n

PREYR1YR2

(c)

-0.5 0.0 0.5

-12

-10

-8-6

-4-2

0

1 kT − 1 kT c

Fung

al b

iom

ass-

spec

ific

resp

iratio

n

PREYR1YR2

(d)

129

Fig. 4.3a,b.

-1.0 -0.5 0.0 0.5 1.0

-20

24

6

1 kT − 1 kT c

Fung

al b

iom

ass

(mg/

g A

FDM

)

(a)

PREYR1YR2

-1.0 -0.5 0.0 0.5 1.0

-20

24

6

1 kT − 1 kT c

Fung

al b

iom

ass

(mg/

g A

FDM

)

(b)

PREYR1YR2

130

Fig. 4.4a-d.

MP

O

R

W

80 100 120 140 160

1.0

1.5

2.0

Initial C:N

Temperature

dependence , E

(eV)

(a)

MP

O

R

W

2000 4000 6000 8000 10000

1.0

1.5

2.0

Intial C:PTemperature

dependence , E

(eV)

(b)

MP

O

R

W

0 5 10 15 20

1.0

1.5

2.0

Early fungal biomass (mg/g AFDM)

Temperature

dependence , E

(eV)

(c)

131

CHAPTER 5

NUTRIENTS ARE MORE IMPORTANT THAN DOC FOR INCREASING LEAF LITTER

DECOMPOSITION DESPITE THEIR COMBINED EFFECTS ON MICROBIAL BIOMASS

AND ACTIVITY1

1 David W. P. Manning, Amy D. Rosemond, Vladislav Gulis, John C. Maerz, Jenna L. Martin and Katie G. Norris. To be submitted to Freshwater Science.

132

Abstract. Increased nutrient and dissolved organic carbon (DOC) availability can stimulate

microbial decomposer activity in aquatic ecosystems, but relatively little is known about their

combined effects on microbial-driven detrital carbon (C) processing. Here, we tested the effects

of nutrient and labile (as dextrose) vs. recalcitrant (as leaf leachate) DOC additions on detrital

decomposition rates in stream mesocosms. We targeted elevated DOC and nutrient (nitrogen [N]

and phosphorus [P]) concentrations that were 2-3× ambient, and measured responses of

microbial decomposer (i.e., fungi) biomass, respiration rates, and red maple (Acer rubrum L.)

and rhododendron (Rhododendron maximum L.) leaf litter decomposition rates. Nutrient

addition, but not DOC, increased decay rates for both maple and rhododendron leaf litter by

1.5×, on average, relative to controls. We found no detectable effects of DOC on decay rates, or

interactions between elevated DOC and nutrients. Respiration rates also responded more to

nutrients compared to DOC, but DOC and nutrients affected fungal biomass in different ways.

Specifically, nutrients increased fungal biomass (up to 1.3×), but DOC suppressed fungal

biomass (~1.4× lower than controls in labile or recalcitrant DOC treatments). As a result,

respiration rates per gram fungal biomass were highest in labile DOC treatments, suggesting

relatively greater contributions of bacteria to respiration rates and/or microbial carbon use

efficiencies in the presence of elevated labile DOC. Leaf litter decomposition rates followed a

similar pattern, with relatively greater decomposition per unit fungal biomass when both labile

DOC and nutrients were available together. Our results provide evidence for greater importance

of streamwater nutrients compared to DOC for detrital decay rates in streams, as well as some

support for the possibility of altered detrital C processing through microbial pathways when both

labile DOC and nutrients are plentiful.

133

Introduction

Dissolved organic carbon (DOC) represents the largest fraction of organic C in streams,

but concentrations can vary according to several factors, including seasonal pulses of leaf litter

and subsequent leaching (Meyer et al. 1998), and land use change such as deforestation due to

logging or agriculture (Yamashita et al. 2011, Stanley et al. 2012). In the case of land use

change, altered hydrology and soil characteristics due to conversion of forested watersheds to

urban or agricultural land use can lead to increased DOC availability (e.g., Aitkenhead-Peterson

et al. 2009, Molinero and Burke 2009, Giling et al. 2014). Further, land use change can induce

DOC quality shifts from primarily recalcitrant compounds to greater proportions of labile DOC,

as a result of increased autochthonous- and/or microbially-derived DOC in streams impacted by

intense agriculture (Wilson and Xenopoulos 2009, Lu et al. 2014), although human disturbances

may predominantly mobilize aged C at larger scales (McCallister and del Giorgio 2012, Butman

et al. 2015).

Increased DOC concentrations may interact with other C pools in streams such as coarse

particulate organic matter (i.e., leaf litter) via ‘priming’ effects (e.g., Kuzyakov et al. 2000).

Several potential mechanisms are thought to control priming effects, which may enhance

recalcitrant C degradation. For example, labile C likely amplifies recalcitrant C decomposition

via increased microbial biomass and activity due to alleviation of C (energy) limitation resulting

in enhanced ‘mining’ of substrate-derived nutrients due to increased nutrient limitation (Guenet

et al. 2010). In contrast, priming may not be observed if microbial decomposers preferentially

use labile, dissolved C in lieu of recalcitrant, particulate C (Kuzyakov 2002). So far, priming

effects in aquatic ecosystems have largely been observed in terms of producer-decomposer

interactions, where the presence of algae and their labile C exudates can stimulate microbial

134

decomposers and leaf litter processing (e.g., Danger et al. 2013, Kuehn et al. 2014). However,

few studies have examined the potential effects of increased DOC from distinct watershed

sources due to land use change (labile/algal-derived vs. recalcitrant/terrestrially-derived) on key

ecosystem functions such as leaf litter decomposition (but see Bernhardt and Likens 2002).

Land-use change is also associated with increased availability of limiting nutrients such

as nitrogen (N) and phosphorus (P), which may interact with the effects of increased DOC

availability and subsequent priming in complex ways. For instance, both positive (Farjalla et al.

2009, Guenet et al. 2014) and undetectable (Hotchkiss et al. 2014, Guenet et al. 2014) effects of

nutrient availability on recalcitrant C priming have been observed. Two opposing mechanisms

may help explain the contrasting effects of labile C and nutrients on priming effects: 1) nutrients

could facilitate priming via stimulation of microbial biomass and enzyme synthesis to acquire

recalcitrant C substrates (Allison and Vitousek 2005), or 2) nutrients have no effect on priming.

More research is needed to determine if the overall result of these mechanisms is to generally

enhance detrital carbon processing rates when concentrations of N, P and DOC are

experimentally elevated.

The objective of this study was to examine interactions between increased concentrations

N and P with increased labile and recalcitrant DOC on microbial respiration, biomass, and

decomposition of detrital carbon (i.e., leaf litter). We addressed 2 questions: 1) How do increased

DOC concentrations affect microbial respiration and decomposition of leaf litter? and 2) Do

nutrients modify the effects of DOC on litter decomposition? We hypothesized that increasing

DOC concentrations would amplify litter decomposition rates for rhododendron more than maple

because fungi and bacteria could use DOC as an energy source to degrade the more recalcitrant

C compounds (e.g., lignin) found in rhododendron leaf litter (e.g., Klotzbücher et al. 2011).

135

Similarly, we predicted that labile DOC would have relatively greater effects on this process than

recalcitrant DOC. Alternatively, DOC could instead suppress litter processing because fungi and

bacteria may favor using DOC, instead of particulate detrital C, for growth (no priming effect,

Kuzyakov 2002), especially for labile DOC compared to recalcitrant DOC. In terms of nutrient

interactions with DOC, we predicted that N and P would further enhance priming effects when

they occurred by stimulating microbial (particularly fungal) biomass and activity and thus affect

decomposition rates. We explored these questions and hypotheses using a short-term (51 day)

nutrient (N+P) crossed with labile (dextrose) and recalcitrant DOC (leaf leachate) additions to

stream mesocosms, and measured the response of leaf litter decomposition rates and microbial

biomass and activity to elevated DOC and nutrients.

Methods

Site description and experimental design

This study took place at Coweeta Hydrologic Laboratory (CWT) a USDA Forest Service

research station and Long-Term Ecological Research site in Macon Co. North Carolina, USA

(35°03’38” N 83°35’55” W). The CWT basin is in the Blue Ridge Province of the southern

Appalachian Mountains. Stream water was pumped from Shope Fork, a third-order stream that

drains part of the CWT basin, into a 1500-L tank that fed three 378-L tanks connected to

corresponding platforms of ten aluminum stream channels (0.15 × 4 m) via adjustable spouts.

We adjusted flow rates in the channels weekly, and targeted rates of ~0.05 L/s. We covered each

of the channels with landscape cloth screens to minimize light in the channels. Streamwater

temperatures in the channels were monitored every 15 min during the experiment using Onset

HOBO Pendant temperature loggers (Bourne, MA, USA).

136

We added labile DOC (as dextrose; ADM, Decatur, Illinois, USA), recalcitrant DOC (as

leaf leachate) and nutrients (concentrated ammonium nitrate and phosphoric acid) to

corresponding stream channels using multichannel peristaltic pumps (Watson-Marlow,

Wilmington, Massachusetts, USA) that dosed DOC and/or nutrients to treatment channels from

20-L carboys filled with concentrated DOC and/or nutrient solutions. Control channels received

water from Shope Fork only. We used a fully factorial design comprising all combinations of

DOC and N+P additions with controls (i.e., we had six treatments: control, recalcitrant DOC

alone, labile DOC alone, N+P alone, labile DOC × N+P, and recalcitrant DOC × N+P). We had

four replicates for each treatment such that we used 24 stream channels for this study.

Our target concentrations for elevated DOC was 1.0 mg/L or ~2× ambient concentrations,

and target concentrations for N+P treatments were 13 µg/L SRP and 95 µg/L DIN or ~3× and 2×

ambient concentrations, respectively, with corresponding molar ratio of added N:P = 16. Our

target DOC concentrations reflect the lower range observed DOC concentrations associated with

agriculture in the southeast U.S. (0.78 – 4 mg/L DOC; Molinero and Burke 2009), and are ~20×

lower than previous experimental DOC additions at Coweeta (Wilcox et al. 2005). The target

concentrations of DIN and SRP are relatively low compared to human-impacted streams in the

region (Scott et al. 2002), but were associated with increased leaf litter decomposition rates in a

previous study conducted in these stream channels (Kominoski et al. 2015). We sampled

streamwater from central points in the stream channels for DOC, DIN, and SRP concentrations

on days 14, 35, and 51 of the experiment. Streamwater was filtered in the field using 0.45 µm

nitrocellulose filters into 20 mL polyethylene scintillation vials (for DIN and SRP), or 60 mL

brown polyethylene bottles (DOC). Dissolved nutrients were analyzed using an Alpkem Rapid

Flow Analyzer 300 (for DIN), or spectrophotometrically using the ascorbic acid method (for

137

SRP; Shimadzu UV-1700, Japan; APHA 1998). We measured DOC concentrations (mg/L) using

a TOC-5000A total organic carbon analyzer (Shimadzu, Japan).

Leaf leachate

Mixed leaf litter was dried for 24 h, coarsely ground using a Wiley mill (Thomas

Scientific, Swedesboro, NJ, USA) and soaked in water for 24 h (approximately 100 g leaf

litter/L). The resulting slurry was filtered through progressively smaller mesh sizes (final mesh

size = 63 µm), and then centrifuged (Allegra X-22R, Beckman Coulter, Indianapolis, IN, USA)

for 5-7 min at 3500 rpm to minimize fine particles in the leachate. The centrifuged leachate was

Tyndallized via sustained heating for ~30 min at 90°C to reduce future microbial growth in the

leachate (Kaplan et al. 2008). The final filtered and centrifuged leaf leachate had an average

concentration of 2.3 g/L C, and therefore was undiluted in our treatments (Recal. DOC, and

Recal. DOC * N+P). We assessed the molecular weight and aromatic content of the water in

each mesocosm using specific UV absorbance of the sample at 254 nm (SUVA254) with a

spectrophotometer (Shimadzu UV-1700) using a quartz cuvette with 1-cm path length.

Absorbance values were corrected for DOC concentration, such that SUVA254 = Abs254/DOC

(mg/L) (Weishaar et al. 2003).

