PUBLICATIONS
Geophysical Research Letters
RESEARCH LETTER
10.1002/2016GL067732
Key Points:
• A substantial portion of the CO2
released from three lakes was
produced exogenously
• Despite positive net ecosystem
production, lakes were carbon sources
to the atmosphere
Supporting Information:
• Text S1, Data Set S1, and Figure S1
Correspondence to:
G. M. Wilkinson,
[email protected]
Citation:
Wilkinson, G. M., C. D. Buelo, J. J. Cole,
and M. L. Pace (2016), Exogenously
produced CO2 doubles the CO2 efflux
from three north temperate lakes,
Geophys. Res. Lett., 43, 1996–2003,
doi:10.1002/2016GL067732.
Received 11 JAN 2016
Accepted 10 FEB 2016
Accepted article online 13 FEB 2016
Published online 4 MAR 2016
Exogenously produced CO2 doubles the CO2 efflux
from three north temperate lakes
Grace M. Wilkinson1, Cal D. Buelo1, Jonathan J. Cole2, and Michael L. Pace1
1
Department of Environmental Sciences, University of Virginia, Charlottesville, Virginia, USA, 2Cary Institute of Ecosystem
Studies, Millbrook, New York, USA
Abstract It is well established that lakes are typically sources of CO2 to the atmosphere. However, it
remains unclear what portion of CO2 efflux is from endogenously processed organic carbon or from
exogenously produced CO2 transported into lakes. We estimated high-frequency CO2 and O2 efflux from
three north temperate lakes in summer to determine the proportion of the total CO2 efflux that was
exogenously produced. Two of the lakes were amended with nutrients to experimentally enhance
endogenous CO2 uptake. In the unfertilized lake, 50% of CO2 efflux was from exogenous sources and
hydrology had a large influence on efflux. In the fertilized lakes, endogenous CO2 efflux was negative (into
the lake) yet exogenous CO2 made the lakes net sources of CO2 to the atmosphere. Shifts in hydrologic
regimes and nutrient loading have the potential to change whether small lakes act primarily as reactors or
vents in the watershed.
1. Introduction
The majority of lakes are supersaturated in carbon dioxide (CO2) [Cole et al., 1994; Sobek et al., 2005] and are a
source of greenhouse gases to the atmosphere. The CO2 emitted from lakes comes from two potential
sources. The first is within-lake (endogenous) mineralization of aquatic and terrestrial organic matter.
Terrestrial material dominates the organic matter pools in many north temperate lakes [Wilkinson et al.,
2013], the mineralization of which can lead to respiration being higher than gross primary production and
CO2 supersaturation [del Giorgio et al., 1999; Jonsson et al., 2003; Duarte and Prairie, 2005]. The other major
source of carbon emitted from lakes is advective (exogenous) input of CO2 from soil respiration and mineral
weathering within the catchment [Striegl and Michmerhuizen, 1998; Finlay et al., 2010; McDonald et al., 2013].
Exogenous CO2 dominates the carbon efflux from streams and rivers [Hotchkiss et al., 2015] and hydrologic
modeling in larger lakes (>4 ha) has revealed that exogenous inputs of pCO2 also drive CO2 efflux from larger
lakes [McDonald et al., 2013]. However, there is very little information about the proportional contribution of
endogenous and exogenous carbon to CO2 efflux from small lakes which are numerically dominant [Downing
et al., 2006]. Partitioning the sources of CO2 efflux is not only necessary for lake and watershed carbon budgets, but it also aides in understanding how CO2 efflux may change with future climate conditions.
Partitioning the exogenous and endogenous contributions to total lake CO2 efflux can be accomplished
through measurements of inflows and outflows used in mass balance modeling [Stets et al., 2009; Knoll
et al., 2013; Perkins et al., 2015], hydrologic modeling of groundwater and surface water movement
[McDonald et al., 2013], or inference from point measurements of stoichiometric ratios of O2 and CO2 concentrations [Torgersen and Branco, 2008; Holgerson, 2015]. With the advent of affordable environmental sensor
technology, researchers are able to collect high-frequency (on the order of minutes) dissolved oxygen
(DO) data in lakes and rivers to assess endogenous carbon processing. Many environmental sensors also
measure pH and temperature at a high frequency. In soft water lakes, pH, dissolved inorganic carbon (DIC)
concentration, and temperature can be used to calculate high-frequency pCO2 and subsequently, highfrequency CO2 efflux. The high-frequency time series of O2 and CO2 efflux can then be stoichiometrically
compared to assess endogenous and exogenous contributions to CO2 efflux from lakes.
