SEARCH

SEARCH BY CITATION

Keywords:

  • carbon cycling;
  • freshwater;
  • large lakes

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Assembly of the Data Set
  5. 3. Environmental Controls on Carbon Fluxes in Large Lakes
  6. 4. Human Impacts: Eutrophication and Carbon Burial
  7. 5. Patterns at the Global Scale
  8. 6. Summary and Conclusions
  9. Acknowledgments
  10. References
  11. Supporting Information

[1] We present a synthesis of available estimates of primary production, organic carbon burial, and lake-atmosphere carbon dioxide exchange data for large lakes of the world. All three fluxes showed significant relationships with latitude and related climate variables, with lower production, higher evasion of carbon dioxide, and higher burial efficiency at higher latitudes. There was no relationship between raw organic carbon mass accumulation rates and latitude. Our estimates suggest that an order of magnitude more carbon is lost to the atmosphere by evasion than is buried in sediments at a global scale, with total global production, evasion, and burial fluxes of approximately 250, 90, and 7 Tg C yr−1. Finally, the data suggest a trend from autotrophy in low-latitude large lakes to heterotrophy and increasing reliance on allochthonous carbon sources in lakes at higher latitudes.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Assembly of the Data Set
  5. 3. Environmental Controls on Carbon Fluxes in Large Lakes
  6. 4. Human Impacts: Eutrophication and Carbon Burial
  7. 5. Patterns at the Global Scale
  8. 6. Summary and Conclusions
  9. Acknowledgments
  10. References
  11. Supporting Information

[2] Large lakes (>500 km2) of the world serve a number of critical roles to the societies surrounding them: they provide freshwater, transportation, and moderate climate in the surrounding regions [e.g., Bonan, 1995]; supply a large percentage of the protein to the local diet in many parts of the developing world [e.g., Molsa et al., 1999; Sarch and Birkett, 2000]; and often house significant fish and invertebrate biodiversity [e.g., Martens et al., 1994]. However, for the most part, fundamental aspects of large lake ecology and carbon cycling, such as the magnitude of carbon fluxes through primary production, sediment burial, and gas exchange, remain poorly documented in most of the 250+ large lakes in the world.

[3] The rate of organic carbon burial in lake basins has received increasing attention during the last decade [e.g., Dean and Gorham, 1998; Stallard, 1998; Einsele et al., 2001; Kortelainen et al., 2004]. Collectively, nearly half as much OC is buried in lakes globally as in the world's oceans (42 vs. ∼100 Tg C yr−1) [Dean and Gorham, 1998]. Small lakes (<500 km2) account for 60–70% of this total OC burial, but large freshwater and saline lakes still sequester an estimated 6–13% as much OC annually as the world ocean, despite having a collective basin area only 0.4% as large (calculated from data from Herdendorf [1982] and Dean and Gorham [1998]). However, on longer timescales, total organic carbon burial in large lake basins is important, as many lakes of tectonic origin are persistent features on the landscape, whereas many smaller lakes are more ephemeral. Organic carbon buried in the deposits of large lakes is substantial and likely to be sequestered out of contact with the atmosphere for long geological periods and thus represents a linkage between the short-term and long-term carbon cycles. As an illustration of this, approximately 20% of the oil and natural gas currently under production globally is derived from ancient large-lake basins [Bohacs et al., 2000]. Thus refining our understanding of the magnitude of carbon sequestration in large-lake basins will illuminate the contribution of this long-term OC sink to the global carbon cycle.

[4] Planktonic primary production serves as the basis for the food web in most large lakes. Photosynthetic uptake of dissolved inorganic carbon by phytoplankton consumes carbon dioxide (CO2) ultimately derived from the atmosphere and yields dissolved and particulate organic carbon (OC) in the water column, which fuels autotrophic and heterotrophic respiration within the lake basin as well as OC burial. Respiration may occur either within the surface mixed layer of lakes, with the resulting CO2 available for gas exchange with the atmosphere, or at greater depths, where the respired CO2 may remain out of contact with the atmosphere for significant periods of time in perennially stratified lakes. Particulate OC may also sink and undergo long-term burial in sediments. Inputs of allochthonous organic carbon may also follow these loss pathways but will not be discussed extensively here owing to the limited availability of well-quantified data on allochthonous OC inputs to large lakes [cf. Ramlal et al., 2003]. The magnitude of photosynthetic CO2 uptake and predominant mechanism of subsequent OC loss in a large lake, whether through burial or gas exchange, will determine whether the ecosystem constitutes a net source or sink of CO2 in the global carbon cycle. In the ocean, the removal of OC from surface waters to deep waters or sediments through the “biological pump” is a critical carbon sink and represents a key control on atmospheric CO2 concentrations through time [e.g., Falkowski et al., 2003]. We expect that the balance among the fluxes associated with primary production, respiration and gas exchange, and sedimentary OC burial will shift as climate conditions change with latitude.