Leaf litter decomposition rates

We measured leaf litter decomposition rates of red maple (Acer rubrum L.) and

rhododendron (Rhododendron maximum L.) in response to experimental additions of DOC, N

and P. Freshly abscised leaf litter was collected in October 2012 from the CWT basin, and air-

dried for several weeks in the laboratory. We weighed leaf litter into 3±0.1 g litterbags

138

constructed using 30×23 cm pieces of 1-mm fiberglass windowscreen that were folded into

14×20 cm envelopes and closed with staples. Initial leaf litter mass was weighed to the nearest

0.01 g for each litterbag. We placed seven litterbags of each species corresponding to seven

sampling dates into each of the 24 channels (n = 2 leaf species × 7 sampling dates × 24 channels

= 288 litterbags). We collected litterbags after incubation in the stream channels on days 0, 7, 14,

21, 35, 42, and 51 (day 0 = 11 June 2013). Day 0 litterbags were used to calculate handling

losses (Benfield 2006). Litterbags were placed into individual plastic bags, and transported to the

laboratory on ice. Within 12 h, leaf litter was removed from the litterbags, and dried at 55°C for

24 h. Dried leaf litter was weighed to the nearest 0.01 g, and a ~0.5 g subsample of the litter was

combusted for 4.5 h at 500°C to determine ash-free dry mass (AFDM).

Microbiological Analyses

We measured respiration rates associated with the incubated litter on days 7, 21, and 35

using methods outlined by Gulis and Suberkropp (2003). We measured respiration rates on ten

2×2 cm leaf pieces from each litterbag collected on that sampling date. The litter pieces were

placed into 30 mL chambers in streamwater in an incubator set to stream temperature. We

measured dissolved oxygen every 5-7 min for 30 min using YSI 5100 bench top dissolved

oxygen meters (Yellow Springs, OH, USA). Respiration was calculated as the slope of the

decline in dissolved oxygen concentration during the 30-min incubation and expressed as the

absolute value of the oxygen (mg) consumed per gram AFDM per hour. We then converted

respiration rates from mg O2 consumed per hour to g C respired per day using a respiration

quotient of 0.85, and correcting for the molar weight of O2 and C (Bott 2006) Thus, respiration

139

rates are comparable in terms of units to decomposition rates, k, and can be interpreted as g C

lost per day.

We also measured fungal biomass as the concentration of ergosterol associated with leaf

litter on day 35 of the decomposition experiment. We used standard methods outlined in Gulis

and Suberkropp (2006). We subsampled and froze pieces from rinsed leaf litter until analysis.

Lipids were extracted from freeze-dried ~2 × 2 cm, weighed leaf litter pieces using liquid-to-

liquid extraction. We used HPLC to determine ergosterol concentrations (LC-10VP, Shimadzu,

Columbia, Maryland, USA) with a Kinetex C18 column (Phenomenex, Torrance, California,

USA) and a UV detector set at 282 nm. We used external ergosterol standards (Acros Organics,

Geel, Belgium), and ergosterol concentrations were converted to fungal biomass using a standard

conversion factor of 5.5 µg of ergosterol per mg of fungal dry mass (Gessner and Chauvet 1993)

Data analyses

All statistical analyses were conducted using the statistical software R v. 3.0.1 (R

Development Core Team 2013). We used Analysis of Variance (ANOVA) to test for differences

in three leaf litter-associated responses (decomposition rate [k], respiration rate [mg C/g

AFDM/day], and fungal biomass [mg/g AFDM]) between treatments. Decomposition rate, k,

was determined according to the negative exponential model mt =m0 × e-kt where mt is the mass

remaining at time t, m0 is the initial mass, and k is the decay constant. Therefore, we used the

slope of the ln-transformed % litter mass remaining through time to determine k (Benfield 2006);

we excluded % mass remaining vs. time relationships that exhibited R2 < 0.4 (n = 6).

For our analysis of decomposition rates, microbial respiration and fungal biomass, we

used ANOVA with an overall model containing DOC treatment, N+P treatment, leaf litter

140

species and their interactions as predictors. We used multiple comparisons (Tukey’s Honestly

Significant Difference [Tukey’s HSD]) to explore differences between DOC and N+P treatments

or between leaf litter species and their interactions with the DOC and N+P treatments where

appropriate. Distribution of the data and model residuals were visually inspected to confirm

linearity and normality of the data to address the assumptions of linear models.

Results

DOC, N and P concentrations in mesocosms

Concentrations of DOC were significantly elevated (between 1.3-1.4× compared to

controls) in DOC treatments (F2,18 = 7.2, P = 0.005); labile and recalcitrant DOC treatments had

similar DOC concentrations (Tukey’s HSD, P = 0.98). In the N+P treatment without DOC

added, DOC concentrations were lower than controls (Table 1). Nutrient concentrations (N+P)

were elevated in N+P treatments, and ranged from 1.3-2.3× control concentrations for DIN, and

1.3-3.0× control concentrations for SRP (Table 1). Although nutrients and DOC were generally

elevated in accordance with our treatments, nutrient concentrations did not match target

concentrations in some cases. Specifically, nutrient concentrations were higher on average than

controls when either labile or recalcitrant DOC was added without nutrients, and nutrients were

lower than targeted concentrations in labile DOC * N+P treatments (Table 1, Appendix H: Table

G1).

Mean SUVA254 was statistically indistinguishable among DOC treatments (F2,18 =1.88,

P = 0.181), and nutrient treatments (F1,18 = 0.042, P = 0.84) although SUVA254 showed a

consistent trend toward lower values in all treatments (i.e., with nutrients or DOC) compared to

controls (Table 1).

141

Streamwater temperature in the channels was 18.3°C on average for the duration of the

experiment. Maximum temperature was 24.8°C, and minimum temperature was 13.9°C. Average

temperature differed slightly between the platforms containing each set of ten mesocosms (mean

difference = 0.2°C). Temperature of the water in the stream channels was generally higher than

streamwater temperatures at Coweeta for the same time period.

Leaf litter decomposition rates

Although nutrient concentrations were lower than targets, nutrient additions significantly

increased leaf litter decomposition rates (F1,30 =6.7, P = 0.01) but DOC addition had no effect

(F2,30 = 1.9, P > 0.05). Decomposition rates were approximately 1.4× higher than controls on

average when considering both maple and rhododendron together (Fig. 1a,b). Maple leaf litter

decomposition rates were 1.8× higher than rhododendron leaf litter decomposition rates (F1,30 =

11.5, P = 0.001), but this difference between leaf litter species was not altered by DOC or

nutrient addition (P > 0.05 for all interactions).

Microbial respiration rates

Nutrient additions significantly increased microbial respiration rates on both maple and

rhododendron leaf litter (F1,36 = 7.6, P < 0.05, Fig. 2a-c), but DOC additions had no effect (F2,36

= 0.56, P > 0.05). Nutrient treatments corresponded to 1.7× and 1.8× higher respiration rates on

average compared to controls for maple and rhododendron, respectively (Fig. 2a,b). Respiration

rates tended to be significantly higher on maple compared to rhododendron (F1,36 =13.0, P <

0.05; Fig. 2a,b), but this difference between leaf litter species was not altered by DOC or nutrient

addition (F2,36 = 2.4, P > 0.05).

142

Fungal biomass

There were significant, but opposing effects of DOC and nutrient addition on fungal

biomass associated with leaf litter. Specifically, nutrient additions significantly increased fungal

biomass (F1,36 = 13.0, P < 0.05, Fig. 3a,b), but DOC additions significantly decreased fungal

biomass (F2,36 = 17.9, P < 0.05, Fig 3a,b). Nutrient treatments alone increased fungal biomass

the most (1.3×) compared to controls, with smaller increases for both DOC * N+P treatments

compared to treatments when DOC was added alone (1.1× and 1.2×, for labile and recalcitrant

DOC additions, respectively, Fig. 3a,b). When DOC was added without nutrients, fungal

biomass tended to decrease in both labile and recalcitrant DOC treatments (1.4× and 1.3× lower

than control treatments, respectively; Fig. 3a,b). We found no evidence for differences in DOC

treatments (i.e., labile vs. recalcitrant) for this effect, or differences in either nutrient or DOC

effects for fungal biomass on maple vs. rhododendron leaf litter (P > 0.05 for both interaction

terms).

Decomposition rates vs. fungal biomass and microbial respiration

We normalized decomposition and rates by both fungal biomass and microbial

respiration to determine how nutrients and DOC affected the contributions of microorganisms to

leaf litter mass loss (i.e., via biomass accrual and respiration of C). We also normalized

respiration rates by fungal biomass to make similar inferences about respiration rates. In terms of

decomposition rates per mg fungal biomass, we found significant effects of DOC additions (F2,36

=9.8, P = 0.0003), but no effects of nutrient additions (F1,36 = 0.033, P =0.85). Labile DOC

additions significantly increased decomposition rates per mg fungal biomass compared to

controls (Tukey’s HSD, P = 0.0002, Fig. 4a). Decomposition rates per gram fungal biomass were

143

also significantly higher on rhododendron leaf litter compared to maple (F1,36 = 16.5, P =

0.0002), but this difference was not changed by DOC or nutrient additions (P > 0.05 for both

interaction terms). Despite nutrient effects on respiration rates, leaf litter decomposition rates

normalized by respiration rates were no different for controls vs. nutrient additions, or for DOC

treatments, and were equivalent for both leaf litter types (all P > 0.05, Fig. 4b).

Respiration rates per unit fungal biomass increased with labile, but not recalcitrant DOC

additions (F2,35 = 8.9, P = 0.0007, Fig. 4c), and were not statistically different than controls with

nutrient additions (F1,35 = 0.12, P = 0.73). Fungal biomass-corrected respiration rates in labile

DOC treatments were 2.2×, and 1.6× higher than controls and recalcitrant DOC treatments,

respectively (Tukey’s HSD, P = 0.005 and 0.03, respectively). Fungal biomass-corrected

respiration rates were significantly higher on rhododendron litter compared to maple (F1,35 =

18.8, P = 0.0001), but this difference was unchanged by nutrient or DOC additions (P > 0.05 for

both interaction terms).

Discussion

Leaf litter decomposition is an important ecosystem process in aquatic habitats that can

be increased by nutrient availability due to stimulation of microbial biomass and activity (e.g.,

Gulis et al. 2003, Greenwood et al. 2007, Suberkropp et al. 2010). Much less is known about the

effects of increased DOC availability, despite the potential for increased DOC to ‘prime’ the

degradation of recalcitrant leaf litter C (Guenet et al. 2010), or how elevated nutrients could alter

the occurrence of such priming effects. Our study emphasizes the importance of nutrient

availability for leaf litter decomposition, but provides little evidence for priming effects via labile

or recalcitrant DOC on leaf litter processing. Increased leaf litter decomposition via nutrient-

144

mediated effects on microbial biomass and activity are well documented (Ferreira et al. 2006,

Kominoski et al. 2015). Our study reveals that these nutrient effects on microbial biomass and

activity are unlikely to interact with parallel increases in DOC availability to affect

decomposition rates. We found that microbial respiration rates were also affected more by

nutrient availability compared to elevated DOC, consistent with stronger nutrient effects on

decomposition rates. Notably, nutrient and DOC additions had significant, but opposing effects

on fungal biomass associated with leaf litter, where nutrients increased fungal biomass on leaf

litter, and DOC suppressed fungal biomass. This result corresponded to higher decomposition

and respiration rates per unit of fungal biomass, particularly when labile DOC and nutrients were

added together. These results provide some evidence that DOC and nutrient additions could alter

microbial processing of detritus in complex ways, which may warrant greater attention in future

studies.

Our treatments effectively increased DOC and/or nutrient concentrations in

corresponding mesocosms, but some treatments exhibited mismatches between nutrient

concentrations and target concentrations, potentially reflecting complex microbial nutrient

uptake dynamics in our mesocosms. For instance, DIN and SRP were lower in the labile DOC *

N+P treatment compared to when labile DOC was added alone. Nutrient concentrations were

elevated compared to controls in both labile and recalcitrant DOC only treatments. Leaf leachate

likely contained some dissolved nutrients in addition to DOC, which could partially explain why

we observed higher nutrient concentrations in the recalcitrant DOC treatments. Further, previous

studies have observed increased demand for NH4 and NO3 with labile DOC additions (Bernhardt

and Likens 2002); thus, the conflicting target vs. measured nutrient concentrations may be

partially explained by enhanced nutrient uptake throughout the experient in the labile DOC *

145

N+P treatments (Table 1). This effect may have been more important in nutrient treatments in

particular, given observed increases in fungal biomass and thus demand for streamwater N and P.

