DISCERNING THE ROLE OF DENITRIFICATION, ANAMMOX,
AND N2O PRODUCTION IN AQUATIC SYSTEMS
By
Kateri Rose Salk-Gundersen

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Integrative Biology—Doctor of Philosophy
2017

ABSTRACT
DISCERNING THE ROLE OF DENITRIFICATION, ANAMMOX,
AND N2O PRODUCTION IN AQUATIC SYSTEMS
By
Kateri Rose Salk-Gundersen
This dissertation investigates (1) the magnitude and relative importance of denitrification
and anammox in aquatic nitrogen (N) loss, (2) microbial pathways of nitrous oxide (N2O)
production, (3) impact of ecosystem restoration on N loss and export, and (4) linkages of N loss
with harmful algal blooms (HABs). Study ecosystems included a mesotrophic lake, a wetland
undergoing restoration, an embayment of Lake Erie, and a coastal seagrass bed. A variety of
stable isotopic techniques were employed to achieve these research objectives.
Chapter one focuses on Muskegon Lake, a Great Lakes Area of Concern that is targeted
for water quality improvements. This system experiences seasonal hypolimnetic hypoxia that is
driven by water column stratification and respiration enhanced by high rates of epilimnetic
primary production. Hypoxic conditions were associated with production of N 2O, which was
produced predominantly by nitrification as evidenced by site preference, the difference in 15N
between the central and outer N atoms in N 2O. Upon seasonal water column mixing, atmospheric
N2O emissions from Muskegon Lake were among the highest reported in estuarine and lacustrine
environments, suggesting that lakes are an underappreciated source of N2O.
Chapter two evaluates the role of restoration activities on N loss in a floodplain wetland
upstream of Muskegon Lake. Current restoration efforts of this former agricultural field include
dredging of organic-rich sediment and hydrologic reconnection to adjacent Bear Creek. 15N
tracer incubations and modeling via the isotope pairing technique (IPT) indicate that restoration
activities will likely result in a twofold reduction in denitrification rate in wetland sediments,

which are driven largely by a decrease in coupled nitrification-denitrification. Despite a decrease
in the N loss capacity of the wetland sediments, hydrologic reconnection has the capacity to
decrease the downstream N load by as much as 10 %. However, a reduction in N availability
under restoration conditions resulted in a threefold increase in N 2O:N2 yield, potentially
increasing the N2O footprint of this Great Lakes Area of Concern following restoration.
Chapter three highlights the development and empirical testing of a new revision to the
IPT that accounts for the activity of dissimilatory NO 3- reduction to NH4+ (DNRA) as well as
denitrification and anammox. This work, conducted in eastern Australia, provides the first
complete accounting of NO3- reduction in a seagrass environment. Without employing the
revised IPT, denitrification in this system would have been overestimated by as much as 600 %.
DNRA outcompeted denitrification for NO3-, indicating that this high C:N system functions to
recycle and retain N rather than facilitate its loss. Furthermore, the suppression of denitrification
allowed anammox to dominate N loss, an unexpected result in comparison to other coastal
systems and suggests anammox may be important in other seagrass systems.
Chapter four evaluates the drivers of dramatic shifts in nutrient abundance and N cycling
in Sandusky Bay, Lake Erie, and the impact of these transitions on HAB formation. Isotope
tracer assays quantified rates of N loss and N fixation. Denitrification and anammox were found
to drive large swings in N:phosphorus ratios from > 10,000 to < 16, the threshold for N
limitation. N limitation stimulated N fixation, which indirectly supported the development of a
HAB in the late summer. Denitrification and nitrification contributed equally to N2O production,
and emissions of N2O were highest when N supply was greatest. Climate change is expected to
increase the frequency of episodic rain events, which will likely (1) enhance N 2O emissions and
nitrate export from Sandusky Bay and (2) stimulate offshore HABs in Lake Erie.

This dissertation is dedicated to Knute,
who has supported my work in countless ways.

iv

ACKNOWLEDGEMENTS

The completion of this dissertation was made possible by the efforts of many. In
particular, I wish to extend my gratitude to the following people:
My doctoral advisor, Nathaniel Ostrom, whose expertise and guidance allowed me to
progress immensely as a scientist. Your criticism, encouragement, and trust have given me
countless opportunities to develop my conceptual expertise, analytical skills, writing,
grantsmanship, and network of scientific peers. Thank you for helping me join the small but
mighty band of isotope biogeochemists.
My graduate committee, Peggy Ostrom, Al Steinman, Bopi Biddanda, and Steve
Hamilton. Each of you have brought unique expertise and insight into my graduate work. The
products of this dissertation (including myself) are richer as a result of your efforts. Thank you
for your edits, questions, and support.
Scientists who have welcomed me into their research groups, showing me the value of
strong collaboration. Thank you to Bopi Biddanda, Al Steinman, Dirk Erler, Brad Eyre, George
Bullerjahn, Mike McKay, Mark McCarthy, and Silvia Newell.
My various coauthors and helpers in the field and laboratory: Anthony Weinke, Leon
Gereaux, Adam McMillan, Alex Dutcher, James Smit, Maggie Oudsema, Mary Ogdahl, Natasha
Carlson-Perret, Jamie David, Iain Alexander, Matheus Carvalho, Taylor Tuttle, and Emily
Davenport. Without your assistance, this work would not have been pushed forward with such
force as it was. Most of all, thank you to Hasand Gandhi, whose knowledge of instrumentation
and patience in training an eager student know no limits.
My family and friends, who have enthusiastically supported my work but have also

v

shown me the value of a good laugh, a good adventure, and a good beer. Thank you to my
parents, who have encouraged my fascination for lakes and academics from the very beginning.
Thank you to my companions at Michigan State, who will no doubt become lifelong friends.
Lastly, thank you to my husband, Knute Gundersen, who continuously motivates me to do my
best work. Thank you for your unconditional support and celebration of my accomplishments.
This dissertation was supported by the NSF Graduate Research Fellowship [No.
DGE1424871], EPA Great Lakes Restoration Initiative [No. GL00E00460-1], NSF/AAS East
Asia and Pacific Summer Institutes [No. IIA-1414842], ARC Grants [DP160100248,
LP110200975, and LP150100519], Michigan State University (MSU) Water Research Initiation
Grant, MSU College of Natural Science Doctoral Recruiting Fellowship, MSU Department of
Integrative Biology John R. Shaver Research Fellowship, MSU Rose Graduate Fellowship in
Water Research, MSU College of Natural Science Marvin Hensley Endowed Fellowship, and an
Annis Water Resources Institute research grant. Any opinions, findings, and conclusions or
recommendations expressed in this material are those of the author and do not necessarily reflect
the views of the Environmental Protection Agency, the National Science Foundation, or the
Australian Research Council.

vi

TABLE OF CONTENTS

LIST OF TABLES ......................................................................................................................... ix
LIST OF FIGURES .........................................................................................................................x
CHAPTER 1: ECOSYSTEM METABOLISM AND GREENHOUSE GAS PRODUCTION IN
A MESOTROPHIC NORTHERN TEMPERATE LAKE EXPERIENCING SEASONAL
HYPOXIA........................................................................................................................................1
ABSTRACT .........................................................................................................................1
INTRODUCTION ...............................................................................................................2
METHODS ..........................................................................................................................5
Study site ..................................................................................................................5
Buoy observatory measurements .............................................................................6
Field sampling procedure ........................................................................................7
Ecosystem metabolism measurements .....................................................................8
Nutrient concentrations and trace gas analysis.....................................................10
Statistical analyses .................................................................................................12
RESULTS ..........................................................................................................................12
DISCUSSION ....................................................................................................................19
Hypoxia development and ecosystem metabolism .................................................19
Seasonal variation in Îľobs .......................................................................................21
Greenhouse gas concentrations and emissions .....................................................22
Determination of microbial origin of N2O .............................................................24
Conclusion .............................................................................................................28
APPENDIX ........................................................................................................................29
LITERATURE CITED ......................................................................................................33
CHAPTER 2: WETLAND RESTORATION AND HYDROLOGIC RECONNECTION
RESULT IN ENHANCED WATERSHED NITROGEN RETENTION AND REMOVAL .......42
ABSTRACT .......................................................................................................................42
INTRODUCTION .............................................................................................................42
METHODS ........................................................................................................................45
Study site ................................................................................................................45
Intact sediment cores .............................................................................................47
Analyses .................................................................................................................49
Modeling ................................................................................................................49
Statistical analysis .................................................................................................50
RESULTS ..........................................................................................................................51
DISCUSSION ....................................................................................................................53
LITERATURE CITED ......................................................................................................61

vii

CHAPTER 3: UNEXPECTEDLY HIGH DEGREE OF ANAMMOX AND DNRA IN
SEAGRASS SEDIMENTS: DESCRIPTION AND APPLICATION OF A REVISED ISOTOPE
PAIRING TECHNIQUE ...............................................................................................................67
ABSTRACT .......................................................................................................................67
INTRODUCTION .............................................................................................................68
METHODS ........................................................................................................................73
Study site ................................................................................................................73
Sediment slurry incubations ...................................................................................73
Intact sediment incubations ...................................................................................74
Analyses .................................................................................................................76
IPT modeling ..........................................................................................................76
Statistical modeling ................................................................................................80
RESULTS ..........................................................................................................................80
Application of R-IPT-DNRA ..................................................................................80
Rates of NO3- reduction processes .........................................................................83
DISCUSSION ....................................................................................................................85
Evaluation of R-IPT-DNRA when DNRA is active ................................................85
Anammox, denitrification, and DNRA in seagrass ecosystem sediments ..............87
Effect of organic carbon quality on N cycling .......................................................92
Conclusion .............................................................................................................93
APPENDIX ........................................................................................................................95
LITERATURE CITED ......................................................................................................99
CHAPTER 4: DRAMATIC SHIFTS IN N CYCLING DRIVE HAB FORMATION AND
MITIGATION OF N LOADING IN SANDUSKY BAY, LAKE ERIE ....................................106
ABSTRACT .....................................................................................................................106
INTRODUCTION ...........................................................................................................107
METHODS ......................................................................................................................109
Field sampling .....................................................................................................109
Isotope tracer assays............................................................................................112
Analyses ...............................................................................................................113
Modeling ..............................................................................................................116
RESULTS ........................................................................................................................116
DISCUSSION ..................................................................................................................123
Nutrient stoichiometry .........................................................................................123
N removal processes ............................................................................................124
N2O production and emissions.............................................................................126
N fixation ..............................................................................................................128
Harmful algal blooms ..........................................................................................129
Conclusions ..........................................................................................................130
LITERATURE CITED ....................................................................................................132

viii

LIST OF TABLES

Table 1. Îľobs values in Muskegon Lake measured by the in situ (2012) and bottle incubation
(2013) methods ..............................................................................................................................15
Table 2. Isotopic composition of N2O in the troposphere (Yoshida and Toyoda 2000),
epilimnion/mixed water column in Muskegon Lake, and hypolimnion of Muskegon Lake .........18
Table 3. DIN concentrations in the Bear Lake Wetland Restoration Area and adjacent Bear
Creek ..............................................................................................................................................47
Table 4. Posterior estimates of denitrification, N2O production, and NH4+ flux under restoration
scenarios. HR = hydrologic reconnection, HR+D = hydrologic reconnection and dredging, CI =
credible interval .............................................................................................................................53
Table 5. Parameters and equations used in the R-IPT (Risgaard-Petersen et al. 2003) and the RIPT-DNRA .....................................................................................................................................79
Table 6. Nitrogen loss rates calculated using the IPT for intact sediments and sediment slurries.
ND = not detected ..........................................................................................................................90

ix

LIST OF FIGURES

Figure 1. Muskegon Lake. Inset shows the location of Muskegon Lake (star) and its watershed
(light gray) within the state of Michigan, USA. The white circle in the middle of the lake
represents the location of the Muskegon Lake Observatory Buoy
(MLO;www.gvsu.edu/buoy/) ...........................................................................................................7
Figure 2. (a-b) Temperature and (c-d) dissolved O2 concentration from April to November in
2012 (left panel) and 2013 (right panel). Data were collected by the MLO sensors placed at
various depths in the water column. Hypoxia (dissolved O 2 <63 ÎźM) is indicated by a dotted
line..................................................................................................................................................13
Figure 3. Water column profiles at the MLO in (a-d) 2012 and (e-h) 2013. Temperature and
dissolved O2 concentration were measured continuously throughout the water column on each
sampling date. CH4 and N2O concentrations were measured at four depths on each sampling
date. Error bars represent standard error ........................................................................................14
Figure 4. Primary production (GPP), respiration (R), and P/R in Muskegon Lake in 2012 and
2013. P and R rates (open and closed symbols, respectively) were calculated according to
Bocaniov et al. (2012) on four occasions each year from May through September. P/R ratios
(bars) were calculated by dividing P by R. The horizontal line represents P/R=1 ........................15
Figure 5. (a) CH4 and N2O emissions and wind speeds for each sampling date in 2012 and 2013
and (b) CH4 and N2O emissions converted to CO2 equivalent for each date. Wind speed was
measured by the MLO. CH4 and N2O emissions were multiplied by their radiative forcing to
obtain CO2 equivalents (IPCC 2007). The vertical dashed line represents the division between
the two sampling years...................................................................................................................17
Figure 6. Linear regression of SP, δ18O, and δ15N of measured N2O with the inverse
concentration of N2O at each sampling point. In situ N2O concentrations and isotopic values are
plotted with the atmospheric N2O concentration and isotopic composition (rightmost value;
Yoshida and Toyoda 2000). The isotopic composition of microbially produced N 2O is
represented as the y-intercept value (Yamagishi et al. 2007) ........................................................19
Figure A1. Nutrient concentrations in 2013. (a) NH4+, (b) NO3-, and (c) SRP concentrations were
measured at four depths on each sampling date.............................................................................31
Figure A2. Principal component analysis for 2013 data. Points represent principal component
scores for each sampling depth on a given date. Vectors represent loadings for each variable on a
scale of zero to one and are plotted for illustrative purposes only. CH4 = CH 4 concentration,
N2O = N2O concentration, Temp = temperature, DO = dissolved O 2, pH = pH, Chlor =
chlorophyll a fluorescence, NO3 = NOx concentration, NH4 = NH4+ concentration, SRP =
soluble reactive phosphorus concentration ....................................................................................31

x

Figure A3. Relationship between ΔN2O and apparent oxygen utilization (AOU). AOU was
calculated as the difference between the saturation concentration and the in situ concentration of
of dissolved O2. ΔN2O was calculated as the difference between the in situ concentration and the
saturation concentration of N2O at each sampling point. A positive correlation (R2 = 0.58)
supports the interpretation of a N2O production source from nitrification ....................................32
Figure A4. Trends in N2O concentration and isotopic composition. (a) δ 18O-N2O and δ15N-N2O
are negatively correlated (R2 = 0.68). (b) δ18O-N2O and isotopic site preference (SP) are
positively correlated (R2 = 0.60) with a slope of 0.78 (solid line). N2O reduction produces a
characteristic δ18O-N2O vs. SP slope of 0.45 (dotted line) ............................................................32
Figure 7. Map of the Bear Lake Wetland Restoration Area (BLWRA). Restoration activities in
the West Pond include removal of the berm that separates Bear Creek from the wetland and
sediment dredging of. Insets show the location of the ponds in the downstream portion of the
watershed (left) and in Michigan, USA (right) ..............................................................................46
Figure 8. Posterior estimates of (a) denitrification, (b) coupled nitrification-denitrification, (c)
N2O production rates, and (d) NH4+ flux. Rates were measured from incubations of 6 and 12 h
duration under three restoration scenarios: (1) no restoration (control), (2) simulated hydrologic
reconnection (HR), and (3) simulated hydrologic reconnection and dredging (HR+D). Open
circles represent measured rates and closed circles represent the mean of the posterior estimate
for each treatment-duration combination. Lines represent the 95 % credible interval around each
posterior mean estimate. In (d), positive values indicate a net release of NH4+ to the water
column from the sediment, and negative values indicate a net uptake of NH 4+ by the sediment
from the water column. Note the different scales for the axes among panels. Statistical groupings
of posterior estimates at the p < 0.05 level are indicated by letters ...............................................52
Figure 9. (a) NO3- and (b) NH4+ mitigation capacity of the wetland with respect to the potential
load delivered by Bear Creek. Denitrification and NH4+ flux were measured from incubations of
6 and 12 h duration under restoration scenarios of simulated hydrologic reconnection (HR) and
simulated hydrologic reconnection with dredging (HR+D). Closed circles represent the mean of
the posterior estimate for each treatment-duration combination. Lines represent the 95 % credible
interval around each posterior mean estimate. Fractions of Bear Creek water reaching the
wetland sediment (0.1, 0.5, and 1.0) are represented by light gray, dark gray, and black lines,
respectively ....................................................................................................................................59
Figure 10. (a) The original IPT model (shaded; Nielsen 1992) that quantifies N2 production via
denitrification and a subsequent revision to the IPT (R-IPT; Risgaard-Petersen et al., 2003)
that includes the production of N2 via anammox as well as denitrification. (b) The
IPTanaN2O model (Hsu and Kao 2013), which adds the capacity to quantify N2O production via
denitrification to the R-IPT. (c) A revision to the IPT model (Song et al. 2016)
which enables quantification of the production of NH4+ via DNRA to the R-IPT. (d) The R-IPTDNRA model (this study), which adds the ability to quantify N2O production via denitrification
and NH4+ production via DNRA to the R-IPT while simultaneously distinguishing canonical
anammox and coupled DNRA-anammox ......................................................................................70

xi

Figure 11. Increase in the fraction of 15N in the NH4+ pool by treatment. F15NH4+ increased
linearly over time for control (df = 8, p < 0.01, R2 = 0.58) and seagrass (df = 8, p < 0.05, R2 =
0.41) treatments. Although there was a trend, there was insufficient replication to confirm a
linear relationship for the phytoplankton treatment (df = 1, p = 0.14, R 2 = 0.90) .........................81
Figure 12. Comparison of (a) anammox and (b) denitrification rates from the R-IPT (RisgaardPetersen et al. 2003) to rates from the R-IPT-DNRA method presented here for control,
phytoplankton, and seagrass treatments. The solid line represents a 1:1 relationship. Insets
represent the total range of under/overestimation by anammox and denitrification across
treatments, with the mean represented by the thick line ................................................................82
Figure 13. Overestimation of denitrification rate (D 14) by the R-IPT model compared to the RIPT-DNRA model as a function of the production of 30N2 by anammox relative to total 30N2
production (A30/P30) .......................................................................................................................82
Figure 14. Rates of (a) DNRA, (b) N2 production by anammox, (c) N2 production by
denitrification, and (d) N2O production by denitrification. Rates were determined from
incubations of three different durations and under three organic carbon treatments (control – no
addition, addition of phytoplankton detritus, and addition of seagrass detritus). Error bars
represent +1 standard deviation. DNRA rate data were not available at the 6 h time point with
phytoplankton addition. Note the difference in y-axis scales among panels .................................84
Figure 15. N2 production by anammox and denitrification, N2O production by denitrification,
and DNRA as a proportion of (a) total NO3- reduction and (b) total N loss to N2 and N2O.
Proportions of each process were measured at three different incubation durations under three
carbon treatments (control – no addition, addition of phytoplankton detritus, and addition of
seagrass detritus) ............................................................................................................................85
Figure A5. Isotope fraction of (a) 29N2 and (b) 30N2 in anoxic slurry incubations following
addition of 15NH4+, or 15NH4+ + 14NO3-, or 15NO3- over the course of 18 h. The control treatment
received no amendment. An increase in F29 indicates anammox activity, whereas an increase in
F30 indicates denitrification activity .............................................................................................96
Figure 16. Sampling locations in Sandusky Bay (circles) and Sandusky River monitoring station
(USGS site 04198000, cross) .......................................................................................................110
Figure 17. (a) NO3-, (b) NH4+, (c) PO43-, (d) N:P (NO3- + NH4+ : PO43-) ratio, and (e) chlorophyll
a concentration in six sites in Sandusky Bay in 2015. (f) Discharge of the Sandusky River at
Fremont obtained from USGS monitoring station 04198000. The dotted line in (d) indicates a
N:P ratio of 16 (note log scale on y axis).....................................................................................118
Figure 18. (a) Denitrification, (b) anammox, and (c) N2O production at Sandusky Bay station
1163 in 2015. Rates were measured by the isotope pairing technique. Error bars represent +1 SD.
Letters indicate groupings of statistically significant differences at p < 0.05 .............................120

xii

Figure 19. (a) N2O saturation and (b) diffusive emissions of N2O to or from the atmosphere at six
stations in Sandusky Bay. A saturation of 100% is represented by the horizontal dotted line in
(a). Data on May 28 and October 12 were available only for station 1163. Error bars represent +1
SD ................................................................................................................................................121
Figure 20. N2O isotopomers in Sandusky Bay. (a) 15N and 18O in black and gray, respectively,
(b) SP values, and (c) 15Nι and 15Nβ in gray and black, respectively ......................................122
Figure 21. N fixation rates at Sandusky Bay station 1163 in 2015. Error bars represent +1 SD.
Letters indicate groupings of statistically significant differences at p < 0.05 .............................122

xiii

CHAPTER 1
ECOSYSTEM METABOLISM AND GREENHOUSE GAS PRODUCTION IN A
MESOTROPHIC NORTHERN TEMPERATE LAKE EXPERIENCING SEASONAL
HYPOXIA
ABSTRACT
Many lacustrine systems, despite management efforts to control eutrophication, are
hypoxic during stratified periods. Hypoxia is a major concern, not only for its impact on aquatic
life but also for its potential to stimulate production of the greenhouse gases, methane (CH 4) and
nitrous oxide (N2O). We investigated the drivers of hypoxia in Muskegon Lake, a temperate
dimictic freshwater estuary that experiences frequent hypolimnetic mixing due to atmospheric
forces, riverine inputs, and intrusion of oxic water from coastal upwelling in Lake Michigan.
Primary production (GPP) and respiration (R) rates obtained from a 18O mass balance model
were similar to other mesotrophic environments (0.56-26.31 and 0.57-13.15 mmol O2 m-3 d-1,
respectively), although high P/R ( 2 in mid-summer) indicated there is sufficient autochthonous
production to support hypoxia development and persistence. The isotopic enrichment factor for
respiration (Îľobs) varied markedly and was least negative in August of both sampling years,
consistent with high R rates. Hypoxic conditions were associated with accumulation of N 2O but
not CH4, and emissions of N2O are among the highest reported from lakes. The average N2O site
preference (SP) value of 25.4‰ indicates that the majority of N2O was produced by nitrification
via hydroxylamine oxidation, despite the presence of resilient hypoxia. While it has been
hypothesized that denitrification acts as a sink for N 2O in hypoxic lakes, it is clear that
Muskegon Lake functions as a strong source of N 2O via nitrification. Further considerations of
lakes as global sources of N2O thus warrant a closer evaluation of nitrification-fueled N2O
production.

1

INTRODUCTION
Hypoxia is a major water quality concern for inland waters and is often accompanied by
nuisance algal blooms, declines in fisheries, and decreased recreational value (Diaz and
Rosenberg 2008). The development of hypoxia is primarily a product of several physical and
biological factors in lakes and coastal oceans: (1) a stratified water column, (2) basin shape, and
(3) respiration (R) in excess of O2 supply. Stratification occurs as surface waters become warmer
than bottom waters, and a thermocline isolates the cooler lower water column (hypolimnion)
from wind-driven mixing with the atmosphere. When stratification minimizes the influx of O 2 to
the hypolimnion, sinking primary production or sediment organic matter drives respiratory O 2
demand. This leads to a depletion of O2 to hypoxic levels (<63 ÎźM) (CENR 2000). Nutrient-rich
agricultural runoff stimulates primary production (GPP) and subsequent R in lakes and coastal
environments, which has led to an increase in the extent, intensity, and duration of hypoxia in
aquatic systems worldwide (Livingstone and Imboden 1996; Jankowski et al. 2006; Diaz and
Rosenberg 2008; Rabalais et al. 2010; Friedrich et al. 2014). Lake Erie and the Gulf of Mexico
are examples of two systems in which hypoxia is enhanced when the water column is stratified in
the shallow portions of their basins and GPP is high (Rabalais et al. 2010; Scavia et al. 2014).
Ecosystem metabolism, often characterized by the ratio of GPP to R (P/R), can thus be an
important indicator of the propensity of a system to develop hypoxia (Hanson et al. 2003;
Williamson et al. 2008). Systems with high P/R generally display a high degree of eutrophy (del
Giorgio and Peters 1994), and given requisite physical conditions, are prone to hypoxia.
Changes in land use and nonpoint source pollution have resulted in increases in the
incidence of hypoxia in freshwaters, raising concerns about the production of greenhouse gases,
particularly CH4 and N2O (Kaushal et al. 2014). Because both CH4 and N2O have anaerobic

2

production pathways, there is potential for these gases to accumulate in the hypoxic or anoxic
hypolimnia of stratified lakes and be emitted rapidly during storm-driven or seasonal periods of
mixing (Michmerhuizen et al. 1996; Riera et al. 1999; Bastviken et al. 2004). CH4 and N2O have
global warming potentials that are 25 and 298 times higher than CO 2 over a period of 100 years,
respectively (IPCC 2007). Thus, small increases in the atmospheric emissions of one or both of
these gases could have a dramatic impact on the greenhouse gas budget from lakes. Despite this
possibility, CH4 and N2O emissions from lakes are markedly understudied relative to CO 2
(Seitzinger et al. 2006; IPCC 2013). Measurements of the rates of atmospheric N2O emissions
from lakes are particularly sparse, with fewer than ten published studies (Lemon and Lemon
1981; Huttunen et al. 2003; Wang et al. 2006; Whitfield et al. 2011; Miettinen et al. 2015; Yang
et al. 2015). However, N2O emissions from lakes have the potential to rival those from
reservoirs, rivers, streams, and wetlands (Whitfield et al. 2011; Morse et al. 2012; Audet et al.
2014; Beaulieu et al. 2014), which merits a closer examination of the role of lakes in the global
N2O budget.
Determining the N2O budget from lakes depends on an understanding of the microbial
pathway by which it is produced. N2O is produced via three microbial processes that occur under
a range of redox conditions: denitrification, nitrifier-denitrification, and hydroxylamine oxidation
(Wrage et al. 2001). Denitrification is generally restricted to very low O2 levels, and as a
heterotrophic process, requires a carbon supply (Knowles 1982). The microbial denitrification
pathway involves a stepwise reduction of nitrate (NO 3-) to N2 gas, in which nitrite (NO2-), nitric
oxide (NO) and N2O are produced as intermediates. It is estimated that 7-16% of nitrogen
applied to the terrestrial biosphere is denitrified in lakes (Seitzinger et al. 2006), of which a small
yet variable portion (typically 0.1-6.0%) is not completely reduced to N2 and released as N2O

3

(Seitzinger and Kroeze 1998). Nitrifiers reduce NO2- by the same pathway as denitrifiers
(nitrifier-denitrification), producing N2O as an intermediate product (Poth and Focht 1985;
Wrage et al. 2001). In the third pathway, nitrifiers produce N2O as a product of the
decomposition of hydroxylamine (NH2OH; Wrage et al. 2001). This process is autotrophic and,
in contrast to denitrification, can occur in oxic environments. In pure cultures of nitrifiers, the
production of N2O is maximized as O2 declines, although production ceases under anoxic
conditions (Goreau et al. 1980; Frame and Casciotti 2010). Therefore, hypoxia provides the ideal
conditions for maximal N2O production, although the precise microbial production mechanism
involved is often uncertain.
Stable isotope analyses of N2O are used to distinguish N2O production mechanisms
(Ostrom and Ostrom 2011). The relative abundance of a stable isotope within a particular
material or reservoir is reported in standard delta notation:
𝛿=

𝑅𝑠𝑎𝑚 −𝑅𝑠𝑡𝑑
𝑅𝑠𝑡𝑑

×1000

(1)

where Rsam is the isotope ratio of the sample, R std is the isotope ratio of the standard, and  is
reported as per mil (‰). Because N2O is a linear asymmetric molecule, the isotopic composition
of the N atom in the central position (ι) and the N atom in the outer position (β) may be distinct.
The difference between 15Nι and 15Nβ in N2O is referred to as site preference (SP). In contrast
to bulk 15N and 18O values, SP is a conservative tracer of microbial N 2O production
mechanisms that is independent of the substrate isotopic composition (Sutka et al. 2003; Toyoda
et al. 2005; Sutka et al. 2006; Sutka et al. 2008). In general, denitrification and nitrifierdenitrification produce N2O with constant SP values of -10 to 0 ‰, whereas hydroxylamine
oxidation produces N2O with a SP value of 33 to 37 ‰ (Toyoda et al. 2005; Sutka et al. 2006;
Frame and Casciotti 2010; Ostrom and Ostrom, 2011). N2O derived from denitrification and

4

nitrifier-denitrification is indistinguishable on the basis of SP, likely because the enzymes
catalyzing these processes are identical (Stein 2011). Therefore, production of N2O from
denitrification and nitrifier-denitrification will henceforth be referred to collectively as
denitrification. Fractionation in SP has been demonstrated with a purified fungal nitric oxide
reductase enzyme but not within microbial culture which likely reflects maintenance of NO in
cells at low concentration and steady state (Yang et al. 2014). Therefore, SP is expected to hold
as a conservative, non-fractionating measure of N2O production mechanism under in situ
environmental conditions. SP analysis has been demonstrated as an effective approach for
distinguishing microbial N2O production from denitrification vs. hydroxylamine oxidation in a
variety of terrestrial and aquatic environments (e.g., Westley et al. 2006; Yamagishi et al. 2007;
Opdyke et al. 2009; Sasaki et al. 2011).
In this study, we examined the relationships among ecosystem metabolism, hypoxia, and
greenhouse gas production and atmospheric emissions in Muskegon Lake, MI. The relationship
between P/R, the oxygen isotopic enrichment factor for respiration (Îľobs), and the development of
hypoxia was also determined. In addition, we evaluated the influence of hypoxia and water
column mixing on the production and atmospheric fluxes of CH 4 and N2O. Further, we
determined the microbial pathway of N2O production in Muskegon Lake through SP analysis
and related biogeochemical indicators.

METHODS
Study site
Muskegon Lake is a 17 km2 drowned river-mouth lake in western Michigan with a mean
depth of 7 m and a maximum depth of 23 m (Freedman et al. 1979; Carter et al. 2006). The lake

5

drains the Muskegon River Watershed, a 7,032 km2 basin comprised mainly of forest and
agricultural land (Tang et al. 2005), and discharges directly into Lake Michigan by means of a
narrow shipping channel (Figure 1). The hydraulic residence time of Muskegon Lake is highly
dependent on the rate of water inflow from the Muskegon River, ranging from 14 to 70 days with
a mean of 23 days (Freedman et al. 1979; Carter et al. 2006). Historical nutrient loading led to
eutrophication and water quality degradation of Muskegon Lake, and as a consequence it was
designated a Great Lakes Area of Concern in 1985 (Steinman et al. 2008; U.S. EPA 2013). There
has been subsequent improvement in water quality, but Muskegon Lake remains characterized
by relatively high nutrient concentrations, periodic nuisance algal blooms, and summer hypoxia
in the hypolimnion (Steinman et al. 2008; Biddanda 2012).

Buoy observatory measurements
Time series data for water temperature (2, 4, 6, 7, 9, and 11 m), dissolved O 2, pH,
chlorophyll a fluorescence (2, 5, 8, and 11 m), and wind speed (1 m above water) were obtained
for the period of May to November in 2012 and 2013 from the Muskegon Lake Observatory
(MLO). The MLO buoy was deployed in the central part of the lake (Figure 1) and collected
water and meteorological data during the ice-free periods. The water depth at the MLO was
approximately 12 m and 11.25 m in 2012 and 2013, respectively, due to differences in lake water
level between the two years. Water sensors at the MLO collected data every 15 to 30 min, and
meteorological sensors collected data every 5 min. Additional information on the MLO can be
found in Biddanda (2012), McNair et al. (2013), and Vail et al. (2015).

6

43.27

lat 43.2324
lon -86.2945

Muskegon
River outlet

lat

43.25

43.23

43.21

Lake Michigan

-86.325

-86.300

-86.275

-86.250

Figure 1. Muskegon Lake. Inset shows the location
lon of Muskegon Lake (star) and its watershed
(light gray) within the state of Michigan, USA. The white circle in the middle of the lake
represents the location of the Muskegon Lake Observatory Buoy (MLO; www.gvsu.edu/buoy/).

Field sampling procedure
Water samples for nutrient, P/R, and greenhouse gas analysis were collected at the MLO
(Figure 1) between 9:00 am and 12:00 noon on an approximately monthly basis from May to
September in 2012 and 2013. In addition, intense temporal sampling was conducted three times
over a four-day period in August 2013. Water samples were taken by Niskin bottle (General

7

Oceanics, Inc., Miami, FL) at 2, 5, 8, and 11 m (2012) or 10.25 m (2013) depth. For nutrient
analysis, water was filtered (0.45 Îźm) and frozen upon return to the lab. Water for determination
of 18O-O2 and greenhouse gas analyses was transferred into 250 mL glass serum bottles and
sealed without headspace with butyl rubber septa. Biological activity was halted by adding 1 mL
of saturated HgCl2 solution to each bottle. Water column profiles of temperature, dissolved O 2,
pH, and chlorophyll a fluorescence were taken using a YSI 6600 sonde (Yellow Springs
Instruments, Inc., Yellow Springs, OH).