[5] Despite widely differing rates of primary production and climatic milieus, most lakes across a wide range of latitudes have been reported to be supersaturated with carbon dioxide and undersaturated with oxygen with respect to the atmosphere most of the time, which suggests that lakes are a net source of CO2 to the atmosphere [e.g., Cole et al., 1994; Kling et al., 1991, 1992]. These observations have been interpreted as an indication that lakes are generally heterotrophic, i.e., that more carbon is consumed than fixed inside their basins, such that in-lake respiration must be subsidized by terrestrial inputs [e.g., Cole and Caraco, 2001]. However, respiration rates have generally been shown to be more constant through space and time in aquatic ecosystems than primary production, which occurs in distinct pulses or blooms [e.g., Biddanda and Cotner, 2002; Karl et al., 2003], and gas exchange acts to restore surface concentrations of dissolved gases toward atmospheric equilibrium concentrations. Low-frequency sampling of dissolved gases will likely fail to reflect major production events. Thus, by examining patterns of CO2 saturation as it varies with latitude and primary production, we may gain more insight into carbon cycle dynamics and controls than by considering the depauperate data available for most individual lake basins.

[6] Anthropogenic enrichment of nutrient inputs and changes in lake thermal/mixing regimes associated with global climate change stand to dramatically alter the magnitude of and controls on these fundamental ecological processes in large lake basins. Many large lake ecosystems are already starting to show signs of the impacts of global climate change on the magnitude of carbon cycle fluxes [e.g., O'Reilly et al., 2003, 2004; Verburg et al., 2003]. We present here a compilation of results-to-date on primary production, carbon burial, and lake-atmosphere CO2 exchange in 41 of the 250+ (∼16%) large lake basins around the world, undertaken in an effort to establish a “baseline” data set for comparison with future work and to identify significant gaps in our understanding of large lake ecology and biogeochemistry. While a number of recent studies have done elegant analyses of key carbon fluxes within individual large lake basins [e.g., McManus et al., 2003; Ramlal et al., 2003; Russ et al., 2004; Müller et al., 2005; Straškrábová et al., 2005; Urban et al., 2005], this is the first study to integrate carbon cycling observations across large lake basins at regional to global scales. Moreover, large lakes can be viewed as natural experiments in aquatic ecosystem dynamics, each with a unique set of physical and chemical properties, yet large enough to behave like miniature oceans, with potential to provide new insight into carbon cycling in the marine realm.

2. Assembly of the Data Set

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Assembly of the Data Set
  5. 3. Environmental Controls on Carbon Fluxes in Large Lakes
  6. 4. Human Impacts: Eutrophication and Carbon Burial
  7. 5. Patterns at the Global Scale
  8. 6. Summary and Conclusions
  9. Acknowledgments
  10. References
  11. Supporting Information

[7] We compiled estimates based on published data for carbon fluxes associated with production, sediment burial, and lake-atmosphere gas exchange in large lakes across a range of latitudes (17°S–67°N, Figure 1, auxiliary material Table S1). Published data on phytoplankton production, water chemistry, sedimentation, and lake basin attributes were gathered from peer-reviewed literature, books based on scientific literature, reports from government agencies, and reputable Internet sources.

image

Figure 1. Locations of lakes included in this study. Lakes are numbered in order of distance from the equator of their central latitude: 1, Edward; 2, Victoria; 3, Albert; 4, Kivu; 5, Turkana; 6, Tanganyika; 7, Maracaibo; 8, Malawi; 9, Nicaragua; 10, Managua; 11, Chad; 12, Titicaca; 13, Chapala; 14, Biwa; 15, Van; 16, Tahoe; 17, Great Salt; 18, Issyk Kul; 19, Caspian Sea; 20, Erie; 21, Ontario; 22, Huron; 23, Michigan; 24, Superior; 25, Constance; 26, Nipigon; 27, Winnipeg; 28, Baikal; 29, Peipus; 30, Vättern; 31, Becharof; 32, Ilmen; 33, Naknek; 34, Vänern; 35, Mälaren; 36, Iliamna; 37, Ladoga; 38, Onega; 39, Päijänne; 40, Great Slave; 41, Great Bear. All additional data on lake morphometric and limnological variables can be found in auxiliary materials.