©2016. American Geophysical Union.
All Rights Reserved.
WILKINSON ET AL.
Using high-frequency time series of CO2 and O2 concentrations, we evaluated the contribution of endogenous
and exogenous sources to total CO2 efflux during the summer in three north temperate lakes. In order to understand the role that eutrophication plays in endogenous CO2 contributions to total efflux, two of the lakes were
amended with nutrients in order to predictably manipulate the endogenous CO2 signal without changing the
EXOGENOUS CO2 DOUBLES EFFLUX FROM LAKES
1996
Geophysical Research Letters
10.1002/2016GL067732
Table 1. Mean (± Standard Deviation) of Limnological Variables in the
Three Study Lakes (See Supporting Information for Sampling Details)
Variable
Surface area
Max depth
DOC
Color
pH
DIC
TN
TP
Chlorophyll a
Unit
Paul
Peter
Tuesday
ha
m
1
mg L
1
m
n.a.
1
μmol L
1
μmol L
1
μmol L
1
μg L
1.9
12
4.3 ± 0.4
1.3 ± 0.2
6.4 ± 0.3
111.1 ± 8.1
13.7 ± 1.7
0.4 ± 0.1
2.1 ± 0.7
2.5
18
8.3 ± 7.3
2.2 ± 0.3
6.6 ± 0.3
90.5 ± 28.2
23.7 ± 6.1
0.5 ± 0.4
5.8 ± 2.7
0.9
15
12.3 ± 4.7
5.3 ± 1.0
5.8 ± 0.4
46.6 ± 12.7
25.0 ± 1.8
0.7 ± 0.3
11.0 ± 4.5
exogenous CO2 signal. The goal was
to determine to what degree lakes
act as reactors (releasing endogenously produced CO2) and vents to
the atmosphere (releasing exogenously produced CO2).
2. Methods
2.1. Study Lakes and Sampling
Peter, Paul, and Tuesday Lakes are
small (<3 ha), soft water seepage
lakes located in Gogebic County,
Michigan, USA (Table 1), with a water residence time of approximately 1 year. The groundwater input to seepage lakes in this area is on average 57% of the total annual water input [Cole and Pace, 1998]. Beginning on
day of year (DoY) 153 in 2014, ammonium nitrate (NH4NO3) and phosphoric acid (H3PO4) were dissolved in
lake water and added daily to Peter and Tuesday Lakes at a molar N:P ratio of 15:1 and a phosphorous loading
rate of 3 mg m2 d1. The nutrients were distributed into the epilimnion of the lakes in the propeller wash of
an underway electric motor. Paul Lake served as a reference with no nutrient amendment.
High-frequency measurements (every 5 min) of DO, pH, and temperature were made using a Hydrolab DS 5X
(OTT Hydromet, Loveland, Colorado, USA) deployed at 0.75 m in each lake. Brief gaps in the data record were
interpolated in R (version 3.1.30) [R Core Development Team, 2015] using a maximum likelihood multivariate
autoregressive state-space model (MARSS package version 3.9) [Holmes et al., 2015] with a simultaneous data
record taken with a YSI 6600 V2-4 multiparameter sonde (YSI Incorporated, Yellow Springs, Ohio, USA). The
MARSS-interpolated pH time series was then calibrated with the weekly manual measurements of surface
water pH using the more accurate long equilibration technique of Stauffer [1990].
Water temperature, pCO2, and dissolved inorganic carbon (DIC) samples were taken weekly at the surface of
each lake. For DIC concentration, 10 mL of water were acidified with 200 μL of 2N H2SO4 and combined with
20 mL of helium in a gas tight syringe and equilibrated through vigorous shaking. The 0.5 mL sample loop
was flushed with 10 mL of headspace and automatically injected into a Shimadzu GC-8A gas chromatograph
with a 2 m column, packed with Porapak-Q, and connected to a thermal conductivity detector. Air samples
for the atmospheric concentration of CO2 were taken weekly at 1.5 m above the surface on the aweather side
of the boat. Samples of surface water pCO2 were taken from 60 mL of atmospheric headspace equilibrated
with surface water in a 2 L container. Staff gauge height was recorded daily in each lake to track changes
in lake level (ΔLL) which was compared to the daily change in mean CO2 efflux (ΔCO2) in each lake.