Conversely, we observed increases in nutrient concentrations through time in our Labile DOC

only treatments (Table 1), potentially as a result of suppressed fungal biomass in these

treatments. Consistent with these two contrasting patterns, we typically observed nutrient

concentrations that were above target concentrations when recalcitrant DOC or labile DOC was

added alone, and lower nutrient concentrations when nutrients were added alone, or with labile

DOC.

Leaf litter decomposition rates are largely driven by fungal vs. bacterial decomposers in

streams, even after nutrient enrichment (Baldy et al. 2007, Tant et al. 2013). The differential

importance of fungal vs. bacterial decomposers on leaf litter (Gessner and Chauvet 1994, Hieber

and Gessner 2002) and the greater effects of nutrients on litter decomposition rates compared to

DOC addition in our study suggests that nutrients are the more important driver of fungal

degradation of leaf litter in our study system. Nutrient availability may be more important for

fungi because fungi are known to preferentially use streamwater nutrients to satisfy their nutrient

requirements (Suberkropp 1995, Cheever et al. 2013), and nutrients are important for producing

enzymes, which are then used degrade complex detrital C rather than monomeric forms of DOC

(Erikkson 1984). Therefore, nutrient availability likely outweighs the importance of DOC

availability for the decomposition of coarse detritus, but DOC availability may have more

important effects where bacteria are more prevalent, such as on fine particles (i.e., allochthonous

soil organic matter, Guenet et al. 2014), or recalcitrant DOC (Hotchkiss et al. 2014).

While we found little evidence for DOC effects on decomposition rates in our study, we

were able to detect some evidence for DOC effects on microbial parameters that could influence

146

C processing in streams. Despite greater effects of nutrients on respiration rates, we found

respiration rates increased more per gram of fungal biomass in labile DOC treatments, which

suggests that labile DOC additions with or without nutrients either alters fungal growth and

carbon use efficiency due to preferential use of labile C (e.g., Manzoni et al. 2012), and/or

increases the amount of bacterial biomass contributing to increased respiration rates. We did not

assess the effects of bacteria specifically in this study, therefore we cannot rule out their potential

contributions to increased respiration rates, vs. effects on microbial carbon use efficiency.

Previous labile DOC additions in streams have documented increased bacterial biomass

and the development of thick microbial mats consisting of sheathed filamentous bacteria (e.g.,

bacteria of the genus Sphaerotilus; Hedin 1989, Bernhardt and Likens 2002, Wilcox et al. 2005).

Such microbial mats were also observed in our study, especially within and on the outside of our

litterbags (D. W. P. Manning, personal observation). It is plausible that these bacteria

contributed to increased respiration rates, and may have inhibited fungal biomass associated with

leaf litter as well (e.g., Romani et al. 2006). Bacteria growing on leaf litter without fungi

typically produce few enzymes, and rely on fungi to produce these enzymes (Romani et al.

2006). Therefore, nutrient stimulation of fungal biomass could have also enhanced the ability of

bacteria to respire C, which contributed to greater stimulation of respiration rates. This

mechanism could partly explain why labile DOC * N+P treatments showed greater

decomposition rates and respiration rates per unit of fungal biomass compared to controls, given

that increased fungal biomass due to nutrients potentially facilitated increased bacterial

respiration of C. Future work to fully explore the interactions between bacteria and fungi in

response to elevated DOC will require isotopic tracers (e.g., Guenet et al. 2014) coupled with

measurements of exoenzyme activity and bacterial biomass to partition the role of specific

147

mechanisms (preferential substrate use vs. increased microbial biomass) that drive altered

microbial processing of detrital C due to elevated DOC and nutrients.

We computed estimates of the proportion and total amount of C respired daily to explore

the potential for DOC-modulated microbial respiration to alter overall C processing in our

mesocosms. We used average respiration measurements on maple litter, daily C loads (based on

measured DOC concentrations × discharge [L/day], and mean particulate C in litterbags on day

35). We found that the amount of C respired per day was highest in the DOC * N+P treatments

was highest, followed closely by N+P treatments. Similarly, the proportion of available C that

was respired on a daily basis was highest in the labile DOC * N+P treatments, followed by the

nutrient treatments. These effects are consistent with relatively greater stimulation of respiration

by nutrients alone, and nutrients combined with labile DOC. In addition, this pattern occurred

despite 2× greater daily C loads in the DOC * N+P treatments compared to nutrient only

treatments. Thus, similar proportions of C were respired in both nutrient only and DOC * N+P

treatments despite differences in overall C availability, implying that that more DOC was

respired when it was added along with nutrients, and more detrital C was respired when nutrients

were added alone.

Examples of priming effects in aquatic ecosystems are often associated with increased

light, and the presence of primary producers on detritus (e.g., Danger et al. 2013, Kuehn et al.

2014, Rier et al. 2014). Our experiment aimed to minimize light and isolate the effects of

heterotrophic microbial decomposers such as fungi and their response to elevated DOC and

nutrients. This difference in experimental design may be a key reason for our inability to detect

priming effects. Previous studies that have detected priming effects in aquatic ecosystems have

invoked the presence of DOC from algal-exudates, or increased pH from photosynthesis as

148

potential mechanisms explaining the occurrence of priming effects. In particular, Kuehn et al.

(2014) show that algal-derived C is readily incorporated into microbial biomass; these exudates

are likely more complex polymers than the dextrose used in our study, and may require

extracellular enzymes to degrade. In addition, pH induced by photosynthetic activity (i.e. >9) is

considered optimal for several extracellular enzymes used to acquire C, such that photosynthetic

activity by itself could increase the effectiveness of enzymes and thus increase microbial

processing of detrital C (Francoeur and Wetzel 2003). Therefore, according to these two

potential mechanisms driving algae-induced priming effects in aquatic ecosystems, our study

likely had sub-optimal conditions for enzyme-producing fungi to induce priming effects, and

suggests that producer-decomposer interactions could be more important for predicting the

occurrence of priming effects than DOC availability per se in aquatic systems.

Our study focused on the effects of microbial processes on detrital decomposition rates in

response to elevated DOC and nutrients, but DOC and nutrient effects on microbial decomposers

could subsequently affect detritivorous macroinvertebrates (i.e., shredders), and thus

decomposition rates. The effects of nutrient enrichment on shredder-mediated effects on leaf

litter decomposition are well known (Woodward et al. 2012), and are generally attributed to

increased microbial biomass and nutrient immobilization on leaf litter (Rosemond et al. 2010,

Manning et al. 2015). Previous whole-stream DOC additions in detritus-based streams have

shown that labile DOC can support macroinvertebrate consumers including shredders (Wilcox et

al. 2005), but the potential effects of DOC on detritivores and subsequent detrital decomposition

are relatively unknown, and were not tested in our stream mesocosms that lacked shredders.

Overall, the elevated nutrient and DOC concentrations we achieved in this study were

moderate (~30% increase in DOC concentrations on average), and lower than some DOC

149

concentrations observed across watersheds or continents (e.g., Molinero and Burke 2009,

Bechtold et al. 2012). Our study adds to evidence for the importance of modest increases in

nutrient availability as a driver of increased detrital C loss (e.g., Rosemond et al. 2015), and

suggests that increasing some sources of DOC (i.e. dextrose, leaf leachate) will have negligible

effects on detrital C pools. Nevertheless, we show that elevated labile DOC and nutrients may

interact to affect microbial C processing in important, but complex ways when both are added

together. These effects are likely to be manifested in terms of greater importance of bacteria on

leaf litter, or potentially altered microbial C use efficiency. A more robust understanding of the

underlying mechanisms driving these patterns could be important for determining the response of

detrital C processing rates to increased DOC and nutrient availability from widespread watershed

disturbances.

Acknowledgements

We thank John Kominoski, Thomas Parr, and Nina Wurzburger for insightful comments

regarding our experimental design. This work was supported by NSF-REU supplemental funding

awarded to ADR and JCM (NSF). Phillip Bumpers, Jason Coombs, Kait Farrell, Meghan

Manning, James Wood, Tom Maddox, and Emmy Deng helped in the laboratory or in the field.

This study also leveraged logistical support from the CWT LTER Program at the University of

Georgia, which is supported by NSF award DEB-0823293 from the Long Term Ecological

Research Program (JCM co-PI). Rob Case, Daniel Hutcheson, and Kevin Simpson of YSI

Integrated Systems and Services constructed the stream channel system. Ammonium nitrate was

provided by The Andersons, Inc. through David Plank. We thank Jonathan Benstead, Alan

Covich, and Nina Wurzburger for critical feedback on earlier versions of this manuscript.

150

Literature Cited

Aitkenhead-Peterson, J. A., M. K. Steele, N. Nahar, and K. Santhy. 2009. Dissolved organic

carbon and nitrogen in urban and rural watersheds of south-central Texas: land use and

land management influences. Biogeochemistry 96:119–129.

Alexander, R. B., and R. A. Smith. 2006. Trends in the nutrient enrichment of U. S. rivers during

the late 20th century and their relation to changes in probable stream trophic conditions.

Limnology and Oceanography 51:639-654.

Allison, S. D., and P. M. Vitousek. 2005. Responses of extracellular enzymes to simple and

complex nutrient inputs. Soil Biology and Biochemistry 37:937–944.

American Public Health Association (APHA). 1998. Standard Methods for the

Examination of Water and Wastewater. 20th edition. American Public Health

Association, Washington, D.C.

Arsuffi, T. L., and K. Suberkropp. 1985. Selective feeding by stream caddisfly (Trichoptera)

detritivores on leaves with fungal-colonized patches. Oikos 45:50-58.

Arsuffi, T., and K. Suberkropp. 1989. Selective feeding by shredders on leaf-colonizing stream

fungi: comparison of macroinvertebrate taxa. Oecologia 79:30–37.

Baldy, V., V. Gobert, F. Guerold, E. Chauvet, D. Lambrigot, and J.-Y. Charcosset. 2007. Leaf

litter decomposition budgets in streams of various trophic status: effects of dissolved

inorganic nutrients on microorganisms and invertebrates. Freshwater Biology 52:1322–

1335.

Bechtold, H. A., A. M. Marcarelli, C. V. Baxter, and R. S. Inouye. Effects of N, P, and organic

carbon on stream biofilm nutrient limitation and uptake in a semi-arid watershed.

Limnology and Oceanography 57:1544-1554.

151

Benfield, E. F. 2006. Decomposition of leaf material. Pages 711-720 in F. R. Hauer and G. A.

Lamberti, editors. Methods in Stream Ecology. Second Edition. Academic Press.

Bernhardt, E. S., and G. E. Likens. 2002. Dissolved organic carbon enrichment alters nitrogen

dynamics in a forest stream. Ecology 83:1689–1700.

Bott, T. L. 2006. Primary productivity and community respiration. Pages 663-690 in F. R. Hauer

and G. A. Lamberti, editors. Methods in Stream Ecology. Second Edition. Academic

Press.

Butman D. E., H. F. Wilson, R. T. Barnes, M. A. Xenopoulos, and P. A. Raymond. 2015.

Increased mobilization of aged carbon to rivers by human disturbance. Nature

Geoscience 8:112–116.

Cheever, B. M., J. R. Webster, E. E. Bilger, and S. A. Thomas. 2013. The relative importance of

exogenous and substrate-derived nitrogen for microbial growth during leaf

decomposition. Ecology 94:1614–1625.

Danger, M., J. Cornut, E. Chauvet, P. Chavez, A. Elger, and A. Lecerf. 2013. Benthic algae

stimulate leaf litter decomposition in detritus-based headwater streams: a case of aquatic

priming effect? Ecology 94:1604–1613.

Eriksson, K. 1984. Enzymes involved in cellulose and lignin degradation. Applied Biochemistry

and Biotechnology 9:315–316.

Farjalla, V. F., C. C. Marinho, B. M. Faria, A. M. Amado, F. de A. Esteves, R. L. Bozelli, and D.

Giroldo. 2009. Synergy of Fresh and Accumulated Organic Matter to Bacterial Growth.

Microbial Ecology 57:657–666.

Ferreira, V., V. Gulis, and M. A. S. Graça. 2006. Whole-stream nitrate addition affects litter

decomposition and associated fungi but not invertebrates. Oecologia 149:718–729.