Ecosystem metabolism measurements
Respiration rates were determined from lake water collected at 2 m using a Niskin bottle.
Water from the Niskin bottle was placed into a 20 L carboy and subsequently dispensed into 300
mL acid-washed BOD bottles that were sealed with glass stoppers without headspace. The BOD
bottles were incubated in the dark at in situ temperature. Dissolved O2 concentrations were then
measured in triplicate each day for four days from sacrificed bottles. O 2 concentration was
measured via Winkler autotitration using a Radiometer Analytical Titralab 650 with platinum
combined Ag/AgCl reference electrode (Weinke et al. 2014).
The 18O of dissolved O2 was determined using water from 2, 5, 8, and 11 m into which
HgCl2 was added at the time of collection (2012) or from water at 2 m that was preserved with
HgCl2 at the same time points as the BOD bottles (2013). Water was transferred from BOD
bottles into pre-evacuated 200 mL glass vessels fitted with high vacuum stopcocks according to
Emerson et al. (1991) and Roberts et al. (2000). The vessels were stored at room temperature for
at least 4 h to allow the water and headspace to come to equilibrium. The headspace was then
introduced to an evacuated 3 mL sampling loop and then onto a 5 m packed molecular sieve (5

8

Å) column (Alltech, Inc., Deerfield, IL) using He carrier gas within a gas chromatograph (HP5980, Hewlett Packard, Ramsey, MN) interfaced to an Isoprime isotope ratio mass spectrometer
(Elementar Americas, Inc., Mount Laurel, NJ) for determination of the 18O of O2. Analytical
reproducibility of standards was 0.3 ‰.
The oxygen isotopic enrichment factor for respiration, Îľobs, was determined using a
Rayleigh model from two sets of data, (1) in situ observations of 18O-O2 by depth in 2012 and
(2) four-day bottle incubations from which 18O-O2 was measured in 2013:
𝛿 −𝛿

𝑠𝑜
𝑠
𝜀𝑜𝑏𝑠 = 𝑙𝑛(𝑂
𝑆𝑎𝑡)

(2)

2

Where and s is the in situ 18O-O2 value or the 18O-O2 value at a given time point in a bottle
incubation, so is the initial 18O-O2 value, and O2Sat is the fractional saturation of O2 (Mariotti
et al. 1981; Ostrom et al. 2014). When ln(O2Sat) is regressed against s, the slope is equivalent to
Îľobs.
Rates of GPP, R, and P/R ratios in the epilimnion were estimated using a steady-state,
mass balance model (hereafter the 18O model; Bocaniov et al. 2012):
𝑂2 (𝛼𝑔 ∙ 18:16𝑂 − 𝛼𝑟 ∙ 18:16𝑂 ) − 𝑂2𝑠 (𝛼𝑔 ∙ 𝛼𝑠 ∙ 18:16𝑂𝑎 − 𝛼𝑟 ∙ 18:16𝑂)
𝐹
𝐺𝑃𝑃 = ( ) ∙
𝑍𝑚
𝛼𝑝 ∙ 18:16𝑂𝑤 − 𝛼𝑟 ∙ 18:16𝑂
(3)
𝑅=(

𝑂2 (𝛼𝑔 ∙
𝐹
)∙
𝑍𝑚

18:16

𝑂 − 𝛼𝑝 ∙

18:16

18:16

𝑂𝑤 ) − 𝑂2𝑠 (𝛼𝑔 ∙ 𝛼𝑠 ∙
𝛼𝑝 ∙ 18:16𝑂𝑤 − 𝛼𝑟 ∙ 18:16𝑂

𝑂𝑎 − 𝛼𝑝 ∙

18:16

𝑂𝑤 )
(4)

where Zm is the depth of the mixed layer, O2 is the measured concentration of dissolved O2, O2s
is the concentration of O2 at atmospheric saturation, 18:16O is the measured oxygen isotope ratio
of dissolved O2, 18:16Ow is the measured oxygen isotope ratio of H2O, and 18:16Oa is the isotopic
ratio of atmospheric oxygen (23.5 ‰). α g, αs, αp, and αr are the fractionation factors associated
with gas transfer (0.9972), gas solubility in water (Benson and Krause 1984), photosynthetic

9

reaction rates of 18O-H2O to 16O-H2O (1.000), and respiration (1 + ξobs/1000), respectively. 18OH2O was determined by off-axis integrated cavity output spectroscopy using a Los Gatos
Research Liquid Water Isotope Analyzer (Lis et al. 2008). Analytical reproducibility of standards
was 0.2 ‰. When direct measurements of 18O-H2O were not available, mean values from the
remainder of the sampling period were used.
The oxygen gas transfer rate (F) was calculated according to Wanninkhof (1992):
𝐹 = 𝑘𝑤 (𝐶𝑤 − 𝐶𝑎 )

(5)

where Cw is the dissolved O2 concentration at the surface (measured at 2 m depth) and Ca is the
calculated dissolved O2 concentration in equilibrium with the atmosphere (Wanninkhof 1992;
Walker et al. 2010). The gas transfer coefficient (kw, in m s-1) is calculated as:
𝑆𝑐

2
𝑘𝑤 = 0.31𝑈10
(600)

−1/2

(6)

where Sc is the Schmidt number for O2 determined by the kinematic viscosity of freshwater
divided by the diffusion coefficient of O2 (Wanninkhof 1992) and U10 is the wind speed 10 m
above the surface determined using the measured wind speed 1 m above the surface and the
power law relationship outlined in Walker et al. (2010). Wind speed was measured by the MLO
and averaged over the dissolved O2 residence time in the mixed layer (2-5 d). Because buoy data
were not available on 6 May 2013, wind speed data were obtained from instruments at the
Muskegon County Airport (8 km from MLO) and adjusted by the roughness lengths of flat land
and water (World Meteorological Organization 2008).

Nutrient concentrations and trace gas analysis
Samples collected in 2013 were analyzed for the concentration of nitrate + nitrite (NO 3-),
ammonium (NH4+), and soluble reactive phosphorus (SRP). All nutrient analyses were

10

performed according to Standard Methods (APHA 2005). SRP was analyzed
spectrophotometrically using the ascorbic acid method, NO3- was determined colorimetrically
after cadmium reduction, and NH4+ was analyzed using the phenate method. NO3- and
NH4+ were analyzed using a wet chemistry continuous flow analyzer (Skalar Analytical B.V.,
Breda, the Netherlands).
Water samples for CH4 and N2O concentration analysis were introduced by syringe
injection at atmospheric pressure into glass serum bottles that had been flushed with He prior to
sample introduction. The water sample was equilibrated with the remaining headspace overnight
by gentle shaking. The headspace was then analyzed by GC-ECD-FID (Shimadzu Greenhouse
Gas Analyzer GC-2014, Shimadzu Scientific Instruments, Columbia, MD) for N 2O and CH4
concentration. The dissolved concentration was calculated based on the headspace equilibrium
concentration (Hamilton and Ostrom 2007). Diffusive atmospheric emissions of CH4 and N2O
(F) were calculated by equations 5 and 6, substituting the measured concentration, saturation
concentration, and Schmidt number of CH4 and N2O for those of dissolved O2.
The isotopic composition of N2O was analyzed upon introduction of sample water into an
enclosed 0.75 L glass vessel that was previously purged of atmospheric air using a gentle flow of
He. Dissolved gases were subsequently stripped from the water by sparging the sample with He
(Sansone et al. 1997), which carried sample gases into a Trace Gas sample introduction system
interfaced to an Isoprime isotope ratio mass spectrometer (Elementar Americas, Inc., Mount
Laurel, NJ). Analytical reproducibility for replicate samples was 0.5 ‰ for bulk 15N and 18O,
0.75 ‰ for 15Nα 15Nβ, and 1.3 ‰ for SP.
The microbial origin of N2O produced in Muskegon Lake was evaluated by several
approaches. First, a Keeling plot was employed to calculate the isotopic composition of

11

microbially-produced N2O (end-member) by regressing the isotopic ratios of N2O vs. the inverse
concentration of N2O in measured samples and the atmospheric end-member (Pataki et al. 2003;
Yamagishi et al. 2007). Second, apparent oxygen utilization (AOU; O2saturation – O2measured) and
ΔN2O (N2Omeasured – N2Osaturation) were compared to provide evidence for nitrification-fueled N2O
production (Yoshinari 1976; Nevison et al. 2003; Bange et al. 2010). Third, Δ18O (18O-N2O 18O-O2) was calculated to evaluate the relative contributions of hydroxylamine oxidation and
denitrification (Ostrom et al. 2000).

Statistical analyses
All statistical analyses were conducted using R statistical software (version 3.0.2).
Principal component analysis was performed on 2013 data to analyze correlation among
temperature, pH, greenhouse gas concentrations, dissolved O 2, nutrient concentrations, and
chlorophyll a fluorescence (Appendix). Prior to principal component analysis, each variable was
standardized to equalize variance across variables. Differences in N 2O isotopic composition
between the epilimnion and hypolimnion were evaluated by Welch 2-sample t-tests following
tests for distribution normality and equal variance. The relationship between AOU and ΔN2O
was evaluated by linear regression.

RESULTS
Muskegon Lake displays a dimictic stratification pattern, with periods in the spring and
fall of homogenous temperature and O2 concentration throughout the water column (Figure 2).
The summer period is characterized by higher surface water temperatures than in the spring and
fall and low concentrations of O2 in the hypolimnion. When the water column is stratified, the

12

thermocline is located between 6 and 8 m. A weak to strong thermocline was present on all
sampling dates except 6 June 2012, 20 September 2012, and 11 June 2013 (Figures 3a, 3e). June
sampling dates coincided with wind-driven episodic mixing events, and 20 September 2012
coincided with the fall mixing period. O2 concentrations in the hypolimnion during stratified
periods regularly decreased to hypoxic levels (Figures 2, 3b, 3f).

25
Temperature (°C)

30

2m
4m
6m
9m
11m

a

20

15

2m
4m
6m
9m
11m

20

15

10

10
A

M

J

J

A
Date

S

O

M

N
2m
5m
8m
11m

c
400

300

200

100

J

J

A
Date

S

O

N

d

2m
5m
8m
11m

400

Dissolved O2 (ÎźM)

Dissolved O2 (ÎźM)

b

25
Temperature (°C)

30

300

200

100

0

0
A

M

J

J

A
Date

S

O

M

N

J

J

A
Date

S

O

N

Figure 2. (a-b) Temperature and (c-d) dissolved O2 concentration from April to November in
2012 (left panel) and 2013 (right panel). Data were collected by the MLO sensors placed at
various depths in the water column. Hypoxia (dissolved O2 <63 ÎźM) is indicated by a dotted line.

13

Temperature (°C)
15
20
25

10

Depth (m)

0

0

a

0.0
0

b

CH4 concentration (ÎźM)
0.5
1.0
1.5
2.0

0
0

c

2

2

2

2

4

4

4

4

6

6

6

8

6-Jun
10-Jul
28-Aug
20-Sept

10

10
0

Depth (m)

O2 concentration (ÎźM)
100
200
300

0

30

15

20

25

8

8

10

10

0

30
0

e

100

200

300

f

0.5

1.0

1.5

8
10

0

2.0
0

g

2

2

2

2

4

4

4

4

6

6

8

8

10

10

6
8

10

6-May
11-Jun
12-Aug
14-Aug
15-Aug
16-Sept

d

6
6 June
10 July
28 Aug
20 Sept

0.0
0

N2O concentration (nM)
10 20 30 40 50

6 May
11 June
12 Aug
14 Aug
15 Aug
16 Sept

10

20

30

40

50

h

6
8
10

Figure 3. Water column profiles at the MLO in (a-d) 2012 and (e-h) 2013. Temperature and
dissolved O2 concentration were measured continuously throughout the water column on each
sampling date. CH4 and N2O concentrations were measured at four depths on each sampling
date. Error bars represent standard error.
Epilimnetic GPP and R rates obtained by the 18O model ranged from 0.56 - 26.31 mmol
O2 m-3 d-1 and 0.57 - 13.15 mmol O2 m-3 d-1, respectively (Figure 4). The lowest GPP and R rates
occurred during September 2012 and May 2013. The highest GPP rate occurred in August of
both years, and the highest R rate occurred in August 2012 and September 2013. P/R ranged
from 0.79 - 2.36, with the highest ratios occurring in August in each year and the lowest ratios
occurring during May, June, and September (Figure 4).
The isotopic enrichment factor for respiration, εobs, varied by over 10 ‰ within each
sampling season (Table 1). Îľobs was most negative in June 2012, May 2013, and September 2013.
Îľobs was least negative in August of both years. The range in Îľ obs values was less negative in 2012
when the in situ method was used than in 2013 when the bottle incubation was used.
14

30

2.5

P/R
GPP
2013
P 2013

2.0

R 2012
20

R 2013
1.5

P/R=1

P/R

GPP or R rate (mmol O2 m-3 d-1)

GPP
2012
P 2012
25

15
1.0
10
0.5

5

0

0.0
M

J

J

A

S

Date

Figure 4. Primary production (GPP), respiration (R), and P/R in Muskegon Lake in 2012 and
2013. P and R rates (open and closed symbols, respectively) were calculated according to
Bocaniov et al. (2012) on four occasions each year from May through September. P/R ratios
(bars) were calculated by dividing P by R. The horizontal line represents P/R=1.
Table 1. Îľobs values in Muskegon Lake measured by the in situ (2012) and bottle incubation
(2013) methods.

15

NH4+, NO3-, and SRP concentrations ranged from 2.3 - 44.5 Îźg L-1, 163.3 - 373.7 Îźg L-1,
and 0.8 - 17.5 Îźg L-1, respectively (Appendix Figure A1). NH4+, NO3-, and SRP concentrations
were negatively correlated with O2 concentration and chlorophyll a fluorescence (Appendix
Figure A2), indicating an accumulation of nutrients in the hypolimnion.
CH4 and N2O were supersaturated in the water column at all times (1.1 - 4.8 and 3,700 30,000 times atmospheric equilibration concentration, respectively). CH4 was fairly
homogeneous throughout the water column on each sampling date with the exception of a single
extremely high concentration at the lowest sampling depth in September 2012 (Figure 3c).
Excluding this point, the highest concentrations of CH 4 throughout the water column occurred in
May 2013 (Figure 3g). N2O concentrations were also high throughout the water column in May
2013 (Figure 3h). Hypolimnetic N2O concentration were negatively associated with O2
concentration, displaying accumulation in the hypolimnion during hypoxic periods (Figure 3).
CH4 and N2O atmospheric emissions by diffusive evasion were low yet variable
throughout the majority of sampling dates in both years with the exception of two dates during
which emissions were 10 - 100 times higher than other sampling dates (Figure 5a). The first
period of high emissions occurred in September 2012 during a period of strong wind, and the
second occurred in May 2013 when concentrations of both gases were high throughout the water
column. When converted to CO2 equivalents by multiplying emissions by 298 and 25 for N2O
and CH4, respectively (IPCC 2007), the radiative forcing of CH4 emission was 2.1 - 22.9 times
that of N2O (Figure 5b).

16

140

a

Methane
120

CH4 emission (mmol m-2 d-1)
or wind speed (m s -1)

10

Nitrous oxide
Wind speed

8

100
80

6
60
4

40

2

20

0

0

N2O emission (Îźmol m-2 d-1)

12

b

Emission in CO2 equivalent (mmol m-2 d-1)

6000
5000
4000
3000
2000
1000
0

Figure 5. (a) CH4 and N2O emissions and wind speeds for each sampling date in 2012 and 2013
and (b) CH4 and N2O emissions converted to CO2 equivalent for each date. Wind speed was
measured by the MLO. CH4 and N2O emissions were multiplied by their radiative forcing to
obtain CO2 equivalents (IPCC 2007). The vertical dashed line represents the division between
the two sampling years.

The isotopic composition of N2O (mean Âą SD) in the epilimnion and mixed water
column (15N: 4.5 ± 0.7 ‰, 18O: 50.1 ± 1.9 ‰, SP: 20.0 ± 2.4 ‰) and hypolimnion (15N: 0.0 ±
0.3 ‰, 18O: 54.3 ± 0.9 ‰, SP: 23.1 ± 1.8 ‰) were distinct (Table 2). Relative to the epilimnion
and mixed water column, Welch 2 sample t-tests indicated that hypolimnetic 15N-N2O values

17

were significantly lower (t = 15.02, df = 9.9, p < 0.0001), 18O-N2O values were significantly
higher (t = -5.5, df = 10.4, p < 0.0001), and SP values were significantly higher (t = -2.59, df =
10.3, p < 0.01). The 15N, 18O, and SP of microbially-produced N2O in Muskegon Lake,
determined by the y-intercepts of the Keeling plot, were -3.2 ‰, 58.8 ‰, and 25.4 ‰,
respectively (Figure 6). AOU was positively correlated with ΔN2O (Appendix Figure A3; linear
regression, R2 = 0.58, t = 6.97, df = 35, p < 0.0001). The mean Δ18O value for N2O was 22.1 
6.0 ‰.

Table 2. Isotopic composition of N2O in the troposphere (Yoshida and Toyoda 2000),
epilimnion/mixed water column in Muskegon Lake, and hypolimnion of Muskegon Lake.
Date

Depth
m

O2
uM

N2O
nM

Tropospheric N2 O

δ18O-N2O
‰

δ15N-N2O
‰

δ15Nι-N2 O δ15 Nβ-N2O
‰
‰

43.7 Âą 0.9 7.0 Âą 0.6

N2 O SP
‰
18.7 Âą 2.2

Epilimnion/Mixed water column
11-Jun-13
2
281
11-Jun-13
5
259
11-Jun-13
10.25
188
12-Aug-13
2
317
12-Aug-13
5
297
14-Aug-13
2
264
14-Aug-13
5
262
16-Sep-13
8
209
Mean
260
SD
43

10.96
11.47
13.05
10.61
13.09
13.11
10.47
11.42
11.77
1.14

47.9
48.6
51.3
48.4
50.3
53.6
49.7
51.0
50.1
1.9

5.4
4.8
3.5
5.1
4.5
5.2
4.1
3.4
4.5
0.7

14.4
12.7
13.8
14.8
16.4
16.2
14.1
13.5
14.5
1.3

-3.7
-3.1
-6.7
-4.6
-7.4
-5.9
-5.9
-6.7
-5.5
1.5

18.1
15.8
20.6
19.4
23.8
22.0
20.0
20.2
20.0
2.4

Hypolimnion
12-Aug-13
14-Aug-13
15-Aug-13
15-Aug-13
16-Sep-13
Mean
SD

22.10
29.80
20.09
29.91
27.66
25.91
4.54

54.9
54.7
55.2
53.4
53.4
54.3
0.9

-0.1
-0.4
0.3
-0.3
0.2
0.0
0.3

11.0
11.2
12.1
11.5
10.2
11.2
0.7

-12.6
-11.9
-11.5
-12.1
-9.8
-11.6
1.0

25.0
23.1
23.6
23.6
20.0
23.1
1.8

8
10.25
8
10.25
10.25

146
42
99
57
53
79
43

18

y = -106.1x + 58.8
R² = 0.80
55

SP
δ18O
δ 18O

Isotope Ratio (‰)

45
35

15N
δ
δ15N

y = -61.6x + 25.4
R² = 0.43

25
15
5
-5
0.00

y = 86.7x - 3.2
R² = 0.90

0.03

0.06

0.09

0.12

1/N2O (1/nM)

Figure 6. Linear regression of SP, δ18O, and δ15N of measured N2O with the inverse
concentration of N2O at each sampling point. In situ N2O concentrations and isotopic values are
plotted with the atmospheric N2O concentration and isotopic composition (rightmost value;
Yoshida and Toyoda 2000). The isotopic composition of microbially produced N 2O is
represented as the y-intercept value (Yamagishi et al. 2007).

DISCUSSION
Hypoxia development and ecosystem metabolism
Hypoxia is a consequence of extensive nutrient loading, requires the development of
stratification, and is predominant in lakes with shallow hypolimnia. In Muskegon Lake, hypoxia
was first documented during a period of extensive nutrient loading the late 1970s (Freedman
1979), but hypoxia may be a natural phenomenon in this system. The water column is stratified
from June through September or October, but episodic intrusion of oxygenated Lake Michigan
water during westerly seiches is demonstrated by increases in hypolimnetic O 2 concentrations by
as much as 150 ÎźM on daily to weekly timescales (Figure 2; B. Biddanda, A. Weinke, S.
Kendall, and D. Koopmans, personal communication). The observed decline of hypolimnetic
temperature from the onset of stratification to early August, an unusual trend in stratified

19

systems, provides additional evidence of upwelling-induced intrusion of cold Lake Michigan
water into Muskegon Lake. Following such hypolimnetic oxygenation events, O2 concentrations
decline rapidly (often within 12 h) to hypoxic conditions (Figure 2). Thus, hypoxia exhibits
resilience in Muskegon Lake despite a short hydraulic residence time and evidence of the
episodic intrusion of cold, oxic water. The persistence of hypoxia in Muskegon Lake points to a
strong driving factor for O2 consumption in the hypolimnion.
The balance of GPP and R and its seasonal variation provides insight into the
development of hypoxia. Although rates of GPP and R are lower than many hypoxic systems
(Ostrom et al. 2005; Bocaniov et al. 2012), observations of P/R in Muskegon Lake suggest there
is sufficient excess primary production to fuel hypolimnetic respiration and hypoxia
development. P/R in the epilimnion was  1 throughout the majority of the sampling period
(Figure 4), demonstrating net autotrophy during the growing season between May and
September. Two earlier studies of pelagic metabolism in Muskegon Lake similarly demonstrated
a consistent trend of P/R > 1 during this time period as well (Weinke et al. 2014; Dila and
Biddanda, 2015). In particular, GPP was two-fold higher than R in August of both sampling
years (2.00 and 2.36 in 2012 and 2013, respectively). A portion of this surplus production from
the epilimnion likely sinks into the hypolimnion and fuels water column or sediment respiration
in subsequent days to years. Thus, despite evidence of episodic intrusion of oxic water into the
hypolimnion, high P/R values indicate a supply of organic matter that reaches the hypolimnion to
fuel strong O2 consumption and persistent hypoxia.
While water column respiration has been demonstrated as the primary driver of hypoxia
in systems such as Lake Erie and the Gulf of Mexico (Conroy et al. 2011; McCarthy et al. 2013;
Ostrom et al. 2014), respiration in organic matter-rich sediments can enable hypoxia to establish

20

at more moderate nutrient loading than would be expected (Turner et al. 2008). A combination of
water column and sediment respiration thus likely fuels hypoxia establishment and resilience in
Muskegon Lake. While water quality indicators have neared or exceeded remedial action plan
targets for reducing eutrophication (Michigan DEQ 2011), summer hypoxic conditions have
persisted and will likely persist for many years. Two potential explanations are that (1)
remediation of eutrophication in Muskegon Lake is insufficient to reduce hypoxia or (2) hypoxia
is a natural feature of this system.

Seasonal variation in Îľobs
The magnitude of isotopic fractionation of O2 during respiration, Îľobs, is generally
assumed to be a constant, but substantial variation has been observed in this study and other
estuarine and coastal marine ecosystems (QuiĂąones-Rivera et al. 2007; Lehman et al. 2009; Fry
and Boyd 2010; Ostrom et al. 2014). Isotopic fractionation during respiration takes place when
O2 is consumed by oxidative respiratory enzymes such as cytochrome oxidase (Feldman et al.
1959). If diffusion limits the supply of O2 to the respiratory enzyme, then the expression of
isotopic fractionation by the enzyme is reduced (Ostrom et al. 2014). For instance, isotopic
discrimination approaching 0 ‰ has been observed for respiration in sediments where diffusion
is expected to be slow, limiting the movement of O 2 into the cell (Brandes and Devol 1997). In
contrast, isotopic discrimination in the water column is generally quite large and assumed to be
constant (e.g., -21.2 ‰), because diffusion is not expected to limit the supply of O2 to the
respiratory enzyme (QuiĂąones-Rivera et al. 2007; Fry and Boyd 2010). In Muskegon Lake, Îľobs
was evaluated by two approaches: bottle incubation and the in situ method. Îľobs values obtained
by bottle incubation reflect only water column respiration, whereas Îľ obs values obtained by the in

21

situ method reflect both water column and sediment respiration. Indeed, the Îľ obs values obtained
from the in situ method were less negative than those obtained from the bottle incubation method
(Table 1), likely reflecting a greater influence of sediment respiration on the Îľobs values than in
the bottle method. Nonetheless, values for Îľobs obtained by the bottle incubation method varied
by over 10 ‰, even in the absence of sediment respiration and were least negative during the
summer stratified period in Muskegon Lake. This is in agreement with Ostrom et al. (2014), who
proposed that εobs in the water column might in fact approach 0 ‰ when R rates are high and that
Îľobs may correlate with rates of total respiration beneath the thermocline (water column and
sediment respiration). Low Îľobs values in August are thus indicative of high R, supported by the
highest water column R observed at this time in 2012 (Figure 4).
The observation of marked seasonal variation in Îľobs indicates that Îľobs in the water
column should not be assumed to be constant, as is often the case. For example, if an Îľ obs value
of -21.2 ‰ had been incorporated into the 18O model rather than the measured value of -11.3 ‰
in August 2013, the GPP and R rates for Muskegon Lake would have been overestimated by
factors of 2.0 and 3.3, respectively. Further, the subsequent calculation of P/R would have
yielded a value of 1.41 rather than 2.36. Accurate assessments of ecosystem metabolism by
models that utilize Îľobs thus require that variation in Îľobs be incorporated into P/R models.

Greenhouse gas concentrations and emissions
On a global basis, freshwater ecosystems support a variety of microbial processes that
emit CH4 and N2O, important greenhouse gases (IPCC, 2007). Fluxes of greenhouse gases
arising from the lower water column and sediments vary as a function of wind mixing and with
the development and deterioration of thermal stratification (e.g., Fallon et al. 1980). In many

22

systems, CH4 accumulates in anoxic waters beneath the thermocline as a result of anaerobic
methanogenesis in both the water column and sediments (Bastviken et al. 2004). This suggests
that O2 and CH4 should be negatively correlated. Contrary to observations in other temperate
dimictic lakes (Fallon et al. 1980; Striegl and Michmerhuizen 1998), profiles of CH4 in
Muskegon Lake revealed no apparent connection between hypoxic conditions and high CH 4
concentrations (Figure 3). With the exception of one date that occurred during a seasonal mixing
event near the sediment-water interface (20 September 2012), accumulation of CH4 within the
hypolimnion was not observed in Muskegon Lake. This lack of hypolimnetic accumulation could
relate to low CH4 production, elevated oxidation of CH4, or evolution of CH4 to upper water
column and atmosphere (Fallon et al. 1980; Striegl and Michmerhuizen 1998; Huttunen et al.
2006).
Despite the lack of hypolimnetic accumulation, diffusive CH4 emissions from Muskegon
Lake were comparable in magnitude and range to those in other lakes of similar size (Bastviken
et al. 2004; Ortiz-Llorente and Alvarez-Cobelas, 2012). Owing to strong winds and high
epilimnetic concentrations, high emissions of CH4 were observed in September 2012 and May
2013 (Figures 3c, 3g, 5a). While CH4 emissions were measured at a single location in Muskegon
Lake and thus provide limited ability to extrapolate to the lake as a whole, our results indicate
that sampling on a seasonal basis is important to capture variability in CH4 emissions from lakes.
In contrast to CH4, hypolimnetic production of N2O was evident by a negative correlation
between N2O and O2 (Appendix Figure A2), indicating that periods of hypoxia and water column
turnover could be strong regulators of N2O emissions from this system. Indeed, N2O was
markedly supersaturated in the hypoxic hypolimnion, potentially setting the stage for high
emissions to the atmosphere during periods of water column turnover. Diffusive emissions of

23

N2O measured at the MLO were lowest during stratified periods and highest during the
development or breakdown of stratification. Periods of exceptionally high emissions, up to two
orders of magnitude greater than those during the stratified period, are among the highest
reported for natural lakes (Lemon and Lemon 1981; Huttunen et al. 2003; Wang et al. 2006;
Whitfield et al., 2011; Miettinen et al. 2015; Yang et al. 2015). In fact, only one study has
reported higher atmospheric fluxes of N2O than those observed in Muskegon Lake (Wang et al.
2006). While limited spatial sampling restricts the ability to evaluate the absolute N 2O emissions
from Muskegon Lake, our results illustrate wide variation in N 2O fluxes on seasonal time scales.
Total annual fluxes may thus be dominated by a few brief periods of high emissions during
mixing events following hypolimnetic accumulation (Figure 5a), emphasizing the importance of
temporal sampling to capture important periods of N2O emissions from lakes.
When diffusive emissions of CH4 and N2O are converted into CO2 equivalents (IPCC
2007), the radiative forcing of CH4 always exceeded that of N2O in Muskegon Lake (Figure 5b).
However, the radiative forcing of N2O in this study closely approached that of CH 4 on several
occasions, indicating that N2O is an important contributor to total greenhouse gas emissions from
this environment. Nonetheless, the number of publications reporting lacustrine CH 4 emissions is
over six-fold greater than those reporting N2O emissions (Whitfield et al. 2011; Ortiz-Llorente
and Alvarez-Cobelas 2012), reflecting a need for greater attention to N2O in studies that measure
greenhouse gas emissions.

Determination of microbial origin of N2O
The marked differences in the 15N, 18O, and SP of N2O between epilimnetic/mixed
water column samples and hypolimnetic samples point to distinct sources, namely atmospheric

24

exchange and microbial N2O production, respectively (Table 2). It is therefore expected that the
isotope composition of N2O samples collected from Muskegon Lake reflects a mixture between
atmospheric and microbially derived N2O, with microbial production stimulated by hypoxic
conditions in the hypolimnion (Goreau et al. 1980; Knowles 1982). Treating the atmospheric
isotopic composition (Table 2; Yoshida and Toyoda 2000) and the microbial isotopic N2O
composition (unknown) as isotopic end members in Muskegon Lake, regressing the isotopic
composition of measured N2O by the inverse N2O concentration yields the isotope composition
of microbially-produced N2O as the y-intercept (Figure 6; Pataki et al. 2003; Yamagishi et al.
2007). While 15N-N2O (-3.2 ‰) and 18O-N2O (58.8 ‰) values obtained from the mixing
model are heavily dependent on the isotopic composition of source material, SP (25.4 ‰) has
been shown to be a conservative tracer, enabling the distinction of production from
hydroxylamine oxidation (nitrification) from that by denitrification (Ostrom and Ostrom 2011).
SP was therefore used as the primary indicator of the microbial N 2O production mechanism in
this study. When SP values of 33 ‰ and -10 to 0 ‰ are considered as end-members for N2O
production via hydroxylamine oxidation and denitrification, respectively (Sutka et al. 2006;
Frame and Casciotti 2010), the observed SP of 25.4 ‰ indicates that 77-82% of in situ
production of N2O is derived from hydroxylamine oxidation.
Because reduction of N2O by denitrification is likely under low-oxygen conditions and
has a marked fractionation effect on its isotopic composition (Popp et al. 2002; Westley et al.
2006; Ostrom et al. 2007; Yamagishi et al. 2007), reduction must be considered as a possible
influence on SP in Muskegon Lake that may bias interpretations of microbial sources. If
substantial reduction occurs, 15N and 18O values of N2O tend to positively correlate because
the conversion of N2O to N2 preferentially leaves 15N and 18O in the residual N2O. Similarly,N2O

25

reduction results in elevated SP values in the residual N 2O pool because the isotopic
discrimination is more pronounced in the ι than the β position (Westley et al. 2006). However, a
negative correlation was found between 15N-N2O and 18O-N2O in Muskegon Lake (Appendix
Figure A4a; linear regression, R2 = 0.68, t = 4.87, df = 11, p < 0.001), which suggests that N2O
reduction is not an important process. Moreover, while the 18O-N2O and SP values were
positively correlated, the observed slope of 0.78 differed substantially from the expected slope of
0.45 arising from the fractionation of N2O during reduction (Appendix Figure A4b; Ostrom et al.
2007; Opdyke et al. 2009). Therefore, we conclude on the basis of stable isotopic indicators that
N2O reduction was not important in removing N2O from Muskegon Lake, supporting our
interpretation of hydroxylamine oxidation as the predominant N2O production pathway on the
basis of SP. Further, while it has been posited that lakes may act as sinks for N 2O under anoxic
conditions (Lemon and Lemon 1981; Beaulieu et al. 2014; 2015), N2O reduction was not
demonstrated as an appreciable pathway in Muskegon Lake. A potential explanation for the
absence of appreciable N2O reduction is the evidence of intrusion of oxic water into the
hypolimnion; although hypoxia reestablishes quickly in this system, complete anoxia does not
develop for long periods, which could inhibit the complete reduction of N 2O by denitrification.
Stoichiometric relationships can also be examined to determine N2O production
pathways. A positive relationship between AOU and ΔN2O has been attributed to N2O
production by nitrification, as AOU is a tracer of organic matter remineralization that produces
the substrates for nitrification (Yoshinari 1976; Nevison et al. 2003; Bange et al. 2010). Indeed,
AOU was positively correlated with ΔN 2O, and NO3- and NH4+ concentrations increased in the
hypolimnion throughout the stratified period (Appendix Figure A1, A2), providing a further

26

indication that nitrification is the primary N2O production mechanism under hypoxic conditions
in Muskegon Lake.
Additional insight into the microbial origin of N2O can be provided by its 18O
composition. The oxidation of NH4+ to hydroxylamine incorporates an initial oxygen atom from
O2, and further oxidation to NO3- derives oxygen from H2O (Dua et al., 1979; Hollocher et al.,
1981; Andersson and Hooper, 1983; Kumar et al., 1983). N2O production by hydroxylamine
oxidation therefore reflects the isotopic composition of O 2, whereas denitrification produces N2O
with a 18O value influenced both by O2 and H2O. Given that O2 is generally enriched in 18O by
at least 23.5 ‰ relative to H2O, Δ18O values are high when hydroxylamine oxidation dominates
and are lower when denitrification dominates. For example, Ostrom et al. (2000) observed a shift
in Δ18O values from approximately 23 ‰ at the surface to approximately 13 ‰ at 300 m depth in
the Pacific Ocean that was interpreted as a transition in N 2O production from hydroxylamine
oxidation to denitrification. Similarly, Muskegon Lake Δ18O values (22.1  6.0 ‰) are consistent
with production via hydroxylamine oxidation.
The microbial origin of N2O from hydroxylamine oxidation was confirmed by multiple
lines of isotopic and stoichiometric evidence, demonstrating that Muskegon Lake is as a
nitrification-driven source of N2O to the atmosphere. Previous studies have shown that
denitrification can dominate N2O production in some eutrophic lakes (Lemon and Lemon 1981;
Wang et al. 2006) and that increased availability of organic carbon and nutrients has the capacity
to stimulate denitrification (Taylor and Townsend 2010). Further, Beaulieu et al. (2014; 2015)
have demonstrated that reservoirs have the capacity to alternate between a source and a sink of
N2O depending on the degree of hypolimnetic anoxia and the timing of water column turnover.
Our results conversely indicate the capacity for ample supplies of organic matter and nutrients to

27

fuel N2O production via denitrification lakes is not ubiquitous. Together, this body of work
demonstrates that N2O dynamics in lakes are more complicated than previously thought, and
constraining the role of lakes in the global N 2O budget will entail (1) distinguishing between
nitrification and denitrification pathways of N2O production, (2) evaluating N2O reduction as a
potential sink for N2O, and (3) investigating water column turnover events as opportunities for
brief yet intense periods of N2O emissions.