2.2. CO2 and O2 Efflux Calculation
Acid neutralizing capacity (ANC) was calculated weekly for each lake using the weekly measurements of DIC,
pH, and temperature (see supporting information for details). We used ANC as it is not altered by gains or
losses of CO2. The weekly ANC time series was interpolated to a 5 min time scale to match the other highfrequency observations (see SI for details). The interpolated ANC time series, high-frequency pH, and temperature were used to calculate high-frequency pCO2. It is common for calculated pCO2 to be greater than
measured pCO2 [Abril et al., 2015]; therefore, the difference between the weekly measured pCO2 and calculated pCO2 values was used to calibrate the calculated pCO2 time series. The DO concentration was measured
directly by the sonde at the same frequency.
Efflux of CO2 and O2 (FXt; movement of gas from the lake to the atmosphere) was calculated using the following equation
(1)
F X t ¼ X surt X eq k Ht kX t
where Xsur (t) is the concentration of the gas (CO2 or O2) in the surface waters at time t and Xeq is the atmospheric concentration of the gas. For CO2, the atmospheric concentration was the average of weekly
measurements on each lake taken during the 99 day sampling period (n = 43). For O2, the atmospheric concentration was calculated using the R package LakeMetabolizer (version 1.3.1) [Winslow et al., 2015]. KH(t) is
WILKINSON ET AL.
EXOGENOUS CO2 DOUBLES EFFLUX FROM LAKES
1997
Geophysical Research Letters
10.1002/2016GL067732
Figure 1. The time series of calculated pCO2 (gray) with the weekly measured pCO2 (black circle) in (a) Paul, (b) Peter, and (c) Tuesday Lakes. The dashed line is the
average atmospheric concentration.
the Henry’s law constant for the gas at time t and kX(t) is the piston velocity (m1) for each gas. The piston
velocity was fixed at the k600 values reported in Cole et al. [2010], measured in the study lakes and adjusted
to be gas and temperature specific based on the high-frequency temperature data [Wanninkhof, 1992].
2.3. Estimating Endogenous and Exogenous CO2 Efflux
Instances of O2 efflux are periods of positive net ecosystem production (NEP) in the lake even if CO2 efflux is also
occurring. Under the assumption that aerobic respiration consumes 1 mole of O2 for every mole of CO2 produced [Mattson and Likens, 1993], O2 influx (the inverse of efflux) is equal to the expected CO2 efflux if only
endogenous, aerobic production of
CO2 was contributing to efflux from
the lakes. Any CO2 efflux greater than
the O2 influx is therefore assumed to
be exogenously produced CO2 or anaerobically produced endogenous CO2
from methanogenesis, sulfate reduction, or denitrification.
Figure 2. The high-frequency (lighter line) and daily mean (darker line)
CO2 and O2 efflux for (a) Paul, the reference lake and (b) Peter and (c)
Tuesday, the fertilized lakes.
WILKINSON ET AL.
EXOGENOUS CO2 DOUBLES EFFLUX FROM LAKES
We estimated anaerobically produced
endogenous CO2 from previous studies. The ebullative flux of methane,
which ranges between 0.27 and
1.02 mmol C m2 d1 in these lakes
[Bastviken et al., 2004], is an additional
source of DIC without a concomitant
loss of O2. The daily flux of methane
from epilimnion sediments (calculated
from hypsographic curves) must be
accounted for as a source of anaerobic,
endogenous carbon processing. The
occurrence of methanogenesis is indicative of a depletion of other anaerobic
anions such as nitrate and sulfate in the
sediment. Thus, the production of DIC
from denitrification and sulfate reduction is temporary. These sources of DIC
can be accounted for by calculating
the difference in spring and fall concentrations of nitrate (measured directly,
1998
Geophysical Research Letters
10.1002/2016GL067732
plus the 63.8 mmol m2 of nitrate
added to Peter and Tuesday Lakes),
assuming that exchange is occurring
between the sediments and water column to a depth of 10 cm and applying
the molar conversions for DIC production from sulfate and nitrate reduction
(2 mol and 1.25 mol, respectively)
[Mattson and Likens, 1993]. The springtime epilimentic concentrations of sulfate in the three lakes were taken from
Houser [2001] and the fall concentrations were assumed to be zero. The calculated DIC from nitrate and sulfate
reduction (DICNO3 - and DICSO4 2-, respectively) was then added to the sum of
the O2 influx and the sum of the ebullative methane flux to calculate the endogenously produced CO2 efflux ðCO2endo Þ
as follows
X
CO2endo ¼
O2 influx
X
þ
CH4 Flux
Figure 3. The high-frequency (lighter points) and daily mean (darker
points) of CO2 efflux and the O2 influx for (a) Paul, (b) Peter, and (c)
Tuesday Lakes. The dashed line is the 1:1 line.