152

Francoeur, S. N., and R. G. Wetzel. 2003. Regulation of periphytic leucine-aminopeptidase

activity. Aquatic Microbial Ecology 31:249–258.

Gessner, M. O., and E. Chauvet. 1993. Ergosterol-to-biomass conversion factors for aquatic

hyphomycetes. Applied and environmental microbiology 59:502–507.

Gessner, M. O., and E. Chauvet. 1994. Importance of stream microfungi in controlling

decomposition rates of leaf litter. Ecology 76:1807–1817.

Giling, D. P., M. R. Grace, J. R. Thomson, R. Mac Nally, and R. M. Thompson. 2014. Effect of

Native Vegetation Loss on Stream Ecosystem Processes: Dissolved Organic Matter

Composition and Export in Agricultural Landscapes. Ecosystems 17:82–95.

Guenet, B., M. Danger, L. Abbadie, and G. Lacroix. 2010. Priming effect: bridging the gap

between terrestrial and aquatic ecology. Ecology 91:2850–2861.

Guenet, B., M. Danger, L. Harrault, B. Allard, M. Jauset-Alcala, G. Bardoux, D. Benest, L.

Abbadie, and G. Lacroix. 2014. Fast mineralization of land-born C in inland waters: first

experimental evidences of aquatic priming effect. Hydrobiologia 721:35–44.

Gulis, V., and K. Suberkropp. 2003. Leaf litter decomposition and microbial activity in nutrient-

enriched and unaltered reaches of a headwater stream. Freshwater biology 48:123–134.

Gulis, V., and K. Suberkropp. 2006. Fungi: Biomass, production, and sporulation of aquatic

hyphomycetes. Pages 311-326 in F.R. Hauer and G.A. Lamberti, editors. Methods in

Stream Ecology. Second Edition. Academic Press.

Hieber, M., and M. O. Gessner. 2002. Contribution of stream detrivores, fungi, and bacteria to

leaf decomposition based on biomass estimates. Ecology 83:1026.

153

Hotchkiss, E. R., R. O. Hall, M. A. Baker, E. J. Rosi-Marshall, and J. L. Tank. 2014. Modeling

priming effects on microbial consumption of dissolved organic carbon in rivers. Journal

of Geophysical Research: Biogeosciences 119:982–995.

Kaplan, L. A., T. N. Wiegner, J. D. Newbold, P. H. Ostrom, and H. Gandhi. 2008. Untangling

the complex issue of dissolved organic carbon uptake: a stable isotope approach.

Freshwater Biology 53:855-864.

Klotzbücher, T., K. Kaiser, G. Guggenberger, C. Gatzek, and K. Kalbitz. 2011. A new

conceptual model for the fate of lignin in decomposing plant litter. Ecology 92:1052–

1062.

Kuehn, K. A., S. N. Francoeur, R. H. Findlay, and R. K. Neely. 2014. Priming in the microbial

landscape: periphytic algal stimulation of litter-associated microbial decomposers.

Ecology 95:749–762.

Kuzyakov Y., J. K. Friedel, and K. Stahr. 2000. Review of mechanisms and quantification of

priming effects. Soil Biology and Biochemistry 32:1485–1498.

Kuzyakov, Y. 2002. Review: factors affecting rhizosphere priming effects. Journal of Plant

Nutrition and Soil Science 165:382–396.

Kominoski, J. S., A. D. Rosemond, J. P. Benstead, V. Gulis, J. C. Maerz, and D. W. Manning.

2015. Low-to-moderate nitrogen and phosphorus concentrations accelerate microbially

driven litter decomposition rates. Ecological Applications 25:856–865.

Lu, Y. H., J. E. Bauer, E. A. Canuel, R. M. Chambers, Y. Yamashita, R. Jaffé, and A. Barrett.

2014. Effects of land use on sources and ages of inorganic and organic carbon in

temperate headwater streams. Biogeochemistry 119:275–292.

154

Manning, D. W. P., A. D. Rosemond, J. S. Kominoski, V. Gulis, J. P. Benstead, and J. C. Maerz.

2015. Detrital stoichiometry as a critical nexus for the effects of streamwater nutrients on

leaf litter decomposition rates. Ecology 96: 2214-2224.

Manzoni, S., P. Taylor, A. Richter, A. Porporato, and G. I. Ågren. 2012. Environmental and

stoichiometric controls on microbial carbon-use efficiency in soils: Research review.

New Phytologist 196:79–91.

McCallister, S. L., and P. A. del Giorgio. 2012. Evidence for the respiration of ancient terrestrial

organic C in northern temperate lakes and streams. Proceedings of the National Academy

of Sciences 109:16963–16968.

Meyer, J. L., J. B. Wallace, and S. L. Eggert. 1998. Leaf litter as a source of dissolved organic

carbon in streams. Ecosystems 1:240–249.

Molinero, J., and R. A. Burke. 2009. Effects of land use on dissolved organic matter

biogeochemistry in piedmont headwater streams of the Southeastern United States.

Hydrobiologia 635:289–308.

Moorhead, D. L., and R. L. Sinsabaugh. 2006. A theoretical model of litter decay and microbial

interaction. Ecological Monographs 76:151–174.

R Development Core Team. 2013. R: a language and environment for statistical computing. R

Foundation for Statistical Computing, Vienna, Austria.

Rier, S. T., J. M. Shirvinski, and K. C. Kinek. 2014. In situ light and phosphorus manipulations

reveal potential role of biofilm algae in enhancing enzyme-mediated decomposition of

organic matter in streams. Freshwater Biology 59:1039–1051.

155

Romaní, A. M., H. Fischer, C. Mille-Lindblom, and L. J. Tranvik. 2006. Interactions of bacteria

and fungi on decomposing litter: differential extracellular enzyme activities. Ecology

87:2559–2569.

Rosemond, A. D., C. M. Swan, J. S. Kominoski, and S. E. Dye. 2010. Non-additive effects of

litter mixing are suppressed in a nutrient-enriched stream. Oikos 119:326–336.

Scott, M. C., G. S. Helfman, M. E. McTammany, E. F. Benfield, and P. V. Bolstad. 2002.

Multiscale influences on physical and chemical stream conditions across Blue Ridge

landscapes. Journal of the American Water Resources Association 38:1379–1392.

Stanley, E. H., S. M. Powers, N. R. Lottig, I. Buffam, and J. T. Crawford. 2012. Contemporary

changes in dissolved organic carbon (DOC) in human-dominated rivers: is there a role for

DOC management? Freshwater Biology 57:26–42.

Suberkropp, K. 1995. The influence of nutrients on fungal growth, productivity, and sporulation during

leaf decomposition in streams. Canadian Journal of Botany 73:S1361-S1369.

Suberkropp, K., V. Gulis, A. D. Rosemond, and J. P. Benstead. 2010. Ecosystem and

physiological scales of microbial responses to nutrients in a detritus-based stream:

Results of a 5-year continuous enrichment. Limnology and Oceanography 55:149–160.

Tant, C. J., A. D. Rosemond, and M. R. First. 2013. Stream nutrient enrichment has a greater

effect on coarse than on fine benthic organic matter. Freshwater Science 32:1111–1121.

Wilcox, H. S., J. Bruce Wallace, J. L. Meyer, and J. P. Benstead. 2005. Effects of labile carbon

addition on a headwater stream food web. Limnology and Oceanography 50:1300–1312.

Weishaar, J. L., G. R. Aiken, B. A. Bergamaschi, M. S. Fram, R. Fujii, and K. Mopper. 2003.

Evaluation of specific ultraviolet absorbance as an indicator of the chemical Composition

and reactivity of dissolved organic carbon. Environmental Science & Technology

37:4702–4708.

156

Wilson H. F., and M. A. Xenopoulos. 2009. Effects of agricultural land use on the composition

of fluvial dissolved organic matter. Nature Geoscience 2:37–41.

Woodward, G., M. O. Gessner, P. S. Giller, V. Gulis, S. Hladyz, A. Lecerf, B. Malmqvist, B. G.

McKie, S. D. Tiegs, and H. Cariss. 2012. Continental-scale effects of nutrient pollution

on stream ecosystem functioning. Science 336:1438–1440.

Yamashita, Y., B. D. Kloeppel, J. Knoepp, G. L. Zausen, and R. Jaffé. 2011. Effects of

Watershed History on Dissolved Organic Matter Characteristics in Headwater Streams.

Ecosystems 14:1110–1122.

Supplementary Material

Appendix H. Target vs. measured concentrations (µg/L) of dissolved organic carbon (DOC),

dissolved inorganic nitrogen (DIN), and soluble reactive phosphorus (SRP) in each treatment

mesocosm.

157

Tables

Table 5.1. Stream channel experimental design, and mean (±SE, n = 12) water chemistry

(dissolved organic carbon [DOC], dissolved inorganic nitrogen [DIN], and soluble reactive

phosphorus [SRP]) during the experiment. Treatments included control channels, and additions

of N and P, labile DOC (as dextrose), recalcitrant DOC (as leaf leachate), and combinations of

labile DOC with nutrients (Labile DOC * N+P and Recal. DOC * N+P).

DOC (mg/L)

Treatment Early Middle Late Mean (±SE)

Control 0.19 0.40 0.62 0.40 (0.09)

N+P 0.07 0.19 0.43 0.26 0.05)

Labile DOC 0.35 0.47 0.89 0.57 (0.08)

Recal. DOC 0.53 0.25 0.86 0.55 (0.1)

Labile DOC * N+P 0.30 0.66 0.60 0.54 (0.06)

Recal. DOC * N+P 0.35 0.34 0.92 0.54 (0.09)

DIN (µg/L)

Treatment Early Middle Late Mean (±SE)

Control 31.18 31.44 45.52 36.05 (5.04)

N+P 59.97 85.87 42.72 62.85 (12.06)

Labile DOC 44.94 69.41 101.45 71.94 (16.91)

Recal. DOC 62.30 35.16 40.58 46.01 (8.91)

Labile DOC * N+P 62.26 45.48 31.61 46.45 (7.49)

Recal. DOC * N+P 108.19 63.55 81.85 84.53 (12.91)

158

SRP (µg/L)

Treatment Early Middle Late Mean (±SE)

Control 4.56 4.93 5.35 4.95 (0.74)

N+P 7.11 17.26 6.28 10.22 (2.40)

Labile DOC 5.35 8.97 14.11 9.48 (2.19)

Recal. DOC 17.07 11.04 7.77 11.96 (2.29)

Labile DOC * N+P 5.35 8.97 14.11 6.64 (0.86)

Recal. DOC * N+P 21.64 9.56 14.34 15.18 (2.35)

SUVA254

Treatment Early Middle Late Mean (±SE)

Control 10.61 1.58 4.08 4.08 (1.70)

N+P 10.43 1.78 5.40 2.82 (1.68)

Labile DOC 5.38 0.85 3.03 2.09 (0.95)

Recal. DOC 4.85 1.91 3.01 2.26 (0.86)

Labile DOC * N+P 6.99 0.29 3.83 2.03 (1.38)

Recal. DOC * N+P 7.44 1.63 3.16 3.03 (1.17)

159

Figure Legends

Fig. 5.1a-c. Mean (±SE) litter decomposition rates for maple (a) and rhododendron (b) in each

DOC treatment (labile vs. recalcitrant [recal.]). Closed symbols represent treatments where

nutrients were added alone (closed control symbol) or in tandem with DOC. There were no

significant main effects of DOC treatment for either leaf litter species, but nutrients significantly

increased decomposition rates for both maple and rhododendron leaf litter (corresponding with

the asterisk in each legend denoting significant N+P treatment).

Fig. 5.2a-c. Mean (n = three dates [day 7, 21, 35] x 4 replicate treatments) respiration rates (g

C/g AFDM/day) for maple (a) and rhododendron (b) leaf litter or both leaf litter species (c) in

each DOC treatment (labile vs. recalcitrant [recal.]). Error bars signify ±1 SE. Closed symbols

and solid lines represent treatments where nutrients were added alone (closed control symbol) or

also with DOC. There were no significant main effects of DOC treatment for either leaf litter

species, but nutrients significantly increased respiration rates for both maple and rhododendron

leaf litter (corresponding with the asterisk in each legend denoting significant N+P treatment).