Conclusion
Hypoxia persists in Muskegon Lake despite episodic intrusions of oxic water from Lake
Michigan into the hypolimnion. We demonstrate net autotrophy in the epilimnion and substantial
variation in Îľobs in late summer, consistent with delivery of autotrophic material to the
hypolimnion followed by high rates of respiration. The presence of hypoxia in Muskegon Lake
not only supports N2O production via nitrification but also results in exceptionally high N2O
emissions to the atmosphere, among the highest reported from lakes, during seasonal water
column mixing events. A complete understanding of the impact of hypoxic lakes on the global
greenhouse budget will rely on quantification of atmospheric fluxes, particularly during seasonal
transitions in stratification. This study adds to the emerging narrative of inland waters as sources
of globally significant greenhouse gas emissions to the atmosphere.

28

APPENDIX

29

Principal component analysis was performed on 2013 data (N 2O and CH4 concentration,
temperature, dissolved O2 concentration, pH, chlorophyll a concentration, and nutrient
concentrations (NO3-, NH4+, and SRP) to identify principal components (PCs) that maximize the
correlation among measured variables. Observations were separated by date and depth. Data
were standardized to eliminate scale differences. The loadings for each variable were calculated
as the correlation of the original variable with each PC, and the scores for each observation were
calculated by multiplying by the eigenvector for each PC.
Principal component analysis revealed that two PCs explained 74% of the variance in the
dataset (Fig. A2). N2O concentration was positively associated with PC1 and PC2, whereas CH 4
was positively associated with PC2 only. The loadings for chlorophyll a and dissolved O2
concentration were similar in magnitude and direction, indicating these two variables were
closely related in Muskegon Lake. Similarly, temperature and pH were also closely related to
one another. Nutrient variables (NO3-, NH4+, and SRP) were all positively correlated to PC1 but
were not strongly correlated with PC2. When plotted by their scores for the two PCs,
observations sorted into four apparent groupings: (1) all depths on 6 May, characterized by high
greenhouse gas concentrations, low temperature and nutrient concentrations, and high dissolved
O2 concentrations; (2) epilimnion samples, characterized by low greenhouse gas concentrations
and high temperature, pH, chlorophyll, and dissolved O 2 concentrations; (3) hypolimnion
samples during the stratified period, characterized by high N 2O concentrations, high nutrient
concentrations, and low chlorophyll and dissolved O 2 concentrations; and (4) metalimnion
samples, characterized by intermediate characteristics between epilimnion and hypolimnion
samples.

30

NH4+ (Îźg L-1)
0
0

10

20

30

40

50

150
0
b

a

200

NO3- (Îźg L-1)
250
300

350

400

0
0

2

2

4

4

4

6

6

6

8

8

8

10

10

10

Depth (m)

2

5

SRP (Îźg L-1)
10
15

20

c

6-May
11-June
12-Aug
14-Aug
15-Aug
16-Sept
2m
5m
8m
10.25 m

1

CH4
N2O

Chlor
DO

SRP

0

PC2 (23%)

2

3

Figure A1. Nutrient concentrations in 2013. (a) NH4+, (b) NO3-, and (c) SRP concentrations were
measured at four depths on each sampling date.

NO3
NH4

pH
-1

Temp

-2

0

2

4

PC1 (51%)

Figure A2. Principal component analysis for 2013 data. Points represent principal component
scores for each sampling depth on a given date. Vectors represent loadings for each variable on a
scale of zero to one and are plotted for illustrative purposes only. CH4 = CH 4 concentration,
N2O = N2O concentration, Temp = temperature, DO = dissolved O2, pH = pH, Chlor =
chlorophyll a fluorescence, NO3 = NOx concentration, NH4 = NH4+ concentration, SRP =
soluble reactive phosphorus concentration.

31

35
30
25

ΔN2O (nM)

20
15
10
5
0
-50

0

50

100

150

200

250

300

AOU (ÎźM)

Figure A3. Relationship between ΔN2O and apparent oxygen utilization (AOU). AOU was
calculated as the difference between the saturation concentration and the in situ concentration of
of dissolved O2. ΔN2O was calculated as the difference between the in situ concentration and the
saturation concentration of N2O at each sampling point. A positive correlation (R2 = 0.58)
supports the interpretation of a N2O production source from nitrification.
6

25

a

b

5

3

SP (‰)

δ15N (‰)

4

2

20

1
0
15

-1
47

49

51
δ18O (‰)

53

46

55

48

50

52

54

56

δ18O (‰)

Figure A4. Trends in N2O concentration and isotopic composition. (a) δ18O-N2O and δ15N-N2O
are negatively correlated (R2 = 0.68). (b) δ18O-N2O and isotopic site preference (SP) are
positively correlated (R2 = 0.60) with a slope of 0.78 (solid line). N2O reduction produces a
characteristic δ18O-N2O vs. SP slope of 0.45 (dotted line).

32

LITERATURE CITED

33

LITERATURE CITED

Andersson KK, Hooper AB (1983) O2 and H2O are each the source of one O in NO2- produced
from NH3 by Nitrosomonas: 15NMR evidence. FEBS Lett 164(2):236-240.
doi:10.1016/0014-5793(83)80292-0
APHA (2005) Standard methods for the examination of water and wastewater. American Public
Health Association, Washington, D.C.
Audet J, Hoffmann CC, Andersen PM, Baattrup-Pedersen A, Johansen JR, Larsen SE,
Kjaergaard, Elsgaard L (2014) Nitrous oxide fluxes in undisturbed riparian wetlands located in
agricultural catchments: Emission, uptake, and controlling factors. Soil Biol Biochem 68:291299. doi: 10.1016/j.soilbio.2013.10.011
Bange H, Freing A, Kock A, LĂśscher C (2010) Marine pathways to nitrous oxide. In: Smith K
(ed) Nitrous Oxide and Climate Change. Earthscan, London, pp 36–62.
Bastviken D, Cole J, Pace M, Tranvik L (2004) Methane emissions from lakes: Dependence of
lake characteristics, two regional assessments, and a global estimate. Global Biogeochem Cy
18:GB4009. doi:10.1029/2004GB002238
Beaulieu JJ, Smolenski RL, Nietch CT, Townsend-Small A, Elovitz MS, Schubauer-Berigan JP
(2014) Denitrification alternates between a source and sink of nitrous oxide in the hypolimnion
of a thermally stratified reservoir. Limnol Oceanogr 59(2):495-506.
doi: 10.4319/lo.2014.59.2.0495
Beaulieu JJ, Nietch CT, Young JL (2015) Controls of nitrous oxide production and consumption
in reservoirs of the Ohio River Basin. J Geophys Res Biogeosci 120:1995-2010.
doi:10.1002/2015JG002941
Benson BB, Krause Jr. D (1984) The concentration and isotopic fractionation of oxygen
dissolved in freshwater and seawater in equilibrium with the atmosphere. Limnol Oceanogr
29(3):620-632.
Biddanda BA (2012) Lake sentinel: Observatory for ecosystem changes in Muskegon Lake.
InterChange, Newsletter of the Regional Math and Science Center, Grand Valley State
University, http://www.gvsu.edu/rmsc/interchange/2012-may-connections-awri-651.htm
Bocaniov SA, Schiff SL, Smith REH (2012) Plankton metabolism and physical forcing in a
productive embayment of a large oligotrophic lake: Insights from stable oxygen isotopes.
Freshwater Biol 57:481-496. doi:10.1111/j.1365-2427.2011.02715.x
Brandes JA, Devol AH (1997) Isotopic fractionation of oxygen and nitrogen in coastal marine
sediments. Geochim Cosmochim Ac 61(9):1793-1801. doi:10.1016/S0016-7037(97)00041-0

34

Carter GS, Nalepa TF, Rediske RR (2006), Status and trends of benthic populations in a coastal
drowned river mouth lake of Lake Michigan. J Great Lakes Res 32:578-595
CENR (2000) Integrated assessment of hypoxia in the Northern Gulf of Mexico. National
Science and Technology Council Committee on Environment and Natural Resources,
Washington, D.C.
Conroy JD, Boegman L, Zhang H, Edwards WJ, Culver DA (2011) “Dead Zone” dynamics in
Lake Erie: The importance of weather and sampling intensity for calculated hypolimnetic oxygen
depleted rates. Aquat Sci 73(2):289-304. doi:10.1007/s00027-010-0176-1
del Giorgio PA, Peters RH (1994) Patterns in planktonic P:R ratios in lakes: Influence of lake
trophy and dissolved organic carbon. Limnol Oceanogr 39(4):772-787.
doi:10.4319/lo.1994.39.4.0772
Diaz RJ, Rosenberg R (2008) Spreading dead zones and consequences for marine ecosystems.
Science 321:926-929. doi:10.1126/science.1156401
Dila DK, Biddanda BA (2015) From land to lake: Contrasting microbial processes across a Great
Lakes gradient of carbon and nutrient inventories. J Great Lakes Res 41:75-85.
Dua RD, Bhandari B, Nicholas DJD (1979) Stable isotope studies on the oxidation of ammonia
to hydroxylamine by Nitrosomonas europaea. FEBS Lett 106(2):401-404.
doi:10.1016/0014-5793(79)80541-4
Emerson S, Quay P, Stump C, Wilbur D, Knox M (1991) O2, Ar, N2, and 222Rn in surface waters
of the subarctic ocean: Net biological O2 production. Global Biogeochem Cy 5(1):49-69.
doi:10.1029/90GB02656
Fallon RD, Harrits S, Hanson RS, Brock TD (1980) The role of methane in internal carbon
cycling in Lake Mendota during summer stratification. Limnol Oceanogr 25(2):357-360.
doi:10.4319/lo.1980.25.2.0357
Feldman DE, Yost Jr. HT, Benson BB (1959) Oxygen isotope fractionation in reactions
catalyzed by enzymes. Science 129:146-147.
Frame CH, Casciotti KL (2010) Biogeochemical controls and isotopic signatures of nitrous oxide
production by a marine ammonia-oxidizing bacterium. Biogeosciences 7:2695-2709.
doi:10.5194/bg-7-2695-2010
Freedman P, Canale R, Auer M (1979) The impact of wastewater diversion spray irrigation on
water quality in Muskegon County lakes, Rep. 905/9–79–006-A. U.S. Environmental Protection
Agency, Washington, D.C.
Friedrich J, Janssen F, Aleynik D et al (2014) Investigating hypoxia in aquatic environments:

35

Diverse to addressing a complex phenomenon. Biogeosciences 11:1215-1259.
doi:10.5194/bg-11-1215-2014
Fry B, Boyd B (2010) Oxygen concentration and isotope studies of productivity and respiration
on the Louisiana continental shelf, July 2007. In: Ohkouchi N et al (ed) Earth, Life, and Isotopes.
Kyoto University Press, Kyoto, pp 223-241.
Goreau TJ, Kaplan WA, Wofsy SC, McElroy MB, Valois FW, Watson SW (1980) Production of
NO2- and N2O by nitrifying bacteria at reduced concentrations of oxygen. Appl Environ Microb
40(3):526–532.
Hamilton SK, Ostrom NE (2007) Measurements of the stable isotope ratio of dissolved N 2 in 15N
tracer experiments. Limnol Oceanogr Methods 5:233-240.
Hanson, PC, Bade DL, Carpenter SR, Kratz TK (2003) Lake metabolism: Relationships with
dissolved organic carbon and phosphorus. Limnol Oceanogr 48(3):1112-1119.
doi:10.4319/lo.2003.48.3.1112
Hollocher TC, Tate ME, Nicholas DJ (1981) Oxidation of ammonia by Nitrosomonas europaea.
Definite 18O-tracer evidence that hydroxylamine formation involves a mono-oxygenase. J Biol
Chem 256:10834-10836.
Huttunen JT, Juutinen S, Alm J, Larmola T, Hammar T, Silvola J, Martikainen PJ (2003) Nitrous
oxide flux to the atmosphere from the littoral zone of a boreal lake. J Geophys Res
108(D14):4421. doi:10.1029/2002JD002989
Huttunen JT, Väisänen TS, Hellsten SK, Martikainen PJ (2006) Methane fluxes at the sedimentwater interface in some boreal lakes and reservoirs. Boreal Environ Res 11:27-34.
IPCC (2007) Climate Change 2007: The Physical Science Basis. In: Solomon S et al (eds)
Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental
Panel on Climate Change. Cambridge University Press, New York.
IPCC (2013) Climate Change 2013: The Physical Science Basis. In: Stocker TF et al (eds)
Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel
on Climate Change. Cambridge University Press, New York.
Jankowski T, Livingstone DM, BĂźhrer H, Forster R, Niederhauser P (2006) Consequences of the
2003 European heat wave for lake temperature profiles, temperature stability, and hypolimnetic
oxygen depletion: Implications for a warmer world. Limnol Oceanogr 51(2):815-819.
doi:10.4319/lo.2006.51.2.0815
Kaushal SS, Mayer PM, Vidon PG, Smith RM, Pennino MJ, Newcomer TA, Duan S, Welty C,
Belt KT (2014) Land use and climate variability amplify carbon, nutrient, and contaminant
pulses: A review with management implications. U.S. Environmental Protection Agency Paper
234.

36

Knowles R (1982) Denitrification. Microbiological Rev 46(1):43-70.
Kumar S, Nicholas DJD, Williams EH (1983) Definitive 15N NMR evidence that water serves as
a source of ‘O’ during nitrite oxidation by Nitrobacter agilis. FEBS Lett 152(1):71-74.
doi:10.1016/0014-5793(83)80484-0
Lemon E, Lemon D (1981) Nitrous oxide in freshwaters of the Great Lakes Basin. Limnol
Oceanogr 26(5):867-879. doi:10.4319/lo.1981.26.5.0867
Lehman MF, Barnett B, GĂŠlinas Y, Gilbert D, Maranger R, Mucci A, Sundby B, Thibodeau B
(2009) Aerobic respiration and hypoxia in the Lower St. Lawrence Estuary: Stable isotope ratios
of dissolved oxygen constrain oxygen sink partitioning. Limnol Oceanogr 54(6):2157-2169.
Lis G, Wassenaar LI, Hendry MJ (2008) High-precision laser spectroscopy D/H and 18O/16O
measurements of microliter natural water samples. Anal Chem 80(1):287-293.
doi:10.1021/ac701716q
Livingstone DM, Imboden DM (1996) The prediction of hypolimnetic oxygen profiles: A plea
for a deductive approach. Can J Fish Aquat Sci 53:924-932. doi:10.1139/f95-230
Mariotti A, Germon JD, Hubert P, Kaiser P, Letolle R, Tardieux A, Tardieux P (1981)
Experimental determination of nitrogen kinetic isotope fractionation: Some principles;
Illustration for the denitrification and nitrification processes. Plant and Soil 62(3):413-430.
doi:10.1007/BF02374138
McCarthy MJ, Carini SA, Liu Z, Ostrom NE, Gardner WS (2013) Oxygen consumption in the
water column and sediments of the Gulf of Mexico hypoxic zone. Estuar Coast Shelf S 123:4653. doi:10.1016/j.ecss.2013.02.019
McNair JN, Gereaux LC, Weinke AD, Sesselmann MR, Kendall ST, Biddanda BA (2013) New
methods for estimating components of lake metabolism based on free-water dissolved-oxygen
dynamics. Ecol Model 263:251-263. doi:10.1016/j.ecolmodel.2013.05.010
Michigan DEQ (2011) Stage 2 remedial action plan Muskegon Lake Area of Concern. Michigan
Department of Environmental Quality, Lansing, Michigan
Michmerhuizen CM, Striegl RG, McDonald ME (1996) Potential methane emission from northtemperate lakes following ice melt. Limnol Oceanogr 41(5):985-991.
doi:10.4319/lo.1996.41.5.0985
Miettinen H, Pumpanen J, Heiskanen JJ, Aaltonen H, Mammarella I, Ojala A, Levula J,
Rantakari M (2015) Towards a more comprehensive understanding of lacustrine greenhouse gas
dynamics – two-year measurements of concentrations and fluxes of CO 2, CH4, and N2O in a
typical boreal lake surrounded by managed forests. Boreal Env Res 20:75-89.

37

Morse JL, ArdĂłn M, Bernhardt ES (2012) Greenhouse gas fluxes in southeastern U.S. coastal
plain wetlands under contrasting land uses. Ecol Appl 22(1):264-280.
Nevison C, Butler JH, Elkins JW (2003) Global distribution of N 2O and the ΔN2O-AOU yield in
the subsurface ocean. Global Biogeochem Cy 17(4):1119. doi:10.1029/2003GB002068
Opdyke MR, Ostrom NE, Ostrom PH (2009) Evidence for the predominance of denitrification as
a source of N2O in a temperate agricultural soils based on isotopologue measurements. Global
Biogeochem Cy 23:GB4018. doi:10.1029/2009GB003523
Ortiz-Llorente MJ, Alvarez-Cobelas M (2012) Comparison of biogenic methane emissions from
unmanaged estuaries, lakes, oceans, rivers and wetlands. Atmos Environ 59:328-337.
doi:10.1016/j.atmosenv.2012.05.031
Ostrom NE, Carrick HJ, Twiss MR, Piwinski L (2005) Evaluation of primary production in Lake
Erie by multiple proxies. Oecologia 144:115-124. doi:10.1007/s00442-005-0032-5
Ostrom NE, Gandhi H, Kamphuis B, DeCamp S, Liu Z, McCarthy MJ, Gardner WS (2014)
Oxygen metabolism and water mass mixing in the northern Gulf of Mexico hypoxic zone in
2010. Geochim Cosmochim Ac 140:39-49. doi:10.1016/j.gca.2014.05.020
Ostrom NE, Ostrom PH (2011) The isotopomers of nitrous oxide: Analytical considerations and
application to resolution of microbial production pathways. In: Baskaran M (ed) Handbook of
Environmental Isotope Geochemistry. Springer-Verlag, New York, pp 453-476.
Ostrom NE, Pitt A, Sutka R, Ostrom PH, Grandy AS, Huizinga KM, Robertson GP (2007)
Isotopologue effects during N2O reduction in soils and in pure cultures of denitrifiers. J Geophys
Res 112:G02005. doi:10.1029/2006JG000287
Ostrom NE, Russ ME, Popp B, Rust TM, Karl DM (2000) Mechanisms of nitrous oxide
production in the subtropical North Pacific based on determinations of the isotopic abundances
of nitrous oxide and di-oxygen. Chemosphere - Global Change Sci 2:281-290.
doi:10.1016/S1465-9972(00)00031-3
Pataki DE, Ehleringer JR, Flanagan LB, Yakir D, Bowling DR, Still CJ, Buchmann N, Kaplan
JO, Berry JA (2003) The application and interpretation of Keeling plots in terrestrial carbon
cycle research. Global Biogeochem Cy 17(1):1022. doi:10.1029/2001GB001850
Popp BN, Westley MB, Toyoda S, Miwa T, Dore JE, Yoshida N, Rust TM, Sansone FJ, Russ
ME, Ostrom NE, Ostrom PH (2002) Nitrogen and oxygen isotopomeric constraints on the
origins and sea-to-air flux of N2O in the oligotrophic subtropical North Pacific gyre. Global
Biogeochem Cy 16(4):1064, doi:10.1029/2001GB001806
Poth M, Focht DD (1985) 15N kinetic analysis of N2O production by Nitrosomonas europaea:
An examination of nitrifier denitrification. Appl Environ Microb 49(5):1134-1141.

38

Quiñones-Rivera ZJ, Wissel B, Justić D, Fry B (2007) Partitioning oxygen sources and sinks in a
stratified, eutrophic coastal ecosystem using stable oxygen isotopes. Mar Ecol Prog Ser 342:6983. doi:10.3354/meps342069
Rabalais NN, Diaz RJ, Levin LA, Turner RE, Gilbert D, Zhang J (2010) Dynamics and
distribution of natural and human-caused hypoxia. Biogeosciences 7:585-619.
doi:10.5194/bg-7-585-2010
Riera JL, Schindler JE, Kratz TK (1999) Seasonal dynamics of carbon dioxide and methane in
two clear-water lakes and two bog lakes in northern Wisconsin, U.S.A. Can J Fish Aquat Sci
56:265-274. doi:10.1139/f98-182
Roberts BJ, Russ ME, Ostrom NE (2000) Rapid and precise determination of the 18O of
dissolved gaseous dioxygen via gas chromatography-isotope ratio mass spectrometry. Environ
Sci Technol 34(11):2337-2341. doi:10.1021/es991109d
Sansone FJ, Popp BN, Rust TM (1997) Stable carbon isotopic analysis of low-level methane in
water and gas. Anal Chem 69(1):40-44. doi:10.1021/ac960241i
Sasaki Y, Koba K, Yamamoto M, Makabe A, Ueno Y, Nakagawa M, Toyoda S, Yoshida N, Yoh
M (2011) Biogeochemistry of nitrous oxide in Lake Kizaki, Japan, elucidated by nitrous oxide
isotopomers analysis. J Geophys Res 116:G04030. doi:10.1029/2010JG001589
Scavia D, Allan JD, Arend KK et al (2014) Assessing and addressing the re-eutrophication of
Lake Erie: Central basin hypoxia. J Great Lakes Res 40:226-246. doi:10.1016/j.jglr.2014.02.004
Seitzinger S, Harrison JA, BĂśhlke JK, Bouwman AF, Lowrance R, Peterson B, Tobias C, Van
Drecht G (2006) Denitrification across landscapes and waterscapes: A synthesis. Ecol Appl
16(6):2064-2090.
Seitzinger SP, Kroeze C (1998) Global distribution of nitrous oxide production and N inputs in
freshwater and coastal marine ecosystems. Global Biogeochem Cy 12(1):93-113.
Stein LY (2011) Surveying N2O-producing pathways in bacteria. In: Klotz MG (ed) Methods in
Enzymology, vol 486. Elsevier Academic Press, San Diego, California, pp 131-152.
Steinman AD, Ogdahl M, Rediske R, Ruetz III CR, Biddanda BA, Nemeth L (2008) Current
status and trends in Muskegon Lake, Michigan. J Great Lakes Res 34(1):169-188.
doi:10.3394/0380-1330(2008)34[169:CSATIM]2.0.CO;2
Striegl RG, Michmerhuizen CM (1998) Hydrologic influence on methane and carbon dioxide
dynamics at two north-central Minnesota lakes. Limnol Oceanogr 43(7):1519-1529.
Sutka RL, Adams GC, Ostrom NE, Ostrom PH (2008) Isotopologue fractionation during N 2O
production by fungal denitrification. Rapid Commun Mass Spectrom 22(24):3989-3996.
doi:10.1002/rcm.3820

39

Sutka RL, Ostrom NE, Ostrom PH, Breznak JA, Gandhi H, Pitt AJ, Li F (2006) Distinguishing
nitrous oxide production from nitrification and denitrification on the basis of isotopomer
abundances. Appl Environ Microb 72(1):638-644. doi:10.1128/AEM.72.1.638-644.2006
Sutka RL, Ostrom NE, Ostrom PH, Gandhi H, Breznak JA (2003) Nitrogen isotopomer site
preference of N2O produced by Nitrosomonas europaea and Methylococcus capsulatus Bath.
Rapid Commun Mass Spectrom 17(7):738-745. doi:10.1002/rcm.968
Tang Z, Engel BA, Pijanowski BC, Lim KJ (2005) Forecasting land use change and its
environmental impact at a watershed scale. J Environ Manage 76:35-45.
doi:10.1016/j.jenvman.2005.01.006
Taylor PG, Townsend AR (2010) Stoichiometric control of organic carbon-nitrate relationships
from the soil to the sea. Nature 464: 1178-1181. doi:10.1038/nature08985
Toyoda S, Mutobe H, Yamagishi H, Yoshida N, Tanji Y (2005) Fractionation of N 2O
isotopomers during production by denitrifier. Soil Biol Biochem 37(8):1535-1545.
doi:10.1016/j.soilbio.2005.01.009
Turner RE, Rabalais NN, Justic D (2008) Gulf of Mexico hypoxia: Alternate states and a legacy.
Environ Sci Tech 42:2323-2327. doi:10.1021/es071617k
U.S. EPA (2013) Great Lakes Areas of Concern: Muskegon Lake. U.S. Environmental
Protection Agency, Washington, D.C.
Vail J, Meyer A, Weinke A, Biddanda BA (2015) Water quality monitoring: Lesson plan for
exploring time-series data. J Mich Sci Teach Assoc 60.1:37-48.
Walker JT, Stow CA, Geron C (2010) Nitrous oxide emissions from the Gulf of Mexico hypoxic
zone. Environ Sci Technol 44(5):1617-1623. doi:10.1021/es902058t
Wang H, Wang W, Yin C, Wang Y, Lu J (2006) Littoral zones as the “hotspots” of nitrous oxide
(N2O) emission in a hyper-eutrophic lake in China. Atmos Environ 40:5522-5527.
doi:10.1016/j.atmosenv.2006.05.032
Wanninkhof R (1992) Relationship between wind speed and gas exchange over the ocean. J
Geophys Res 97(C5):7373-7382. doi:10.1029/92JC00188
Weinke AD, Kendall ST, Kroll DJ, Strickler EA, Weinert ME, Holcomb TM, Defore AA, Dila
DK, Snider MJ, Gereaux LC, Biddanda BA (2014) Systematically variable planktonic carbon
metabolism along a land-to-lake gradient in a Great Lakes coastal zone. J Plankton Res 0:1-15.
doi:10.1093/plankt/fbu066

40

Westley MB, Yamagishi H, Popp BN, Yoshida N (2006) Nitrous oxide cycling in the Black Sea
inferred from stable isotope and isotopomers distributions. Deep-Sea Res Pt II 53:1802-1816.
doi:10.1016/j.dsr2.2006.03.012
Whitfield CJ, Aherne J, Baulch HM (2011) Controls on greenhouse gas concentrations in
polymictic headwater lakes in Ireland. Sci Total Environ 410-411:217-225.
doi:10.1016/j.scitotenv.2011.09.045
Williamson CE, Dodds W, Kratz TK, Palmer MA (2008) Lakes and streams as sentinels of
environmental change in terrestrial and atmospheric processes. Front Ecol Environ 6(5):247-254.
doi:10.1890/070140
World Meteorological Organization (2008) Guide to meteorological instruments and methods of
observation, WMO-No. 8. World Meteorological Organization, Geneva, Switzerland
Wrage N, Velthof GL, van Beusichem ML, Oenema O (2001) Role of nitrifier denitrification in
the production of nitrous oxide. Soil Biol Biochem 33(12-13):1723-1732.
doi:10.1016/S0038-0717(01)00096-7
Yamagishi H, Westley MB, Popp BN, Toyoda S, Yoshida N, Watanabe S, Koba K, Yamanaka Y
(2007) Role of nitrification and denitrification on the nitrous oxide cycle in the eastern tropical
North Pacific and Gulf of California. J Geophys Res 112:G02015. doi:10.1029/2006JG000227
Yang H, Andersen T, DĂśrsch P, Tominaga K, Thrane J-E, Hessen DO (2015) Greenhouse gas
metabolism in Nordic boreal lakes. Biogeochemistry 126:211-225.
doi:10.1007/s10533-015-0154-8
Yang H, Gandhi H, Ostrom NE, Hegg EL (2014) Isotopic fractionation by a fungal P450 nitric
oxide reductase during the production of N2O. Environ Sci Technol 48:10707-10715.
doi:10.1021/es501912d
Yoshida N, Toyoda S (2000) Constraining the atmospheric N 2O budget from intramolecular site
preference in N2O isotopomers. Nature 405:330-334. doi:10.1038/35012558
Yoshinari T (1976) Nitrous oxide in the sea. Mar Chem 4:189-202.
doi:10.1016/0304-4203(76)90007-4

41

CHAPTER 2
WETLAND RESTORATION AND HYDROLOGIC RECONNECTION RESULT IN
ENHANCED WATERSHED NITROGEN RETENTION AND REMOVAL
ABSTRACT
Restoration of wetlands presents a potential water quality benefit via removal of
nutrients, including excess nitrogen (N) and phosphorus (P), but there is potential for complex
and unresolved changes in nutrient cycling following restoration. In this study, we evaluated N
removal and release in a former agricultural wetland under scenarios of hydrologic reconnection
to the watershed and sediment dredging. We modeled potential downstream impacts of
restoration activities in the context of N management. Denitrification, N2O production, and
anammox were measured via the isotope pairing technique in intact sediment cores, and
probabilistic estimates of N cycling under restoration scenarios were modeled in a Bayesian
framework. Anammox was not detected in this system, but denitrification and N2O production
were stimulated shortly following simulated dredging, indicating a temporary response to
sediment disturbance. Owing to a decrease in NH 4+ availability under hydrologic reconnection,
coupled nitrification-denitrification was inhibited. Despite a decrease in denitrification activity,
the wetland has the capacity to remove up to 10 % of stream NO 3= via denitrification following
hydrologic reconnection. Restoration activities are predicted to mitigate NH4+ delivery to a
downstream eutrophic lake, yet permanent N removal will be reduced due to the decoupling of
nitrification and denitrification.

INTRODUCTION
Wetlands have the potential to provide significant ecosystem services, including wildlife
habitat, nutrient retention, and groundwater recharge (Keddy 2000). Many natural wetlands in

42

the United States have been converted to agricultural production (Dahl 2000) but are now being
targeted for restoration that in many cases is supported by financial incentives such as the
Wetlands Reserve Program (USDA 2012). Agricultural wetland restoration presents benefits
such as the creation of wildlife habitat (Bunn and Arthington 2002; Zelder and Kercher 2005).
However, many restored wetland sediments in agricultural landscapes bear residual nutrients that
impact the capacity for the wetland to retain phosphorus (P) and nitrogen (N). Restoration efforts
must thus consider not only habitat benefits but also the implications for water quality in the face
of eutrophication of many watersheds (ArdĂłn et al. 2010a; Welti et al. 2012).
Wetlands are well known for their ability to remove substantial quantities of P and N
(Fennessy and Cronk 1997; Reddy et al. 1999; Saunders and Kalff 2001). Although wetlands
have a tremendous capacity to remove nutrients, restoration may induce a decrease in water
quality, at least temporarily. Nutrient release can occur following reflooding and hydrologic
reconnection (Aldous et al. 2005; Lindenberg and Wood 2009; Morse et al. 2012), particularly in
systems with a history of nutrient loading such as agricultural soils (Newman and Pietro 2001;
Pant and Reddy 2003; Steinman and Ogdahl 2011). Downstream release of P is common,
whereas N has a greater capacity to be removed by sediments in restored wetlands (ArdĂłn et al.
2010a; 2010b; Steinman and Ogdahl 2011). While mitigation of P loading has received greater
attention owing to its role as a primary limiting nutrient for phytoplankton growth in freshwater
environments (Schindler 1974; Carpenter et al. 1998), N is emerging as an important co-limiting
nutrient that has recently been implicated in the growth and persistence of harmful algal blooms
in lakes (Conley et al. 2009; Chaffin and Bridgeman 2014; Steffen et al. 2014; Paerl et al. 2016).
The potential for wetlands to mitigate or exacerbate downstream N delivery is thus a key

43

determinant of downstream water quality, and N management should be incorporated alongside
P management into restoration activities.
Denitrification is the predominant process that permanently removes N from wetlands
(Seitzinger et al. 2006). Denitrification is the stepwise reduction of NO 3- to N2, with
intermediates of NO2-, NO, and N2O (Knowles 1982). Of particular concern is the proportion of
N that is not completely reduced to N2 and is lost as N2O, a potent greenhouse gas (IPCC 2007).
Thus, enhanced denitrification may have the unintended consequence of increasing greenhouse
gas emissions. N2O production rates and the N2O:N2 production ratio are likely to change
following wetland restoration, owing to differences in NO 3- supply, quality and quantity of
organic carbon, and hydrologic pulses (Welti et al. 2012; Sun et al. 2014). Denitrification can be
augmented by direct coupling with nitrification, which oxidizes NH4+ to NO3-. Coupled
nitrification-denitrification can make up a significant proportion of total denitrification in
wetlands (Scott et al. 2008).
Anammox, the transformation of NH4+ and NO2- to N2 (van de Graaf et al. 1995), is an
additional N removal process that is not well understood in wetland systems. This process is
expected to be minor compared to denitrification in wetlands owing to the inferior ability of
anammox bacteria to compete for substrates in organic-rich sediments (Thamdrup and Dalsgaard
2002; Humbert et al. 2012). However, anammox genes have been detected in wetlands (Zhu et
al. 2011; Humbert et al. 2012; Waki et al. 2015) and hotspots of anammox activity have been
observed at land-freshwater interfaces (Zhu et al. 2013). A complete understanding of the N
removal regime of wetlands and their potential impact on water quality will therefore benefit
from examination of both denitrification and anammox.

44

In this study, we aimed to quantify N removal and release by wetland sediments under
restoration scenarios, using statistical modeling to produce probabilistic estimates of the rates of
each N cycling process. Additionally, we evaluated potential downstream impacts of wetland
restoration activities in the context of watershed N and P management. Together, this work
provides an analysis of wetland N removal and release capacity before restoration activities
occur, which can help guide management decisions and expectations.