CO2exo ¼
X
þ DICNO3 þ DICSO4 2-
(2)
for the entire 99 day data record in
each lake. The total exogenous CO2
efflux CO2exo was then calculated as
ðCO2 FluxÞ CO2endo
(3)
where CO2endo can be either positive (efflux from the lake to the atmosphere) or negative (influx to the lake).
3. Results
3.1. Patterns of Gas Efflux
The measured and calculated pCO2 values are in strong agreement after calibration (Figure 1) with R2 = 0.99
in a regression between measured and calibrated pCO2 (Figure S1 in the supporting information). Surface
water pCO2 in all three lakes was substantially higher than the atmosphere (389 ± 44.6 ppmv) for large periods of the data record (Figure 1). On average, all three lakes were net sources of CO2 to the atmosphere even
though there were brief periods of CO2 influx (Figure 2). Over the course of the summer, the unfertilized system, Paul Lake, had the highest average (± SD) CO2 efflux (8.0 ± 3.3 mmol m2 d1) followed by the fertilized
lakes Tuesday (4.0 ± 3.5 mmol m2 d1) and Peter (2.0 ± 4.9 mmol m2 d1; Figure 2). Conversely, Paul Lake
had the lowest average O2 efflux (3.7 ± 5.0 mmol m2 d1) followed by Peter Lake (5.3 ± 7.2 mmol m2 d1)
and Tuesday Lake (6.5 ± 7.9 mmol m2 d1).
A phytoplankton bloom composed largely of Anabaena spp. peaked on DoY 225 in Peter Lake. This bloom
coincided with the peak daily average O2 efflux (18.8 mmol m2 d1; DoY 220) and the minimum daily average CO2 efflux (8.2 mmol m2 d1; DoY 220). Similarly, in Tuesday Lake, there was a Chrysochromulina spp.
bloom peaking on DoY 234 which coincided with the only period of intermittent negative daily average CO2
efflux in the lake (DoY 231234), although there was not a peak in O2 efflux at this time. In Paul Lake, 13% of
the days had a positive NEP (O2 efflux > 0), the majority of which were during the spring bloom period. In the
fertilized lakes NEP was usually positive, with 78% and 88% of the days having a positive O2 efflux in Peter
and Tuesday Lakes, respectively. The range in ΔLL in all lakes was 5.6 cm to +5.1 cm. In all three lakes there
WILKINSON ET AL.
EXOGENOUS CO2 DOUBLES EFFLUX FROM LAKES
1999
Geophysical Research Letters
10.1002/2016GL067732
was a positive relationship between daily
ΔLL and ΔCO2 (linear regression by lake, p
value < 0.05 for all regressions) such that
when lake level increased CO2 efflux also
significantly increased.
3.2. Endogenous and Exogenous
CO2 Efflux
The daily average influx of O2 was less than
the efflux of CO2 86% of the time in Paul,
85% in Peter, and 95% in Tuesday
(Figure 3). The endogenous contribution
to the total CO2 efflux was negative in
the two fertilized lakes (547.4 mmol m2
in Tuesday Lake and 413.3 mmol m2 in
Peter Lake) and the positive in Paul Lake
(394.9 mmol m2; Figure 4). However, the
net CO2 efflux was positive in all three
lakes due to exogenous CO2 contributions.
Figure 4. The endogenous, exogenous, and net CO2 efflux summed
The exogenous contribution to the total
for the summer for all three lakes.
CO2 efflux ranged from 400.4 mmol m2
in Paul Lake to 938.3 mmol m2 in
Tuesday Lake. The percent contribution of exogenous CO2 to net CO2 efflux was 50% in Paul Lake and
100% in the fertilized lakes. The negative endogenous CO2 efflux in the fertilized lakes lowered the net
CO2 efflux.
4. Discussion
The exogenous contribution to total CO2 efflux more than doubled the total CO2 efflux in Paul Lake and caused
the fertilized lakes to be net sources of CO2 to the atmosphere instead of sinks. Considering the hydrologic residence times of these lakes [Cole and Pace, 1998] and the calculated exogenous CO2 efflux, incoming ground water
would need to have pCO2 concentrations of approximately 1000 μmol L1 to support the estimate. Average
groundwater pCO2 from an area near the study lakes ranged from 0.8 to 955 μmol L1 with maximum measurements of 2080 μmol L1 [Crawford et al., 2014]. This indicates that the estimated input of pCO2 from groundwater
is within the range of possibility. Based on the significant but weak relationships between daily mean CO2 efflux
and changes in lake level, there was evidence that changes in groundwater input to the lake may play a role in the
within lake variability in CO2 efflux over the course of the summer. Increases in lake level from precipitation can
indicate pulses of exogenous CO2 input into the lake [Vachon and del Giorgio, 2014]. In each lake, an increase in
ΔLL from one day to the next coincided with an increase in CO2 efflux. However, the correlation was weak and
without examination on a finer scale with precipitation data should be interpreted cautiously.