Download figure to PowerPoint

2.1. Primary Production

[8] Estimates of primary production were based on in vitro methods using either 14C-CO2 (or occasionally 13C-CO2) uptake or oxygen change in light versus dark bottles. Primary production estimates were used only if the original authors calculated growing season or annual production rates on an areal basis (see auxiliary material Table S2). The two exceptions to this were lakes Iliamna and Becharof, for which only 4-hour midday estimates were available during the growing season; seasonal estimates were extrapolated from these results using reasonable assumptions about euphotic zone depth, growing season length, and the relationship between midday production and diel production (see auxiliary materials for further details).

[9] In vitro methods are prone to error from inconsistent incubation intervals among studies and bottle-related artifacts because they inhibit normal water circulation, which may lead to photoinhibition of phytoplankton within the bottles [e.g., Peterson, 1980; Marra, 2002]. The 14C (or 13C) uptake method, which has been most commonly used in large lake studies, has several additional sources of uncertainty, including recycling of unlabeled CO2 pools and loss of 14C-labeled DOC [e.g., Marra, 2002]. Estimates of primary production from the 14C method are thought to lie somewhere between the values of net and gross primary production (NPP and GPP, respectively); studies comparing methods indicate that the 14C method underestimates gross primary production by at least a factor of two [Grande et al., 1989; Juranek and Quay, 2005]. Further, some evidence suggests that the extent of this underestimation may vary depending on the production rate [cf. Luz et al., 2002]. For these reasons, we report all production estimates simply as primary or planktonic production (PP) and assume that these values fall between NPP and GPP.

[10] For a few lakes, modeled PP values, which attempt to correct for some of the limitations of traditional measurements, have been reported in the literature [e.g., Sarvala et al., 1999; Patterson et al., 2000]. These values are reported here but not included in statistical comparisons because their coverage is too sparse, and the differences from traditional measurements are substantial.

[11] Finally, several lakes included in this study underwent cultural eutrophication during the 20th century. At the basin-wide scale, lakes Victoria, Managua, Tahoe, Erie, Ontario, Constance, Mälaren, and Ladoga have experienced extensive eutrophication. Many of the other lakes show signs of eutrophication on local to subbasin scales (e.g., Biwa, Michigan). A subset of the eutrophied lakes are in various stages of reoligotrophication as a result of phosphorus abatement programs (Erie, Ontario, Constance).

[12] Historical PP measurements tend to be isolated in space and/or time, such that little is known about interannual variations or long-term trends in primary production in most large lakes. A few notable exceptions to this exist, such as Lake Tahoe, where dozens of productivity measurements have been taken in the lake during each growing season since 1959 [Jassby et al., 2001]. For all statistical comparisons of PP data, we made every effort to exclude values reflecting culturally eutrophied conditions by including only values thought to represent preeutrophication or postreoligotrophication PP values. However, effects of cultural eutrophication in data for some lakes undoubtedly remain and may account for some of the noise in the statistical relationships.

2.2. Organic Carbon Burial in Sediments

[13] Organic carbon mass accumulation rates (MAROC) for the past 100–1000 years of deposition in each lake were gathered from papers generally in one of two forms: (1) with previously calculated sediment mass accumulation rates (MARSED) and estimates of weight percent organic carbon in the sediment (%OC) or (2) from raw data including some combination of %OC, linear sedimentation rate (LSR) (cm yr−1), porosity (ϕ), and dry density of sediment (ρSED) (g cm−3) or the components needed to estimate ρSED (i.e., weight percentages of OC, biogenic silica [BSi], and siliciclastic sediments). In the simplest case, MAROC values were calculated using the equation

  • equation image

When needed, sediment MARs were calculated as

  • equation image

where LSR is typically established by either 14C or 210Pb chronology. Porosity can be approximated using water content data (%H2O, by weight) as

  • equation image

Finally, if ρSED required calculation, we used the equation

  • equation image

where 1.1, 2.1 and 2.65 are the assumed densities in g cm−3 of organic matter (2 × %OC), biogenic silica (BSi) and minerals (silicates and carbonates), respectively.