METHODS
Study site
This research was conducted in the Bear Lake Wetland Restoration Area (BLWRA) in
western Michigan, USA that is currently being subject to restoration activities (Figure 7).
Historically, the BLWRA was a deltaic wetland that was hydrologically connected to Bear Creek
on its north side. In the early 1900s, an earthen berm was constructed between Bear Creek and
the wetland, and the wetland was drained and placed into celery production. Agricultural
activities ceased and the pumps were turned off in 2002, allowing the wetland to be re-inundated
as two separate ponds. The main purpose of the restoration was to create habitat for wildlife,
although an additional benefit could be improved water quality. This study focused on the West
Pond (0.09 km2), which had not been subject to sediment dredging prior to 2016. Additional
restoration activities, completed in the two years following this study, were completed to (1)
dewater the pond, (2) dredge the sediment to a depth of approximately 0.5 m to remove the
agriculturally-impacted layer, and (3) re-inundate the pond. Future restoration will remove the
berm to reconnect the surface hydrology of the wetland and Bear Creek, assuming the nutrient

45

concentrations in the refilled ponds are lower than those in adjacent Bear Creek (Steinman and
Ogdahl 2016).
Restoration has the potential to impact water quality in Bear Lake, a eutrophic drowned
river mouth lake located approximately 250 m downstream of the West Pond that is connected
via Bear Creek. Bear Lake is within the geographic boundary of the Muskegon Lake Area of
Concern and is subject to evaluations of eutrophication and nutrient loading (USEPA 2013).
Previous studies have examined the potential effects of hydrologic reconnection on P retention
and release from the BLWRA (Smit and Steinman 2015; Steinman and Ogdahl 2016), and the
examination of N cycling was compared with those results to evaluate potential downstream
nutrient impacts on Bear Lake.

Figure 7. Map of the Bear Lake Wetland Restoration Area (BLWRA). Restoration activities in
the West Pond include removal of the berm that separates Bear Creek from the wetland and
sediment dredging of. Insets show the location of the ponds in the downstream portion of the
watershed (left) and in Michigan, USA (right).

46

Intact sediment cores
Sediment was collected at a sampling location in the northwest side of the pond (Figure
7) using a modified piston corer described by Smit and Steinman (2015). Intact sediment cores
were collected in polycarbonate tubes (7 cm i.d.) to a depth of 25 cm. Cores were then divided
into three treatment groups: “Control,” which simulates conditions without restoration, “HR,”
which simulates a hydrologic reconnection of the wetland to Bear Creek, and “HR+D,” which
simulates hydrologic reconnection and dredging of the top layer of sediment. Water from the
West Pond was added to Control cores and water from Bear Creek was added to HR and HR+D
cores to a depth of 20 cm inside the core tube (Table 3). The top 5 cm of sediment was removed
from the cores in the HR+D treatment. Though 5 cm is shallower than the depth to which
sediments are currently being dredged, it represents the removal of the most likely labile and
microbially active sediment layer. Cores were pre-incubated 12 h in the dark at in situ
temperature prior to the addition of 15N-NO3- under gentle bubbling with air to maintain oxic
conditions in the overlying water.

Table 3. DIN concentrations in the Bear Lake Wetland Restoration Area and adjacent Bear
Creek.
NH4Âą-N
M

NO2--N + NO3--N
M

Undredged Field

29.45

46.40

Bear Creek

1.35

22.34

Location

Following pre-incubation, a water sample was taken for DIN concentration analysis
(NO3-, NO2- and NH4+) by filtering through a 0.45 Âľm Millipore filter and freezing until analysis.
15N-NO 3

was then added to obtain a final concentration of 100 ÂľM in the overlying water of

47

each sediment core. Cores were then capped with a rubber plug and equipped with a motorized
stirring device inserted through a port in the plug. Stirring of the water overlying the core was
maintained just below the sediment resuspension threshold. An equilibration period of 30 min
was employed to allow for homogenization of NO 3- between the overlying water and with the
NO3- reduction zone in the sediment porewater (Dalsgaard et al. 2000). Cores from each
treatment were sacrificed in triplicate at intervals of 0, 6, and 12 h. Upon sacrifice, dissolved O 2
was measured using a YSI 6600 sonde (Yellow Springs Instruments Inc., Yellow Springs, OH).
A DIN sample was collected and processed as described above. Then, the top 2-4 cm of the
sediment core were slurrified with a glass rod to mix the overlying water with porewater, and
samples were collected for dissolved N2 and N2O analysis.
Samples for the determination of 15N2 were collected according to Hamilton and Ostrom
(2007); briefly, dissolved gases were equilibrated with a He atmosphere, and the headspace was
transferred into a pre-evacuated 12 mL Exetainer. Samples for analysis of dissolved N 2
concentration were siphoned into 12 mL Exetainers to overflowing and amended with 200 ÂľL of
saturated ZnCl2 solution to halt biological activity. All Exetainers were stored underwater at
room temperature to minimize diffusion of atmospheric N2 during storage. Samples for analysis
of the 15N2O and N2O concentration were siphoned into 250 and 60 mL serum bottles,
respectively, to overflowing and sealed without a headspace. Biological activity was halted by
addition of 1 and 0.25 mL of saturated HgCl2 solution, respectively, to each serum bottle. Prior
to analysis, a headspace of 20 mL He was injected into each 60 mL serum bottle, with
atmospheric pressure maintained by allowing 20 mL of water to exit via a vent needle. Serum
bottles were allowed to equilibrate under gentle shaking for at least 12 h prior to analysis.

48

Analyses
The concentration of NO3- was determined by ion chromatography on a Dionex DX500
(APHA 2005). NH4+ concentrations were analyzed by the automated phenate method (APHA
2005) on a on a Bran + Luebbe Autoanalyzer (SEAL Analytical, Mequon, WI). The isotopic
composition of N2 was analyzed by gas chromatograph (HP-5980, Hewlett Packard, Ramsey,
MN) interfaced to an Isoprime isotope ratio mass spectrometer (Elementar Americas, Inc.,
Mount Laurel, NJ). Analytical reproducibility was 0.3 ‰. Concentrations of dissolved N 2 were
analyzed by membrane inlet mass spectrometry according to Eyre et al. (2002). The headspace of
the 60 mL serum bottles was analyzed for the concentration of N2O by GC-ECD (Shimadzu
Greenhouse Gas Analyzer GC-2014, Shimadzu Scientific Instruments, Columbia, MD). The
original dissolved concentration of N2O in each sample was calculated based on the headspace
equilibrium concentration (Hamilton and Ostrom 2007). The isotopic composition of N 2O was
analyzed by introducing the sample into a Trace Gas sample introduction system following He
sparging and cryogenic trapping interfaced to an Isoprime isotope ratio mass spectrometer
(Elementar Americas, Inc., Mount Laurel, NJ). Analytical reproducibility for replicate samples
was 0.5 ‰ for bulk 15N and 18O, 0.75 ‰ for 15Nα 15Nβ, and 1.3 ‰ for SP.

Modeling
Rates of N removal were determined according to the isotope pairing technique (IPT).
The original IPT calculates the denitrification rate (including coupled nitrificationdenitrification) based on the production of 29N2 and 30N2 following addition of 15NO3- (Nielsen
1992). Subsequent revisions of the IPT include calculation of rates of anammox and N 2O
production by incorporating the analysis of 45N2O and 46N2O production (Risgaard-Petersen et al.

49

2003; Trimmer et al. 2006; Hsu and Kao 2013). Briefly, N 2 production by denitrification and
anammox, respectively, were calculated as:
(r14-N2O + 1) ¡ 2 ¡ r14-N2O ¡ P30

(7)

2 · r14-N2O · (P29 – 2 · r14-N2O · P30)

(8)

and

where r14-N2O is the ratio of 14N to 15N in N2O, P29 is the production of 29N2, and P30 is the
production of 30N2. N2O production in the absence of 15N amendment was calculated as:
r14-N2O ¡ (2 ¡ P46 + P45)

(9)

where P45 and P46 are the production of 45N2O and 46N2O, respectively. Coupled nitrificationdenitrification, a subset of total denitrification, was calculated as:
P14 · (1 – r14w/r14-N2O)

(10)

where P14 is the total production of 14N-N2 and 14N-N2O and r14w is the ratio of 14N to 15N in the
water overlying the core. NH4+ flux into or out of the sediment was calculated as the NH 4+
concentration in the overlying water at the end of the incubation minus the NH 4+ concentration in
the overlying water at the start of the incubation.

Statistical analysis
A two-way interaction effects ANOVA was used to test for differences in denitrification,
N2O production, and NH4+ flux among the different restoration scenarios and incubation
durations. Statistical models and posterior distributions were analyzed in a Bayesian framework
using R (version 3.2.4) and JAGS (Su and Yajima 2015). For each test, 1,200 Markov-chain
Monte Carlo iterations were run on three chains after a burn-in of 200 and thinned by two,
yielding 1,500 samples in the posterior distribution. Credible intervals (95 %) were constructed

50

around the mean of each group. The probability of a significant difference between groups was
determined by calculating the proportion of iterations in which the posterior estimate of one
group was smaller than another (Turner et al. 2010). The difference between groups was
considered significant when less than 5 % of the posterior estimates were smaller than those of
another group. Posterior estimates of 95 % credible intervals for denitrification and N2O
production generally encompassed measured rates, whereas posterior estimates for NH 4+ flux
tended to occupy a narrower range than the observed fluxes.

RESULTS
Anammox was not detected in wetland sediments. The absence of anammox activity was
confirmed by comparing the proportion of 15N in N2 with that of N2O (qN2 and qN2O,
respectively; Trimmer et al. 2006). Anammox activity is suggested if qN2 is lower than qN2O,
illustrating disproportionate production of 29N2 relative to 45N2O (Erler et al. 2008). In our study,
qN2 values were consistently higher than qN2O values, indicating production of 45N2O that
outweighed production of 29N2.
Denitrification was the only appreciable mechanism of N removal in Bear Creek wetland
sediments. Posterior estimates generated by statistical modeling indicated that denitrification
rates were lower in the HR and HR+D scenarios than in the Control scenario across both
incubation durations (Figure 8; Table 4). Incubation duration had a significant effect on the
denitrification rate in the HR+D scenario, where denitrification rates in the 6 h incubation were
significantly higher than in the 12 h incubation (Figure 8a, Table 4). Coupled nitrificationdenitrification (mean  SD), a subset of total denitrification, was significantly higher in the

51

Control scenario (60.4  10.9 µmol N m-2 h-1) than in the HR and HR+D scenarios (18.6  10.3
and 16.6  11.3 µmol N m-2 h-1, respectively; Figure 8b).

a

b

Denitrification rate (mmol N m-2 h-1)
0

20

60

80

100

0

40

60

80

Control

A
BC
BD
C
D

100

0.00

0.05

0.10

0.15

A

HR

B

HR+D

B

d

N2O production rate (mmol N m-2 h-1)

Control 6h
Control 12h
HR 6h
HR 12h
HR+D 6h
HR+D 12h

20

A

Control 6h
Control 12h
HR 6h
HR 12h
HR+D 6h
HR+D 12h

c

40

Coupled nit.-denit. rate (mmol N m-2 h-1)

0.20

NH4+ flux (mmol N m-2 h-1)
-400

Control 6h
Control 12h
HR 6h
HR 12h
HR+D 6h
HR+D 12h

A
B
B
AB
C
A

0

200

400

600

800

A
B
B
C
C
C

Figure 8. Posterior estimates of (a) denitrification, (b) coupled nitrification-denitrification, (c)
N2O production rates, and (d) NH4+ flux. Rates were measured from incubations of 6 and 12 h
duration under three restoration scenarios: (1) no restoration (control), (2) simulated hydrologic
reconnection (HR), and (3) simulated hydrologic reconnection and dredging (HR+D). Open
circles represent measured rates and closed circles represent the mean of the posterior estimate
for each treatment-duration combination. Lines represent the 95 % credible interval around each
posterior mean estimate. In (d), positive values indicate a net release of NH4+ to the water
column from the sediment, and negative values indicate a net uptake of NH 4+ by the sediment
from the water column. Note the different scales for the axes among panels. Statistical groupings
of posterior estimates at the p < 0.05 level are indicated by letters.
The N2O production rate was below 0.10 Âľmol N m-2 h-1 for all treatments and incubation
durations except the HR+D scenario in the 6 h incubation, which was significantly greater than
any other treatment and duration (Figure 8c, Table 4). As a percentage of total denitrification,
N2O production was similar between incubation durations within a single restoration scenario. In

52

the control, HR, and HR+D scenarios, N2O made up (mean  SD) 0.06  0.05 %, 0.19  0.04 %,
and 0.21  0.08 % of total denitrification, respectively.
NH4+ flux data indicated a net release (positive flux) or net uptake (negative flux) of
NH4+ from the sediment depending on the restoration scenario. The NH 4+ fluxes for the Control
scenario for both incubation durations and for the HR scenario for the 6 h duration were positive.
The NH4+ fluxes for the HR scenario for the 12 h incubation durations and for the HR+D
scenario for both incubation scenarios were negative and were significantly different from the
positive fluxes but not from one another. In terms of magnitude, NH4+ flux was the dominant N
cycling process of those studied (Figure 8d, Table 4).

Table 4. Posterior estimates of denitrification, N2O production, and NH4+ flux under restoration
scenarios. HR = hydrologic reconnection, HR+D = hydrologic reconnection and dredging, CI =
credible interval.
Denitrification
(Âľmol N m-2 h-1)

NH4+ flux
(Âľmol N m-2 h-1)

N2O production
(Âľmol N m-2 h-1)

5% CI

Mean

95% CI

5% CI

Mean

95% CI

5% CI

Mean

95% CI

59.1
67.0

78.2
87.2

93.1
102.0

-0.008
0.034

0.029
0.070

0.065
0.105

59.8
239.4

168.7
351.2

275.0
456.1

6h
12 h

24.7
7.0

40.9
24.8

56.0
41.1

0.045
0.017

0.082
0.052

0.118
0.090

289.6
-185.93

396.0
-77.4

505.6
31.7

HR+D
6h
12 h

37.9
-2.4

56.0
14.7

71.5
30.1

0.119
-0.010

0.154
0.024

0.190
0.060

-286.9
-173.9

-180.3
-66.0

-76.4
44.1

Control
6h
12 h
HR

DISCUSSION
This study evaluated N cycling pathways in intact sediment cores from a restored wetland
and the potential response of restoration activities on these processes. By analyzing N cycling in

53

experiments that represented different restoration scenarios, we modeled the potential effects of
restoration on N cycling in the context of watershed nutrient management. N cycling was
evaluated with a Bayesian analysis, which offers the advantage of providing probabilistic
estimates of environmental processes by quantifying the uncertainty associated with modeled
parameters (KĂŠry 2010).
We demonstrate that the west pond of the BLWRA, like other wetland systems, has a
substantial capacity to remove N from the watershed via N 2 production (Saunders and Kalff
2001; Seitzinger et al. 2006). While the IPT model has the capacity to quantify both
denitrification and anammox, denitrification was the only N 2 production process detected in this
system. The absence of anammox activity is not surprising, as denitrification tends to
outcompete anammox in organic-rich sediments (Ahn 2006; Algar and Vallino 2014). Indeed,
sediment in this location in the BLWRA contains a high amount of organic matter (~20 % by
mass), and this organic content is consistent from 0 to 43 cm depth (Steinman and Ogdahl 2016).
Further, while anammox is a prominent N removal process in marine environments, it has been
observed at a much lower frequency in freshwaters (Burgin and Hamilton 2007; Hu et al. 2011).
Denitrification rates differed significantly among experimental treatments, which
included introduction of water from Bear Creek and removal of sediments from intact cores.
Across treatments, posterior estimates of denitrification rates varied over three orders of
magnitude, indicating strong responsiveness of denitrification activity to hydrologic changes that
is common to other wetland systems (ArdĂłn et al. 2010a). When sediment cores were exposed to
water from Bear Creek under simulated hydrologic reconnection (HR and HR+D treatments),
there was a two-fold decrease in the mean denitrification rate relative to the control (Figure 8a,
Table 4). This decreased rate could be due to a difference in NO 3- supply via two mechanisms.

54

First, the concentration of NO3- in Bear Creek water was ~½ of that present in the wetland water
column (Table 3). If these results are indicative of the BLWRA, they suggest that hydrologic
reconnection would decrease the NO3- supply to denitrifiers in the sediment, resulting in lower
rates of denitrification (Tiedje et al. 1982). Second, the NH 4+ concentration in the stream was
~20 times lower than that of the wetland, and the rate of coupled nitrification-denitrification was
significantly reduced under hydrologic reconnection scenarios (Figure 8b). This suggests that a
dramatic decrease in NH4+ supply following hydrologic reconnection could limit nitrification
activity, which in turn could limit the supply of NO 3- to denitrification. Reconnection therefore
may result in a decoupling of nitrification and denitrification, reducing the capacity for
denitrification to remove N from the wetland.
Incubation length had a significant effect on denitrification and N 2O production under
simulated hydrologic reconnection with dredging. Denitrification and N2O production rates were
greater for the 6 h duration than for the 12 h duration in the HR+D scenario. While the removal
of wetland sediments throughout the range of depths encompassed by dredging activities (5 cm
this study, ~0.5 m expected) is not expected to alter the quantity of organic matter in the surface
sediment (Steinman and Ogdahl 2016), physical disturbance and reoxygenation of the sediment
has been shown liberate labile pools of particulate and dissolved carbon and N (Aller 1994). The
stimulation of denitrification and N2O production shortly following dredging could represent
rapid utilization of newly exposed substrates followed by a return to more modest rates as ratelimiting substrates are consumed. In the context of restoration, we may expect dredging to result
in a temporary spike in N2 and N2O production prior to a return to a quasi-equilibrium
characterized by more moderate rates (Morse et al. 2012).

55

In general, the BLWRA displayed N2O production typical of wetland environments
(Huttunen et al. 2003; Morse et al. 2012; Audet et al. 2014), and changes in N2O production
were suggestive of changes in substrate availability in the context of restoration activities. For a
wetland that has historically been subject to high levels of nutrient loading, denitrification was
remarkably efficient; the proportion of N lost as N 2O (0.06 to 0.21 %) is typical of benthic
sediments that are not heavily nutrient-impacted (Seitzinger and Kroeze 1998). While the
N2O:N2 ratio remained comparatively low across all restoration scenarios, a threefold increase in
the proportion of N lost as N2O and a spike in N2O production in the HR+D scenario is indicative
of a decrease in the degree to which denitrification proceeds to completion. This observation
could relate to a change in the energetic yield of N 2O reduction compared to other steps in
denitrification following sediment disturbance (Tiedje et al. 1982), which could result in a brief
yet substantial increase in greenhouse gas production by the wetland following restoration.
Restoration scenarios resulted in a shift from net NH 4+ release to net NH4+ uptake from
BLWRA sediments, providing insight into the behavior of nitrification in this system. In the
absence of hydrologic reconnection, NH4+ was released from the sediment and high NH4+
concentrations were present in the water, indicating there was ample NH 4+ supply provided by
mineralization to fuel nitrification. With a non-limiting supply of NH4+, nitrification supplied a
steady source of NO3- for denitrification as indicated by high coupled nitrification-denitrification
rates in the Control scenario. Under conditions of hydrologic reconnection, however, the supply
of NH4+ from the overlying water decreased by 20-fold, and the sediments shifted from net NH4+
release to net NH4+ uptake from the 6 h to the 12 h incubation duration in the HR scenario and
displayed net uptake of NH4+ throughout both incubation durations in the HR+D scenario. This
shift suggests that demand for NH4+ by nitrification exceeded its supply, resulting in less NO3-

56

generated by nitrification. Coupled nitrification-denitrification was thus hindered and total
denitrification rates decreased.
To estimate the possible changes in the NO 3- and NH4+ load to Bear Lake that would
result from reconnecting the wetland to Bear Creek, we modeled several scenarios relating the N
cycling capacity of the wetland sediment to the N load from Bear Creek. These scenarios are
hypothetical, as our experiments were conducted prior to current restoration efforts. However,
the modeling does offer opportunities to evaluate the impacts of management decisions, develop
a baseline for expectations about downstream nutrient delivery, and monitor progress in the
context of expected impacts (Maack 2001). We evaluated NO3- mitigation capacity as the
proportion of stream nitrate removed by the wetland (NO 3-rem):
NO−
3−rem =

denitrification rate ∙ wetland sediment surface area
stream NO3 − ∙ stream discharge ∙ fraction
(11)

where the denitrification rate under restoration scenarios (Âľmol m-2 h-1) was incorporated as the
posterior estimate of the mean and 95 % credible interval, wetland sediment surface area (m 2)
and measured stream NO3- concentration (Âľmol L-1) were incorporated as constant parameters,
and stream discharge (L h-1) was incorporated as the mean discharge rate of Bear Creek (Cadmus
Group and AWRI 2007). A “fraction” term was added to represent the portion of Bear Creek
water that will enter the wetland following hydrologic reconnection, which is currently unknown
but will be engineered to maximize hydrologic exchange. A range of fractions between 0.1 and 1
were thus incorporated to represent the maximum probable range of hydrologic exchange
between Bear Creek and the wetland. NH4+ retention capacity was modeled similarly by
replacing the denitrification rate and stream NO 3- concentration with NH4+ flux rate and stream
NH4+ concentration.

57

The BLWRA has the potential to intercept a substantial amount of NO 3- and NH4+ from
Bear Creek. Wetland denitrification can remove up to 10 % of stream NO3- (Figure 9a),
providing a potential water quality benefit to downstream Bear and Muskegon Lakes given
sufficient diversion of Bear Creek water through the wetland. Although several studies have
reported the capacity of riparian wetlands to remove up to 100 % of NO3- inputs (Fennessy and
Cronk 1997; ArdĂłn et al. 2010a), clearly this assumption cannot be made across all wetland
restoration settings. Although BLWRA denitrification rates are typical of many wetland
environments, the NO3- load and discharge from Bear Creek are sufficiently high to limit the
capacity for denitrifying bacteria to fully consume stream NO3-. An a priori assessment of the
wetland N removal capacity, as was conducted here, can therefore be helpful in establishing
expectations for the potential impact of restoration activities on water quality. Three out of four
restoration scenarios project a net uptake of NH 4+, translating to a capacity of the wetland to
retain all incoming NH4+ from Bear Creek, potentially with an initial pulse of NH4+ released
from the wetland following hydrologic reconnection (Figure 9b). These results are consistent
with previous studies, which suggest is not uncommon for wetlands to shift back and forth
between the retention and release of NH4+ following restoration (ArdĂłn et al. 2010a).
Restoration activities in the BLWRA are predicted to result in major changes in N
cycling, illustrating a probable shift in the biogeochemistry of this watershed. However,
restoration activities must be evaluated in the context of P cycling in addition to N cycling.
Analysis by Smit and Steinman (2015) indicates that hydrologic reconnection could result in
enhanced release of P to downstream Bear Lake (20 to 81 Âľmol P m-2 h-1 across treatments),
while dredging activities could substantially limit this release owing to the exposure of sediments
with higher P sorption capacity. While it appears that P loading will likely remain an ongoing

58

issue for the eutrophication of Bear Lake, restoration activities are predicted to decrease
downstream N loading from the BLWRA. Our analyses indicate that N retention and removal in
this wetland depend on nitrification; hydrologic reconnection could reduce the supply of NH 4+
available for nitrification, resulting in a shift from net NH 4+ release to uptake and a decoupling of
nitrification and denitrification. Despite the decreased rate of permanent N removal from the
wetland under restoration scenarios, the total projected amount of N removed from the wetland
via denitrification is roughly equivalent to the total projected amount of P released. Furthermore,
the magnitude of NH4+ uptake under restoration scenarios is tenfold that of denitrification and P
release, demonstrating a high capacity for the wetland to reduce downstream N loading. The
marked shift in N cycling displayed here indicates that nitrification is a strong regulator of
permanent N removal that can be heavily influenced by NH 4+ availability.

a

% stream NO3- removed via denitrification
0

2

4

6

8

% stream NH4+ retained or released

b

10

-1000

HR 6h

HR 6h

HR 12h

HR 12h

HR+D 6h

HR+D 6h

HR+D 12h

HR+D 12h

-500

0

500

1000

Figure 9. (a) NO3- and (b) NH4+ mitigation capacity of the wetland with respect to the potential
load delivered by Bear Creek. Denitrification and NH4+ flux were measured from incubations of
6 and 12 h duration under restoration scenarios of simulated hydrologic reconnection (HR) and
simulated hydrologic reconnection with dredging (HR+D). Closed circles represent the mean of
the posterior estimate for each treatment-duration combination. Lines represent the 95 % credible
interval around each posterior mean estimate. Fractions of Bear Creek water reaching the
wetland sediment (0.1, 0.5, and 1.0) are represented by light gray, dark gray, and black lines,
respectively.

59

Changes in wetland N cycling carry implications for downstream receiving water bodies.
In the case of Bear Lake, as well as other lakes that are either limited by N or co-limited by P and
N (Paerl et al. 2016), reducing N loads through reconnection may drive the lake to stronger N
limitation, which may contribute to larger blooms of toxic cyanobacteria (Chaffin and
Bridgeman 2014). However, because reduced forms of N are needed for toxin production genes
to be expressed (Kuniyoshi et al. 2011), N cycling in the BLWRA may actually inhibit bloom
toxicity, even as the reduction in N concentrations stimulates bloom formation. These results, in
combination with those of Smit and Steinman (2015) and Steinman and Ogdahl (2016) suggest
that nutrient management must consider both N and P, not just a single nutrient (Conley et al.
2009; Paerl et al. 2016). The current study provides a mechanistic understanding of N mitigation
in a wetland undergoing restoration, and reveals that future studies need to investigate not only
denitrification but also the processing of NH 4+ and its effects on nitrification.

60

LITERATURE CITED

61

LITERATURE CITED

Ahn Y-H (2006) Sustainable nitrogen elimination biotechnologies: A review. Process Biochem
41:1709-1721. doi:10.1016/j.procbio.2006.03.033
Aldous A, McCormick C, Ferguson C, Graham S, Craft C (2005) Hydrologic regime controls
soil phosphorus fluxes in restoration and undisturbed wetlands. Restor Ecol 13:341-347.
Algar CK, Vallino JJ (2014) Predicting microbial nitrate reduction pathways in coastal
sediments. Aquat Microb Ecol 71:223-238. doi:10.3354/ame01678
Aller RC (1994) Bioturbation and remineralization of sedimentary organic matter: Effects of
redox oscillation. Chem Geol 114:331-345.
APHA (2005) Standard methods for the examination of water and wastewater. American Public
Health Association, Washington, D.C.
ArdĂłn M, Morse JL, Doyle MW, Bernhardt ES (2010a) The water quality consequences of
restoring wetland hydrology to a large agricultural watershed in the southeastern coastal plain.
Ecosystems 13:1060-1078. doi:10.1007/s10021-010-9374-x
ArdĂłn M, Montanari S, Morse JL, Doyle MW, Bernhardt ES (2010b) Phosphorus export from a
restored wetland ecosystem in response to natural and experimental hydrologic fluctuations. J
Geophys Res 115:G04031. doi:10.1029/2009JG001169
Audet J, Hoffman CC, Andersen PM, Baattrup-Pedersen A, Johansen JR, Larsen SE, Kjaergaard
C, Elsgaard L (2014) Nitrous oxide fluxes in undisturbed riparian wetlands located in
agricultural catchments: Emission, uptake and controlling factors. Soil Biol Biochem 68:291299. doi:10.1016/j.soilbio.2013.10.011
Bunn SE, Arthington AH (2002) Basic principles and ecological consequences of altered flow
regimes for aquatic biodiversity. Environ Manage 30(4):492-507.
doi:10.1007/s00267-002-2737-0
Burgin AJ, Hamilton SK (2007) Have we overemphasized the role of denitrification in aquatic
ecosystems? A review of nitrate removal pathways. Front Ecol Environ 5(2):89-96.
Cadmus Group, AWRI (2007) Bear Lake Nutrient Study; U.S. Environmental Protection
Agency, Contract No. 68-C-02-109.
Carpenter SR, Caraco NF, Correll DL, Howarth RW, Sharpley AN, Smith VH (1998) Nonpoint
pollution of surface waters with phosphorus and nitrogen. Ecol Appl 8(3):559-568.

62

Chaffin JD, Bridgeman TB (2014) Organic and inorganic nitrogen utilization by nitrogenstressed cyanobacteria during bloom conditions. J Appl Phycol 26:299-309.
doi:10.1007/s10811-013-0118-0
Conley DJ, Paerl HW, Howarth RW, Boesch DF, Seitzinger SP, Havens KE, Lancelot C, Likens
GE (2009) Controlling eutrophication. Science 323:1014−1015.
Dahl TE (2000) Status and trends of wetlands in the conterminous United States 1986-1997. U.S.
Department of the Interior, Fish and Wildlife Service; Washington, D.C.
Dalsgaard T, Nielsen LP, Brotas V, Viaroli P, Underwood G, Nedwell D, Sundbäck K, Rysgaard
S, Miles A, Bartoli M, Dong L, Thornton DCO, Ossosen LDM, Castaldelli G, Risgaard-Petersen
N (2000) Protocol handbook for NICE- Nitrogen cycling in estuaries. National Environmental
Research Institute, Silkeborg, Denmark.
Erler DV, Eyre BD, Davidson L (2008) The contribution of anammox and denitrification to
sediment N2 production in a surface flow constructed wetland. Environ. Sci. Technol. 42:91449150. doi:10.1021/es801175t
Eyre BD, Rysgaard S, Dalsgaard T, Christensen PB (2002) Comparison of isotope pairing and
N2:Ar methods for measuring sediment denitrification: Assumptions, modifications and
implications. Estuaries 25(6A):1077-1087.
Fennessy MS, Cronk JK (1997) The effectiveness and restoration potential of riparian ecotones
for the management of nonpoint source pollution, particularly nitrate. Crit Rev Env Sci Tec
27(4):285-317. doi:10.1080/10643389709388502
Hamilton SK, Ostrom NE (2007) Measurements of the stable isotope ratio of dissolved N 2 in 15N
tracer experiments. Limnol Oceanogr Methods 5:233-240.
Hsu T-C, Kao S-J (2013) Technical note: Simultaneous measurement of sedimentary N 2 and
N2O production and new 15N isotope pairing technique. Biogeosciences 10:7847-7862.
doi:10.5194/bg-10-7847-2013
Hu B-I, Shen L, Xu X, Zheng P (2011) Anaerobic ammonium oxidation (anammox) in different
natural ecosystems. Biochem Soc Trans 39:1811-1816. doi:10.1042/BST20110711
Humbert S, Zopfi J, Tarnawski S-E (2012) Abundance of anammox bacteria in different wetland
soils. Environ. Microbiol. Rep. 4(5):484-490. doi:10.1111/j.1758-2229.2012.00347.x
Huttunen JT, Juutinen S, Alm J, Larmola T, Hammar T, Silvola J, Martikainen PJ (2003) Nitrous
oxide flux to the atmosphere from the littoral zone of a boreal lake. J Geophys Res
108(D14):4421. doi:10.1029/2002JD002989

63

IPCC (2007) Climate Change 2007: The Physical Science Basis. In: Solomon S et al (eds)
Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental
Panel on Climate Change. Cambridge University Press, New York.
Keddy PA (2000) Wetland Ecology: Principles and Conservation, 1st ed.. Cambridge University
Press, Cambridge, U.K.
KĂŠry M (2010) Introduction to WinBUGS for Ecologists: Bayesian Approach to Regression,
ANOVA, Mixed Models and Related Analyses. Academic Press, Cambridge, MA, U.S.
Knowles R (1982) Denitrification. Microbiological Rev 46(1):43-70
Kuniyoshi TM, Gonzalez A, Lopez-Gomollon S, Valladares A, Bes MT, Fillat MF, Peleato ML
(2011) 2-oxoglutarate enhances NtcA binding activity to promoter regions of the microcystin
synthesis gene cluster. FEBS Lett 585(24):3921−3926.
Lindenberg MK, Wood TM (2006) Water quality of a drained wetland, Caledonia Marsh on
upper Klamath Lake, Oregon, after flooding in 2006. U. S. Geological Survey, Oregon.
Maack JN (2001) Scenario analysis: a tool for task managers. In: Krueger RA, Casey MA,
Donner J, Kirsch S, Maack JN (eds) Social Analysis Selected Tools and Techniques. The World
Bank, Washington, D.C. pp 62-86.
Morse JL, ArdĂłn M, Bernhardt ES (2012) Greenhouse gas fluxes in southeastern U.S. coastal
plain wetlands under contrasting land uses. Ecol Appl 22(1):264-280.
Newman S, Pietro K (2001) Phosphorus storage and release in response to flooding: implications
for Everglades stormwater treatment areas. Ecol Engineer 18:23-38.
Nielsen LP (1992) Denitrification in sediment determined from nitrogen isotope pairing. FEMS
Microbiol Ecol 86:357-362.
Paerl HW, Scott JT, McCarthy MJ, Newell SE, Gardner WS, Havens KE, Hoffman DK,
Wilhelm SW, Wurtsbaugh WA (2016) It takes two to tango: When and where dual nutrient (N &
P) reductions are needed to protect lakes and downstream ecosystems. Environ Sci Technol
50(20):10805-10813. doi:10.1021/acs.est.6b02575
Pant HK, Reddy KR (2003) Potential internal loading of phosphorus in a wetland constructed in
agricultural land. Water Res 37:965–72.
Reddy KR, Kadlec RH, Flaig E, Gale PM (1999) Phosphorus retention in streams and wetlands:
A review. Crit Rev Env Sci Tec 29(1):83-146.
Risgaard-Petersen N, Nielsen LP, Rysgaard S, Dalsgaard T, Meyer RL (2003) Application of the
isotope pairing technique in sediments where anammox and denitrification coexist. Limnol
Oceanogr Methods 1:63-73.