As expected, the fertilization of Peter and Tuesday Lakes influenced the endogenous CO2 efflux making it
negative. The fertilization not only provided a check on the modeling method used in this study by modifying
the endogenous signal, the results are also consistent with previous metabolism estimates in the two fertilized lakes. In previous years, both lakes have had either negative or near-zero NEP [Van de Bogert et al.,
2007, 2012; Coloso et al., 2011a, 2011b], which would correspond to a small or negative endogenous CO2
efflux. In a previous fertilization of Peter Lake in 2002, NEP was highly positive as it was in this study
[Carpenter et al., 2005]. However, while endogenous CO2 efflux was negative in 2014, the internal processing
was still overwhelmed by the net exogenous inputs and made both fertilized lakes net sources of CO2 to the
atmosphere, a phenomenon which has been reported in other north temperate lakes [Stets et al., 2009].
The positive endogenous CO2 efflux in Paul Lake is also consistent with previous studies of its metabolic
behavior. Paul Lake has had a negative summertime NEP over the past decade [Carpenter et al., 2005;
Coloso et al., 2011a, 2011b; Batt et al., 2015] indicating that it is a consistent source of both endogenously
and exogenously produced CO2 to the atmosphere. Using high-frequency DO data in Paul Lake but only
WILKINSON ET AL.
EXOGENOUS CO2 DOUBLES EFFLUX FROM LAKES
2000
Geophysical Research Letters
10.1002/2016GL067732
weekly pCO2 measurements, Cole et al. [2000] concluded that CO2 efflux was entirely endogenous. In this
study, the once per week measurements of CO2 missed the diel and day to day variability in pCO2, illustrating
the value of the continuous CO2 approach used here. The large exogenous contribution to CO2 efflux in Paul
Lake estimated here is consistent with estimates for other lakes [Striegl and Michmerhuzien, 1998; Stets et al.,
2009; McDonald et al., 2013; Weyhenmeyer et al., 2015] leading to an emerging consensus that exogenous CO2
inputs contribute significantly to lake CO2 efflux.
In order to calculate the exogenous and endogenous contributions to the total CO2 efflux, a number of assumptions had to be made. The first assumption was that the stoichiometry of aerobic respiration was one mole of
DIC produced for every mole of O2 consumed [Mattson and Likens, 1993], which has also been found in other
studies [Kelly et al., 1988; Matthews et al., 2008]. If the ratio of DIC produced to O2 consumed is less than one,
then the estimate of exogenous CO2 contribution would be slightly lower. Similar assumptions were made concerning the stoichiometry of DIC produced during anaerobic metabolism [Mattson and Likens, 1993; Hedin et al.,
1998]. However, the contribution of anaerobic metabolism to the endogenous CO2 efflux was much smaller
than aerobic metabolism and therefore uncertainty in the ratios used in this calculation has a minimal impact.
A second assumption was that accounting for epilimnetic anaerobic respiration using previous measurements of methane ebullition [Bastviken et al., 2004] and assuming that springtime concentrations of nitrate
and sulfate accounted for the total availability of these electron acceptors. In Peter and Tuesday Lakes this
was not the case for nitrate as it was added to the epilimnion daily. However, the total added mass of nitrate
was included in the calculation, likely leading to an overestimate of anaerobic metabolism as some nitrate
was incorporated into algal biomass instead of being denitrified. Additionally, with the continual nitrate
amendment, sulfate reduction and methanogenesis may have been suppressed and, therefore, overaccounted for in the anaerobic metabolism calculation. Iron reduction was not estimated in the anaerobic
metabolism calculation as measured iron concentrations in these lakes are low (1.56.6 μM; unpublished
data) and only 0.25 moles of DIC are produced for every mole of iron reduced. Although the anaerobic metabolism processes were not directly measured and only constrained estimates, they are important to include
as anaerobic metabolism can contribute significantly to CO2 efflux [Mattson and Likens, 1993; Torgersen and
Branco, 2008; Holgerson, 2015], although this was not the case in the lakes considered here.