Fig. 5.3a-c. Mean fungal biomass (±SE) measured on day 35 of the experiment for maple (a) and

rhododendron (b) leaves by DOC treatment. Error bars correspond to ±1 SE. Closed circles and

solid lines denote treatments where N+P were added alone (closed control symbol) or in tandem

with DOC. Nutrients and DOC affected fungal biomass in opposing directions, with increased

and decreased fungal biomass due to nutrients and DOC additions, respectively.

Fig. 5.4a-c. Mean decomposition rates normalized by either fungal biomass (a, [mg/g AFDM])

or respiration rates (b, [g C/g AFDM/day]), and respiration rates normalized by fungal biomass

(c), for each DOC treatment and N+P additions. Error bars correspond to ±1 SE. Closed circles

and solid lines denote treatments where N+P were added alone (closed control symbol) or in

160

tandem with DOC. Decomposition rates per gram fungal biomass were significantly greater in

labile DOC * N+P treatments compared to controls (Tukey’s HSD, P < 0.05), but there was no

effect of DOC, nutrients or leaf litter type on decomposition rates per g C respired. Respiration

rates per mg fungal biomass were significantly higher in both DOC treatments (P < 0.05), and

nutrient addition had no effect on this parameter.

161

Fig. 5.1a,b.

0.000

0.005

0.010

0.015

0.020

Maple

DOC Treatment

Breakdown rate

(k)

Control Labile Recal

N+P *Control

(A)

0.000

0.005

0.010

0.015

0.020

Rhododendron

DOC Treatment

Breakdown rate

(k)

Control Labile Recal

N+P *Control

(B)

162

Fig. 5.2a,b.

0.000

0.001

0.002

0.003

0.004

Maple

DOC Treatment

Respiration rate

(g C

/ g AFDM

/ day)

Control Labile Recal

N+P *Control

(A)

0.000

0.001

0.002

0.003

0.004

Rhododendron

DOC Treatment

Respiration rate

(g C

/ g AFDM

/ day)

Control Labile Recal

N+P *Control

(B)

163

Fig. 5.3a,b.

010

2030

4050

60

Maple

DOC Treatment

Fungal

biomass (mg / g

AFDM)

Control Labile Recal

N+P *Control (A)

* *

010

2030

4050

60

Rhododendron

DOC Treatment

Fungal

biomass (mg / g

AFDM)

Control Labile Recal

N+P *Control (B)

* *

164

Fig. 5.4a-c.

0.0000

0.0006

0.0012

DOC Treatment

Bre

akdo

wn

rate

(k)/F

unga

l bio

mas

s

Control Labile Recal

N+PControl

(A)

24

68

10

DOC Treatment

Bre

akdo

wn

rate

(k)/R

espi

ratio

n ra

te

Control Labile Recal

N+PControl

(B)

0.004

0.008

0.012

DOC Treatment

Res

pira

tion

rate

/Fun

gal b

iom

ass

Control Labile Recal

N+PControl

(C)

165

CHAPTER 6

CONCLUSIONS

Nutrient pollution and the role of streams

Mobilization of N and P in watersheds is associated with several deleterious effects on

receiving aquatic ecosystems, including impairment of freshwater resources used for drinking

water, irrigation, and recreation (USEPA 2013). In lakes and coastal ecosystems, excess N and P

from upstream sources can lead to harmful algal blooms (Paerl et al. 2011) and large areas of

hypoxia (i.e., ‘dead zones’), which can negatively affect wildlife, human health, and coastal food

web production (Diaz and Rosenberg 2008, Smith and Schindler 2009).

Meeting the challenge of mitigating nutrient pollution requires considering the important

role of streams as the ‘first-line’ of response to nutrients mobilized in watersheds. For example,

streams can receive nutrient loads from human-modified upland or riparian nutrient inputs (e.g.,

agriculture, urbanization) via groundwater and/or overland flow (Mulholland 1992, Sudduth et

al. 2013); these nutrients are disproportionately removed or retained in small streams compared

to larger rivers, but some nutrients are ultimately transported downstream to lakes or coastal

oceans (Peterson et al. 2001, Wollheim et al. 2008). Importantly, stream ecosystem functions

likely respond differently to excess nutrients compared to lakes or coastal oceans because they

depend on detrital C, rather than autotrophic C (Webster et al. 1997, Wallace et al. 1999).

Currently, strategies to detect and mitigate the effects of nutrient loading typically rely on

stressor-response relationships between nutrient concentrations and autotrophic responses (i.e.,

algal biomass; Evans-White et al. 2013), despite the importance of detritus in streams and rivers.

166

For this reason, a stronger mechanistic understanding of elevated N and P effects on ecosystem

functions such as detrital processing is needed to better integrate nutrient management strategies

across watersheds, and account for distinct responses to nutrient loading in streams vs. lakes and

coastal zones.

The importance of detrital C in streams

This collection of studies represents a significant step toward a stronger mechanistic

understanding of N and P enrichment effects on a fundamental ecosystem process: detrital C

loss. Detritus comprises >99% of the energy budget supporting food webs in many streams

(Fisher and Likens 1973, Webster and Meyer 1997); therefore, the findings from our study

streams are likely extendable to similar streams in other areas of the globe. Detritus is also a

critical connection between adjacent ecosystems in both time and space; allochthonous detritus

produced in terrestrial ecosystems is fundamental for in-stream food web production (Wallace et

al. 1997, Walther and Whiles 2011), and local biogeochemical cycling and nutrient retention

(Mulholland 1992, Webster et al. 2009). In terms of longitudinal connections, reach-scale C

retention rates can partly determine the downstream abundance, form, and function of detrital C,

contributing to the overall role of river networks in the global C cycle (Cole et al. 2007).

Therefore, altered detrital C processing due to increased N and P availability is important to

consider in the context of local and global C cycling in streams and rivers.

Dissertation summary

There is a critical need to fully consider the effects of both N and P on detrital C

processing, because both N and P are often elevated simultaneously, and/or one nutrient may be

167

found in relatively greater supply than the other (i.e. high N, low P and vice versa; Arbuckle and

Downing 2001, Peñuelas et al. 2012). Our findings indicate that both increased N and increased

P were associated with increased detrital breakdown rates (Chapter 2). Litter breakdown rates for

maple and rhododendron increased up to 2.7 and 6.4× compared to pretreatment conditions,

respectively. These increases in breakdown rates via N and P effects occurred through subtly

different pathways. First, our path analysis revealed that both N and P were important predictors

of increasing fungal biomass. However, N content of detritus was driven by fungal biomass, and

P content of detritus was driven by a combination of P concentrations and fungal biomass. These

findings suggest that N immobilization is largely via fungal biomass accrual, but P may be

immobilized due to other factors, including storage in fungal biomass, increased bacterial

biomass, or abiotic sorption to sediments in the detrital matrix (e.g., Beever and Burns 1980,

Mehring et al. 2015). We cannot rule out any of these potential mechanisms with our data, but

microcosm studies that were complementary to this study have shown that P storage is possible

in fungal biomass associated with leaf litter (V. Gulis, unpublished data). Given the results of

this microcosm study, the importance of fungi on coarse detritus such as leaf litter, and the strong

relationships between fungal biomass, breakdown rates, and litter C:N:P, we suggest that the

storage mechanism is the most likely explanation for N and P effects on leaf litter breakdown

through different pathways in our study. Future research could further examine the mechanisms

by which N vs. P is differentially immobilized on leaf litter.

Our study targeted a range of N:P ratios corresponding to crossed N and P concentration

gradients. We initially hypothesized that the ratio of N:P could be a useful predictor of

breakdown rates, where relatively greater importance of N vs. P on this process would be

detected in terms of higher breakdown rates relative to pretreatment in our high N:P treatments

168

vs. low N:P treatments or vice versa. In general, we found little evidence for differing responses

across the N:P treatment gradient in terms of breakdown rates. That is, we found that leaf litter

breakdown (average of all four litter species for both enrichment years) responded similarly to

high N treatments (N:P = 128 response: 2.8×) and high P treatments (N:P = 2 response: 2.9×)

relative to pretreatment. As leaf litter breakdown integrates the action of several important

components of stream ecosystems, including aquatic fungi and macroinvertebrate activity, these

findings also suggest that these drivers of leaf litter breakdown rates could be responding across

the N:P gradient in a similar way. As a result, there was not a significant relationship between

N:P treatment and leaf litter breakdown response to nutrient enrichment, suggesting leaf litter

breakdown rates will increase if either N or P are slightly elevated relative to each other.

As a result of increased N and P immobilization, detrital C:N, and C:P were reduced,

which was associated with increased shredder biomass and increased detrital breakdown rates.

Based on our path analysis, we suggest that detrital stoichiometry is a crucial piece for predicting

where and when detrital C loss rates will be increased, given the finding that both N and P were

important for reducing detrital C:N or C:P content. Overall, nutrients had the effect of reducing

and homogenizing detrital nutrient content across all of the distinct detrital substrates used in this

study. We used these results to examine how reduced and homogenized detrital stoichiometry

could be used to predict the occurrence of increased breakdown rates (Chapter 3). In this regard,

our data indicated that N and P content of detritus increased relatively more for nutrient poor

detritus, compared to nutrient-rich detritus. Specifically, wood, and nutrient-poor leaf litter such

as rhododendron gained the most nutrients compared to other leaf litter species such as maple,

oak and poplar. Furthermore, reduced and homogenized detrital stoichiometry was associated

with increased breakdown rates, particularly when detrital C:nutrient content approached

169

shredder nutrient requirements (i.e., threshold elemental ratios [TERs], Frost et al. 2006). As a

result, breakdown rates tended to increase more below breakpoint C:N and C:P ratios

corresponding to closer matches between consumer nutrient requirements and detrital C:N or

C:P. Based on our breakpoint ratios for increased detrital breakdown rates, we propose that

detrital stoichiometry could be used as a management tool for detecting the effects of nutrient

pollution in detritus-based streams, particularly reduced C retention. Detrital stoichiometry may

be especially useful in this regard, as we demonstrate that it integrates the effects nutrients via

microbial decomposers, and subsequent shredder activity.

Microbial breakdown rates were also stimulated by nutrients, and microbial breakdown

was greater for detritus with lower initial C:N or C:P content. Much of the detrital C loss

attributed to microbial pathways is due to conversion of detrital C from leaf litter or wood to CO2

via respiration (Gulis and Suberkropp 2003). Consistent with nutrient enrichment increasing

microbial breakdown rates (Chapter 3), nutrient enrichment also stimulated microbial respiration

rates on leaf litter and wood (Chapter 4). We found little evidence for nutrient effects on fine

benthic organic matter (FBOM). Our results suggest that microbial respiration rates generally

increased by ~1.3× for naturally occurring leaf litter (YR1 only), and between 1.25× and 1.5× for

naturally occurring wood.

Rising temperatures due to climate change or land use change in aquatic ecosystems are

likely to impact carbon processing (Kaushal et al. 2010, Ferreira et al. 2014), but the interactions

between temperature and nutrients on detrital C have only recently begun to be investigated.

Therefore, we examined how nutrient enrichment might modify the temperature dependence of

microbial respiration rates (Chapter 4) in the context of the metabolic theory of ecology (MTE;

Gillooly et al. 2001, Brown et al. 2004). We found that respiration rates increased predictably

170

with temperature for both naturally occurring detritus, and deployed detritus, corresponding to

activation energies (E) that were either slightly lower than, or above MTE predictions for

naturally occurring and deployed detritus, respectively (average E = 0.43 eV, 1.15 eV,

respectively). However, nutrient enrichment had complex effects on temperature dependence of

microbial respiration depending on the substrate in question. For example, nutrient enrichment

had no detectable effects on the temperature dependence of respiration on naturally occurring

detritus, while nutrient enrichment altered the temperature dependence of respiration on

deployed detritus. The difference in the response between these two types of detritus was likely

due to the timing of fungal biomass accrual. We sampled submerged, colonized leaf litter and

wood material that had time to develop intact fungal communities and biomass, while deployed

detritus likely comprised fungal communities at different stages of succession along the decay

sequence (e.g., Gessner 1993, Gessner et al. 2007, Dang et al. 2009). Therefore, our data suggest

that nutrient enrichment may have negligible effects on the activation energy of respiration rates

on detritus at similar stages of decay, and nutrients could decrease the apparent activation energy

of respiration via increased fungal biomass and respiration rates at early stages of decay

coinciding with cold stream temperatures. These two results suggest that further consideration of

the responses of cold-adapted fungi vs. warm-adapted fungi to both nutrient enrichment and

increasing temperatures could be important for predicting in-stream microbial processing of

detritus and CO2 flux.