64

Saunders DL, Kalff J (2001) Nitrogen retention in wetlands, lakes, and rivers. Hydrobiologia
443:205-212. doi:10.1023/A:1017506914063
Schindler DW (1974) Eutrophication and recovery in experimental lakes: implications for lake
management. Science 184:897-899.
Scott JT, McCarthy MJ, Gardner WS, Doyle RD (2008) Denitrification, dissimilatory nitrate
reduction to ammonium, and nitrogen fixation along a nitrate concentration gradient in a created
freshwater wetland. Biogeochemistry 87:99-111. doi:10.1007/s10533-007-9171-6
Seitzinger SP, Kroeze C (1998) Global distribution of nitrous oxide production and N inputs in
freshwater and coastal marine ecosystems. Global Biogeochem Cy 12(1):93-113.
Seitzinger S, Harrison JA, BĂśhlke JK, Bouwman AF, Lowrance R, Peterson B, Tobias C, Van
Drecht G (2006) Denitrification across landscapes and waterscapes: A synthesis. Ecol Appl
16(6):2064-2090.
Smit JT, Steinman AD (2015) Wetland sediment phosphorus flux in response to proposed
hydrologic reconnection and warming. Wetlands 35:655-665. doi:10.1007/s13157-015-0655-1
Steffen MM, Belisle BS, Watson SB, Boyer GL, Wilhelm SW (2014) Status, causes and controls
of cyanobacterial blooms in Lake Erie. J Great Lakes Res 40:215-225.
doi:10.1016/j.jglr.2013.12.012
Steinman AD, Ogdahl ME (2011) Does converting agricultural fields to wetlands retain or
release phosphorus? J N Am Benthol Soc 30:820-830.
Steinman AD, Ogdahl ME (2016) From wetland to farm and back again: phosphorus dynamics
of a proposed restoration project. Environ Sci Pollut Res 23(22):22596-22605.
doi:10.1007/s11356-016-7485-4
Su Y-S, Yajima M (2015) R2jags: A Package for Running jags from R. R package version 0.0501. http://CRAN.R-project.org/package=R2jags
Sun WG, Sun ZG, Gan ZT, Sun WL, Wang W (2014) Contribution of different processes in
wetland soil N2O production in different restoration phases of the Yellow River estuary, China.
Huanjing Kexue 35(8):3110-3119.
Thamdrup B, Dalsgaard T (2002) Production of N2 through anaerobic ammonium oxidation
coupled to nitrate reduction in marine sediments. Appl Environ Microbiol 68:1312-1318.
doi:10.1128/AEM.68.3.1312-1318.2002
Tiedje JM, Sexstone AJ, Myrold DD, Robinson JA (1982) Denitrification: Ecological niches,
competition and survival. Antonie van Leeuwenhoek 48:569-583.

65

Trimmer M, Risgaard-Petersen N, Nicholls JC, EngstrĂśm P (2006) Direct measurement of
anaerobic ammonium oxidation (anammox) and denitrification in intact sediment cores. Mar
Ecol Prog Ser 326:37-47.
Turner TF, Collyer ML, Krabbenhoft TJ (2010) A general hypothesis-testing framework for
stable isotope ratios in ecological studies. Ecology 91(8):2227-2233.
[USDA] U. S. Department of Agriculture (2012) Restoring America’s Wetlands : A Private
Lands Conservation Success Story. U.S. Department of Agriculture, Washington, D.C.
[USEPA] U.S. Environmental Protection Agency (2013) Great Lakes Areas of Concern:
Muskegon Lake. U.S. Environmental Protection Agency, Washington, D.C.
van de Graaf AA, Mulder A, de Bruijn P, Jetten MSM, Robertson LA, Kuenen JG (1995)
Anaerobic oxidation of ammonium is a biologically mediated process. Appl Environ Microbiol
61:1246-1251.
Waki M, Yasuda T, Suzuki K, Komada M, Abe K (2015) Distribution of anammox bacteria in a
free-water-surface constructed wetland with wild rice (Zizania latifolia). Ecol Eng 81:165-172.
doi:10.1016/j.ecoleng.2015.04.005
Welti N, Bondar-Kunze E, Mair M, Bonin P, Wanek W, Pinay G, Hein T (2012)Mimicking
floodplain reconnection and disconnection using 15N mesocosm incubations. Biogeosciences
9:4263-4278. doi:10.5194/bg-9-4263-2012
Zelder JB, Kercher S (2005) Wetland resources: Status, trends, ecosystem services, and
restorability. Annu Rev Environ Resour 30:39-74.
doi:10.1146/annurev.energy.30.050504.144248
Zhu G, Wang S, Feng X, Fan G, Jetten MSM, Yin C (2011) Anammox bacterial abundance,
biodiversity and activity in a constructed wetland. Environ Sci Tech 45:9951-9958.
doi:10.1021/es202183w
Zhu G, Wang S, Wang W, Wang Y, Zhou L, Jian B, Op den Camp HJM, Risgaard-Petersen N,
Schwark L, Peng Y, Hefting MM, Jetten MSM, Yin C (2013) Hotspots of anaerobic ammonium
oxidation at land-freshwater interfaces. Nature Geoscience 6:103-107. doi:10.1038/NGEO1683

66

CHAPTER 3
UNEXPECTEDLY HIGH DEGREE OF ANAMMOX AND DNRA IN SEAGRASS
SEDIMENTS: DESCRIPTION AND APPLICATION OF A REVISED ISOTOPE PAIRING
TECHNIQUE
ABSTRACT
Understanding the magnitude of nitrogen (N) loss and recycling pathways is crucial for
coastal N management efforts. However, quantification of denitrification and anammox by a
widely-used method, the isotope pairing technique, is challenged when dissimilatory NO3reduction to NH4+ (DNRA) occurs. In this study, we describe a revised isotope pairing technique
that accounts for the influence of DNRA on NO 3- reduction (R-IPT-DNRA). The new calculation
procedure improves on previous techniques by (1) accounting for N 2O production, (2)
distinguishing canonical anammox from coupled DNRA-anammox, and (3) including the
production of 30N2 by anammox in the quantification of DNRA. We apply this technique to
simultaneously quantify rates of anammox, denitrification, and DNRA in intact sediments
adjacent to a seagrass bed in subtropical Australia. The effect of organic carbon lability on NO 3reduction was also addressed by adding detrital sources with differing C:N (phytoplankton- or
seagrass-derived). DNRA was the predominant pathway, contributing 49-74% of total NO3reduction (mean 0.42 Âľmol N m-2 h-1). In this high C:N system, DNRA outcompetes
denitrification for NO3-, functioning to recycle rather than remove N. Anammox exceeded
denitrification (mean 0.18 and 0.04 Âľmol N m -2 h-1, respectively) and accounted for 64-86% of N
loss, a rare high percentage in shallow coastal environments. Owing to low denitrification
activity, N2O production was ~100-fold lower than in other coastal sediments (mean 7.7 nmol N
m-2 h-1). All NO3- reduction pathways were stimulated by seagrass detritus but not by
phytoplankton detritus, suggesting this microbial community is adapted to process organic matter

67

that is typically encountered. Our new model, the R-IPT-DNRA, is widely applicable in other
environments where the characterization of co-existing NO3- reduction pathways is desirable.

INTRODUCTION
Coastal ecosystems are important regions of nitrogen (N) loss, the conversion of
bioavailable N to gaseous N, particularly as N loading to marine environments has intensified in
response to agricultural activities (Schlesinger 2009). While it is known that coastal systems
remove significant quantities of inorganic N from watershed loading (e.g., Seitzinger et al. 2006;
Murray et al. 2015; Eyre et al. 2016), the magnitudes, pathways, and environmental controls on
N loss are not completely resolved. The potential for natural microbial processes to mitigate the
detrimental effects of N loading such as eutrophication and harmful algal blooms will depend on
a more complete understanding of N loss mechanisms in coastal environments (Howarth 2008).
Denitrification via the respiratory reduction of NO3- to N2 is generally recognized as the
predominant mechanism for N loss in coastal marine systems (Seitzinger et al. 2006). This
heterotrophic process occurs in hypoxic to anoxic environments and requires a supply of NO 3and organic carbon (Knowles 1982). A variable but usually small fraction of reduced NO 3- is not
fully reduced to N2 and is released as N2O, a potent greenhouse gas with nearly 300 times the
global warming potential of CO2 over a 100-year timespan (IPCC 2007). The factors regulating
the N2O:N2 production ratio include O2 concentration, carbon availability, and substrate
concentration (Betlach and Tiedje 1981; Knowles 1982; Davidson and Verchot 2000; Hwang et
al. 2006), although the relative importance of these factors in concert is not well understood
(Murray et al. 2015).

68

An alternate N loss pathway, anaerobic ammonium oxidation (anammox), was
discovered only two decades ago (van de Graaf et al. 1995) and has since been recognized as a
globally important process that rivals denitrification in its removal of N from offshore marine
systems. A chemoautotrophic process, anammox produces N 2 via the reaction of NO2- and NH4+
with no direct organic carbon demand. Anammox does not produce an appreciable amount of
N2O (Kartal et al. 2012). Anammox is now estimated to contribute 1/3 to 1/2 of global marine N
loss from sediments and anoxic water columns (Dalsgaard et al. 2005). However, owing to its
recent discovery, the environmental controls of anammox in a variety of natural environments
are not well constrained, particularly as they relate to the interplay between anammox and
denitrification.
The isotope pairing technique (IPT) is a widely used technique for measuring N loss rates
(i.e., those that would naturally occur in the absence of 15N addition) from aquatic sediments.
The original isotope pairing technique quantifies N 2 production via denitrification (Fig. 10a;
Nielsen 1992). Following the discovery of anammox as an additional N2 production pathway, a
revision to the isotope pairing technique (R-IPT) that distinguished N2 production by
denitrification and anammox was developed (Fig. 10a; Risgaard-Petersen et al. 2003).
Contemporary applications of the isotope pairing technique utilize the R-IPT rather than the IPT.
An addition to the isotope pairing technique has enabled the quantification of N 2O production
(IPTanaN2O; Fig. 10b; Hsu and Kao 2013). In practice, the isotope pairing technique is carried out
by adding 15NO3- to the ambient NO3- pool of a sediment slurry, water sample, or intact sediment
core. Supposing random isotope pairing, rates of denitrification and anammox are then calculated
based on the 14N/15N ratio in the NO3- pool undergoing reduction (r14) and the isotopic
composition of nitrogenous end products (N2 and N2O).

69

a IPT (shaded; Nielsen, 1992) and
R-IPT (Risgaard-Petersen et al., 2003)

b IPTanaN2O (Hsu and Kao, 2013)

N2

N2

N2
N2 O

Anammox

N2

NO
NH4+

NO2-

N2
N2 O

Anammox
NO
NO2-

Canonical Anammox

N2

d R-IPT-DNRA (this study)
N2

Coupled DNRA-Anammox

c IPT including DNRA (Song et al., 2016)
N2

NO3-

NO2NO3-

NO3-

NH4+

N2 O

Anammox

NO
NH4+

N2 O

N2

N2

N2 O

N2

N2
N2 O

Anammox
NO
NH4+

NO2NO3-

Figure 10. (a) The original IPT model (shaded; Nielsen 1992) that quantifies N2 production via
denitrification and a subsequent revision to the IPT (R-IPT; Risgaard-Petersen et al., 2003)
that includes the production of N2 via anammox as well as denitrification. (b) The
IPTanaN2O model (Hsu and Kao 2013), which adds the capacity to quantify N2O production via
denitrification to the R-IPT. (c) A revision to the IPT model (Song et al. 2016)
which enables quantification of the production of NH4+ via DNRA to the R-IPT. (d) The R-IPTDNRA model (this study), which adds the ability to quantify N2O production via denitrification
and NH4+ production via DNRA to the R-IPT while simultaneously distinguishing canonical
anammox and coupled DNRA-anammox.

70

In addition to N loss via denitrification and anammox in coastal sediments, N recycling
processes can compete with N loss pathways for inorganic N substrates. The dissimilatory
reduction of NO3- to NH4+ (DNRA) has a similar energetic yield as denitrification and also
occurs under low O2 conditions (Tiedje et al. 1982). DNRA makes up a significant proportion of
total NO3- reduction in coastal sediments (Burgin and Hamilton 2007; Dong et al. 2011; Giblin et
al. 2013) and tends to outcompete denitrification in high carbon, low NO 3- systems (Algar and
Vallino 2014; Kraft et al. 2014; Brin et al. 2015; Hardison et al. 2015; van den Berg et al. 2015).
In coastal marine systems where DNRA dominates, N is retained rather than lost, which could
result in sustained or enhanced eutrophication (Hardison et al. 2015).
A key assumption of the isotope pairing technique is that NO3- is the only appreciable
source of 15N, but this assumption is violated when 15NH4+ is produced by DNRA. The coupling
of DNRA-anammox, in this case, results in production of 30N2 that is incorrectly attributed to
denitrification, resulting in an underestimate of anammox and an overestimate of denitrification.
Song et al. (2016) have recently proposed a revision to the isotope pairing technique that
simultaneously quantifies the NO3- reduction processes of denitrification, anammox, and DNRA,
thus resolving the issues from the isotope pairing technique that occur when DNRA is active
(Fig. 10c). This approach hinges on the evaluation of the ratio of 14N/15N in the NH4+ pool (r14a).
Further improvements to these models can be made by (1) quantifying N2O production, (2)
distinguishing coupled DNRA-anammox from uncoupled (canonical) anammox, and (3)
including the production of 30N2 by anammox in the quantification of DNRA (Fig. 10d).
Simultaneous quantification of denitrification, anammox, N2O production, and DNRA
will enable an examination of the competitive interactions among these processes under varying
environmental conditions. For instance, organic carbon has been identified as a predominant

71

control on the relative importance of NO3- reduction pathways in marine systems (Algar and
Vallino 2014). It is expected that anammox should dominate NO 3- reduction at low C:N ratios,
denitrification at intermediate C:N ratios, and DNRA at high C:N ratios (Thamdrup and
Dalsgaard 2002; Dalsgaard et al. 2005; Algar and Vallino 2014; Hardison et al. 2015). However,
this relationship is not definitive (Trimmer et al. 2003; Trimmer and Nicholls 2009), and few
studies have simultaneously measured these processes (e.g., Erler et al. 2013), particularly with
regard to carbon quality (i.e., C:N ratio and organic carbon lability).
There have been a number of N cycling studies in subtropical seagrass environments
(e.g., Eyre et al. 2011a; 2011b; 2013; 2016), but to our knowledge, there have been no published
studies on anammox rates or N2O production in intact sediments from seagrass ecosystems.
Further, seagrass ecosystems are exemplified by a largely refractory pool of organic carbon
(Khan et al. 2015) that could serve as the basis of an investigation of the effect of carbon quality
on N loss and recycling processes. While microbial activity is expected to be dampened in bare
sediments compared to vegetated sediments, microbial communities tend to have similar
taxonomic composition and metabolism in both settings (Boschker et al. 2000; Jones et al. 2003;
Holmer et al. 2004). In addition, sediments adjacent to seagrass stands tend to have stable redox
conditions (EnrĂ­quez et al. 2001), making them an ideal environment in which to apply a new
revision to the IPT that includes denitrification, anammox, N2O production, and DNRA.
This study has three objectives: (1) revise the isotope pairing technique to simultaneously
quantify rates of anammox, denitrification, N2O production, and DNRA; (2) generate the first
estimates of anammox and N2O production in intact sediments in a seagrass ecosystem; and (3)
assess the competitive interactions among DNRA, denitrification, and anammox under differing
organic carbon quality.

72

METHODS
Study site
This study was conducted in Shaws Bay, New South Wales, Australia (28.8675 S,
153.5871 E), a subtropical estuarine system. Throughout the course of the study, non-vegetated
(i.e., no plant or root biomass) sediments were collected from in and around the edge of a
seagrass bed (consisting primarily of Zostera muelleri) that covers approximately 0.02 km2 of the
eastern part of the bay (Ballina Shire Council 2000). Sediment and water collections took place
in June and July 2014.

Sediment slurry incubations
The presence of anammox and denitrification activity in Shaws Bay sediments was evaluated by
sediment slurry assays following the approach outlined in Thamdrup and Dalsgaard (2002) and
Trimmer et al. (2003). Briefly, sediments from 0 – 10 cm depth were collected from the edge of
the seagrass bed and homogenized. Sediments were then transferred to an anoxic glove box,
combined in a 1:1 volumetric ratio with He-sparged site water, and pre-incubated overnight.
Following pre-incubation, sediment slurries were incubated in 12 mL Exetainers TM without a
headspace either unamended or after addition of 15NH4+, 15NH4+ + 14NO3-, or 15NO3- at a
concentration of 100 M. Biological activity was halted with addition of 100 L of saturated
solution of HgCl2 at 6 h intervals for a total of 18 h. The isotope ratio of N2 was then analyzed
using a Thermo Trace GC Ultra with a 30 m x 0.32 mm CarboPLOT column interfaced to a
Thermo Delta V Plus isotope ratio mass spectrometer (IRMS). Anammox activity was confirmed
by an increase in the isotope fraction of 29N2 in the 15NH4+ + 14NO3- and 15NO3- incubations, and

73

denitrification activity was confirmed by an increase in the isotope fraction of 30N2 in the 15NO3incubation (Appendix Fig. A5).

Intact sediment incubations
Intact sediment cores (21-30 cm depth) were collected in 9 cm i.d. transparent PVC
tubes, leaving approximately 1 cm of water overlying each core. Upon return to the lab, sediment
core tubes were placed into an enclosure containing site water held at in situ temperature (16.5
C). Oxic conditions throughout the enclosure were maintained by bubbling, and water was
circulated throughout the enclosure with an aquarium pump. The water inside the sediment core
tubes was circulated throughout the experiment by magnetic stir bars placed 5 cm above the
sediment surface and rotated just below the resuspension threshold by an external magnet.
Sediment core tubes were left uncapped and pre-incubated overnight (Ferguson et al. 2004).
Immediately prior to initiation of the incubation, a sample for dissolved inorganic N
(NH4+ and NO3- + NO2-, hereafter NO x) analysis were taken from each core, filtered through a
0.45 m cellulose acetate syringe filter into a 10 mL polyethylene vial, and frozen until analysis.
Cores were then sealed with a cap equipped with a syringe port and a vent port with on-off
valves. 15N-NO3- was added to obtain a final concentration of 100 M in the overlying water of
each core via the syringe port. Cores were left either unamended (control) or amended with 1.5 g
m-2 of dried seagrass or phytoplankton detritus, which was added through the syringe port.
Seagrass detritus consisted of dead seagrass and epiphytes was collected from the benthos of
Shaws Bay, rinsed with deionized water, dried at 60 C, and ground finely with a mortar and
pestle. This represented the type of detritus regularly encountered in the system. Phytoplankton

74

detritus consisted of laboratory-cultured phytoplankton that was rinsed and dried at 60 C. The C
and N contents of Shaws Bay sediment, seagrass detritus, and phytoplankton detritus were
measured using a Thermo Finnigan Flash EA 1112 interfaced with a Thermo Conflo III and a
Thermo Delta V Plus IRMS. Sediment from Shaws Bay had a C:N molar ratio of 20.1 (0.15% C;
0.008% N). Seagrass detritus had a C:N molar ratio of 40.7 (34.9% C; 1.0% N), and
phytoplankton detritus had a C:N molar ratio of 7.7 (40.3% C; 6.1% N).
Cores were allowed to equilibrate for 30 min to ensure homogenization of the NO 3- pool
in the overlying water and NO3- reduction zone at the sediment-water interface (Dalsgaard et al.
2000). At intervals of 0, 6, 12, and 19.5 h, cores from each organic carbon treatment were
sacrificed in triplicate. Dissolved oxygen concentration and pH of the overlying water were
measured with a Hach HQ40D dissolved O2/pH probe that was calibrated to manufacturer
specifications prior to use. A final dissolved nutrient sample was collected, as above, and an
additional 30 mL sample was filtered and frozen for 15N-NH4+ analysis. The top 2-4 cm of the
core were then slurrified with a glass rod, and samples were collected for dissolved gas analysis.
Samples for isotopic and concentration analysis of dissolved N2 were collected in 12 mL
Exetainers and 7 mL glass vials, respectively, to overflowing. 20 ÂľL of a saturated solution of
HgCl2 was then added to halt biological activity, and the container was capped with a screw top
septum (Exetainer) or a ground-glass stopper (glass vial) without a headspace. NO 2- was not
detected in samples, and thus inorganic N2O production in response to HgCl2 addition was not
considered a concern. A 2 mL He headspace was added to the Exetainers at atmospheric pressure
and allowed to equilibrate for at least 12 h prior to analysis. N 2 samples were stored underwater
at in situ temperature until analysis to minimize diffusion of atmospheric N 2 during storage.
Samples for the isotopic and concentration analysis of N 2O were introduced by siphon into 300

75

mL glass serum bottles to overflowing and capped without a headspace with screw cap septa.
One mL of a saturated solution of HgCl2 was added to halt biological activity. A 25 mL
headspace of He was added at atmospheric pressure using a vent needle, and N 2O samples were
allowed to equilibrate for least 12 h prior to analysis.

Analyses
Concentrations of NH4+ and NOx were measured colorimetrically on a Lachat
QuickChem 8000 four channel Flow Injection Analyzer according to Eyre (2000). Headspace
gas samples from the Exetainers were analyzed for the isotopic composition of N 2 on a Thermo
Trace GC Ultra with a 25 m x 0.32 mm PoraPLOT Q column interfaced to a Thermo Delta V
Plus IRMS. Concentrations of dissolved N2 were analyzed by membrane inlet mass
spectrometry (Eyre et al. 2002). The concentration and isotopic composition of N 2O were
analyzed by a Thermo Fisher GasBench II interfaced to a Thermo Delta V Plus IRMS following
He sparging and cryogenic trapping. The isotopic composition of NH 4+ was analyzed following
oxidation to NO2- by hypobromite and subsequent reduction to N2O by sodium azide and acetic
acid according to Zhang et al. (2007). The isotope ratio of the N 2O was analyzed by IRMS as
above. NO2- was not detected in the samples and did not require removal prior to 15NH4+
analysis.

IPT modeling
N2 production via denitrification and anammox (P14) was calculated according to the RIPT (Figure 10a, Table 5; Risgaard-Petersen et al. 2003). N2O production via denitrification was

76

calculated according to the IPTanaN2O and was also included in P14 (Figure 10b, Table 5; Hsu and
Kao 2013). The ratio of 14NO3-:15NO3- in the NO3- reduction zone (r14) was estimated by
determining the ratio of 14N:15N in N2O. Unlike r14 derived from the NO3- pool, r14 derived from
the N2O pool produces values for N2 and N2O production that are independent from the amount
of 15NO3- added and allows for variation among individual cores (Trimmer et al. 2006). The r14
derived from N2O thus represents a more direct representation of N undergoing reduction than
r14 derived from NO3-.
If DNRA occurs, 15NO3- is reduced to 15NH4+, introducing 15N into the NH4+ pool. If 15N
labeling of the NH4+ pool is high enough, this violates a central assumption of the isotope pairing
technique, as both denitrification and anammox are capable of producing 30N2 if DNRA is active.
The isotope pairing technique equations were therefore revised to account for the influence of
DNRA, which we call the R-IPT-DNRA (Fig. 10d; Table 5). A key addition of this approach is
the quantification of the degree of 15N labeling of the NH4+ pool, namely the ratio of
14NH +:15NH +
4
4

(r14a). The R-IPT-DNRA utilizes r14a to quantify production of 28N2, 29N2 (via

coupled DNRA-anammox and canonical anammox), and 30N2 by anammox, respectively:
A28 = (r14 * A29)/(r14/r14a + 1)

(12)

A29-DNRA = A29/(1 + r14a/r14)

(13)

A29-ana = A29 – A29-DNRA

(14)

A30 = (P29 – 2P30 * r14)/(r14a – r14)

(15)

Quantification of A29-DNRA and A29-ana enables the distinction of production of N2 via coupled
DNRA-anammox from canonical anammox, respectively:
A14-DNRA = A29-DNRA + 2A28 * (A29-DNRA/A29)

77

(16)

A14-ana = 2A28 * (A29-ana/A29)

(17)

Detailed derivations of equations are listed in the Appendix.
Genuine DNRA rates were calculated as:
DNRA = r14 (P15NH4+ + A30)

(18)

where P15NH4+ indicates the production in 15NH4+ over time and A30 indicates the production of
30N

2

by anammox (Table 5). Given the short length of the incubations and the fact that

incubations were run in the dark, it is unlikely that significant assimilation of NO3- and
subsequent remineralization of organic matter to NH 4+ occurred. However, it is not possible to
distinguish NO3- assimilation and subsequent remineralization to NH4+ from DNRA using the RIPT-DNRA, so the rate of DNRA could include this coupled process.
Rates of N2 production by anammox and denitrification and N 2O production by
denitrification are together termed “total N loss.” Adding the rate of DNRA to these processes is
termed “total NO3- reduction.”

78

Table 5. Parameters and equations used in the R-IPT (Risgaard-Petersen et al. 2003) and the R-IPT-DNRA.
Parameter

Designation

P29

Production of 29N2 (measured directly)

P30

Production of 30N2 (measured directly)

P45

Production of 45N2O (measured directly)

P46

Production of 46N2O (measured directly)

r14*

Ratio of 14N/15N in NO3- (measured directly in N2O)

r14a

Ratio of 14N/15N in NH4+ (measured directly)

D28
D29
D30

R-IPT

R-IPT-DNRA

Production of 28N2 via denitrification

r142 ¡ P30

r142 · (P30 – A30)

Production of 29N2 via denitrification

2r14 ¡ P30

2r14 · (P30 – A30)

30

P30

P30 – A30

28

Production of N2 via denitrification

A28

Production of N2 via anammox

r14 ¡ A29

r14 ¡ A29
r14/r14a+1

A29

Production of 29N2 via anammox

P29 – D29

P29 – D29

29

A29-DNRA

Production of N2 via coupled DNRA-anammox

-

A29
1 + r14a/r14

A29-ana

Production of 29N2 via canonical anammox

-

A29 – A29-DNRA

A30**

Production of 30N2 via anammox

-

P29 – 2P30 · r14
r14a – r14

D14

Genuine N2 production via denitrification

2D28 + D29

2D28 + D29

A14

Genuine N2 production via anammox

2A28

2A28 + A29-DNRA

A14-DNRA

Genuine production of N2 via coupled DNRA-anammox

-

A29-DNRA + 2A28 ¡ (A29-DNRA/A29)

A14=ana

Genuine production of N2 via canonical anammox

-

2A28 ¡ (A29-ana/A29)

D14-N2O***

Genuine N2O production via denitrification

r14 ¡ (2P46 + P45)

r14 ¡ (2P46 + P45)

P14
Total genuine N2 + N2O production
D14 + A14 + D14-N2O
*Estimated from the direct measurement of 14N/15N in N2O (Trimmer et al. 2006)
**Modified from Song et al. (2013)
***From Hsu and Kao (2013)

79

i

D14 + A14 + D14-N2O

Statistical modeling
A simple linear regression was used to evaluate a potential linear increase in the fraction
of 15N in NH4+ over time. The effect of carbon treatment and incubation duration on each N
cycling rate was tested by main effects two-way ANOVA and analyzed in a Bayesian framework
using the programs R (version 3.1.2) and JAGS (Su and Yajima, 2015). For each test, 1,200
Markov-chain Monte Carlo iterations were run on three chains after a burn-in of 200 and thinned
by two, yielding 1,500 samples in the posterior distribution. Credible intervals (95%) were
constructed around the mean of each treatment-time group. The probability of a significant
difference between treatment-time groups was determined by calculating the proportion of
iterations in which the posterior estimate of one group was smaller than another (Turner et al.,
2010). The difference between groups was considered significant when less than 5% of the
posterior estimates were smaller than those of another group.

RESULTS
Application of R-IPT-DNRA
An increase in 15N was detected in the NH4+ pool in intact sediment cores, indicating
DNRA was occurring in Shaws Bay sediments. The fraction of 15N in the NH4+ pool increased
linearly over time for control (df = 8, p < 0.01, R2 = 0.58) and seagrass (df = 8, p < 0.05, R2 =
0.41) treatments. Although there was a trend, there was insufficient replication to confirm a
linear relationship for the phytoplankton treatment (df = 1, p = 0.14, R 2 = 0.90) (Fig. 11). The
degree of 15NH4+ labeling was sufficient to generate the production of 30N2 from anammox (A30),
which made up 6 – 87% of the total production of 30N2 (P30). When rates modeled by the R-IPT

80

were compared to those modeled by the R-IPT-DNRA, anammox rates across treatments were
underestimated by 1.7 – 15.0% and denitrification rates across treatments were overestimated by
6.8 – 690.4% by the R-IPT (Fig. 12). The degree of overestimation of denitrification was tightly
controlled by the proportion of 30N2 (P30) attributed to anammox (A30) (Fig. 13). As a percentage
of total anammox rates, DNRA-dependent anammox rates were 4.85  1.85% for the control,
13.51  1.96% for the phytoplankton treatment, and 3.17  0.91% for the seagrass treatment.

Treatment
Control
Phytoplankton
Seagrass

0.20

15

F NH4

+

0.15

0.10

0.05

0.00
0

5

10

15

20

Time (h)

Figure 11. Increase in the fraction of 15N in the NH4+ pool by treatment. F15NH4+ increased
linearly over time for control (df = 8, p < 0.01, R2 = 0.58) and seagrass (df = 8, p < 0.05, R2 =
0.41) treatments. Although there was a trend, there was insufficient replication to confirm a
linear relationship for the phytoplankton treatment (df = 1, p = 0.14, R 2 = 0.90).

81

Treatment
Control
Phytoplankton
Seagrass

1.5

1.0

0.5

20
15
10

0.0

5
0
Control Seagrass Phyto.

0.0

0.5

1.0

1.5

b
Treatment
0.4

Control
Phytoplankton
Seagrass

0.3

0.2
% overestimation

R-IPT-DNRA denitrification rate (mmol N m-2 h-1)

a

% underestimation

R-IPT-DNRA anammox rate (mmol N m-2 h-1)

2.0

0.1

600
400
200
0

0.0

2.0

Control Seagrass Phyto.

0.0

R-IPT anammox rate (mmol N m-2 h-1)

0.1

0.2

0.3

0.4

R-IPT denitrification rate (mmol N m-2 h-1)

Overestimation of denitrification (%)

Figure 12. Comparison of (a) anammox and (b) denitrification rates from the R-IPT (RisgaardPetersen et al. 2003) to rates from the R-IPT-DNRA method presented here for control,
phytoplankton, and seagrass treatments. The solid line represents a 1:1 relationship. Insets
represent the total range of under/overestimation by anammox and denitrification across
treatments, with the mean represented by the thick line.
x = 7.17 e5.28 y
R2 = 0.98
600

400

200

0
0.25

0.50

0.75

Proportion of 30N2 produced by anammox (A30/P30)

Figure 13. Overestimation of denitrification rate (D14) by the R-IPT model compared to the RIPT-DNRA model as a function of the production of 30N2 by anammox relative to total 30N2
production (A30/P30).

82

Rates of NO3- reduction processes
DNRA rates (mean Âą SD) were 0.42 Âą 0.22, 0.88 Âą 0.60, and 1.03 Âą 0.90 Âľmol N m -2 h-1
for control, phytoplankton, and seagrass treatments, respectively (Fig. 14a). Anammox rates
were 0.18 Âą 0.10, 0.26 Âą 0.10, and 0.60 Âą 0.52 Âľmol N m-2 h-1 for control, phytoplankton, and
seagrass treatments, respectively (Fig. 14b). Denitrification rates were 0.04 Âą 0.04, 0.08 Âą 0.08,
and 0.16 Âą 0.13 Âľmol N m-2 h-1 for control, phytoplankton, and seagrass treatments, respectively
(Fig. 14c). N2O production was 0.008 Âą 0.004, 0.012 Âą 0.014, and 0.018 Âą 0.018 Âľmol N m -2 h-1
for control, phytoplankton, and seagrass treatments, respectively (Fig. 14d). However, rates of
NO3- reduction were not uniform across incubation duration, and figures are divided by
incubation duration to illustrate these differences. The addition of seagrass detritus resulted in
significantly greater rates of anammox, denitrification, and DNRA in the first 6 h of the
incubation relative to the control (p < 0.05), whereas the addition of phytoplankton detritus did
not result in greater rates of N cycling compared to the control. DNRA was the most important
NO3- reduction process, comprising 49-74% of total NO3- reduction across carbon treatments and
incubation durations (Fig. 15a). Proportions of N loss processes were also similar across
treatment and incubation duration (Fig. 15b). N2 production by anammox comprised 13-38% of
NO3- reduction and 64-86% of total N loss, whereas N2 production by denitrification comprised
6-13% of NO3- reduction and 11-31% of total N loss. N2O production comprised < 3% of NO3reduction < 9% of total N loss.

83

Figure 14. Rates of (a) DNRA, (b) N2 production by anammox, (c) N2 production by
denitrification, and (d) N2O production by denitrification. Rates were determined from
incubations of three different durations and under three organic carbon treatments (control – no
addition, addition of phytoplankton detritus, and addition of seagrass detritus). Error bars
represent +1 standard deviation. DNRA rate data were not available at the 6 h time point with
phytoplankton addition. Note the difference in y-axis scales among panels.

84

a

Control

Phytoplankton

Seagrass

Control

Phytoplankton

Proportion of total NO3- reduction

1.00

Seagrass

0.75

Process
+

NH4 -DNRA
N2O-denitrification
N2-denitrification
N 2-anammox

0.50

0.25

0.00
06 h

12 h

19.5 h

06 h

12 h

19.5 h

06 h

12 h

19.5 h

Incubation duration (h)

Process
NH4+-DNRA
N2O-denitrification
N2-denitrification
N 2-anammox

b

Control

Phytoplankton

Seagrass

h

12 h

19.5 h

06 h

12 h

19.5 h

06 h

Incubation duration (h)

12 h

19.5 h

Proportion of N loss

1.00

0.75

0.50

0.25

0.00
06 h

12 h

19.5 h

06 h

12 h

19.5 h

06 h

12 h

19.5 h

Incubation duration (h)

Figure 15. N2 production by anammox and denitrification, N2O production by denitrification,
and DNRA as a proportion of (a) total NO3- reduction and (b) total N loss to N2 and N2O.
Proportions of each process were measured at three different incubation durations under three
carbon treatments (control – no addition, addition of phytoplankton detritus, and addition of
seagrass detritus).