A final assumption involved nonrespiratory oxygen consumption. While photodegradation would not affect
the endogenous CO2 efflux accounting, photolysis (the photochemical mineralization of dissolved organic
matter without O2 consumption) was not accounted for in the model. Photolysis of dissolved organic matter
ranges from undetectable [Laurion and Mladenov, 2013] to 10% of planktonic respiration or greater [Bélanger
et al., 2006; Granéli et al., 1996]. For more than 1000 lakes across Sweden, the average DIC production from
photolysis was approximately 12% of total C emissions [Koehler et al., 2014]. If these rates held for the lakes
in this study, that would reduce the estimated exogenous CO2 contribution to the total CO2 efflux but not
alter the overall pattern of exogenous CO2 contribution. Conversely, if the groundwater is undersaturated
in O2, the exogenous signal would be higher.
Recent research has demonstrated that CO2 efflux from streams and rivers is dominated by exogenous CO2
inputs [Hotchkiss et al., 2015], whereas for large lakes the endogenous contribution likely dominates due to
longer hydraulic residence times. In small lakes (<0.025 km2) we found that hydrologic inputs of CO2 more
than double the CO2 efflux from these inland water bodies, even in fertilized lakes. In the future, changes
in precipitation and nutrient runoff could shift whether small lakes are primarily processors of organic carbon
or vents of inorganic carbon to the atmosphere.
Acknowledgments
We thank Jason Kurtzweil and the
Cascade field crew for assistance with
data collection and the University of
Notre Dame for providing access to the
lakes. This project was funded by the
National Science Foundation Division of
Environmental Biology grant 1144624.
Weekly data are available in the supporting information and high-frequency
data are available upon request.
WILKINSON ET AL.
References
Abril, G., et al. (2015), Technical note: Large overestimation of pCO2 calculated from pH and alkalinity in acidic, organic-rich freshwaters,
Biogeosciences, 12, 67–78, doi:10.5194/bg-12-67-2015.
Bastviken, D., J. Cole, M. Pace, and L. Tranvik (2004), Methane emissions from lakes: Dependence of lake characteristics, two regional
assessments, and a global estimate, Global Biogeochem. Cycles, 18, GB4009, doi:10.1029/2004GB002238.
Batt, R. D., S. R. Carpenter, J. J. Cole, M. L. Pace, R. A. Johnson, J. K. Kurtzweil, and G. M. Wilkinson (2015), Altered energy flow in the food web of
an experimentally darkened lake, Ecosphere, 6, 1–23, doi:10.1890/ES14-00241.1.
Bélanger, S., H. Xie, N. Krotkov, P. Larouche, W. F. Vincent, and M. Babin (2006), Photomineralization of terrigenous dissolved organic matter
in Arctic coastal waters from 1979 to 2003: Interannual variability and implications of climate change, Global Biogeochem. Cycles, 20,
GB4005, doi:10.1029/2006GB002708.
EXOGENOUS CO2 DOUBLES EFFLUX FROM LAKES
2001
Geophysical Research Letters
10.1002/2016GL067732
Carpenter, S. R., J. J. Cole, M. L. Pace, M. C. Van de Bogert, D. L. Bade, D. Bastviken, C. M. Gille, J. R. Hodgson, J. F. Kitchell, and E. S. Kritzberg (2005),
13
Ecosystem subsidies: Terrestrial support of aquatic food webs from C addition to contrasting lakes, Ecology, 86, 2737–2750, doi:10.1890/04-1282.
Cole, J. J., and M. L. Pace (1998), Hydrologic variability of small, northern lakes measured by the addition of tracers, Ecosystems, 1, 310–320,
doi:10.1007/s100219900024.
Cole, J. J., N. F. Caraco, G. W. Kling, and T. K. Kratz (1994), Carbon dioxide supersaturation in the surface waters of lakes, Science, 265,
1568–1570, doi:10.1126/science.265.5178.1568.
Cole, J. J., M. L. Pace, S. R. Carpenter, and J. F. Kitchell (2000), Persistence of net heterotrophy in lakes during nutrient addition and food web
manipulations, Limnol. Oceanogr., 45, 1718–1730, doi:10.4319/lo.2000.45.8.1718.
Cole, J. J., D. L. Bade, D. Bastviken, M. L. Pace, and M. Van de Bogert (2010), Multiple approaches to estimating air-water gas exchange in small
lakes, Limnol. Oceanogr. Methods, 8, 285–293, doi:10.4319/lom.2010.8.285.
Coloso, J. J., J. J. Cole, and M. L. Pace (2011a), Difficulty in discerning drivers of lake ecosystem metabolism with high frequency data,
Ecosystems, 14, 935–948, doi:10.1007/s10021-011-9455-5.
Coloso, J. J., J. J. Cole, and M. L. Pace (2011b), Short-term variation in thermal stratification complicates estimation of lake metabolism, Aquat.
Sci., 73, 305–315, doi:10.1007/s00027-010-0177-0.