[14] The calculated MAROC values apply to the depositional zones of each lake from which cores are typically collected. In order to facilitate comparison between OC production and burial rate estimates, we generated basin-wide MAROC averages by multiplying the depositional zone MAROC by the percentage of the basin constituting the depositional zone (see auxiliary material Table S3). The percentage of a lake basin in the depositional zone varied between 30 and 85%, with deep, steep-sided lakes like Tanganyika and Malawi having ∼80–85% of their area in depositional zones, compared to 30–50% in shallower lakes with more gradual bathymetry, such as the Laurentian Great Lakes. For those lakes where depositional areas within the basin were not indicated in publications, we either estimated the percentage of lake area deeper than 100 m in the case of deep lakes (e.g., Tanganyika, Malawi), 50 m in the case of moderately deep lakes (e.g., Victoria), or applied a depositional zone percentage based on the lake's area and known relationship to surface wind wave dynamics [Johnson, 1980; Rowan et al., 1992].

[15] Finally, we estimated the burial efficiency of organic matter in large lakes. In the marine literature, burial efficiency is defined as the proportion of organic matter flux to the seafloor that is ultimately buried below the diagenetically active zone [Hedges and Keil, 1995]. Because data on the rates of organic matter settling to the basin floor are sparse for large lakes, we instead compared the annual water column primary production rate with the annual OC accumulation rate below the diagenetically active surficial sediments (i.e., below 5 cm depth). Thus our burial efficiency corresponds to the percent of annual PP buried in the sediment. Compared to estimates in the marine literature, our estimates of burial efficiency should be smaller, as substantial losses from sinking particulate OC would be expected between the time of its production in surface waters and its arrival at the sea/lake floor.

[16] We also calculated the total organic carbon content in the entire sediment column of a subset of the study lakes. The eight lakes chosen represent four tectonic basins: Tanganyika, Malawi, Biwa (including paleo-Biwa sediments in its watershed), and Baikal, with sediment packages of up to 8 km underlying their water columns, as well as four glacial lake basins, Erie, Ontario, Michigan, and Superior, each with only a thin veneer of postglacial sediments (0–20 m). The total volume of sediment packages in each lake basin was estimated from sources listed in Table 1, and modern organic carbon content and dry densities were assumed to be representative. For comparative purposes, we also estimated the total organic carbon content of the soils and aboveground and belowground biomass in the watershed of each lake on the basis of a World Resources Institute map (http://earthtrends.wri.org/text/climate-atmosphere/map-227.html).

Table 1. Total Amount of Organic Carbon in Watershed Soils Compared to Lake Sediments for Several Large Tectonic and Glacial Lake Basins
LakeWatershed OC Content,a Tg CLake OC Content,b Tg CLake:Watershed OC Ratio
Tectonic Lakes
Tanganyika440014,000,0003273
Malawi15005,500,0003729
Biwa713,80053
Baikal10004,500,000441
 
Glacial Lakes
Erie13002,1001.6
Ontario13002700.21
Michigan270017000.63
Superior48002,3000.47

2.3. Carbon Dioxide Concentrations and Lake-Atmosphere Exchange

[17] Partial pressures of carbon dioxide (pCO2, in μatm = 10−6 atm) were calculated from published measurements of pH, total alkalinity (TA, in μeq L−1) or dissolved inorganic carbon (DIC, in μmol L−1), and water temperature (°C) at 0–1 m depth in the water column, using the program CO2SYS [Lewis and Wallace, 1998]. Freshwater dissociation constants were used for all lakes except Van and Great Salt [Millero, 1979], where salinity exceeded 15 p.p.t. so the seawater constants of Roy et al. [1993, 1994, 1996] were deemed more appropriate. Because the salinity of lakes Turkana and Issyk Kul falls between the salinity ranges appropriate for analysis with freshwater and seawater constants, some error is introduced to the pCO2 and gas exchange calculations for these lakes. The quality of published water chemistry data used to calculate pCO2 is variable, with assessments of data quality shown in auxiliary material Table S4. An exception to this was Lake Baikal, where pCO2 and flux values were published without a description of methods [Kozhova and Izmest'eva, 1998].