DISCUSSION
Evaluation of R-IPT-DNRA when DNRA is active
In Shaws Bay sediments where DNRA was a substantial N cycling pathway, quantifying
DNRA simultaneously with anammox and denitrification yielded more accurate rate estimates
than quantifying anammox and denitrification alone. High proportional rates of DNRA resulted
in overestimation of denitrification and underestimation of anammox rates by the R-IPT model
relative to the R-IPT-DNRA model. The degree of over/underestimation was an order of
85

magnitude greater, on average, for denitrification than for anammox. Further, the overestimation
of denitrification increased exponentially as the proportion of 30N2 produced by anammox
(A30/P30) increased. This result is in agreement with Song et al. (2016), who propose that rates of
denitrification are more greatly overestimated when anammox makes up the majority of N loss.
Although previous studies have suggested that the majority of N loss in marine sediments is
dominated by denitrification (Trimmer and EngstrĂśm 2011) and the overestimation of
denitrification is unlikely to be appreciable in shallow marine systems (Song et al., 2016), our
results indicate this assumption is not always valid. Thus, denitrification rates determined by the
R-IPT model in marine sediments where DNRA is active could be significantly overestimated.
The variability in this overestimation cannot be identified without quantifying DNRA via
analysis of the isotopic composition of NH 4+ (r14a). Published studies of R-IPT-derived
denitrification rates, in sediments where appreciable DNRA is suspected, should thus be
interpreted with caution, and future studies should utilize the R-IPT-DNRA to obtain accurate
anammox and denitrification measurements.
The R-IPT-DNRA model calculated anammox rates in Shaws Bay sediments, but future
applications of the R-IPT-DNRA model could be impeded if substantial 15N labeling of the NH4+
pool occurs. If the in situ NH4+ concentration is low and the DNRA rate is large, the abundance
of NH4+ could increase to the point of stimulating anammox beyond the rate accounted for in the
model. While DNRA rates were considerable in this study, 15NH4+ made up 7.7  5.4% of the
total NH4+ pool at the end of the incubation and was therefore unlikely to have stimulated
anammox activity beyond the rates accounted for in the R-IPT. The concentration and isotopic
composition of NH4+ should be carefully monitored to evaluate if anammox rates derived from
the R-IPT-DNRA model may have been enhanced.

86

The application of the R-IPT-DNRA model allows coupled DNRA-anammox and
canonical anammox to be distinguished, a unique attribute of this modeling approach in
comparison to that presented by Song et al. (2016). Despite high relative rates of both DNRA
and anammox, coupled DNRA-anammox (A14-DNRA) made up a small proportion (< 15%) of total
anammox in Shaws Bay. While DNRA generated 22 Âą 14% of ambient NH 4+ on an hourly basis,
the lack of strong coupling between DNRA and anammox suggests that remineralization supplies
the majority of NH4+ to anammox in Shaws Bay.

Anammox, denitrification, and DNRA in seagrass ecosystem sediments
In this study we have quantified, for the first time, the co-occurrence of DNRA,
anammox and denitrification in intact sediments from a seagrass-dominated environment.
Typical of seagrass ecosystems, sediment in Shaws Bay has a moderately high C:N ratio
compared to other coastal ecosystems (Eyre et al. 2013; Khan et al. 2015) and receives detrital
organic matter with a higher C:N ratio than that of the sediment. DNRA was the predominant
NO3- reduction pathway in Shaws Bay sediments, illustrating that this high C:N environment
functions to retain N that might otherwise be denitrified. Similar N retention has been reported
previously in seagrass communities (Eyre et al. 2013). Although the energy yield of
denitrification is higher than that of DNRA, the stoichiometry of organic C and NO 3consumption for each reaction is such that high C:N environments favor DNRA over
denitrification (Tiedje et al. 1982; Burgin and Hamilton 2007; Kraft et al. 2014; van den Berg et
al. 2015).
Anammox contributed an average of 74% of total N loss in Shaws Bay sediments, an
unusually high proportion compared to other coastal sediments (Table 6). Proportions this high
87

have previously been reported only in deep sea sediments and were attributed to the limitation of
denitrification by low organic carbon availability (Thamdrup and Dalsgaard 2002; Dalsgaard et
al. 2005; Trimmer and Nicholls 2009; Trimmer et al. 2013). While seagrass ecosystems are
highly productive and contain large amounts of organic matter, there are driving factors that may
function to suppress denitrification in Shaws Bay, namely the refractory nature of seagrass
organic matter (Eyre et al. 2013) and competition for limited NO3- supply by the high rates of
DNRA observed. Competition between DNRA and denitrification points to a potential
mechanism that allows anammox to be the prominent N loss pathway.
Sediments in Shaws Bay were characterized by low rates of N loss that are atypical of
shallow sediments (Table 6). However, rates of other microbial processes demonstrate an active
microbial community in Shaws Bay that functions to recycle rather than remove N from the
system. The slope of O2 demand vs. N2 production in Shaws Bay was an order of magnitude
lower than that found across a range of Australian sand-, mud-, seagrass-, and macroalgaedominated sediments (Eyre et al. 2013). This trend emphasizes that while the sediment microbial
community is actively respiring organic matter, N loss rates are diminished in Shaws Bay
relative to other coastal sediments. Several factors could contribute to the suppression of N loss
in this system. Owing to the high C:N ratios of organic matter in the system, DNRA likely
maintains NO3- limitation by outcompeting denitrifiers for substrate (Burgin and Hamilton 2007;
van den Berg et al. 2015). In addition, the assimilative demand for N by heterotrophic
microorganisms processing high C:N organic matter could outcompete denitrifiers for NO3(Oakes et al. 2011). Furthermore, sulfide is found in seagrass systems owing to the activity of Nfixing sulfate reducers (Hemminga and Duarte 2000) and could inhibit nitrification and suppress
coupled nitrification-denitrification (Joye and Hollibaugh 1995). Finally, although the organic

88

matter source is similar, there may be dampened microbial activity in bare sediments adjacent to
seagrasses compared to sediments within the seagrass bed (Boschker et al. 2000; Jones et al.
2003; Holmer et al. 2004).
N2O production comprised between 1.1 and 8.6% of total N loss, a typical range for
aquatic sediments (Seitzinger and Kroeze 1998; McCrackin and Elser 2010). However, given
that denitrification rates were markedly low, the total rate of N 2O production was ~100-fold
lower than production as indicated by atmospheric emission rates measured in other coastal
environments (Murray et al. 2015 and references therein). The total range of N 2O production
across carbon treatments in Shaws Bay was 0.002 to 0.008 mol N2O m-2 h-1 (median = 0.003),
whereas previous measurements in mangroves, salt marshes, and estuaries ranged from 0.1 to
6.0, -2.5 to 8.9, and 0.1 to 8.3 mol N2O m-2 h-1, respectively (Murray et al. 2015 and references
therein). To our knowledge, this is the first published measurement of N 2O production in
sediments from a seagrass-dominated system. While this study does not address the regulation of
N2O production and atmospheric flux by seagrasses themselves, it does provide a representation
of N2O production in a coastal system that receives relatively refractory high C:N organic matter
derived from seagrasses. If our results are representative of other seagrass systems, seagrass
environments could have a much smaller impact on the global N 2O budget than other coastal
environments. However, given that denitrification rates in Shaws Bay seem to depart from the
relatively higher rates reported from other seagrass systems (reviewed in Murray et al. 2015),
additional studies are needed to constrain the role of seagrass systems on the global N 2O budget.

89

Table 6. Nitrogen loss rates calculated by the isotope pairing technique for intact sediments and sediment slurries. ND = not detected.
Location
Intact sediments
Shaws Bay, Australia

12 coastal sites, Greenland
Alsback, Gullmarsfjorden, Sweden
Gravesend, Thames estuary, U.K.
Constructed wetland, Australia
Colne Estuary, U.K.

Irish Sea
North Atlantic
Mae Klong Estuary, Thailand
Cisadane Estuary, Indonesia
Vunidawa-Rewa Estuary, Fiji
St. Lawrence Estuary, Canada
Danshuei River, Thailand
Skagerrak, Denmark

Great Barrier Reef, Australia
Sediment slurries
Norsminde Fjord
Randers Fjord
Young Sound
Skagerrak, Denmark
Skagerrak, Denmark

Logan/Albert River, Australia

Anammox
rate
mol N m-2 h-1

famx
anammox
total N loss

0.04
0.18
0.16
0.60
0.08
0.26
1.38 to 11.04
0.04 to 3.83
6.08
6.64
192.9
48.94
Intake zone
965.9
66.1
Outflow zone
651.5
199.40
Hythe (head)
390 to 400
ND to 157
Alresford Creek (middle)
50 to 80
ND
Brightlingsea (mouth)
5 to 10
ND
Continental shelf, 50-100 m depth
3.44
1.38
Continental slope, 500-2000 m depth
0.42
0.61
Salinity 0-35 PSU
3.3
ND
Salinity 0-35 PSU
22.5
ND
Salinity 0-35 PSU
0.6
ND
345 m depth
11.3
5.5
Intertidal zone
42.3
13
176 m depth
13.25
ND
384 m depth
9.57
ND
610 m depth
0.61
2.01
688 m depth
0.3
0.77
29 m depth
2.08
4.86

0.74
0.74
0.78
0.01 to 0.35
0.48
0.21
0.06
0.23
0.00 to 0.30
0.00
0.00

1 m depth
1 m depth
36 m depth
695 m depth
16 m depth
380 m depth
700 m depth
Upstream to river mouth transect

0.00
0.06
0.19
0.70
0.02
0.24
0.67
0.00 to 0.09

Details

Dentrification
rate
mol N m-2 h-1

Bare sediment
Sediment + seagrass
Sediment + phytoplankton
36 to 100 m depth
116 m depth

340
8
2

0.5 to 8.5

90

22
2
4

0 to 0.8

Reference

This study

Rysgaard et al. 2004
Trimmer et al. 2006
Erler et al. 2008
Dong et al. 2009

Trimmer and Nicholls 2009
0.00
0.00
0.00
0.33
0.12
0.00
0.00
0.77
0.72
0.70

Dong et al. 2011

Crowe et al. 2012
Hsu and Kao 2013
Trimmer et al. 2013

Erler et al. 2013

Risgaard-Petersen et al. 2003

Dalsgaard et al. 2005

Meyer et al. 2005

Table 6 (cont’d)
Heron Island, Australia
9 estuaries, United Kingdom

1 m depth

Location

Details

Intact sediments
Aarhus Bay, Denmark
Skagerrak, Denmark
Thames estuary, United Kingdom

Sediment slurries
East China Sea

16 m depth
380 m depth
695 m depth
Salinity 2 PSU
Salinity 10 PSU
Salinity 10 PSU
Salinity 26 PSU
Salinity 30 PSU
Salinity 30 PSU
19 m depth
85 m depth

34 to 480*

ND

0.00
0.05 to 0.11

Eyre et al. 2008
Nicholls and Trimmer 2009
Reference

Dentrification
rate
mol N L-1 h-1

Anammox
rate
mol N L-1 h-1

famx
anammox
total N loss
0.02
0.24
0.67
0.015
0.075
0.043
0.025
0.017
0.014

Thamdrup and Dalsgaard 2002

100
118
155
116
59
36

3.46
4.13
1.25
1.5
9.5
7
3
1
0.5
2
8

0.125
0.8

Song et al. 2013

14
2

*Denitrification measured by N2:Ar
**Unpublished data from Eyre et al. 2011b

91

Trimmer et al. 2003

Effect of organic carbon quality on N cycling
The addition of seagrass detritus stimulated rates of anammox, denitrification, and
DNRA in Shaws Bay sediments, but low C:N phytoplankton detritus did not. This is in direct
contrast to other results that suggest that labile organic matter stimulates N loss in estuarine
sediments to a greater extent than relatively refractory organic matter does (Brettar and
Rheinheimer 1992; Caffrey et al. 1993; Fulweiler et al. 2008; Oakes et al. 2011; Fulweiler et al.
2013). However, Babbin et al. (2014) demonstrated that N loss in the East Tropical North Pacific
oxygen minimum zone was greater when water was amended with sinking particulate matter
than with highly labile amino acids or sucrose. This suggests that microbial communities may be
preferentially adapted to decompose and recycle organic matter that is commonly encountered.
Given that this system encounters seagrass detritus to a greater extent than phytoplankton
detritus, the sediment microbial community could be predisposed to utilize this refractory
organic matter source.
Alternatively, because seagrass detritus includes the associated epiphytic community, the
stimulation in N cycling in the first 6 h following addition of seagrass could represent selective
consumption of epiphytic material (Dahllof and Karle 2005). Bacteria in seagrass sediments
metabolize organic matter derived from a mix of seagrass and benthic macro- and microalgae,
with a greater reliance on seagrass in more pristine sites and a mixed to algal-dominated
signature in anthropogenically-impacted sites (Boschker et al. 2000; Holmer et al. 2001; Jones et
al. 2003; Holmer et al. 2004; Bouillon and Boschker 2006; Williams et al. 2009). Shaws Bay is
considered only slightly impacted by anthropogenic activities (Ballina Shore Council 2000), so
microbial communities likely rely mainly on seagrass-derived organic matter. This relationship is
expected to hold in both vegetated sediment and adjacent bare sediments (Boschker et al. 2000;

92

Jones et al. 2003; Holmer et al. 2004), meaning that the organic matter utilization measured in
unvegetated sediments we sampled in Shaws Bay is expected to be seagrass-derived. Although
we do not know whether the epiphytic component or seagrass itself enhanced N cycling in the
seagrass detritus addition treatments, our results indicate that the addition of a C source common
to the environment stimulates N cycling.
In this study, the proportions of anammox and denitrification contributing to N loss
remained the same regardless of carbon treatment and total N loss rate. This is somewhat
unexpected given that denitrification, a heterotrophic process, uses organic carbon directly,
whereas anammox, an autotrophic process, does not (Jetten et al. 1999). Although the anammox
reaction itself does not require carbon, a supply of carbon may be crucial to support
heterotrophic nitrogen cyclers that provide substrates for anammox (Trimmer et al. 2003;
Dalsgaard et al. 2005; van Hulle et al. 2012). Further, there is evidence that anammox bacteria
are capable of coupling organic acid oxidation and NO3- reduction and compete with
denitrification for both substrates (Guven et al. 2005; Kartal et al. 2007). Together, these
pathways represent a means by which anammox activity can be stimulated upon addition of
seagrass detritus and explain how the proportion of anammox making up total N loss can remain
consistent across carbon treatments.

Conclusion
We present a revision to the isotope pairing technique (R-IPT-DNRA) that enables the
simultaneous measurement of rates of anammox, denitrification, N2O production, and DNRA in
intact sediments when these processes co-occur. This approach presents a significant advance for
the understanding of N processing in coastal systems, where N mitigation efforts depend on a
93

detailed understanding of N loss and recycling pathways. The R-IPT-DNRA model is likely to
be more accurate than the traditional application of the isotope pairing technique in environments
that receive inputs of carbon-rich, relatively refractory organic matter such as the seagrass
ecosystem in this study. Under these conditions, DNRA is energetically favored over
denitrification, and the introduction of 15N into the NH4+ pool via DNRA can cause a substantial
overestimation of denitrification rates if DNRA is not taken into account.
In Shaws Bay, a carbon-rich and N-poor seagrass-dominated coastal ecosystem, the
dominance of DNRA conserves N in the system, effectively outcompeting denitrifiers for NO3-.
In turn, the NH4+ production by DNRA sets the stage for anammox to dominate overall N loss.
N2O production made up a typical proportion of denitrification, yet absolute rates were markedly
low compared to other coastal ecosystems owing to low overall rates of denitrification. This N
cycling regime was stimulated by the addition of seagrass detritus but not by relatively labile
phytoplankton detritus, suggesting that microbial communities in coastal sediments are adapted
to process the organic matter that is typically encountered.

94

APPENDIX

95

a

b
0.0100

Treatment

0.004

15

NH4+
15
NH4+ + 14NO315
NO3Control

15

NH4+
NH4+ + 14NO315
NO3Control
15

0.003

F30

F29

0.0095

Treatment

0.0090

0.002
0.0085

0.001

0.0080

0.000
0

6

12

0

18

6

12

18

Incubation duration (h)

Incubation duration (h)

Figure A5. Isotope fraction of (a) 29N2 and (b) 30N2 in anoxic slurry incubations following
addition of 15NH4+, or 15NH4+ + 14NO3-, or 15NO3- over the course of 18 h. The control treatment
received no amendment. An increase in F29 indicates anammox activity, whereas an increase in
F30 indicates denitrification activity.

96

R-IPT-DNRA Derivations
Base Equations
𝑥=

𝑦=

14

𝑁𝑂3−
14𝑁𝑂 − + 15𝑁𝑂 −
3
3

1 − 𝐹𝑁
𝑥
= 𝑟14 =
𝐹𝑁
𝑦

15

𝑥 = 𝑟14 ∙ 𝑦

𝑁𝑂3−
14𝑁𝑂− + 15𝑁𝑂 −
3
3
14

𝑢=

𝑁𝐻4+
14𝑁𝐻 + + 15𝑁𝐻 +
4
4

1 − 𝐹𝐴
𝑢
= 𝑟14𝑎 =
𝐹𝐴
𝑣

15

𝑁𝐻4+
𝑢 = 14 +
𝑁𝐻4 + 15𝑁𝐻4+

𝑢 = 𝑟14𝑎 ∙ 𝑣

Derivation of A28
𝐴28
𝑥∙𝑢
=
𝐴29 𝑥 ∙ 𝑣 + 𝑢 ∙ 𝑦
𝐴28 (𝑥 ∙ 𝑣 + 𝑢 ∙ 𝑦) = 𝐴29 ∙ 𝑥 ∙ 𝑢
𝐴28 ∙ 𝑥 ∙ 𝑣 = 𝑢(𝐴29 ∙ 𝑥 − 𝐴29 ∙ 𝑦)
𝐴28 ∙ 𝑥 ∙ 𝑣
= 𝐴29 ∙ 𝑥 − 𝐴28 ∙ 𝑦
𝑟14𝑎 ∙ 𝑣
𝐴28 (𝑟

𝑥

+ 𝑦) = 𝐴29 ∙ 𝑥

14𝑎

𝑟 ∙𝑦

𝐴28 ( 𝑟14 + 𝑦) = 𝐴29 ∙ 𝑟14 ∙ 𝑦
14𝑎

𝑟

𝐴28 ∙ 𝑦 (𝑟 14 + 1) = 𝐴29 ∙ 𝑟14 ∙ 𝑦
14𝑎

𝐴28 =

𝐴29 ∙ 𝑟14
14 + 1)
(𝑟𝑟14𝑎

97

Derivation of A30
𝐴30 =

𝑣(𝑃29 ∙ 𝑦 − 2𝑥 ∙ 𝑃30 )
𝑦∙𝑢−𝑣∙𝑥

𝐴30 =

𝑣 ∙ 𝑃29 ∙ 𝑦 − 𝑣 ∙ 2𝑟14 ∙ 𝑦 ∙ 𝑃30
𝑦 ∙ 𝑟14𝑎 ∙ 𝑣 − 𝑣 ∙ 𝑟14 ∙ 𝑦

𝐴30 =

𝑣 ∙ 𝑦(𝑃29 − 2𝑃30 ∙ 𝑟14)
𝑣 ∙ 𝑦(𝑟14𝑎 − 𝑟14 )

𝐴30 =

𝑃29 − 2𝑃30 ∙ 𝑟14
𝑟14𝑎 − 𝑟14

Derivation of A29-DNRA
A29=DNRA: probability of pairing 14NO3- + 15NH4+ = x*v
A29=ana: probability of pairing 15NO3- + 14NH4+ = u*y
𝐴29−𝐷𝑁𝑅𝐴 = 𝐴29 (
𝐴29
𝐴29−𝐷𝑁𝑅𝐴
𝐴29
𝐴29−𝐷𝑁𝑅𝐴
𝐴29
𝐴29−𝐷𝑁𝑅𝐴

=

𝑥∙𝑣
)
𝑥∙𝑣+𝑢∙𝑦

𝑥∙𝑣+𝑢∙𝑦
𝑥∙𝑣

=1+

𝑢∙𝑦
𝑥∙𝑣

=1+

𝑟14𝑎
𝑟14

𝐴29−𝐷𝑁𝑅𝐴 =

𝐴29
(1 + 𝑟𝑟14𝑎
)
14

𝐴29−𝑎𝑛𝑎 = 𝐴29 − 𝐴29−𝐷𝑁𝑅𝐴

98

LITERATURE CITED

99

LITERATURE CITED

Algar CK, Vallino JJ (2014) Predicting microbial nitrate reduction pathways in coastal
sediments. Aquat Microb Ecol 71:223-238. doi:10.3354/ame01678
Babbin AR, Keil RG, Devol AH, Ward BB (2014) Organic matter stoichiometry, flux, and
oxygen control nitrogen loss in the ocean. Science 344:406-408. doi:10.1126/science.1248364
Ballina Shire Council (2000) Shaws Bay, East Ballina: Estuary management plan. Estuary
Processes Study Report, Vol. 1. Reference no. J3320/R2147.s.
Betlach MR Tiedje JM (1981) Kinetic explanation for accumulation of nitrite, nitric oxide, and
nitrous oxide during bacterial denitrification. Appl Environ Microb 42:1074-1084.
Boschker HTS, Wielemaker A, Schaub BEM, Holmer M (2000) Limited coupling of macrophyte
production and bacterial carbon cycling in the sediments of Zostera spp. meadows. Mar EcolProg Ser 203:181-189.
Bouillon S, Boschker HTS (2006) Bacterial carbon sources in coastal sediments: A cross-system
analysis based on stable isotope data of biomarkers. Biogeosciences 3:175-185.
doi:10.5194/bg-3-175-2006
Brettar I Rheinheimer G (1992) Influence of carbon availability on denitrification in the central
Baltic Sea. Limnol Oceanogr 37(6):1146-1161.
Brin LD, Giblin AE, Rich JJ (2015) Effects of experimental warming and carbon addition on
nitrate reduction and respiration in coastal sediments. Biogeochemistry 125:81-95.
doi:10.1007/s10533-015-0113-4
Burgin AJ, Hamilton SK (2007) Have we overemphasized the role of denitrification in aquatic
ecosystems? A review of nitrate removal pathways. Front Ecol Environ 5(2):89-96.
Caffrey JM, Sloth NP, Kaspar HF, Blackburn TH (1993) Effect of organic loading on
nitrification and denitrification in a marine sediment microcosm. FEMS Microbiol Ecol 12:159167.
Dahllof I, Karle I-M (2005) Effect on marine sediment nitrogen fluxes caused by organic matter
enrichment with varying organic carbon structure and nitrogen content. Mar Chem 94:17-26.
doi:10.1016/j.marchem.2004.07.008
Dalsgaard T, Nielsen LP, Brotas V, Viaroli P, Underwood G, Nedwell D, Sundbäck K, Rysgaard
S, Miles A, Bartoli M, Dong L, Thornton DCO, Ossosen LDM, Castaldelli G, Risgaard-Petersen
N (2000) Protocol handbook for NICE- Nitrogen cycling in estuaries. National Environmental
Research Institute, Silkeborg, Denmark.

100

Dalsgaard T, Thamdrup B, Canfield DE (2005) Anaerobic ammonium oxidation (anammox) in
the marine environment. Res Microbiol 156:45-464. doi:10.1016/j.resmic.2005.01.011
Davidson EA, Verchot LV (2000) Testing the hole-in-the-pipe model of nitric and nitrous oxide
emissions from soils using the TRAGNET database. Global Biogeochem Cy 14(4):1035-1043.
doi:10.1029/1999GB001223
Dong LF, Sobey MN, Smith CJ, Rusmana I, Phillips W, Scott A, Osborn AM, Nedwell DB
(2011) Dissimilatory reduction of nitrate to ammonium, not denitrification or anammox,
dominates benthic nitrate reduction in tropical estuaries. Limnol Oceanogr 56(1):279-291.
doi:10.4319/lo.2011.56.1.0279
EnrĂ­quez S, MarbĂ  N, Duarte CM, van Tussenbroek BI, Reyes-Zavala G (2001) Effects of
seagrass Thalassia testudinum on sediment redox. Mar Ecol Prog Ser 219:149-158.
Erler DV, Trott LA, Alongi DM, Eyre BD (2013) Denitrification, anammox and nitrate reduction
in the sediments of the Southern Great Barrier Reef Lagoon. Mar Ecol Prog Ser 478:57-70.
Eyre BD (2000) Regional evaluation of nutrient transformation and phytoplankton growth in
nine river-dominated sub-tropical east Australian estuaries. Mar Ecol Prog Ser 205:61-83.
Eyre BD, Rysgaard S, Dalsgaard T, Christensen PB (2002) Comparison of isotope pairing and
N2:Ar methods for measuring sediment denitrification: Assumptions, modifications and
implications. Estuaries 25(6A):1077-1087.
Eyre BD, Ferguson AJP, Webb A, Maher D, Oakes JM (2011a) Denitrification, N-fixation and
nitrogen and phosphorus fluxes in different benthic habitats and their contribution to the nitrogen
and phosphorus budgets of a shallow oligotrophic sub-tropical coastal system (southern Moreton
Bay, Australia). Biogeochemistry 102:111-133. doi:10.1007/s10533-010-9425-6
Eyre BD, Maher D, Oakes JM, Erler DV, Glasby TM (2011b) Differences in benthic
metabolism, nutrient fluxes, and denitrification in Caulerpa taxifolia communities compared to
uninvaded bare sediment and seagrass (Zostera capricorni) habitats. Limnol Oceanogr
56(5):1737-1750. doi:10.4319/lo.2011.56.5.1737
Eyre BD, Maher DT, Squire P (2013) Quantity and quality of organic matter (detritus) drives N2
effluxes (net denitrification) across seasons, benthic habitats, and estuaries. Global Biogeochem
Cy 27:1-13. doi:10.1002/2013GB004631
Eyre B., Maher DT, Sanders C (2016) The contribution of denitrification and burial to the
nitrogen budgets of three geomorphically distinct Australian estuaries: Importance of seagrass
habitats. Limnol Oceanogr 61(3):1144-1156. doi:10.1002/lno.10280
Ferguson AJP, Eyre BD, Gay J (2004) Benthic nutrient fluxes in euphotic sediments along
shallow sub-tropical estuaries, northern NSW, Australia. Aquat Microb Ecol 37:219-235.

101

Fulweiler RW, Nixon SW, Buckley BA, Granger SL (2008) Net sediment N2 fluxes in a coastal
marine system - experimental manipulations and a conceptual model. Ecosystems
11:1168−1180.
Fulweiler RW, Brown SM, Nixon SW, Jenkins BD (2013) Evidence and a conceptual model for
the co-occurrence of nitrogen fixation and denitrification in heterotrophic marine sediments. Mar
Ecol Prog Ser 482:57-68. doi:10.3354/meps10240
Giblin AE, Tobias CR, Song B, Weston N, Banta GT, Rivera-Monroy VH (2013) The
importance of dissimilatory nitrate reduction to ammonium (DNRA) in the nitrogen cycle of
coastal ecosystems. Oceanogr 26(3):124–131. doi:10.5670/oceanog.2013.54
Guven D, Dapena A, Kartal B, Schmid MC, Mass B, van de Pas-Schoonen K, Sozen S, Mendez
R, Op den Camp HJM, Jetten MSM, Strous M, Schmidt I (2005) Propionate oxidation by and
methanol inhibition of anaerobic ammonium-oxidizing bacteria. Appl Environ Microbiol
71(2):1066. doi:10.1128/AEM.71.2.1066-1071.2005
Hardison AK, Algar CK, Giblin AE, Rich JJ (2015) Influence of organic carbon and nitrate
loading on partitioning between dissimilatory nitrate reduction to ammonium (DNRA) and N 2
production. Geochim Cosmochim Acta 164:146-160. doi:10.1016/j.gca.2015.04.049
Hemminga MA, Duarte CM (2000) Seagrass Ecology. Cambridge University Press, Cambridge.
Holmer M, Andersen FO, Nielsen SL, Boschker HTS (2001) The importance of mineralization
based on sulfate reduction for nutrient regeneration in tropical seagrass sediments. Aquat Bot
71:1-17.
Holmer M, Duarte CM, Boschker HTS, BarrĂłn C (2004) Carbon cycling and bacterial carbon
sources in pristine and impacted Mediterranean seagrass sediments. Aquat Microb Ecol 36:227237.
Howarth RW (2008) Coastal nitrogen pollution: A review of sources and trends globally and
regionally. Harmful Algae 8:14-20. doi:10.1016/j.hal.2008.08.015
Hsu T-C, Kao S-J (2013) Technical note: Simultaneous measurement of sedimentary N 2 and
N2O production and new 15N isotope pairing technique. Biogeosciences 10:7847-7862.
doi:10.5194/bg-10-7847-2013
Hwang S, Jang K, Jang H, Song J, Bae W (2006) Factors affecting nitrous oxide production: A
comparison of biological nitrogen removal processes with partial and complete nitrification.
Biodegradation 17:19-29. doi:10.1007/s10532-005-2701-9
IPCC (2007) Climate Change 2007: The Physical Science Basis. In: Solomon S et al (eds)
Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental
Panel on Climate Change. Cambridge University Press, New York.

102

Jetten MSM, Strous M, van de Pas-Schoonen KT, Schalk J, van Dongen UGJM, van de Graaf
AA, Logemann S, Muyzer G, van Loosdrecht MCM, Kuenen JG (1999) The anaerobic oxidation
of ammonium. FEMS Microbiol Rev 22:421-437.
Jones WB, Cifuentes LA, Kaldy JE (2003) Stable carbon isotope evidence for coupling between
sedimentary bacteria and seagrasses in a sub-tropical lagoon. Mar Ecol Prog Ser 255:15-25.
Joye SB, Hollibaugh JT (1995) Influence of sulphide inhibition of nitrification on nitrogen
regeneration in sediments. Science 270:623-625.
Kartal B, Rattray J, van Niftrik LA, van de Vossenberg J, Schmid MC, Webb RI, Schouten S,
Fuerst JA, Damsté JS, Jetten MSM, Strous M (2007) Candidatus “Anammoxoglobus
propionicus” a new propionate oxidizing species of anaerobic ammonium oxidizing bacteria.
Syst Appl Microbiol 30:39-49. doi:10.1016/j.syapm.2006.03.004
Kartal B, van Niftrik LA, Keltjens JT, Op den Camp HJM, Jetten MSM (2012) Anammox—
growth physiology, cell biology, and metabolism. Adv Microb Physiol 60:211-262.
doi:10.1016/B978-0-12-398264-3.00003-6
Khan NS, Vane CH, Horton BP (2015) Stable carbon isotope and C/N geochemistry of coastal
wetland sediments as a sea-level indicator. In: Shennan I, Long AJ, Horton BP (eds) Handbook
of Sea-Level Research. Wiley. pp 295-311.
Knowles R (1982) Denitrification. Microbiological Rev 46(1):43-70
Kraft B, Tegetmeyer HE, Sharma R, Klotz MG, Ferdelman TG, Hettich RL, Geelhoed JS, Strous
M (2014) The environmental controls that govern the end product of bacterial nitrate respiration.
Science 345:676-679. doi:10.1126/science.1254070
McCrackin ML, Elser JJ (2010) Atmospheric nitrogen deposition influences denitrification and
nitrous oxide production in lakes. Ecology 91(2):528-539.
Murray RH, Erler DV, Eyre BD (2015) Nitrous oxide fluxes in estuarine environments:
Response to global change. Global Change Biol 21(9):3219-3245. doi:10.1111/gcb.12923
Nielsen LP (1992) Denitrification in sediment determined from nitrogen isotope pairing. FEMS
Microbiol Ecol 86:357-362.
Oakes JM, Eyre BD, Ross DJ (2011) Short-term enhancement and long-term suppression of
denitrification in estuarine sediments receiving primary- and secondary-treated paper and pulp
mill discharge. Environ Sci Technol 45:3400-3406. doi:10.1021/es103636d
Risgaard-Petersen N, Nielsen LP, Rysgaard S, Dalsgaard T, Meyer RL (2003) Application of the
isotope pairing technique in sediments where anammox and denitrification coexist. Limnol
Oceanogr Methods 1:63-73.

103

Schlesinger WH (2009) On the fate of anthropogenic nitrogen. Proc Natl Acad Sci USA
106:203-208. doi:10.1073pnas.0810193105
Seitzinger SP, Kroeze C (1998) Global distribution of nitrous oxide production and N inputs in
freshwater and coastal marine ecosystems. Global Biogeochem Cy 12(1):93-113.
Seitzinger S, Harrison JA, BĂśhlke JK, Bouwman AF, Lowrance R, Peterson B, Tobias C, Van
Drecht G (2006) Denitrification across landscapes and waterscapes: A synthesis. Ecol Appl
16(6):2064-2090.
Song GD, Liu SM, Marchant H, Kuypers MMM, Lavik G (2013) Anammox, denitrification, and
dissimilatory nitrate reduction to ammonium in the East China Sea sediment. Biogeosciences
10:6851-6864. doi:10.5194/bg-10-6851-2013
Song GD, Liu SM, Kuypers MMM, Lavik G (2016) Application of the isotope pairing technique
in sediments where anammox, denitrification, and dissimilatory nitrate reduction to ammonium
coexist. Limnol Oceanogr Methods 14(12):801-815. doi:10.1002/lom3.10127
Su Y-S, Yajima M (2015) R2jags: A Package for Running jags from R. R package version 0.0501. http://CRAN.R-project.org/package=R2jags
Thamdrup B, Dalsgaard T (2002) Production of N2 through anaerobic ammonium oxidation
coupled to nitrate reduction in marine sediments. Appl Environ Microbiol 68:1312-1318.
doi:10.1128/AEM.68.3.1312-1318.2002
Tiedje JM, Sexstone AJ, Myrold DD, Robinson JA (1982) Denitrification: Ecological niches,
competition and survival. Antonie van Leeuwenhoek 48:569-583.
Trimmer M, Nicholls JC, Deflandre B (2003) Anaerobic ammonium oxidation measured in
sediments along the Thames Estuary, United Kingdom. Appl Environ Microbiol 69(11):64476454. doi:10.1128/AEM.69.11.6447–6454.2003
Trimmer M, Risgaard-Petersen N, Nicholls JC, EngstrĂśm P (2006) Direct measurement of
anaerobic ammonium oxidation (anammox) and denitrification in intact sediment cores. Mar
Ecol Prog Ser 326:37-47.
Trimmer M, Nicholls JC (2009) Production of nitrogen gas via anammox and denitrification in
intact sediment cores along a continental shelf to slope transect in the North Atlantic. Limnol
Oceanogr 54(2):577-589.
Trimmer M, EngstrĂśm P (2011) Distribution, activity, and ecology of anammox bacteria in
aquatic environments. In Ward BB, Arp DJ, Klotz MG (eds) Nitrification. American Society of
Microbiology, Washington. pp. 201-235.