Crawford, J. T., N. R. Lottig, E. H. Stanley, J. F. Walker, P. C. Hanson, J. C. Finlay, and R. G. Striegl (2014), CO2 and CH4 emissions from streams in a
lake-rich landscape: Patterns, controls and regional significance, Global Biogeochem. Cycles, 28, 197–210, doi:10.1002/2013GB004661.
del Giorgio, P. A., J. J. Cole, N. F. Caraco, and R. H. Peters (1999), Linking planktonic biomass and metabolism to net gas fluxes in north
temperate lakes, Ecology, 80, 1422–1431, doi:10.2307/177085.
Downing, J. A., et al. (2006), The global abundance and size distribution of lakes, ponds, and impoundments, Limnol. Oceanogr., 51,
2388–3297, doi:10.4319/lo.2006.51.5.2388.
Duarte, C., and Y. Prairie (2005), Prevalence of heterotrophy and atmospheric CO2 emissions from aquatic ecosystems, Ecosystems, 8,
862–870, doi:10.1007/s10021–005-0177-4.
Finlay, K., P. R. Leavitt, A. Patoine, and B. Wissel (2010), Magnitudes and controls of organic and inorganic carbon flux through a chain of hardwater lakes on the northern Great Plains, Limnol. Oceanogr., 55, 1551–1564, doi:10.4319/lo.2010.55.4.1551.
Granéli, W., M. Lindell, and L. Tranvik (1996), Photo-oxidative production of dissolved inorganic carbon in lakes of different humic content,
Limnol. Oceanogr., 41, 698–706, doi:10.4319/lo.1996.41.4.0698.
Hedin, L. O., J. C. von Fischer, N. E. Ostrom, B. P. Kennedy, M. G. Brown, and G. P. Robertson (1998), Thermodynamic constraints on nitrogen
transformations and other biogeochemical processes at soil-stream interfaces, Ecology, 79, 684–703, doi:10.1890/0012-9658(1998)079
[0684:TCONAO]2.0.CO;2.
Holgerson, M. A. (2015), Drivers of carbon dioxide and methane supersaturation in small, temporary ponds, Biogeochemistry, 124, 305–318,
doi:10.1007/s10533-015-0099-y.
Holmes, E., E. Ward, and K. Wills (2015), MARSS: Multivariate autoregressive state space modeling, R package version 3.9. [Available at http://
CRAN.R-project.org/package=MARSS.]
Hotchkiss, E. R., R. O. Hall, R. A. Sponseller, D. Butman, J. Klaminder, H. Laudon, M. Rosvall, and J. Karlsson (2015), Sources of and processes
controlling CO2 emissions change with size of streams and rivers, Nat. Geosci., 8, 696–699, doi:10.1038/ngeo2507.
Houser, J. N. (2001), Dissolved organic carbon in lakes: Effects on thermal structure, primary production, and hypolimnetic metabolism,
PhD thesis, Univ. Wisconsin-Madison.
Jonsson, A., J. Karlsson, and M. Jansson (2003), Sources of carbon dioxide supersaturation in clearwater and humic lakes in northern Sweden,
Ecosystems, 6, 224–235, doi:10.1007/s10021-002-0200-y.
2
3+
4+
Kelly, C. A., J. W. M. Rudd, and D. W. Schindler (1988), Carbon and electron flow via methanogenesis, SO4 , NO3 , Fe , and Mn reduction
in the anoxic hypolimnia of three lakes, Arch. Hydrobiol., 31, 333–344.
Knoll, L. B., M. J. Vanni, W. H. Renwick, E. K. Dittman, and J. A. Gephart (2013), Temperate reservoirs are large carbon sinks and small CO2
sources: Results from high-resolution carbon budgets, Global Biogeochem. Cycles, 27, 52–64, doi:10.1002/gbc.20020.
Koehler, B., T. Landelius, G. A. Weyhenmeyer, N. Machida, and L. J. Tranvik (2014), Sunlight-induced carbon dioxide emissions from inland
waters, Global Biogeochem. Cycles, 28, 696–711, doi:10.1002/2014GB004850.
Laurion, I., and N. Mladenov (2013), Dissolved organic matter photolysis in Canadian arctic thaw ponds, Environ. Res. Lett., 8, 035026,
doi:10.1088/1748-9326/8/3/035026.
Matthews, D. A., S. W. Effler, C. T. Driscoll, S. M. O’Donnell, and C. M. Matthews (2008), Electron budgets for the hypolimnion of a recovering
urban lake, 1989–2004: Response to changes in organic carbon deposition and availability of electron acceptors, Limnol. Oceanogr., 53,
743–759, doi:10.4319/lo.2008.53.2.0743.