Along with nutrients, land-use change is expected to increase concentrations of DOC in

some cases (Stanley et al. 2012). These increased DOC concentrations likely affect microbial

processing of detritus in complex ways, particularly with regard to potential ‘priming’ effects of

DOC on recalcitrant detrital C such as leaf litter or wood (e.g., Guenet et al.2010). As well, land-

171

use change may bring changes to the bioavailability of DOC, based on expected increases in

labile algal- or microbial-derived C when intense agriculture is prevalent (Wilson and

Xenopoulos 2008, Giling et al. 2014, Lu et al. 2014). We predicted that increases in DOC

concentrations and lability would induce priming effects on recalcitrant detrital C such as maple

or rhododendron leaf litter (e.g., Kuehn et al. 2014), and that nutrients would enhance these

priming effects by stimulating microbial activity. The results of our study did not support this

hypothesis, as we found little evidence for DOC effects on breakdown rates of either maple or

rhododendron leaf litter. Instead, we found evidence for significant nutrient enrichment effects,

emphasizing the importance of nutrients for increasing detrital C loss rates. Our results provided

some evidence for increased microbial respiration rates in response to elevated DOC when

nutrients were simultaneously added. These increases in respiration rates with DOC and nutrient

additions were somewhat counterintuitive, because DOC additions suppressed fungal biomass

associated with leaf litter. Therefore, DOC additions most likely stimulated bacterial biomass on

leaf litter (e.g., Wilcox et al. 2005), or potentially increased the efficiency of fungi (e.g.,

Manzoni et al. 2012). Future work using stable isotopes (i.e., 13C), along with quantification of

bacterial biomass and enzyme activity would be a productive way to shed more light on the

mechanisms by which elevated DOC might drive increasing microbial activity that is decoupled

from detrital C loss.

The results of these studies underscore how local, reach-scale alterations to N and P

availability can affect C processing. We found that N and P are both critically important for

driving effects on detrital C loss rates, particularly regarding their dual effects on detrital

stoichiometry and microbial and detritivore pathways. In general, these mechanisms caused

faster detrital C loss rates, indicating that detrital C resources will likely be depleted more rapidly

172

from stream reaches with elevated nutrients (e.g., Suberkropp et al. 2010) and ecosystems

services associated with detritus will be diminished (e.g., nutrient retention, food web

production; Wallace et al. 1997, Webster et al. 2000). Our findings are consistent with the idea

that these three elements (C, N, and P) are coupled through basic biochemical processes (Sterner

and Elser 2002), giving rise to predictable patterns of their storage and flux from the level of

microbial decomposers to entire stream reaches and beyond. In this context, we provide strong

evidence for elevated concentrations of both N and P to significantly alter the flux of detrital C to

adjoining ecosystems, and the contributions of streams to local and global C cycling.

Literature Cited

Brown, J. H., J. F. Gillooly, A. P. Allen, V. M. Savage, and G. B. West. 2004. Toward a

metabolic theory of ecology. Ecology 85:1771-1789.

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

Duarte, P. Kortelainen, J. A. Downing, J. J. Middelburg, and J. Melack. 2007. Plumbing

the global carbon cycle: integrating inland waters into the terrestrial carbon budget.

Ecosystems 10:172-185.

Dang, C. K., M. Schindler, E. Chauvet, and M. O. Gessner. 2009. Temperature oscillation

coupled with fungal community shifts can modulate warming effects on litter

decomposition. Ecology 90:122-131.

Diaz, R. J., and R. Rosenberg. 2008. Spreading dead zones and consequences for marine

ecosystems. Science 321:926-929.

Evans-White, M. A., B. E. Haggard, and J. T. Scott. 2013. A review of stream nutrient criteria

development in the United States. Journal of Environmental Quality 42:1002-1014.

173

Ferreira V., V. Gulis, C. Pascoal, and M. A. S. Graça. 2014. Stream pollution and fungi. Pages

389-412 in G. Jones, K. D. Hyde, and K-L Pang, editors. Freshwater Fungi and Fungal-

like Organisms. De Gruyter, Berlin, Germany.

Finlay, J. C., G. E. Small, and R. W. Sterner. 2013. Human influences on nitrogen removal in

lakes. Science 342:247-250.

Fisher, S. G., and G. E. Likens. 1973. Energy flow in Bear Brook, New Hampshire: an

integrative approach to stream ecosystem metabolism. Ecological Monographs 43:421-

439.

Frost, P. C., J. P. Benstead, W. F. Cross, H. Hillebrand, J. H. Larson, M. A. Xenopoulos, and T.

Yoshida. 2006. Threshold elemental ratios of carbon and phosphorus in aquatic

consumers. Ecology Letters 9:774-779.

Giling, D. P., M. R. Grace, J. R. Thomson, R. Mac Nally, and R. M. Thompson. 2014. Effect of

native vegetation loss on stream ecosystem processes: dissolved organic matter

composition and export in agricultural landscapes. Ecosystems 17:82-95.

Gillooly, J. F., J. H. Brown, G. B. West, V. M. Savage, and E. L. Charnov. 2001. Effects of size

and temperature on metabolic rate. Science 293:2248-2251.

Gessner, M. O., V. Gulis, K. A. Kuehn, E. Chauvet, and K. Suberkropp. 2007. Fungal

decomposers of plant litter in aquatic ecosystems. Pages 301-324 in C. P. Kubicek, and I.

S. Druzhinina, editors. The Mycota. Springer. Berlin.

Guenet, B., M. Danger, L. Abbadie, and G. Lacroix. 2010. Priming effect: bridging the gap

between terrestrial and aquatic ecology. Ecology 91:2850-2861.

174

Kaushal, S. S., G. E. Likens, N. A. Jaworski, M. L. Pace, A. M. Sides, D. Seekell, K. T. Belt, D.

H. Secor, and R. L. Wingate. 2010. Rising stream and river temperatures in the United

States. Frontiers in Ecology and the Environment 8:461-466.

Kuehn, K. A., S. N. Francoeur, R. H. Findlay, and R. K. Neely. 2014. Priming in the microbial

landscape: periphytic algal stimulation of litter-associated microbial decomposers.

Ecology 95:749-762.

Lu, Y. H., J. E. Bauer, E. A. Canuel, R. M. Chambers, Y. Yamashita, R. Jaffé, and A. Barrett.

2014. Effects of land use on sources and ages of inorganic and organic carbon in

temperate headwater streams. Biogeochemistry 119:275-292.

Manzoni, S., P. Taylor, A. Richter, A. Porporato, and G. I. Ågren. 2012. Environmental and

stoichiometric controls on microbial carbon-use efficiency in soils. New Phytologist

196:79-91.

Mehring, A. S., K. A. Kuehn, A. Thompson, C. M. Pringle, A. D. Rosemond, M. R. First, R. R.

Lowrance, and G. Vellidis. 2015. Leaf litter nutrient uptake in an intermittent blackwater

river: influence of tree species and associated biotic and abiotic drivers. Functional

Ecology 29:849-860.

Mulholland, P. J. 1992. Regulation of nutrient concentrations in a temperate forest stream: roles

of upland, riparian, and instream processes. Limnology and Oceanography 37:1512-1526.

Paerl, H. W., N. S. Hall, and E. S. Calandrino. 2011. Controlling harmful cyanobacterial blooms

in a world experiencing anthropogenic and climatic-induced change. Science of The

Total Environment 409:1739-1745.

175

Peterson, B. J., et al. 2001. Control of nitrogen export from watersheds by headwater streams.

Science 292:86-90.

Smith, V. H., and D. W. Schindler. 2009. Eutrophication science: where do we go from here?

Trends in Ecology & Evolution 24:201-207.

Stanley, E. H., S. M. Powers, N. R. Lottig, I. Buffam, and J. T. Crawford. 2012. Contemporary

changes in dissolved organic carbon (DOC) in human-dominated rivers: is there a role for

DOC management? Freshwater Biology 57:26-42.

Sudduth, E. B., S. S. Perakis, and E. S. Bernhardt. 2013. Nitrate in watersheds: straight from

soils to streams? Journal of Geophysical Research:Biogeosciences 118:291-302.

USEPA (US Environmental Protection Agency). 2013. National rivers and streams assessment

2008-2009. EPA/841/D-13/001. Office of Water, US Environmental Protection Agency,

Washington, DC.

Wallace, J. B., S. L. Eggert, J. L. Meyer, and J. R. Webster. 1997. Multiple trophic levels of a

forest stream linked to terrestrial litter inputs. Science 277:102-104.

Wallace, J. B., S. L. Eggert, J. L. Meyer, and J. R. Webster. 1999. Effects of resource limitation

on a detrital-based ecosystem. Ecological Monographs 69:409-442.

Walther, D. A., and M. R. Whiles. 2011. Secondary production in a southern Illinois headwater

stream: relationships between organic matter standing stocks and macroinvertebrate

productivity. Journal of the North American Benthological Society 30:357-373.

Webster, J. R., and J. L. Meyer. 1997. Stream organic matter budgets: an introduction. Journal of

the North American Benthological Society 16:3-13.

176

Webster, J. R., J. L. Meyer, J. B. Wallace, and E. F. Benfield. 1997. Organic matter dynamics in

Hugh White Creek, Coweeta Hydrologic Laboratory, North Carolina, USA. Journal of

the North American Benthological Society 16:74-78.

Webster, J. R., J. L. Tank, J. B. Wallace, J. L. Meyer, S. Eggert, T. P. Ehrman, B. R. Ward, B. L.

Bennett, P. F. Wagner, and M. E. McTammany. 2000. Effects of litter exclusion and

wood removal on phosphorus and nitrogen retention in a forest stream. Verhandlungen

der Internationale Vereinigung für Theoretische und Angewandte Limnologie 27:1337-

1340.

Webster, J. R., J. D. Newbold, S. A. Thomas, H. M. Valett, and P. J. Mulholland. 2009. Nutrient

uptake and mineralization during leaf decay in streams - a model simulation.

International Review of Hydrobiology 94:372-390.

Wilcox, H. S., J. B. Wallace, J. L. Meyer, and J. P. Benstead. 2005. Effects of labile carbon

addition on a headwater stream food web. Limnology and Oceanography 50:1300-1312.

Wollheim, W. M., C. J. Vörösmarty, A. F. Bouwman, P. Green, J. Harrison, E. Linder, B. J.

Peterson, S. P. Seitzinger, and J. P. M. Syvitski. 2008. Global N removal by freshwater

aquatic systems using a spatially distributed, within-basin approach Global

Biogeochemical Cycles 22: GB2026:1-14.

177

Appendix A. Additional path model results, including weight of support for each of the models

tested, model performance when specific parameters were removed from path models, selected

single-year model path coefficients, and unstandardized path coefficients for overall models.

Table A1. Support for N, P or N+P models based on AIC for both overall models tested in this

study. Each model is specified based on the parameters being estimated that are different

between contrasted models (included are all 6 models tested in this analysis, which evaluated N

[as dissolved inorganic nitrogen; DIN], and P [as soluble reactive phosphorus; SRP], litter C:N,

and litter C:P as the predictor variables of interest). The number of parameters (K), AIC, the

change in AIC (ΔAIC), the weight of support (AIC Wt, Cum. Wt), log-likelihood (LL), χ2, and

P-values are reported for each model. The χ2 statistic and corresponding P-value indicate overall

agreement between modeled and observed covariance, giving the first line of evidence for

accepting or rejecting a given model structure. AIC is also indicative of model fit and is

weighted based on LL and model parsimony, allowing comparison of multiple acceptable

models (Burnham and Anderson 2002).

Model K χ2 d.f. P AIC ΔAIC AIC Wt Cum.Wt LL

C:N, N 13 9.3 5 0.10 1600.4 0.0 1 1 -787.2

C:N, P 13 5.1 5 0.4 1625.8 25.4 0 1 -799.9

C:P, P 14 0.7 5 0.95 1833.4 233.3 0 1 -902.7

C:N, N+P 14 11.4 8 0.18 1866.4 266.4 0 1 -919.2

C:P, N+P 15 12.7 7 0.08 2101.2 498.7 0 1 -1034.2

C:P, N 14 44.7 8 0 2130.7 528.2 0 1 -1050.2

178

Table A2. Results of removing specific parameters from the full path models to test for the importance of parameters in explaining

litter breakdown rates. Parameters were removed by fixing path coefficient estimates to zero. We then assessed modeled and observed

covariance structure for each reduced model (using χ2 tests), followed by ranking the importance of each parameter based on ΔAIC

(full – reduced model) and a χ2 difference test (χ2-diff test), where P-values < 0.05 denote significant reduction in model fit between

the reduced and full model. Also reported are the number of parameters in the model (K), the χ2-statistic and associated P-value for

model fit, the degrees of freedom (d.f.), the AIC score, the AIC Wt, cumulative AIC wt, and log-likelihood (LL) of each model.