104

Trimmer M, EngstrĂśm P, Thamdrup B (2013) Stark contrast in denitrification and anammox
across the deep Norwegian Trench in the Skagerrak. Appl Environ Microbiol 79(23):7381-7389.
doi:10.1128/AEM.01970-13
Turner TF, Collyer ML, Krabbenhoft TJ (2010) A general hypothesis-testing framework for
stable isotope ratios in ecological studies. Ecology 91(8):2227-2233.
van den Berg EM, van Dongen U, Abbas B, van Loosdrecht MCM (2015) Enrichment of DNRA
bacteria in continuous culture. The ISME Journal 9:2153-2161.
van de Graaf AA, Mulder A, de Bruijn P, Jetten MSM, Robertson LA, Kuenen JG (1995)
Anaerobic oxidation of ammonium is a biologically mediated process. Appl Environ Microbiol
61:1246-1251.
van Hulle SWH, Callens J, Mampaey KE, van Loosdrecht MCM, Volcke EIP (2012) N2O and
NO emissions during autotrophic nitrogen removal in a granular sludge reactor – a simulation
study. Environ Technol 33:2281-2290. doi:10.1080/09593330.2012.665492
Williams CJ, JaffĂŠ R, Anderson WT, Jochem FJ (2009) Importance of seagrass as a carbon
source for heterotrophic bacteria in a subtropical estuary (Florida Bay). Estuarine Coast Shelf S
85:507-514. doi:10.1016/j.ecss.2009.09.019
Zhang L, Altabet MA, Wu T, Hadas O (2007) Sensitive measurement of NH4+ 15N/14N (δ15NH4+)
at natural abundance levels in fresh and saltwaters. Anal Chem 79:5297-5303.
doi:10.1021/ac070106d

105

CHAPTER 4
DRAMATIC SHIFTS IN N CYCLING DRIVE HAB FORMATION AND MITIGATION OF
N LOADING IN SANDUSKY BAY, LAKE ERIE
ABSTRACT
Recent water crises in Lake Erie point to an urgent need for greater understanding of
harmful algal blooms (HABs) and their drivers. While much of the focus on the control of HABs
in Lake Erie is on phosphorus (P) management, nearshore areas such as Sandusky Bay become
limited by nitrogen (N) in the summer and are characterized by distinct HAB compositions (i.e.,
Planktothrix over Microcystis). We hypothesized that microbial N removal processes drive rapid
N depletion in Sandusky Bay, and N limitation stimulates N fixation that supports HAB
persistence. Stable isotope tracer approaches were employed to quantify rates of denitrification,
anammox, N2O production, and N fixation. During periods of high discharge and nutrient
delivery from the Sandusky River, Sandusky Bay acted as a conduit for nutrient loading into
Lake Erie, which could stimulate offshore HABs. These periods were also exemplified by high
N2O emissions from the Bay, as substrates for denitrification and nitrification were abundant. As
water residence time increased throughout the season, N removal processes drove marked
decreases in NO3- and the development of N limitation. Denitrification, anammox, and N 2O
production made up 84, 14, and 2 % of total N removal, respectively. N fixation was active
under N limitation (maximum 2.2 Âľmol N L-1 h-1), representing bioavailable N generation that
could indirectly support Planktothrix blooms in late summer. Dramatic shifts in N availability in
Sandusky Bay will likely be intensified by climate change, and the observed trends can inform
management of HABs and N2O production in Lake Erie.

106

INTRODUCTION
Harmful algal blooms (HABs) are increasing in frequency on a global scale and are
stimulated by excessive nutrient loading to aquatic systems (Bricker et al. 2008; Heisler et al.
2008). Lake Erie, in particular, has been subject to increased incidence and expansion of
cyanobacterial HABs in recent years (Michalak et al. 2013; Ho and Michalak 2015, Bullerjahn et
al. 2016). These blooms are dominated by cyanobacteria that accumulate the powerful
hepatotoxin, microcystin (Carmichael and Boyer 2016). cHABs are not a monoculture, and
cHAB taxa often show large temporal and spatial variability (Bozarth et al. 2010; Bridgeman et
al. 2012). A greater knowledge of the drivers of cHABs, including nutrient cycling, will allow
for better prediction and management in Lake Erie and other ecosystems.
Phytoplankton biomass and cHABs in Lake Erie have historically been correlated with P
loading from river inflows (Kane et al. 2014; Kim et al. 2014). Calls to control eutrophication in
Lake Erie have proposed targets for reduced P loading but have largely ignored N (Scavia et al.
2014). However, there is a growing dialogue surrounding the dual management of N and P in
lacustrine systems (Gobler at al. 2016; Paerl et al. 2016), particularly as co-limitation of
phytoplankton growth by both N and P has been demonstrated in the late summer in Lake Erie
(Moon and Carrick 2007; Chaffin et al. 2013; Monchamp et al. 2014; Steffen et al. 2014a; Davis
et al. 2015). Further, N concentration and speciation influences toxin production by cHABs
(Davis et al. 2010; Horst et al. 2014; Monchamp et al. 2014; Davis et al. 2015). Delineating the
role of N in controlling HABs, therefore, involves investigation of not only total N availability
but also spatial and temporal variation in multiple species of N.
Several N transformations may drive variation in N availability. Denitrification
permanently removes dissolved inorganic N (DIN) through the stepwise reduction of nitrate

107

(NO3-) to dinitrogen gas (N2; Knowles 1982). Anammox is a competing N removal process that
converts nitrite (NO2-) and ammonium (NH4+) to N2, but this process has not been extensively
studied in freshwater environments (Yoshinaga et al. 2011; Zhu et al. 2013). N removal
processes are expected to be particularly active in aquatic systems that receive high DIN inputs
(Seitzinger et al. 2006). N removal can be accompanied by release of the potent greenhouse gas,
nitrous oxide (N2O), which is produced by both denitrification and nitrification (Wrage et al.
2001). The global warming potential of N2O is 300 times that of carbon dioxide over a 100-year
time span, and its release could exacerbate climate change (IPCC 2007). N fixation, conversely,
is capable of generating bioavailable N when DIN is scarce, and its activity has been inferred in
Lake Erie (MacGregor et al. 2001; Monchamp et al. 2014; Steffen et al. 2014a; Davis et al.
2015) but not quantified for several decades (Howard et al. 1970).
While the colonial cyanobacterium Microcystis dominates the HAB community in
offshore regions of Lake Erie, filamentous Planktothrix has been shown to persist in N-limited
bays and tributaries (Conroy et al. 2007; Davis et al. 2015). Both cHAB taxa are incapable of N
fixation and require combined N for growth. Planktothrix, in particular, is a superior scavenger
for N (Conroy et al. 2007) and responds strongly to additions of DIN (Donald et al. 2011; 2013).
Under N limitation, Planktothrix may be able to outcompete other phytoplankton for small N
inputs or rely on a commensal relationship with diazotrophs. The persistence of Planktothrix in
nearshore zones likely operates under a fundamentally different paradigm than offshore
Microcystis blooms, and mitigation may require attention to the distinct biogeochemical
functioning of these genera in the nearshore vs. offshore.
Sandusky Bay, a south shore embayment of Lake Erie, serves as an ideal location in
which to examine the relationship between N cycling and cHABs. This system is hypereutrophic

108

(Ostrom et al. 2005; Davis et al. 2015), with the cyanobacterium Planktothrix dominating
phytoplankton biomass from May to October (Davis et al. 2015). Growth of Planktothrix in
Sandusky Bay is stimulated by additions of NO3-, NH4+, and urea, indicating that formation of
cHABs in this system relies on a supply of bioavailable N (Chaffin and Bridgeman 2014; Davis
et al. 2015). However, Sandusky Bay experiences large variations in NO 3- concentrations and
N:P ratios throughout the summer (Davis et al. 2015), suggesting that the dynamic N cycling
may influence cHAB formation in this system. Evaluating the mechanisms that promote the
persistence of Planktothrix in this system will benefit from an examination of N removal
processes and N fixation that directly influence the availability of DIN. A thorough
understanding of these processes will also inform the capacity for Sandusky Bay to act as a
nutrient filter, which may curtail the formation of cHABs in offshore regions of Lake Erie.
As N cycling processes have been largely ignored in lieu of P in Lake Erie, we sought to
examine of the potential role of N in cHAB formation. The objectives of this study in Sandusky
Bay were to (1) determine the driving factors influencing N limitation, namely denitrification,
anammox, and N2O production, (2) examine N fixation as a possible source of bioavailable N
under N limitation, and (3) determine how HAB formation and persistence relate to N cycling.
We hypothesized that microbial N removal processes drive rapid N depletion in Sandusky Bay,
and N fixation that occurs under N limitation allows bloom persistence.

METHODS
Field sampling
Sample collection occurred weekly (water column nutrient and chlorophyll (chl) a
concentration analyses) or monthly (water column N 2O analyses, water column and sediment N

109

cycling assays) from May to October, 2015 through partnership with the Ohio Department of
Natural Resources (ODNR). Sampling occurred at six stations: two stations in the inner portion
of Sandusky Bay (ODNR4, ODNR6), three stations in the outer portion of Sandusky Bay
(ODNR2, ODNR1, and EC 1163, hereafter 1163), and one station directly outside Sandusky Bay
in the western basin of Lake Erie (Bells; Figure 16). Tributary discharge data from the primary
water source to Sandusky Bay, the Sandusky River, were obtained from the USGS stream
monitoring station near Fremont, OH (site 04198000; Figure 16). Water residence time in
Sandusky Bay was estimated by dividing the Bay volume (1.6 - 2.6 m depth, 162 km2 area;
Richards and Baker 1985) by the Sandusky River discharge rate.

Figure 16. Sampling locations in Sandusky Bay (circles) and Sandusky River monitoring station
(USGS site 04198000, cross).

110

At each sampling location, water column physical and chemical parameters (pH,
conductivity, temperature, dissolved oxygen) were measured and recorded using a YSI 600QS
sonde (YSI Inc., Yellow Springs, OH). Water samples were collected by Van Dorn bottle at 1 m
depth for analysis of N2O, NO3-, NH4+, phosphate (PO43-), and chl a concentrations. Samples for
N2O concentrations and isotopic analysis were transferred into 60 and 250 mL glass serum
bottles, respectively, by filling from the bottom to overflowing and sealing with no headspace
with butyl rubber septa. Biological activity was halted by adding saturated HgCl2 solution to a
final concentration of 0.4 % by volume. Samples for dissolved nutrient analysis were filtered
(0.2 Îźm) and frozen upon return to lab. Samples for determination of chl a concentrations were
collected on 0.2 Âľm polycarbonate membrane filters and frozen. Station 1163 has an extensive
monitoring history and was chosen for additional water and sediment assays. An additional 20 L
carboy was filled with surface water from station 1163 for N fixation assays and sediment
incubations.
Sediment cores for the evaluation of denitrification, anammox, and N2O production rates
by the isotope pairing technique (IPT) were collected at station 1163 using a modified piston
corer as described by Davis and Steinman (1998). Intact sediment cores were collected in
polycarbonate tubes (7 cm i.d.) to a depth of 25 cm. Water (1 cm) was maintained overlying the
sediment during transport to preserve redox gradients. Upon return to the lab, water from station
1163 was gently added to cores to a depth of 20 cm. Cores were pre-incubated for 12 h in the
dark at in situ temperature under gentle aeration to maintain oxic conditions in the overlying
water.

111

Isotope tracer assays
Following pre-incubation of sediment cores, a sample was taken from the overlying water
for DIN concentration analysis (NO3-, NO2- and NH4+), filtered through a precombusted GF/F
filter and frozen until analysis. 15N-NO3- (100 Âľmol L-1) was then added to the overlying water in
each core. Cores were then capped and gently stirred throughout the duration of the experiment.
An initial equilibration period was employed to allow homogenization of NO 3- between the
overlying water and the NO3- reduction zone in the sediment porewater (Dalsgaard et al. 2000).
Cores were sacrificed in triplicate or quadruplicate at intervals of 0, 3 or 4, and 6 h. Dissolved O2
in the overlying water was monitored to evaluate the maintenance of oxic conditions throughout
the incubation using a YSI 600QS sonde (YSI Inc., Yellow Springs, OH). A final sample for
DIN analysis was collected when each core was sacrificed and processed as described above.
Samples for the determination of 15N2 were collected according to Hamilton and Ostrom
(2007); briefly, dissolved gases were equilibrated with a He atmosphere, and the headspace was
transferred into a pre-evacuated 12 mL Exetainer. Samples for analysis of dissolved N2
concentrations were siphoned into 12 mL Exetainers to overflowing and amended with 200 ÂľL
of saturated ZnCl2 solution to halt biological activity. All Exetainers were stored underwater at
room temperature to minimize diffusion of atmospheric N2 during storage. Samples for analysis
of the 15N2O and N2O concentrations were siphoned into 250 and 60 mL serum bottles,
respectively, to overflowing and sealed without a headspace with a butyl rubber septum.
Biological activity was halted by adding saturated HgCl2 solution to a final concentration of
0.4 % by volume.
N fixation assays were conducted by the dissolution method, which involves the addition
of 15N2-equilibrated water to a water sample rather than a 15N2 bubble (Großkopf et al. 2012).

112

Preparation of 15N2-equilibrated water involved sparging water in a serum bottle equipped with a
butyl rubber septum with He to remove ambient N2, followed by injection of 15N2 (98 % atom
fraction, Sigma-Aldrich Lot #MBBB0968V) while maintaining atmospheric pressure with a vent
needle. Water from station 1163 was transferred into 1.18 L serum bottles and amended with
15N

2-equilibrated

water to a final dissolved atom fraction of 1.14 - 2.33 %. Bottles were inverted

100 times to ensure uniform mixing and then incubated at in situ light and temperature
conditions for 24 h. Samples were then vacuum filtered through precombusted GF/F filters.
Filters were then dried at 60° C, acidified with 10 % HCl to remove carbonates, and dried again.
To evaluate the potential contamination of 15N2 gas with 15NO3- or 15NH4+ (Dabundo et
al. 2014), a mass scan of the isotopically enriched gas was performed on an Isoprime isotope
ratio mass spectrometer (IRMS; Elementar Americas, Inc., Mount Laurel, NJ). The mass scan
revealed that potential impurities made up < 1 % of the enriched gas, and maximum
contamination reported by Dabundo et al. (2014) could have made up less than 5 % of measured
N fixation rates if all available 15NO3- and 15NH4+ was assimilated.

Analyses
Concentrations of NO3- + NO2-, NO2-, NH4+, and PO 43- were measured on field-filtered
sample water using standard U.S. EPA methods (353.1, 353.2, 350.1, and 365.1, respectively) on
a SEAL Analytical QuAAttro continuous segmented flow analyzer (SEAL Analytical Inc.,
Mequon, WI). NO3- concentration was determined as the difference between NO3- + NO2- and
NO2-. Seven known concentration standard solutions (including 0) were used for the standard
curve (R2 > 0.999), and every-tenth sample was spiked with a known amount of analyte to ensure
high accuracy and precision throughout the analysis (> 95 % recovery). Samples with

113

concentrations exceeding the highest standard were diluted and reanalyzed. Values were
averaged over two or three replicates. Method detection limits were 0.165, 0.044, 0.558, and
0.044 Âľmol L-1 for NO3- + NO2-, NO2-, NH4+, and PO43-, respectively. Extractive chl a
concentration was measured following Welschmeyer (1994). Filters containing phytoplankton
seston were extracted with 90% aqueous acetone overnight at -20 °C followed by measurement
of the clarified extract by fluorometry (model TD-700, Turner Designs, Sunnyvale, CA).
Prior to N2O concentration analysis, a headspace of 20 mL He at atmospheric pressure
was created in each 60 mL bottle by removing 20 mL water by syringe while simultaneously
adding He. Serum bottles were allowed to equilibrate under gentle shaking for at least 12 h prior
to analysis. The headspace was then analyzed by GC-ECD (Shimadzu Greenhouse Gas Analyzer
GC-2014, Shimadzu Scientific Instruments, Columbia, MD) for N2O concentration. The
dissolved concentration was calculated based on the headspace equilibrium concentration
(Hamilton and Ostrom 2007). Diffusive atmospheric emissions of N2O (F) were calculated by
equations 19 and 20:
F = kw(Cw – Ca)

(19)

where Cw is the dissolved N2O concentration at the surface (measured at 1 m depth) and Ca is the
calculated dissolved N2O concentration in equilibrium with the atmosphere (Wanninkhof 1992;
Walker et al. 2010). The gas transfer coefficient (kw, in m s-1) is calculated as:
kw = 0.31U102(Sc/600)-1/2

(20)

where Sc is the Schmidt number for N2O determined by the kinematic viscosity of freshwater
divided by the diffusion coefficient of N2O (Wanninkhof 1992) and U10 is the wind speed 10 m
above the surface determined using the measured wind speed 1 m above the surface and the

114

power law relationship outlined in Walker et al. (2010). Wind speed data were collected from
Erie-Ottawa International Airport (KPCW), located 2 km north of ODNR6.
The isotopic composition of N2O was analyzed upon introduction of sample water into an
enclosed 0.75 L glass vessel that was previously purged of atmospheric air using a gentle flow of
He. Dissolved gases were subsequently stripped from the water by sparging the sample with He
(Sansone et al. 1997), which carried sample gases into a Trace Gas sample introduction system
interfaced to an IRMS. The relative abundance of a stable isotope within a particular material or
reservoir is reported in standard delta notation:
𝛿=

𝑅𝑠𝑎𝑚 −𝑅𝑠𝑡𝑑
𝑅𝑠𝑡𝑑

×1000

(21)

where Rsam is the isotope ratio of the sample, R std is the isotope ratio of the standard, and  is
reported as per mil (‰). The isotopic composition of the N atom in the central position (α) and
the N atom in the outer position (β) was analyzed separately. The difference between 15Nι and
15Nβ in N2O, referred to as site preference (SP), was also calculated. SP is a conservative tracer
of N2O production mechanism and can be used to quantitatively apportion N 2O production
sources, primarily through isotope mixing models that consider denitrification (SP = -10 to 0 ‰)
and nitrification via hydroxylamine oxidation (SP = 33 to 37 ‰) as end members (Toyoda et al.
2005; Sutka et al. 2006; Frame and Casciotti 2010). Analytical reproducibility for replicate
samples was 0.5 ‰ for bulk 15N and 18O, 0.75 ‰ for 15Nα and 15Nβ, and 1.3 ‰ for SP.
Concentrations of dissolved N2 were analyzed by membrane inlet mass spectrometry
(MIMS; Eyre et al. 2002). The isotopic composition of N 2 was analyzed by introducing the
sample to an evacuated 800 µL sampling loop and then onto a packed molecular sieve (5 Å)
column (Alltech, Inc., Deerfield, IL) using He carrier gas within a gas chromatograph (HP-5980,

115

Hewlett Packard, Ramsey, MN) interfaced to an Isoprime IRMS. Analytical reproducibility of
standards was 0.3 ‰.
The concentration and isotopic composition of particulate organic matter (POM) from N
fixation assays was analyzed by scraping the contents of dried and acidified filters into tin cups
and introducing samples to an elemental analyzer interfaced to an Isoprime IRMS. Analytical
reproducibility of standards was 0.2 ‰.

Modeling
Denitrification, anammox, and N2O production rates were calculated by the IPT anaN2O
(Hsu and Kao 2013), which builds on the R-IPT (Risgaard-Petersen et al. 2003) by enabling
quantification of N2O production simultaneously with denitrification and anammox. N fixation
rates were calculated according to Montoya et al. (1996).
Statistical modeling was carried out in R (version 3.2.4). Correlations between nutrient
concentrations and discharge as well as N2O isotopes by date were analyzed by linear regression.
Potential differences in denitrification, N2O production, and N fixation by date were analyzed by
one-way ANOVA.

RESULTS
NO3- concentrations ranged from below detection to > 600 Âľmol L-1 across sites during
the sampling period (Figure 17a). The greatest NO3- concentrations occurred in June and July,
followed by a decline to non-detectable levels in late August that continued through October.
The magnitude of these shifts was greater for the inner bay stations (ODNR4, ODNR6) than the
outer bay and coastal Lake Erie stations (ODNR1, ODNR2, 1163, Bells). NO3- concentration

116

was positively correlated with Sandusky River discharge (Fig. 2f; df = 81, R2 = 0.18,
p < 0.0001). NH4+ concentration ranged from below detection to 17.5 Âľmol L -1 across sites
during the sampling period (Figure 17b). The greatest NH4+ concentrations occurred in the inner
bay, but these spikes occurred episodically throughout the sampling season and were not
correlated with Sandusky River discharge or with NO 3- concentration. PO43- concentrations
ranged from below detection to 4.25 Âľmol L -1 across sites during the sampling period (Figure
17c). Higher PO43- concentrations generally occurred in the inner bay, and overall concentrations
were correlated with Sandusky River discharge (df = 81, R2 = 0.06, p = 0.01). The ratio of DIN
to dissolved inorganic P (N:P; NO3- + NH4+ to PO43-) was variable throughout the sampling
period, exceeding 10,000 in July and falling to values below Redfield stoichiometry (16) in late
August and September (Figure 17d). Chl a concentrations ranged from 3.5 to nearly 150 Âľg L-1
across sites during the sampling period (Figure 17e). Maximum chl a concentrations occurred in
late August to early September, approximately one month after the peak in NO 3- and PO43concentrations. Water residence time was as low as eight days when discharge from the
Sandusky River was greatest in June and Early July (Figure 17f). As discharge decreased in late
July and remained low for the remainder of the summer and early fall, water residence time
increased to several months.

117

a

ODNR4
ODNR6
ODNR2
ODNR1
1163
Bells

400

200

Site
ODNR4
ODNR6
ODNR2
ODNR1
1163
Bells

15

NH4+ (mmol L-1)

NO3- (mmol L-1)

b

Site

600

10

5

0

0
Jul

Aug

Sep

Jul

Oct

c

Sep

d

Site
ODNR4
ODNR6
ODNR2
ODNR1
1163
Bells

3

Oct

2

Site
ODNR4
ODNR6
ODNR2
ODNR1
1163
Bells

10000
1000

N:P

PO43- (mmol L-1)

4

Aug

Date (2015)

Date (2015)

100
20
5

1

1
0
Jul

Aug

Sep

Oct

Jul

Aug

Date (2015)

e

Oct

f

Site
ODNR4
ODNR6
ODNR2
ODNR1
1163
Bells

100

50

300

Discharge (m3 d-1)

Chl a (mg L-1)

150

Sep

Date (2015)

200

100

0

0
Jun

Jul

Aug

Sep

Oct

Jun

Date (2015)

Jul

Aug

Sep

Oct

Date (2015)

Figure 17. (a) NO3-, (b) NH4+, (c) PO43-, (d) N:P (NO3- + NH4+ : PO43-) ratio, and (e) chlorophyll
a concentration in six sites in Sandusky Bay in 2015. (f) Discharge of the Sandusky River at
Fremont obtained from USGS monitoring station 04198000. The dotted line in (d) indicates a
N:P ratio of 16 (note log scale on y axis).
118

Sediment denitrification rates at station 1163 ranged from 10.02 - 64.81 Âľmol N m-2 h-1
over all sampled dates, decreasing over time (Figure 18a). Denitrification rates varied
significantly by date (ANOVA, F = 6.53, df = 24, p < 0.01), and rates on June 22 were
significantly higher than other dates (Tukey HSD post-hoc test). Sediment anammox activity was
detected on all sampling dates but not in all replicate sediment cores. Anammox rates at station
1163 ranged from 0.52 - 8.10 Âľmol N m-2 h-1 across sampled dates, displaying no clear temporal
trend (Figure 18b). The majority of measured anammox rates were less than 7 Âľmol N m -2 h-1,
with the exception of two cores on July 27 that displayed anammox rates of 15.15 and 30.75
Âľmol N m-2 h-1. Owing to high variability among replicates, anammox rates did not vary
significantly by date (ANOVA, F = 2.60, df = 24, p = 0.08). N2O production rates at station 1163
ranged from 0.09 - 2.34 Âľmol N m-2 h-1 across sampled dates, decreasing over time (Figure 18c).
N2O production rates varied significantly by date (ANOVA, F = 6.85, df = 24, p < 0.01), and
rates on June 22 were significantly higher than other dates (Tukey HSD post-hoc test).
Denitrification, anammox, and N2O production made up an average of 84 %, 14 %, and 2 % of
total N removal, respectively.
N2O was supersaturated in the surface water by more than 10 % in all sites on June 22
and July 27 and for three sites (ODNR4, 1163, Bells) on August 31 (Figure 19a). The greatest
degree of N2O supersaturation (up to 676 %) occurred on June 22, corresponding to the date with
the greatest measured N2O production rate. N2O saturation was positively correlated with NO 3concentration (df = 16, R2 = 0.42, p < 0.01). N2O fluxes to the atmosphere were positive at all
sites on all sampled dates except ODNR6 and ONDR2 on August 31, with highest fluxes
occurring on June 22 (Figure 19b).

119

Denitrification rate (mmol N m-2 h-1)

a A
100

75

B
50

B
B

25

0
June 22

July 27

August 31

October 12

Anammox rate (mmol N m-2 h-1)

Date (2015)
20

A

b

15

10

A

A

5

A

0
June 22

July 27

August 31

October 12

N2O production rate (mmol N m-2 h-1)

Date (2015)

c

A

4

3

2

B
1

B

B

0
June 22

July 27

August 31

October 12

Date (2015)

Figure 18. (a) Denitrification, (b) anammox, and (c) N2O production at Sandusky Bay station
1163 in 2015. Rates were measured by the isotope pairing technique. Error bars represent +1 SD.
Letters indicate groupings of statistically significant differences at p < 0.05.

120

a

ODNR 4
ODNR 6
ODNR 2
ODNR 1
1163
Bells

400

Station
ODNR 4
ODNR 6
ODNR 2
ODNR 1
1163
Bells

30

N2O Flux (mmol N m-2 d-1)

600

N2O Saturation (%)

b

Station

200

20

10

0

0
May 28

June 22

July 27 August 31October 12

May 28

June 22

Date (2015)

July 27 August 31October 12

Date (2015)

Figure 19. (a) N2O saturation and (b) diffusive emissions of N2O to or from the atmosphere at six
stations in Sandusky Bay. A saturation of 100% is represented by the horizontal dotted line in
(a). Data on May 28 and October 12 were available only for station 1163. Error bars represent +1
SD.
The 15N and 18O of N2O did not exhibit variation as a function of date or station
(Figure 20a). N2O SP data clustered within two distinct groups: (a) four samples from ODNR4
and ODNR1 with values > 45 ‰ and (b) the majority of samples with values ranging within 7.0
to 34.3 ‰ (Figure 20b). N2O samples in the former group were not associated with anomalous
15N or 18O values (Figure 20a) but were associated with the highest 15Nι and lowest 15Nβ
values of the dataset (Figure 20c). SP values in the latter group increased throughout the
sampling period (linear regression, df = 14, R2 = 0.20, p = 0.048), but this increase was not
accompanied by an increase in 15N, 18O, 15Nι, or 15Nβ (Figure 20).
Water column N fixation rates at station 1163 ranged from 0.74 – 2.16 µmol N L-1 h-1
over all sampled dates (Figure 21). N fixation rates varied significantly by date (ANOVA, F =
48.81, df = 8, p < 0.0001), and rates on June 22 and October 12 were significantly higher than
those on July 27 and August 31 (Tukey HSD post-hoc test).

121

d15N, d18O (‰)

60

a

Station
ODNR4
ODNR6
ODNR2
ODNR1
1163
Bells

40

20

0

Jun

Jul

Aug

Sep

Oct

Date (2015)

b

Station
ODNR4
ODNR6
ODNR2
ODNR1
1163
Bells

SP (‰)

50
40
30
20
10
Jun

Jul

Aug

Sep

Oct

Date (2015)

d15Na, d15Nb (‰)

c

Station
ODNR4
ODNR6
ODNR2
ODNR1
1163
Bells

20

0

-20

Jun

Jul

Aug

Sep

Oct

Date (2015)

Figure 20. N2O isotopomers in Sandusky Bay. (a) 15N and 18O in black and gray, respectively,
(b) SP values, and (c) 15Nι and 15Nβ in gray and black, respectively.

122

N fixation rate (mmol N L-1 h-1)

B

2.5

B

2.0

1.5

1.0

A

A

0.5

0.0
June 22

July 27

August 31

October 12

Date (2015)

Figure 21. N fixation rates at Sandusky Bay station 1163 in 2015. Error bars represent +1 SD.
Letters indicate groupings of statistically significant differences at p < 0.05.

DISCUSSION
Nutrient stoichiometry
Sandusky Bay displayed considerable shifts in nutrient concentrations, N:P ratios, and chl
a concentrations, indicative of a system characterized by dynamic changes in hydrology and
biogeochemical activity. Maximum chl a concentrations (> 100 Âľg L-1) were similar to other
hypereutrophic systems (Zhang et al. 2011; Wheeler et al. 2012; Steffen et al. 2014b), as were
large swings in NO3- concentrations (Xu et al. 2010; McCarthy et al. 2016). As discharge from
the Sandusky River decreased throughout the summer, N:P ratios fell from a maximum of over
10,000 to below 16, the threshold for N limitation. This decline was largely driven by decreases
in NO3- concentration, as the range in PO43- and NH4+ concentration was comparatively narrow
(Figure 17). Depletion of NO3- can be attributed to both assimilatory and dissimilatory NO 3reduction. If phytoplankton were solely responsible for the decline in NO 3-, nutrients would be
expected to be drawn down in molar proportions of approximately 16N:1P (Redfield 1934;
Sterner and Elser 2002). However, N:P ratios in Sandusky Bay fell sharply throughout the
summer, while PO43- concentrations were relatively constant by comparison. Although this trend

123

could be influenced by luxury uptake of N by phytoplankton and internal P loading from
sediments (Filbrun et al. 2013; McCarthy et al. 2016), the dramatic depletion in DIN compels an
examination of microbial N removal processes as a major mechanism for N drawdown in
Sandusky Bay.

N removal processes
Denitrification and anammox were evaluated as processes likely responsible for
consuming appreciable quantities of NO3- in Sandusky Bay. The primary N removal mechanism
in Sandusky Bay was denitrification, which made up an average of 84 % of total N removal
across sampling dates. Denitrification rates were positively associated with N availability, which
has also been observed in estuaries, lakes, and continental shelves (Seitzinger et al. 2006).
Denitrification rates measured in Sandusky Bay are among the highest rates reported for the
Laurentian Great Lakes (McCarthy et al. 2007; Small et al. 2014; 2016), reinforcing the capacity
for Sandusky Bay and other shallow embayments and river mouths to act as hotspots of N
removal in the Great Lakes system.
Anammox activity in Sandusky Bay was highly variable, even among replicate sediment
cores from the same site and date. The range of observed anammox rates (0 - 30.75 Âľmol N m-2
h-1) is similar to those reported for shallow estuarine environments (Dong et al. 2009; 2011; Hsu
and Kao 2013). Consequently, marked variability may be characteristic of anammox activity in
freshwater environments, even across small spatial and temporal scales (Yoshinaga et al. 2011;
Zhu et al. 2013; 2015). Anammox made up an average of 14 % of total N removal across the
sampling period, indicating anammox activity in Sandusky Bay, with a mean depth of 1.6 - 2.6 m
(Richards and Baker 1985), may be typical of shallow estuarine (Thamdrup and Dalsgaard 2002;

124

Dalsgaard et al. 2005) and freshwater systems (Schubert et al. 2006; McCarthy et al. 2016). A
recent study using 16s RNA showed that anammox genera were active across western and
central basins of Lake Erie (Small et al. 2016). Together, anammox rate and genomic data
demonstrate that although denitrification is generally the dominant N removal mechanism, there
is potential for anammox to be an significant N removal pathway in Sandusky Bay and other
nearshore regions within the Great Lakes.
While denitrification and anammox were the primary drivers of N depletion in Sandusky
Bay, water residence time was evaluated as an additional contributing factor. Estimates of water
residence time when discharge was greatest in June and early July was as low as eight days. By
late July, however, water residence time increased to several months and continued to increase as
discharge remained low for the remainder of the summer and early fall. Although N removal
rates were greatest in late June when NO3- concentrations were highest, the capacity for N
removal to substantially deplete NO3- was hindered by the short residence time within the Bay. N
removal rates in late June were capable of consuming approximately 2 % of ambient NO 3- daily,
so water residence times of 1 - 2 weeks would allow for the removal of only ~1/4 of NO 3- from
Sandusky Bay prior to release into Lake Erie. Export of nutrients from Sandusky Bay to Lake
Erie is exemplified by increases in NO3- concentrations at the Lake Erie Bells station
concomitant with increases in NO3- in the Bay (Figure 17a). The development of N limitation in
mid-July through August occurred when water residence time lengthened to several months,
which provided the opportunity for N removal processes to extensively consume NO3-. Thus,
Sandusky Bay acts as a conduit for N delivery from the Sandusky River to Lake Erie under short
water residence time but acts as a sink for NO3- under long water residence time. This pattern is

125

consistent with observations in other river mouths and coastal embayments of the Great Lakes
(McCarthy et al. 2007; Larson et al. 2013).