Mattson, M. D., and G. E. Likens (1993), Redox reactions of organic matter decomposition in a soft water lake, Biogeochemistry, 19, 149–172,
doi:10.1007/BF00000876.
McDonald, C. P., E. G. Stets, R. G. Striegl, and D. Butman (2013), Inorganic carbon loading as a primary driver of dissolved carbon dioxide
concentrations in the lakes and reservoirs of the contiguous United States, Global Biogeochem. Cycles, 27, 285–295, doi:10.1002/gbc.20032.
Perkins, A. K., I. R. Santos, M. Sadat-Noori, J. R. Gatland, and D. T. Maher (2015), Groundwater seepage as a driver of CO2 evasion in a coastal
lake (Lake Ainsworth, NSW, Australia), Environ. Earth Sci., 74, 779–792, doi:10.1007/s12665-015-4082-7.
R Development Core Team (2015), R: A language and environment for statistical computing, R Foundation for Statistical Computing, Vienna.
[Available at http://www.R-project.org.]
Sobek, S., L. J. Tranvik, and J. J. Cole (2005), Temperature independence of carbon dioxide supersaturation in global lakes, Global Biogeochem.
Cycles, 19, GB2003, doi:10.1029/2004GB002264.
Stauffer, R. E. (1990), Electrode pH error, seasonal epilimnetic pCO2, and the recent acidification of the Maine lakes, Water Air Soil Pollut., 50,
123–148, doi:10.1007/BF00284788.
Stets, E. G., R. G. Striegl, G. R. Aiken, D. O. Rosenberry, and T. C. Winter (2009), Hydrologic support of carbon dioxide flux revealed by whole-lake
carbon budgets, J. Geophys. Res., 114, G01008, doi:10.1029/2008JG000783.
Striegl, R. G., and C. A. Michmerhuizen (1998), Hydrologic influence on methane and carbon dioxide dynamics at two north-central
Minnesota Lakes, Limnol. Oceanogr., 43, 1519–1529.
Torgersen, T., and B. Branco (2008), Carbon and oxygen fluxes from a small pond to the atmosphere: Temporal variability and the CO2/O2
imbalance, Water Resour. Res., 44, W02417, doi:10.1029/2006WR005634.
Vachon, D., and P. A. del Giorgio (2014), Whole-lake CO2 dynamics in response to storm events in two morphologically different lakes,
Ecosystems, 17, 1338–1353, doi:10.1007/s10021-014-9799-8.
Van de Bogert, M. C., S. R. Carpenter, J. J. Cole, and M. L. Pace (2007), Assessing pelagic and benthic metabolism using free water measurements,
Limnol. Oceanogr. Methods, 5, 145–155, doi:10.4319/lom.2007.5.145.
WILKINSON ET AL.
EXOGENOUS CO2 DOUBLES EFFLUX FROM LAKES
2002
Geophysical Research Letters
10.1002/2016GL067732
Van de Bogert, M. C., D. L. Bade, S. R. Carpenter, J. J. Cole, M. L. Pace, P. C. Hanson, and O. C. Langman (2012), Spatial heterogeneity strongly affects
estimates of ecosystem metabolism in two north temperate lakes, Limnol. Oceanogr., 57, 1689–1700, doi:10.4319/lo.2012.57.6.1689.
Wanninkhof, R. (1992), Relationship between wind speed and gas exchange over the ocean, J. Geophys. Res., 97, 2156–2202, doi:10.1029/92JC00188.
Weyhenmeyer, G. A., S. Kosten, M. B. Wallin, L. J. Tranvik, E. Jeppesen, and F. Roland (2015), Significant fraction of CO2 emissions from boreal
lakes derived from hydrologic inorganic carbon inputs, Nat. Geosci., 8, 933–936, doi:10.1038/ngeo2582.
Wilkinson, G. M., M. L. Pace, and J. J. Cole (2013), Terrestrial dominance of organic matter in north temperate lakes, Global Biogeochem. Cycles,
27, 43–51, doi:10.1029/2012GB004453.
Winslow, L., J. Zwart, R. D. Batt, J. Corman, H. Dugan, P. Hanson, A. Jaimes, J. S. Read, and R. Woolway (2015), LakeMetabolizer: Tools for the
analysis of ecosystem metabolism, R package version 1.3.1. [Available at http://CRAN.R-project.org/package=LakeMetabolizer.]
WILKINSON ET AL.
EXOGENOUS CO2 DOUBLES EFFLUX FROM LAKES
2003