Model K χ2 d.f. P AIC ΔAIC AIC Wt Cum.Wt LL χ2-diff test

C:N/N 13 9.28 4 0.1 1600 0 0.76 0.76 -787.22 n.a.

shredder biomass 12 13.98 5 0.03 1603 3 0.24 1 -789.57 0.03

discharge 10 35.54 8 <0.05 1621 21 0 1 -800.35 <0.05

stoichiometry 11 47.14 7 <0.05 1634 34 0 1 -806.15 <0.05

fungal biomass 10 285.32 8 <0.05 1871 271 0 1 -925.24 <0.05

C:P/P 14 0.71

0.95 1833 0 0.77 0.77 -902.7 n.a.

shredder biomass 13 5.63 5 0.34 1836 3 0.22 1 -905.16 0.03

fungal biomass 11 21.7 7 0.003 1848 15 0 1 -913.19 <0.05

discharge 12 20.97 6 0.002 1849 16 0 1 -912.83 <0.05

179

stoichiometry 12 50.57 6 <0.05 1879 46 0 1 -927.63 <0.05

180

Table A3. Comparison of path coefficients among PRE, YR1, and YR2 for the links among

fungal biomass, shredders, litter stoichiometry, discharge, and litter breakdown rates. Bold text

indicates significant path coefficient estimates. Also reported are predicted effects of fungi and

C:N/C:P, and C:N/C:P and shredders on litter breakdown rates in PRE, YR1 and YR2. In

general, fungi, shredders and discharge became more important for predicting litter breakdown

rates in YR1 and YR2 compared to PRE. The absolute magnitude of the effect of fungal biomass

on breakdown rates through C:N or C:P or C:N/C:P effects on litter breakdown mediated by

shredders tended to increase in YR1 and YR2 compared to PRE.

Model PRE YR1 YR2

C:N/N

fungi 0.05 0.13 0.11

shredders 0.09 0.16 0.15

C:N -0.69 -0.08 -0.59

Q 0.12 0.59 0.26

Compound paths

Fungi through C:N 0.09 0.04 0.47

C:N through shredders -0.03 -0.05 -0.08

C:P/P

fungi 0.09 0.12 0.49

shredders -0.03 0.21 0.29

C:P -0.75 -0.01 -0.09

Q 0.26 0.58 0.28

181

Compound paths

Fungi through C:P -0.03 0.00 0.04

C:P through shredders -0.01 -0.01 -0.07

182

Table A4. Unstandardized path coefficients (±SE) for the two best supported overall models.

Unstandardized path coefficients can be interpreted as the slope of the relationship between the

predictor and response variables (in this case, log-log slopes). All pathways are significant with

P < 0.05 except for the path between fungi and shredders in the C:P/P model.

C:N/N Path Unstandardized coefficient ±SE

N è fungi 0.4 0.08

fungi è C:N -0.25 0.03

fungi è litter breakdown 0.2 0.08

C:N è shredders -1.8 0.52

C:N è litter breakdown -1.0 0.19

shredders è litter breakdown 0.08 0.03

discharge è litter breakdown 0.04 0.01

discharge è shredders -0.12 0.03

C:P/P Path Unstandardized coefficient ±SE

P è fungi 0.39 0.05

P è C:P -0.23 0.04

fungi è C:P -0.21 0.06

fungi è litter breakdown 0.19 0.08

C:P è shredders -1.4 0.31

C:P è litter breakdown -0.71 0.12

shredders è litter breakdown 0.08 0.04

discharge è litter breakdown 0.04 0.012

discharge è shredders -0.08 0.03

183

Appendix B. Supplementary results: Shredder biomass in maple and rhododendron litter bags during PRE, YR1 and YR2 as a

function of litter C:N and C:P.

Figure B1. Shredder biomass (mg g AFDM-1) in maple (circles) and rhododendron (triangles) litter bags during PRE (open symbols),

YR1, and YR2 (gray and black closed symbols) as a function of litter C:N (a) or C:P (b). The vertical dotted line indicates the mean

reported TERC:N (a) and TERC:P (b) based on Tant et al. (2013) and Frost et al. (2006), respectively.

a. b.

20 40 60 80 100 120 140

050

100

150

200

250

300

C:N

Shredder biomass (mg g AFDM−1)

2000 4000 6000 8000

050

100

150

200

250

300

C:P

Shredder biomass (mg g AFDM−1)

184

Appendix C. Table C1. Mean (SE) middle-stage C:N and C:P of maple (M), poplar (P), oak (O), rhododendron (R) and wood (W)

during PRE, YR1, and YR2 (n = 20 for each year and detritus type), and initial C:N and C:P (n = 15 for each leaf litter type, wood n =

3). Also presented are the magnitude of the changes in C:N and C:P compared to both conditioned and initial stoichiometry (computed

as YRx/PRE for conditioned, and Mid C:N/Initial for initial) for each detritus type used in this study.

Detritus Conditioned C:N Δ conditioned C:N Initial C:N Δ initial (Mid C:N/Initial)

PRE YR1 YR2 YR1/PRE YR2/PRE

PRE YR1 YR2

M 51 (1) 44 (1) 43 (1) 0.86 0.84 81 (3) 0.63 0.54 0.53

P 53 (5) 31 (0) 33 (1) 0.58 0.62 66 (2) 0.80 0.47 0.50

O 57 (2) 48 (1) 44 (1) 0.84 0.77 82 (2) 0.70 0.59 0.54

R 100 (4) 64 (1) 63 (1) 0.64 0.63 146 (4) 0.68 0.44 0.43

Mean Δ: 0.73 0.72 Mean Δ: 0.70 0.51 0.50

W 149 (22) 50 (5) 40 (1) 0.34 0.27 167 (10) 0.89 0.30 0.24

Conditioned C:P Δ conditioned C:P Initial C:P Δ initial (Mid C:P/Initial)

PRE YR1 YR2 YR1/PRE YR2/PRE

PRE YR1 YR2

M 2312 (96) 1267 (92) 1416 (121) 0.55 0.61 3195 (273) 0.72 0.40 0.44

P 1804 (59) 1030 (40) 1077 (63) 0.57 0.60 2077 (131) 0.87 0.50 0.52

185

O 3255 (177) 1498 (116) 1573 (124) 0.46 0.48 4437 (260) 0.73 0.34 0.35

R 5337 (245) 1843 (144) 2013 (147) 0.35 0.38 7560 (281) 0.71 0.24 0.27

Mean Δ: 0.48 0.52 Mean Δ: 0.76 0.37 0.40

W 14975 (3475) 2957 (1036) 1077 (81) 0.20 0.07 10215 (1188) 1.47 0.29 0.11

186

Table C2. Linear model estimates (±SE) for differences in mean middle-stage C:N and C:P (data

were ln-transformed to meet assumptions of normality) by year and detritus type. We used

categorical predictors to test for differences in detrital stoichiometry between years (i.e., PRE vs.

YR1 and YR2) and among detritus types by year. Coefficients of this model can be interpreted as

the expected percent difference in the mean C:N or C:P for a given year and detritus type

comparison (e.g., e(3.93+0.88-0.15-0.81) = 0.38 or a 62% decrease in mean C:N for wood veneers in

YR1 compared to maple C:N in PRE). Bold text indicates significant parameter estimates (P <

0.05).

C:N Estimate ±SE

C:P Estimate ±SE

Maple 3.93 (0.05)

Maple 7.73 (0.10)

Poplar -0.03 (0.07)

Poplar -0.24 (0.14)

Oak 0.09 (0.08)

Oak 0.33 (0.14)

Rhododendron 0.64 (0.08)

Rhododendron 0.83 (0.14)

Wood 0.88 (0.08)

Wood 1.37 (0.14)

YR1*Maple -0.15 (0.07)

YR1*Maple -0.62 (0.14)

YR2*Maple -0.19 (0.07)

YR2*Maple -0.52 (0.14)

YR1*Poplar -0.31 (0.10)

YR1*Poplar 0.06 (0.19)

YR2*Poplar -0.23 (0.10)

YR2*Poplar -0.01 (0.19)

YR1*Oak 0.00 (0.11)

YR1*Oak -0.17 (0.19)

YR2*Oak -0.05 (0.10)

YR2*Oak -0.23 (0.19)

YR1*Rhododendron -0.27 (0.11)

YR1*Rhododendron -0.47 (0.20)

YR2*Rhododendron -0.26 (0.11)

YR2*Rhododendron -0.48 (0.20)

YR1*Wood -0.81 (0.11)

YR1*Wood -1.01 (0.20)

YR2*Wood -0.95 (0.11)

YR2*Wood -1.64 (0.20)

187

Table C3. Difference matrices for mean C:N and C:P at middle stages of decay for the five

detrital substrates used in this study (maple [M], poplar [P], oak [O], rhododendron [R], wood

[W]). All ratios are molar.

C:N

M P O R W

51 53 57 100 149

PRE M 51 0

P 53 2 0

O 57 6 4 0

R 100 49 47 43 0

W 149 98 96 92 49 0

44 31 48 64 50

YR1 M 44 0

P 31 -13 0

O 48 4 17 0

R 64 20 33 16 0

W 50 6 19 2 -14 0

43 33 44 63 40

YR2 M 43 0

P 33 -10 0

O 44 1 11 0

R 63 20 30 19 0

W 40 -3 7 -4 -23 0

188

C:P

M P O R W

PRE

2312 1804 3254 5336 14975

M 2312 0

P 1804 -508 0

O 3254 942 1450 0

R 5336 3024 3532 2082 0

W 14975 12663 13171 11721 9639 0

YR1

1267 1030 1498 1843 2958

M 1267 0

P 1030 -237 0

O 1498 231 468 0

R 1843 576 813 345 0

W 2958 1691 1928 1460 1115 0

YR2

1416 1077 1573 2012 1077

M 1416 0

P 1077 -339 0

O 1573 157 496 0

R 2012 596 935 439 0

W 1077 -339 0 -496 -935 0

189

Table C4. Mean ktotal and kmicrobe for all leaf litter types (maple [M], poplar [P], oak [O] and

rhododendron [R]) used in this study during PRE, YR1, and YR2. The magnitude of the

increases in ktotal and kmicrobe are also reported as YRx/PRE. We used analysis of variance

(ANOVA) to test for significant increases in breakdown rates. These significant increases in

breakdown rates by year and leaf litter type based on Tukey’s HSD post hoc tests are indicated

by bold text.

ktotal

Leaf litter type PRE YR1 YR2 YR1/PRE YR2/PRE

M 0.0112 0.0151 0.0204 1.35 1.82

P 0.0104 0.0210 0.0258 2.02 2.48

O 0.0051 0.0110 0.0133 2.17 2.62

R 0.0033 0.0160 0.0099 4.81 2.99

Mean YRx/PRE: 2.53

kmicrobe

Leaf litter type PRE YR1 YR2 YR1/PRE YR2/PRE

M 0.0051 0.0037 0.0075 0.73 1.47

P 0.0058 0.0044 0.0077 0.76 1.32

O 0.0022 0.0024 0.0038 1.09 1.72

R 0.0011 0.0023 0.0024 2.13 2.18

Mean YRx/PRE: 1.42

190

Appendix D. Stream nutrient treatments, measured nutrient concentrations and mean seasonal

temperatures for each study stream during the experiment. Also reported are the mean, minimum

and maxiumum temperatures for each of the streams during litterbag and wood veneer

deployment.