N2O production and emissions
N2O displayed a wide range in production rate, concentration, and atmospheric flux on a
seasonal basis (Figure 18c, 19). The greatest degree of N2O saturation and highest N2O
production rate occurred when NO3- concentrations were highest, suggesting that nutrient
delivery from the Sandusky River drives N2O production in Sandusky Bay. The range of
observed N2O fluxes are typical of other inland lakes (Whitfield et al. 2011; Yang et al. 2015 and
references therein). However, the maximum observed atmospheric emissions of N 2O from
Sandusky Bay were remarkably high, considering the highest published global emissions from
lacustrine environments occur in littoral zones dominated by emergent macrophytes (Wang et al.
2006) and in stratified systems that develop hypoxia or anoxia (Salk et al. 2016; Wenk et al.
2016). The influx of NO3- from the Sandusky River coupled with a shallow water column and
high rates of N cycling drives substantial benthic N 2O production and release (Figure 17a, 18c,
19). Moreover, active N cycling in this and other shallow embayments in the Great Lakes
(McCarthy et al. 2007; Small et al. 2016) suggests that emissions from these locations may make
up the majority of N2O emissions from the Great Lakes.
The isotopic composition of N2O in Sandusky Bay provided insight into its microbial
production pathways. With the exception of four samples that were characterized by remarkably
high SP values (> 45 ‰), N2O produced in Sandusky Bay appeared to be derived from a mixture
of denitrification (SP = -10 to 0 ‰) and nitrification (SP = 33 to 37 ‰) pathways (Toyoda et al.
2005; Sutka et al. 2006; Frame and Casciotti 2010), with SP values ranging from 7 to 34 ‰. SP

126

values increased throughout the sampling period, indicating either a greater proportional
contribution of nitrification to N2O production or reduction of N2O to N2 by denitrification.
Concomitant increases in 15N and 18O with SP are expected when N2O reduction occurs
(Westley et al. 2006; Jinuntuya-Nortman et al. 2008; Well et al. 2012; Wenk et al. 2016), but this
trend was not apparent in Sandusky Bay (Figure 20). Consequently, we conclude that increasing
SP is associated with an increased contribution of nitrification to N 2O production, which ranged
from approximately 50 % up to 100 % of N2O production from the beginning to the end of the
sampling season. This pattern can be attributed to a decrease in overall denitrification activity
throughout the sampling season combined with a small proportion of denitrification end-product
released as N2O (2 % of total N removal, on average). N2O production in Sandusky Bay is thus
governed by variation in denitrification activity, which allows production from nitrification to
dominate when denitrification rates are low. Nitrification has previously been observed as the
primary N2O production mechanism in Muskegon Lake, another lake in the Great Lakes basin
(Salk et al. 2016) as well as several ocean environments (Charpentier et al. 2007; Fujii et al.
2013), illustrating the importance of nitrification as a source of N 2O fluxes from aquatic systems.
The high SP values (> 45 ‰) observed in four N2O samples could result from N2O
reduction or an unusual production signal. Because these samples were not associated with
elevated 15N or 18O and reduction is associated with substantial fractionation in the isotopic
ratios of SP, N, and O (Lewicka-Szczebak et al 2015 and references therein), reduction was
unlikely to have appreciably influenced N2O SP in these samples. High SP, on the other hand,
was associated with high 15Nι and low 15Nβ. This observation is consistent with Breider et al.
(2015), who found that N2O produced in the oxygen minimum layer in the western North Pacific
was characterized by higher SP values than in the oxic layer. This increase in SP was

127

accompanied by a shift toward higher 15Nι and lower 15Nβ values with no overall change in
15N. Similarly, a pattern of increasing SP, increasing 15Nι, and decreasing 15Nβ has also been
observed in purified fungal nitric oxide reductase despite observations of constant SP in fungal
cultures (Sutka et al. 2008; Yang et al. 2014). These authors proposed that SP generally does not
isotopically fractionate under in situ conditions because the internal concentration of the
substrate, nitric oxide, is held constant in cells. If, however, the substrate concentration within
cells is not held constant, as in enzyme cultures, SP could vary. It is possible that during periods
of hydrologic and biogeochemical transition in Sandusky Bay, the capacity for denitrifiers and
nitrifiers to produce N2O changes and intracellular concentrations of hydroxylamine and/or nitric
oxide fluctuate. Therefore, changes in intracellular substrate concentrations could result in the
unusually high SP values observed at ODNR4 and ODNR1 in August and September.

N fixation
N fixation in the water column of Sandusky Bay was an active process not only during
periods of N limitation but also when bioavailable N was replete. The greatest rates of N fixation
occurred in July and August as N limitation developed in Sandusky Bay. Planktothrix made up
the majority of phytoplankton biomass during this period, suggesting that N fixation may
generate bioavailable N that is utilized by non-diazotrophs. Additionally, N fixation was detected
when NO3- and NH4+ was abundant in June, an unusual observation considering high NO 3concentrations are expected to inhibit N fixation (Holl and Montoya 2005). Trace metal colimitation has been proposed as a mechanism for the inhibition of NO 3- uptake in offshore Lake
Erie (North et al. 2007; Havens et al. 2012), although common metal cofactors between nitrate
reductase and nitrogenase render the simultaneous suppression of NO 3- uptake and stimulation of

128

N fixation unlikely (Romero et al. 2013). However, alleviation of trace metal colimitation by
iron addition has been shown to influence NO3- uptake and N fixation differently between
nearshore and offshore zones in Lake Victoria (Guildford et al. 2003), and a similar mechanism
may operate in Sandusky Bay. The data in this study offer a contemporary depiction of N
fixation in an area of Lake Erie exemplified by active N cycling, establishing that N fixation is
active not only under N limitation but also under N-replete conditions. The occurrence of N
fixation also signals a mechanism for the generation of bioavailable N that could allow
Planktothrix to thrive under N limitation.

Harmful algal blooms
Blooms of Planktothrix were evident in Sandusky Bay, particularly during August and
early September (Figure 17e; T. Tuttle, personal communication). Sandusky River discharge
decreased in the mid-summer, leading to an increase in the water residence time in the Bay that
allowed N removal and N uptake processes to deplete nutrient concentrations. Chl a, a proxy for
total phytoplankton abundance (Becker et al. 2009, Millie et al., 2009, Davis et al. 2012),
reached maximum levels approximately one month after maximum NO 3- and PO43concentrations were observed (Figure 17). This pattern suggests that low flushing rates play a
positive role in the production of phytoplankton biomass, perhaps by creating a stable physical
environment in which phytoplankton can prosper (Michalak et al. 2013). Further, low nutrient
availability during this period coupled with highest measured N fixation rates indicates that N
fixation is integral to phytoplankton growth and also cHAB persistence. Analysis of nitrogenase
(nifH) transcripts in metatranscriptomic datasets from Sandusky Bay reveals that the pelagic Nfixing population was comprised largely of N-fixing Aphanizomenon and Dolichospermum (T.

129

Tuttle, personal communication). However, these diazotrophs were minor components of the
cyanobacterial assemblage, while Planktothrix was the major component of phytoplankton in the
summer, even during periods of DIN scarcity (Davis et al. 2015; T. Tuttle, personal
communication). The persistence of Planktothrix blooms in Sandusky Bay when DIN is low may
be supported by uptake of small inputs of N (Chaffin and Bridgeman 2014; Donald et al. 2011;
2013; Davis et al. 2015), perhaps from sediment NH 4+ regeneration (Paerl et al. 2011; McCarthy
et al. 2106), and/or a commensal relationship with diazotrophs that are active during this period.

Conclusions
In 2015, Sandusky Bay underwent a rapid and dramatic transition from excess N
abundance to N limitation by mid-summer, a pattern that has been observed in other years
(Conroy et al. 2007; Davis et al. 2015). During periods of short water residence time, Sandusky
Bay acted as a conduit for nutrient loading into the central basin of Lake Erie despite high N
removal rates. This period was also exemplified by high N 2O emissions from the Bay, owing to
the abundance of N substrates for denitrification and nitrification. As water residence time
increased, N removal (mainly via denitrification) drove marked decreases in NO3- concentrations
and the development of N limitation. N fixation was active during periods of N limitation,
representing a mechanism for bioavailable N generation that could indirectly support blooms of
Planktothrix that reached their maximum during this period. Discharge of nutrients from
Sandusky Bay into Lake Erie in the early summer could stimulate the formation of blooms
offshore, whereas filtering of nutrients later in the season could limit nutrient export.
Climate change-induced shifts in precipitation have the capacity to intensify the strong
transitions in N cycling in Sandusky Bay. Overall precipitation in the summer is decreasing in

130

the Lake Erie catchment, whereas incidences of extreme precipitation are increasing (Prein et al.
2016). Thus, discharge into Sandusky Bay is expected to become more episodic in the future,
with the hydrology of the Bay dominated by periods of long water residence time but punctuated
by pulses of high discharge and nutrient loading. These changing hydrologic patterns have
implications for N cycling and cHAB formation in Lake Erie. High N loading has the potential to
drive brief yet substantial N2O emissions from Sandusky Bay. Periods of low discharge will
likely stimulate cHAB formation in Sandusky Bay, while episodic high discharge events will
favor delivery of nutrients that may support cHABs in the central basin of Lake Erie and
exacerbate water quality issues downstream, such as hypoxia in the St. Lawrence Estuary
(Lehmann et al. 2009; Michalak et al. 2013).

131

LITERATURE CITED

132

LITERATURE CITED

Becker RH, Sultan MI, Boyer GL, Twiss MR, Konopko E (2009) Mapping cyanobacterial
blooms in the Great Lakes using MODIS. J Great Lakes Res 35(3):447–453.
doi:10.1016/j.jglr.2009.05.007
Bozarth CS, Schwartz AD, Shaparson JW, Colwell FS, Dreher TW (2010) Population turnover
in a Microcystis bloom results in predominately nontoxigenic variants late in the season. Appl
Environ Microbiol 76:5207-5213.
Breider F, Yoshikawa C, Abe H, Toyoda S, Yoshida N (2015) Origin and fluxes of nitrous oxide
along a latitudinal transect in western North Pacific: Controls and regional significance. Global
Biogeochem Cy 29:1014-1027. doi:10.1002/ 2014GB004977
Bricker SB, Longstaff B, Dennison W, Jones A, Boicourt K, Wicks C, Woerner J (2008) Effects
of nutrient enrichment in the nation’s estuaries: A decade of change. Harmful Algae 8:21-32.
doi:10.1016/j.hal.2008.08.028
Bridgeman TB, Chaffin JD, Kane DD, Conroy JD, Panek SE, Armenio PM (2012) From river to
lake: phosphorus partitioning and algal community compositional changes in western Lake Erie.
J Great Lakes Res 38:90-97.
Bullerjahn GS, McKay RM, Davis TW, Baker DB, Boyer GL, D’Anglada LV, Doucette GJ, Ho
JC, Irwin EG, Kling CL, Kudela RM, Kurmayer R, Michalak AM, Ortiz JD, Otten TG, Paerl
HW, Qin B, Sohngen BL, Stumpf RP, Visser PM, Wilhelm SW (2016) Global solutions to
regional problems: Collecting global expertise to address the problem of harmful cyanobacterial
blooms. A Lake Erie case study. Harmful Algae 54:223-238. doi: 10.1016/j.hal.2016.01.003
Carmichael WW, Boyer, GL (2016) Health impacts from cyanobacteria harmful algae blooms:
Implications for the North American Great Lakes. Harmful Algae 54:194-212.
Chaffin JD, Bridgeman TB, Bade DL (2013) Nitrogen constrains the growth of late summer
cyanobacterial blooms in Lake Erie. Advances Microbiol 3(6A):16-26.
doi:10.4236/aim.2013.36A003
Chaffin JD, Bridgeman TB (2014) Organic and inorganic nitrogen utilization by nitrogenstressed cyanobacteria during bloom conditions. J Appl Phycol 26:299-309.
doi:10.1007/s10811-013-0118-0
Charpentier J, Farias L, Yoshida N, Boontanon N, Raimbault P (2007) Nitrous oxide distribution
and its origin in the central and eastern South Pacific Subtropical Gyre. Biogeosciences 4:729741.

133

Conroy JD, Quinlan EL, Kane DD, Culver DA (2007) Cylindrospermopsis in Lake Erie: Testing
its association with other cyanobacterial genera and major limnological parameters. J Great
Lakes Res 33(3):519-535. doi:10.3394/0380-1330(2007)33[519:CILETI]2.0.CO;2
Dabundo R, Lehmann MF, Treibergs L, Tobias CR, Altabet MA, Molsander PH, Granger J
(2014) The contamination of commercial 15N2 gas stocks with 15N-labeled nitrate and ammonium
and consequences for nitrogen fixation measurements. PLoS ONE 9(10): e110335.
doi:10.1371/journal.pone.0110335
Dalsgaard T, Nielsen LP, Brotas V, Viaroli P, Underwood G, Nedwell D, Sundbäck K, Rysgaard
S, Miles A, Bartoli M, Dong L, Thornton DCO, Ossosen LDM, Castaldelli G, Risgaard-Petersen
N (2000) Protocol handbook for NICE- Nitrogen cycling in estuaries. National Environmental
Research Institute, Silkeborg, Denmark.
Dalsgaard T, Thamdrup B, Canfield, DE (2005) Anaerobic ammonium oxidation (anammox) in
the marine environment. Res Microbiol 156:45-464. doi:10.1016/j.resmic.2005.01.011
Davis WP, Steinman AD (1998) A lightweight, inexpensive benthic core sampler for use in
shallow water. J. Freshwater Ecology 13:475-479.
Davis TW, Harke MJ, Marcoval MA, Goleski J, Orano- Dawson C, Berry DL, Gobler CJ (2010)
Effects of nitrogenous compounds and phosphorus on the growth of toxic and non-toxic strains
of Microcystis during cyanobacterial blooms. Aquat Microb Ecol 61:149−162.
Davis TW, Koch F, Marcoval MA, Wilhelm SW, Gobler CJ (2012) Mesozooplankton and
microzooplankton grazing during cyanobacterial blooms in the western basin of Lake Erie.
Harmful Algae 15:26–35. doi:10.1016/j.hal.2011.11.002
Davis TW, Bullerjahn GS, Tuttle T, McKay RM, Watson SB (2015) Effects of increasing
nitrogen and phosphorus concentrations on phytoplankton community growth and toxicity during
Planktothrix blooms in Sandusky Bay, Lake Erie. Environ Sci Tech 49:7197-7207.
doi:10.1021/acs.est.5b00799
Donald DB, Bogard MJ, Finlay K, Leavitt PR (2011) Comparative effects of urea, ammonium,
and nitrate on phytoplankton abundance, community composition, and toxicity in hypereutrophic
freshwaters. Limnol Oceanogr 56(6):2161-2175. doi:10.4319/lo.2011.56.6.2161
Donald DB, Bogard MJ, Finlay K, Bunting L, Leavitt PR (2013) Phytoplankton-specific
response to enrichment of phosphorus-rich surface waters with ammonium, nitrate, and urea.
PLoS ONE 8(1):E53277. doi:10.1371/journal.pone.0053277
Dong LF, Smith CJ, Papaspyrou S, Stott A, Osborn AM, Nedwell DB (2009) Changes in benthic
denitrification, nitrate ammonification, and anammox process rates and nitrate reductase gene
abundances along an estuarine nutrient gradient (the Colne Estuary, United Kingdom). Appl
Environ Microbiol 75(10):3171-3179, doi:10.1128/AEM.02511-08

134

Dong LF, Sobey MN, Smith CJ, Rusmana I, Phillips W, Scott A, Osborn AM, Nedwell DB
(2011) Dissimilatory reduction of nitrate to ammonium, not denitrification or anammox,
dominates benthic nitrate reduction in tropical estuaries. Limnol Oceanogr 56(1):279-291.
doi:10.4319/lo.2011.56.1.0279
Eyre BD, Rysgaard S, Dalsgaard T, Christensen PB (2002) Comparison of isotope pairing and
N2:Ar methods for measuring sediment denitrification: Assumptions, modifications and
implications. Estuaries 25(6A):1077-1087.
Filbrun JE, Conroy JD, Culver DA (2013) Understanding seasonal phosphorus dynamics to
guide effective management of shallow, hypereutrophic Grand Lake St. Marys, Ohio. Lake
Reservoir Manage 29:165-178.
Frame CH, Casciotti KL (2010) Biogeochemical controls and isotopic signatures of nitrous oxide
production by a marine ammonia-oxidizing bacterium. Biogeosciences 7:2695-2709.
doi:10.5194/bg-7-2695-2010
Fujii A, Toyoda S, Yoshida O, Watanabe S, Sasaki K, Yoshida N (2013) Distribution of nitrous
oxide dissolved in water masses in the eastern subtropical North Pacific and its origin inferred
from isotopomers analysis. J Oceanogr 69:147-157. doi: 10.1007/s10872-012-0162-4
Gobler CJ, Burkholder JM, Davis TW, Harke MJ, Johengen T, Stow CA, Van de Waal DB
(2016) The dual role of nitrogen supply in controlling the growth and toxicity of cyanobacterial
blooms. Harmful Algae 54:87-97.
Großkopf T, Mohr W, Baustian T, Schunck H, Gill D, Kuypers MMM, Lavik G, Schmitz RA,
Wallace DWR, La Roche J (2012) Doubling of marine dinitrogen-fixation rates based on direct
measurements. Nature 488:361-364. doi:10.1038/nature11338
Guildford SJ, Hecky RE, Taylor WD, Mugidde R, Bootsma HA (2003) Nutrient enrichment
experiments in tropical great lakes Malawi and Victoria. J Great Lakes Res 29:89-106.
doi:10.1016/S0380-1330(03)70541-3
Hamilton SK, Ostrom NE (2007) Measurements of the stable isotope ratio of dissolved N2 in 15N
tracer experiments. Limnol Oceanogr Methods 5:233-240.
Havens SM, Hassler CS, North RL, Guildford SJ, Silsbe G, Wilhelm SW, Twiss MR (2012) Iron
plays a role in nitrate drawdown by phytoplankton in Lake Erie surface waers as observed in
lake-wide assessments. Can J Fish Aquat Sci 69:369-381. doi:10.1139/F2011-157
Heisler J, Glibert P, Burkholder J, Anderson D, Cochlan W, Dennison W, Gobler C, Dortch Q,
Heil C, Humphries E, Lewitus A, Magnien R, Marshall H, Sellner K, Stockwell D, Stoecker D,
Suddleson M (2008) Eutrophication and harmful algal blooms: A scientific consensus. Harmful
Algae 8:3–13. doi:10.1016/j.hal.2008.08.006

135

Ho JC, Michalak AM (2015) Challenges in tracking harmful algal blooms: A synthesis of
evidence from Lake Erie. J Great Lakes Res 41(2):317-325. doi:10.1016/j.jglr.2015.01.001
Holl CM, Montoya JP (2005) Interactions between nitrate uptake and nitrogen fixation in
continuous cultures of the marine diazotroph Trichodesmium (Cyanobacteria). J Phycoloby
41(6):1178-1183. doi: 10.1111/j.1529-8817.2005.00146.x
Horst GP, Sarnelle O, White JD, Hamilton SK, Kaul RB, Bressie JD (2014) Nitrogen availability
increases the toxin quota of a harmful cyanobacterium, Microcystis aeruginosa. Water Res
54:188-198. doi: 10.1016/j.watres.2014.01.063
Howard DL, Frea JI, Pfister RM, Dugan PR (1970) Biological nitrogen fixation in Lake Erie.
Science 169(3940):61-62.
Hsu T-C, Kao S-J (2013) Technical note: Simultaneous measurement of sedimentary N 2 and
N2O production and new 15N isotope pairing technique. Biogeosciences 10:7847-7862.
doi:10.5194/bg-10-7847-2013
Jinuntuya-Nortman M, Sutka RL, Ostrom PH, Gandhi H, Ostrom NE (2008) Isotopologue
fractionation during microbial reduction of N2O within soil mesocosms as a function of waterfilled pore space. Soil Biol Biochem 40:2273-2280. doi:10.1016/j.soilbio.20 08.05.016
Kane DD, Conroy JD, Richards RP, Baker DB, Culver DA (2014) Re-eutrophication of Lake
Erie: Correlations between tributary nutrient loads and phytoplankton biomass. J Great Lakes
Res 40:496-501. doi:10.1016/j.jglr.2014.02.004
Kim D-K, Zhang W, Watson S, Arhonditsis GB (2014) A commentary on the modeling of the
causal links among nutrient loading, harmful algal blooms, and hypoxia patterns in Lake Erie. J
Great Lakes Res 40:117-129. doi:10.1016/j.jglr.2014.02.014
Knowles R (1982) Denitrification. Microbiol Rev 46(1):43-70.
Lehmann MF, Barnett B, GĂŠlinas Y, Gilbert D, Maranger RJ, Mucci A, Sundby B, Thibodeau B
(2009) Aerobic respiration and hypoxia in the Lower St. Lawrence Estuary: Stable isotope ratios
of dissolved oxygen constrain oxygen sink partitioning. Limnol Oceanogr 54(6):2157-2169.
Lewicka-Szczebak D, Well R, Bol R, Gregory AS, Matthews GP, Misselbrook T, Whalley WR,
Cardenas LM (2015) Isotope fractionation factors controlling isotopocule signatures of soilemited N2O produced by denitrification processes of various rates. Rapid Commun Mass
Spectrom 29:269-282. doi:10.1002/rcm.7102
Larson JH, Trebitz AS, Steinman AD, Wiley MJ, Mazur MC, Pebbles V, Braun HA, Seelbach
PW (2013) Great Lakes rivermouth ecosystems: Scientific synthesis and management
implications. J Great Lakes Res 39:513-524. doi:10.1016/j.jglr.2013.06.002

136

MacGregor BJ, Van Mooy B, Baker BJ, Mellon M, Moisander PH, Paerl HW, Zehr J, Hollander
D, Stahl DA (2001) Microbiological, molecular biological and stable isotopic evidence for
nitrogen fixation in the open waters of Lake Michigan. Environ Microbiol 3(3):205-219.
McCarthy MJ, Gardner WS, Lehmann MF, Guindon A, Bird DF (2016) Benthic nitrogen
regeneration, fixation, and denitrification in a temperate, eutrophic lake: Effects on the nitrogen
budget and cyanobacteria blooms. Limnol Oceanogr 61(4):1406-1423. doi: 10.1002/lno.10306
McCarthy MJ, Gardner WS, Lavrentyev PJ, Moats KM, Jochem FJ, Klarer DM (2007) Effects of
hydrological flow regime on sediment-water interface and water column nitrogen dynamics in a
Great Lakes coastal wetland (Old Woman Creek, Lake Erie). J Great Lakes Res 33(1):219-231.
Michalak AM, Anderson EJ, Beletsky D, Bolans S, Bosch NS, Bridgeman TB, Chaffin JD, Cho
K, Confesor R, Daloglu I, DePinto JV, Evans MA, Fahnenstiel GL, He L, Ho JC, Jenkins L,
Johengen TH, Kuo KC, LaPorte E, Liu X, McWilliams MR, Moore MR, Posselt DJ, Richards
RP, Scavia D, Steiner AL, Verhamme E, Wright DM, Zagorski MA (2013) Record-setting algal
bloom in Lake Erie caused by agricultural and meterological trends consistent with expected
future conditions. P Natl Acad Sci USA 110(16):6448-6452. doi:10.1073/pnas.1216006110
Millie DF, Fahnenstiel GL, Bressie JD, Pigg RJ, Rediske RR, Klarer DM, Tester PA, Litaker
RW (2009) Late-summer phytoplankton in western Lake Erie (Laurentian Great Lakes): bloom
distributions, toxicity, and environmental influences. Aquat Ecol 434:915–934.
doi:10.1007/s10452-009-9238-7
Monchamp M-E, Pick FR, Beisner BE, Maraner R (2014) Nitrogen forms influence
microcystinconcentration and composition via changes in cyanobacterial community structure.
PLoS ONE 9(1):e85573. doi:10.1371/journal.pone.0085573
Montoya JP, Voss M, Kähler P, Capone DG (1996) A simple, high-precision, high-sensitivity
tracer assay for N2 fixation. Applied Environ Microbiol 62(3):986-993.
Moon JB, Carrick HJ (2007) Seasonal variation of phytoplankton nutrient limitation in Lake
Erie. Aquat Microb Ecol 48:61-71.
North RL, Guildford SJ, Smith REH, Havens SM, Twiss MR (2007) Evidence for phosphorus,
nitrogen, and iron colimitation of phytoplankton communities in Lake Erie. Limnol
Oceanogr 52(1):315-328. doi:10.4319/lo.2007.52.1.0315
Ostrom NE, Carrick HJ, Twiss MR, Piwinski L (2005) Evaluation of primary production in Lake
Erie by multiple proxies. Oecologia 144:115-124.
Paerl HW, Xu H, McCarthy MJ, Zhu G, Qin B, Li Y, Gardner WS (2011) Controlling harmful
cyanobacterial blooms in a hyper-eutrophic lake (Lake Taihu, China): The need for a dual
nutrient (N & P) management strategy. Water Res 45(5):1973-1983.
doi:10.1016/j.watres.2010.09.018

137

Paerl HW, Scott JT, McCarthy MJ, Newell SE, Gardner WS, Havens KE, Hoffman DK,
Wilhelm SW, Wurtsbaugh WA (2016) It takes two to tango: When and where dual nutrient (N &
P) reductions are needed to protect lakes and downstream ecosystems. Environ Sci Technol
50(20):10805-10813. doi:10.1021/acs.est.6b02575
Prein AF, Rasmussen RM, Ikeda K, Liu C, Clark MP, Holland GJ (2016) The future
intensification of hourly precipitation extremes. Nature Climate Change
doi:10.1038/nclimate3168
Redfield A (1934) On the proportions of organic derivatives in sea water and their relation to the
composition of plankton. In Daniel, R.J. (ed James Johnstone Memorial Volume). University
Press of Liverpool 177–192.

Richards RP, Baker DB (1985) Assimilation and flux of sediments and pollutants in the
Sandusky River Estuary, Sandusky Bay, and the adjacent nearshore zone of Lake Erie. Final
Report, Grant NA80RAD00038.
Risgaard-Petersen N, Nielsen LP, Rysgaard S, Dalsgaard T, Meyer RL (2003) Application of the
isotope pairing technique in sediments where anammox and denitrification coexist. Limnol
Oceanogr Methods 1:63-73.
Romero IC, Klein NJ, SaĂąudo-Wilhelmy SA, Capone DG (2013) Potential trace metal colimitation controls on N2 fixation and NO3- uptake in lakes with varying trophic status. Frontiers
Microbiol 4:54. doi:10.3389/fmicb.2013.00054
Salk KR, Ostrom PH, Biddanda BA, Weinke AD, Kendall ST, Osrom NE (2016) Ecosystem
metabolism and greenhouse gas production in a mesotrophic northern temperate lake
experiencing seasonal hypoxia. Biogeochemistry 131:303-319. doi:10.1007/s10533-016-0280-y
Sansone FJ, Popp BN, Rust TM (1997) Stable carbon isotopic analysis of low-level methane in
water and gas. Anal Chem 69(1):40-44. doi:10.1021/ac960241i
Scavia D, Allan JD, Arend KK, Bartell S, Beletsky D, Bosch NS, Brandt SB, Briland RD,
Daloglu I, DePinto JV, Dolan DM, Evans MA, Farmer TM, Goto D, Han H, Hook TO, Knight
R, Ludsin SA, Mason D, Michalak AM, Richards RP, Roberts JJ, Rucinski DK, Rutherford E,
Schwab DJ, Sesterhenn TM, Zhang H, Zhou Y (2014) Assessing and addressing the reeutrophication of Lake Erie: Central basin hypoxia. J Great Lakes Res 40:226-246.
doi:10.1016/j.jglr.2014.02.004
Schubert CJ, Durisch-Kaiser E, Wehrli B, Thamdrup B, Lam P, Kuypers MMM (2006)
Anaerobic ammonium oxidation in a tropical freshwater system (Lake Tanganyika). Environ
Microbiol 8(10):1857-1863. doi:10.1111/j.1462-2920.2006.001074.x
Seitzinger S, Harrison JA, BĂśhlke JK, Bouwman AF, Lowrance R, Peterson B, Tobias C, Van
Drecht G (2006) Denitrification across landscapes and waterscapes: A synthesis. Ecol Appl
16(6):2064-2090.

138

Small GE, Cotner JB, Finlay JC, Stark RA, Sterner RW (2014) Nitrogen transformations at the
sediment-water interface across redox gradients in the Laurentian Great Lakes.
Hydrobiologia 731:95-108. doi:10.1007/s10750-013-1569-7
Small GE, Finlay JC, McKay RML, Rozmarynowycz MJ, Brovold S, Bullerjahn GS, Spokas K,
Sterner RW (2016) Large differences in potential denitrification and sediment microbial
communities across the Laurentian great lakes. Biogeochemistry 128: 353-368.
doi:10.1007/s10533-016-0212-x
Steffen MM, Belisle BS, Watson SB, Boyer GL, Wilhelm SW (2014a) Status, causes and
controls of cyanobacterial blooms in Lake Erie. J Great Lakes Res 40:215-225.
doi:10.1016/j.jglr.2013.12.012
Steffen MM, Zhu Z, McKay RML, Wilhelm SW, Bullerjahn GS (2014b) Taxonomic assessment
of a toxic cyanobacteria shift in hypereutrophic Grand Lake St. Marys (Ohio, USA). Harmful
Algae 33:12-18.
Sterner RW, Elser JJ (2002) Ecological Stoichiometry: The Biology of Elements from Molecules
to the Biosphere. Princeton University Press, New Jersey.
Sutka RL, Ostrom NE, Ostrom PH, Breznak JA, Gandhi H, Pitt AJ, Li F (2006) Distinguishing
nitrous oxide production from nitrification and denitrification on the basis of isotopomer
abundances. Appl Environ Microb 72(1):638-644. doi:10.1128/AEM.72.1.638-644.2006
Sutka RL, Adams GC, Ostrom NE, Ostrom PH (2008) Isotopologue fractionation during N 2O
production by fungal denitrification. Rapid Comm Mass Spec 22:3989-3996.
doi:10.1002/rcm.3820
Thamdrup B, Dalsgaard T (2002) Production of N2 through anaerobic ammonium oxidation
coupled to nitrate reduction in marine sediments. Appl Environ Microbiol 68:1312-1318.
doi:10.1128/AEM.68.3.1312-1318.2002
Toyoda S, Mutobe H, Yamagishi H, Yoshida N, Tanji Y (2005) Fractionation of N 2O
isotopomers during production by denitrifier. Soil Biol Biochem 37(8):1535-1545.
doi:10.1016/j.soilbio.2005.01.009
Walker JT, Stow CA, Geron C (2010) Nitrous oxide emissions from the Gulf of Mexico hypoxic
zone. Environ Sci Technol 44(5):1617-1623. doi:10.1021/es902058t
Wang H, Wang W, Yin C, Wang Y, Lu J (2006) Littoral zones as the “hotspots” of nitrous oxide
(N2O) emission in a hyper-eutrophic lake in China. Atmos Environ 40:5522-5527.
doi:10.1016/j.atmosenv.2006.05.032
Wanninkhof R (1992) Relationship between wind speed and gas exchange over the ocean. J
Geophys Res 97(C5):7373-7382. doi:10.1029/92JC00188

139

Well R, Eschenbach, Flessa H, von der Heide C, Weymann D (2012) Are dual isotope and
isotopomers ratios of N2O useful indicators for N2O turnover during denitrification in nitratecontaminated aquifers? Geochim Cosmochim Ac 90:265-282. doi:10.1016/j.gca.2012.04.045
Welschmeyer NA (1994) Fluorometric analysis of chlorophyll a in the presence of chlorophyll b
and pheopigments. Limnology and Oceanography 39(8):1985-1992.
Wenk CB, Frame CH, Koba K, Casciotti KL, Veronesi M, Niemann H, Schubert CJ, Yoshida N,
Toyoda S, Makabe A, Zopfi J, Lehmann MF (2016) Differential N 2O dynamics in two oxygendeficient lake basins revealed by stable isotope and isotopomer distributions. Limnol Oceanogr
61:1735-1749. doi:10.1002/lno.10329
Westley MB, Yamagishi H, Popp BN, Yoshida N (2006) Nitrous oxide cycling in the Black Sea
inferred from stable isotope and isotopomers distributions. Deep-Sea Res Pt II 53:1802-1816.
doi:10.1016/j.dsr2.2006.03.012
Wheeler SM, Morrissey LA, Levine SN, Livingston GP, Vincent WF (2012) Mapping
cyanobacterial blooms in Lake Champlain's Missisquoi Bay using QuickBird and MERIS
satellite data. J Great Lakes Res 38:68-75.
Whitfield CJ, Aherne J, Baulch HM (2011) Controls on greenhouse gas concentrations in
polymictic headwater lakes in Ireland. Sci Total Environ 410-411:217-225.
doi:10.1016/j.scitotenv.2011.09.045
Wrage N, Velthof GL, van Beusichem ML, Oenema O (2001) Role of nitrifier denitrification in
the production of nitrous oxide. Soil Biol Biochem 33(12-13):1723-1732.
Xu H, Paerl HW, Qin B, Zhu G, Gao G (2010) Nitrogen and phosphorus inputs control
phytoplankton growth in eutrophic Lake Taihu, China. Limnol Oceanogr 55(1):420-432.
Yang H, Andersen T, DĂśrsch P, Tominaga K, Thrane J-E, Hessen DO (2015) Greenhouse gas
metabolism in Nordic boreal lakes. Biogeochemistry 126:211-225.
doi:10.1007/s10533-015-0154-8
Yang H, Gandhi H, Ostrom NE, Hegg EL (2014) Isotopic fractionation by a fungal P450 nitric
oxide reductase during the production of N2O. Environ Sci Technol 48:10707-10715.
doi:10.1021/es501912d
Yoshinaga I, Amano T, Yamagishi T, Okada K, Ueda S, Sako Y, Suwa Y (2011) Distribution
and diversity of anaerobic ammonium oxidation (anammox) bacteria in the sediment of a
eutrophic freshwater lake, Lake Kiraura, Japan. Microbes Environ 26(3):189-197.
doi:10.1264/jsme2.ME10184

140

Zhang Y, Lin S, Qian X, Wang Q, Qian Y, Liu J, Ge Y (2011) Temporal and spatial variability
of chlorophyll a concentration in Lake Taihu using MODIS time-series data. Hydrobiologia
661(1):235-250. doi:10.1007/s10750-010-0528-9
Zhu G, Wang S, Wang W, Wang Y, Zhou L, Jian B, Op den Camp HJM, Risgaard-Petersen N,
Schwark L, Peng Y, Hefting MM, Jetten MSM, Yin C (2013) Hotspots of anaerobic ammonium
oxidation at land-freshwater interfaces. Nature Geoscience 6:103-107. doi:10.1038/NGEO1683
Zhu G, Wang S, Zhou L, Wang Y, Zhao S, Xia C, Wang W, Zhou R, Wang C, Jetten MSM,
Hefting MM, Yin C, Qu J (2015) Ubiquitous anaerobic ammonium oxidation in inland waters of
China: An overlooked nitrous oxide mitigation process. Nature Scientific Reports 5:17306.
doi:10.1038/srep17306

141