Table D1. Stream Temperatures (°C)

PRE Stream: 2 8 16 32 128

Summer 18.89 15.93 16.80 17.78 18.35

Autumn 11.78 11.20 11.68 10.97 11.15

Winter 2.73 4.64 3.00 3.14 2.21

Spring 10.02 9.68 9.55 10.03 9.19

YR1

Summer 18.20 15.41 16.36 17.15 17.61

Autumn 10.21 10.58 10.87 10.21 10.34

Winter 6.70 7.08 5.44 6.77 5.19

Spring 10.57 10.27 10.15 10.53 9.84

YR2

Summer 16.99 14.57 16.24 15.86 16.64

Autumn 11.92 11.38 11.78 11.76 11.47

Winter 6.97 7.66 6.79 7.37 6.24

Spring 8.96 9.19 8.57 9.36 8.16

Table D2. Leaf litter and wood deployment temperatures (°C)

191

Deployed leaf litter and wood (T°C)

N:P treatment Mean Min Max

2 5.17 0.84 11.53

8 6.12 2.70 10.92

16 4.92 1.00 10.90

32 4.95 0.30 11.19

128 4.45 0.11 10.85

Table D3. Mean (±SE) annual measured nutrient concentrations for the study streams.

Target N:P:

2 8 16 32 128

NO3 (µg/L) PRE 10.4 (1.8) 105 (17.3) 28.5 (6.4) 179.9 (14.9) 49.6 (8.2)

YR1 49.8 (4.5) 167 (9.5) 257 (25.8) 375.6 (30.1) 230.8 (20.9)

YR2 40.1 (3.7) 113.8 (7.3) 179.5 (23.0) 182.2 (9.9) 161.3 (16.4)

NH4 (µg/L) PRE 7.7 (1.0) 6.6 (1.1) 8.9 (1.6) 8.9 (1.5) 7 (0.9)

YR1 50.6 (6.3) 84.2 (7.7) 126.5 (13.8) 135.1 (14.1) 133.6 (11.6)

YR2 26.2 (4.2) 31.4 (4.7) 97.9 (12.7) 33.5 (3.7) 92.7 (10.7)

DIN (µg/L) PRE 18.1(1.5) 111.6 (17.3) 37.4 (5.7) 188.8 (14.4) 56.6 (7.8)

YR1 100.4 (9.1) 251.2 (14.3) 381.7 (35.4) 510.6 (37.3) 364.3 (29.2)

YR2 66.3 (6.2) 145.2 (10.4) 277.4 (32.9) 215.6 (11.5) 254 (25.3)

SRP (µg/L) PRE 2.9 (0.2) 2.5 (0.2) 3 (0.5) 3.1 (0.3) 2.5 (0.2)

YR1 42.4 (3.1) 78.4 (5.3) 40.6 (2.6) 30.4 (2.1) 8.2 (0.6)

YR2 53.3 (5.2) 31.6 (3.9) 30.4 (3.1) 14.2 (1.1) 6.2 (0.5)

N:P PRE 15.3 (1.8) 95.0 (16.3) 30 (4.5) 138.3(10.8) 48.9 (7.1)

YR1 8.4 (1.1) 18.6 (3.5) 25.5 (3.3) 54 (7.5) 159.6 (15.1)

192

YR2 5.6 (1.5) 37.4 (8.7) 24.2 (3.5) 52.9 (7.9) 113.1 (13.5)

193

Appendix E. Parameter estimates and 95% confidence intervals for these estimates for linear

models describing deployed litterbag and wood veneer respiration rates. Shown are intercepts for

all five deployed detrital substrates used in this study (maple, poplar, oak, rhododendron, and

wood). Slopes (all parameters including the term Tc) are interpreted as the apparent activation

energy of respiration for a given substrate and year (E in eV [1 eV = 1.6 × 10-19 J]). Confidence

intervals that did not include zero are emphasized with bold text (i.e., P < 0.05).

Parameter Estimate 95% CI

Intercepts

Maple * PRE -1.931 (-2.260 -1.602)

Poplar * PRE 0.126 (-0.354 0.607)

Oak * PRE -0.451 (-0.916 0.014)

Rhododendron * PRE -0.227 (-0.664 0.210)

Wood * PRE -1.020 (-1.429 -0.610)

Maple * YR1 0.083 (-0.749 0.914)

Poplar * YR1 -0.537 (-1.752 0.679)

Oak * YR1 -0.194 (-1.119 0.731)

Rhododendron *YR1 -0.153 (-1.064 0.757)

Wood * YR1 0.705 (-0.191 1.601)

Maple * YR2 -0.062 (-0.800 0.677)

Poplar * YR2 0.606 (-0.488 1.701)

Oak * YR2 0.954 (-0.098 2.006)

Rhododendron * YR2 -0.785 (-1.845 0.275)

Wood * YR2 1.302 (0.396 2.208)

194

Slopes

Tc * Maple * PRE -0.671 (-1.280 -0.063)

Tc * Poplar * PRE -0.161 (-1.022 0.699)

Tc * Oak * PRE 0.078 (-0.779 0.934)

Tc * Rhododendron* PRE -0.728 (-1.552 0.096)

Tc * Wood * PRE -1.605 (-2.362 -0.847)

Tc * Maple * YR1 0.128 (-1.426 1.682)

Tc * Poplar * YR1 1.041 (-1.204 3.287)

Tc * Oak * YR1 1.117 (-0.652 2.887)

Tc * Rhododendron * YR1 1.151 (-0.607 2.909)

Tc * Wood * YR1 -0.751 (-2.495 0.993)

Tc * Maple * YR2 0.691 (-0.871 2.253)

Tc * Poplar * YR2 -0.955 (-3.224 1.313)

Tc * Oak * YR2 -1.603 (-3.776 0.569)

Tc * Rhododendron * YR2 1.244 (-0.991 3.479)

Tc * Wood * YR2 -0.031 (-2.133 2.071)

195

Appendix F. Linear models describing temperature and nutrient enrichment effects on fungal

biomass-specific respiration rates (ln-transformed) for naturally occurring detritus (A; leaf litter

and wood only) and deployed litterbags and wood veneers (B). Intercepts are equivalent to the

mean respiration rate (ln fungal biomass-specific respiration) at 10°C for a given substrate and

year. Slopes (all parameters including the term Tc) are interpreted as the apparent activation

energy of fungal biomass-specific respiration (E in eV [1 eV = 1.6 × 10-19 J]). Confidence

intervals that did not include zero are emphasized with bold text (i.e., P < 0.05), and can be

interpreted as the mean difference in respiration rate, or E for a given year and substrate

comparison.

Table F1. Naturally occurring leaf litter and wood.

Parameter Estimate 95% CI

Intercepts

Litter -5.639 (-5.831 -5.447)

Wood -1.131 (-1.407 -0.855)

YR1 * Litter 0.276 (-0.018 0.570)

YR2 * Litter 0.035 (-0.251 0.321)

YR1 * Wood -0.164 (-0.580 0.251)

YR2 * Wood 0.203 (-0.207 0.613)

Slopes (E)

Tc * Litter -0.467 (-0.765 -0.170)

Tc * Wood 0.346 (-0.082 0.774)

Tc * Litter * YR1 -0.212 (-0.692 0.269)

196

Table F2. Deployed litterbags and wood veneers.

Parameter Estimate 95% CI

Intercepts

Litter -5.962 (-6.166 -5.759)

Wood 0.501 (0.119 0.883)

YR1 * Litter -0.222 (-0.513 0.068)

YR2 * Litter 0.451 (-0.030 0.932)

YR1 * Wood -0.005 (-0.531 0.520)

YR2 * Wood -1.009 (-1.830 -0.188)

Slopes (E)

Tc * Litter 0.370 (-0.002 0.743)

Tc * Wood 1.973 (1.266 2.680)

Tc * Litter * YR1 -0.017 (-0.602 0.568)

Tc * Litter * YR2 -2.049 (-3.049 -1.050)

Tc * Wood * YR1 -0.272 (-1.486 0.942)

Tc * Wood * YR2 -0.190 (-2.230 1.851)

Tc * Litter * YR2 0.309 (-0.299 0.916)

Tc * Wood * YR1 -0.202 (-0.888 0.484)

Tc * Wood * YR2 -0.268 (-1.126 0.590)

197

Appendix G. We analyzed the temperature dependence of respiration across different

temperature gradients for deployed substrates. We compared respiration across the temperature

gradient for all of the data, as well as for temperatures associated with respiration rates measured

at early and middle stages of decay (d 14, 70 [PRE]; in YR1, YR2 = d 14, 34 [M, P], d 14, 63

[O,R]) and the temperatures for the middle range of all the temperatures observed in this study.

Table G1. Parameter estimates, their standard errors, t-values and P-values for litterbag and

wood veneer respiration rates for all available data, early and middle respiration only, and only

data from the interquartile range (IQR). Slopes (all parameters including the term Tc) are

interpreted as the activation energy of respiration (eV). When considering all available data, we

found significantly lower temperature dependence for respiration rates in YR1 compared to PRE.

This trend did not continue in YR2; instead we found that respiration rates increased overall

(evidenced by higher intercept) but the temperature dependence of respiration remained the same

as for PRE.

All Data Estimate SE t-value P-value

PRE -2.26 0.08 -29.37 0.00

Tc -1.21 0.15 -8.26 0.00

YR1 0.06 0.11 0.55 0.58

YR2 0.35 0.17 1.99 0.05

Tc * YR1 0.95 0.22 4.27 0.00

Tc * YR2 0.56 0.38 1.49 0.14

Early and Mid Estimate SE t-value P-value

PRE -2.20 0.30 -7.30 0.00

198

Tc -1.33 0.47 -2.82 0.01

YR1 0.15 0.34 0.44 0.66

YR2 0.51 0.38 1.35 0.18

Tc * YR1 0.75 0.58 1.30 0.19

Tc * YR2 0.24 0.70 0.34 0.74

IQR Estimate SE t-value P-value

PRE -2.55 0.32 -7.95 0.00

Tc -0.64 0.51 -1.25 0.21

YR1 0.22 0.41 0.53 0.60

YR2 0.66 0.36 1.82 0.07

Tc * YR1 0.64 0.73 0.88 0.38

Tc * YR2 -0.07 0.63 -0.11 0.91

199

Fig. G1a-c. Arrhenius plots for respiration associated with leaf litter and wood veneers incubated

in our study streams for known periods of time. Temperatures were centered at 10°C

(approximate mean temperature for the duration of the study period in all five study streams).

We show all data (a), data from early and middle stages of detrital decay (~d 14, 70 [b]) and the

middle 50% of the data (c). Open circles correspond to PRE respiration rates, gray and black

circles denote YR1 and YR2 respectively.

-0.5 0.0 0.5

-6-4

-20

All Data

1 kT − 1 kT c

ln R

(T)

PREYR1YR2

A

-1.0 -0.5 0.0 0.5 1.0

-6-4

-20

Early and Middle Respiration

1 kT − 1 kT c

ln R

(T)

PREYR1YR2

B

-1.0 -0.5 0.0 0.5 1.0

-6-4

-20

IQR only

1 kT − 1 kT c

ln R

(T)

PREYR1YR2

C

200

Appendix H. Table G1. Target vs. measured concentrations (µg/L) of dissolved organic carbon (DOC), dissolved inorganic nitrogen

(DIN), and soluble reactive phosphorus (SRP) (n = 12 for mean measured values) in each treatment mesocosm (Control, N+P, labile

DOC, recalcitrant [recal.] DOC, labile DOC * N+P, and recal. DOC * N+P). Standard errors for the mean measured concentrations are

reported in Table 1. Target concentrations reflect expected ambient concentrations + added concentrations (µg/L). The differences

between target and measured concentrations are shown in the ΔDOC, ΔDIN, and ΔSRP columns of the table. When target

concentrations were ambient, concentrations from control mesocosms were used to compute the difference between target and

measured concentrations.

Treatment

Meas.

DOC

Target

DOC ΔDOC

Meas.

DIN

Target

DIN ΔDIN

Meas.

SRP

Target

SRP ΔSRP

Control 0.40 ambient N/A 36.05 ambient N/A 4.95 ambient N/A

N+P 0.26 ambient -0.14 62.85 95.00 -32.15 10.22 13.00 -2.78

Labile DOC 0.57 1.00 -0.43 71.94 ambient 35.89 9.48 ambient 4.53

Recal. DOC 0.55 1.00 -0.45 46.01 ambient 9.97 11.96 ambient 7.01

Labile DOC * N+P 0.54 1.00 -0.46 46.45 95.00 -48.55 6.64 13.00 -6.36

Recal. DOC * N+P 0.54 1.00 -0.46 84.53 95.00 -10.47 15.18 13.00 2.18