[18] Using calculated pCO2 values, we calculated rudimentary estimates of lake-atmosphere CO2 exchange as follows. The wind speed-gas exchange relationship of Cole and Caraco [1998] was used to estimate the CO2 gas transfer velocity for freshwater at a temperature of 20°C (i.e., k600, cm h−1):

  • equation image

where U10 is the wind speed (m s−1) normalized to a height of 10 m above the water surface. Wind speed data were obtained from a variety of publications and other sources (listed in auxiliary material Table S4), including the National Data Buoy Center (NDBC) web site for the Laurentian Great Lakes. In order to convert from the k600 value to the gas transfer value at the observed temperature in each lake (kT), we used the equation

  • equation image

where ScT is the Schmidt number at temperature T (°C) and the exponent n depends on water surface conditions and varies from −0.67 to 1 [Cole and Caraco, 1998]. Consistent with numerous other authors, we use a value of −0.5 for n, typical of a wavy water surface free of films, which can inhibit gas exchange [Jähne et al., 1987]. The Schmidt number for CO2 in freshwater varies with temperature according to the following equation [Wanninkhof, 1992]:

  • equation image

Finally, lake-atmosphere exchange fluxes (FCO2) were calculated using the flux equation,

  • equation image

where ΔpCO2 is the water-air pCO2 gradient (pCO2water − pCO2air, in10−6 atm) and α is CO2 solubility in freshwater (mol L−1 atm−1), calculated as

  • equation image

where TK is temperature in degrees Kelvin [Weiss, 1974], and converted to units of mol cm−3 atm−1 by dividing by 1000. Flux values were converted from mol C cm−2 h−1 to units of g C m−2 yr−1 by multiplying by the appropriate conversion factors. Air pCO2 values were not reported, so we used a value of 377.4 μatm for 2004 and corrected it for the year of measurement by subtracting 1.4 μatm per year prior to that datum. Finally, solubility was corrected for atmospheric pressure at the elevation of each lake, calculated using an approximation of the barometric formula

  • equation image

where PZ and PSL (atm) represent atmospheric pressure at height Z (m) above sea level (SL), and 8000 approximates the value of the temperature term and constants in the exponent at 273.15° K [cf. Berberan-Santos et al., 1997].

[19] In all cases, pCO2 and wind data were used to calculate flux estimates at the highest resolution the data would afford (i.e., biweekly, monthly, or seasonal values where sufficient data existed) then combined by weighted averaging to generate an annual flux estimate. For most lakes, sufficiently resolved water chemistry and wind speed data do not exist at this juncture for the flux estimates to adequately reflect interseasonal or even intraseasonal variation in CO2 concentrations and fluxes. These flux estimates should be viewed as rudimentary.

2.4. Related Environmental Data and Statistical Analyses

[20] We also collected data on environmental parameters that may control carbon cycle processes. These parameters include basin morphology characteristics (e.g., surface area, volume, depth, drainage basin area, etc.), climatic data (e.g., insolation), and limnological conditions (e.g., mean annual water temperature) (auxiliary material Table S1). A major resource for these ancillary data was the World Lakes Database of the International Lake Environment Committee (http://www.ilec.or.jp/database/).

[21] We used the NASA Surface Meteorology and Solar Energy web site to generate average annual values for incident solar radiation on the surface of each lake, using the latitude and longitude of the center of each lake as its location (http://eosweb.larc.nasa.gov/sse/). The program returns 10-year averages of monthly values (July 1983 to June 1993) for the 1° × 1° cell centered on each input location, with an accuracy of 13–18%.

[22] Relationships among all carbon cycle variables (PP, MAROC, burial efficiency, pCO2, FCO2) and ancillary environmental data were explored and fitted using JMP 6. Only statistically significant relationships (α < 0.05) are reported (Table 2).

Table 2. Statistical Relationships Between Carbon Cycle Parameters and Environmental Parameters
ScaleResponse Variable (y)Latitude ZoneFactor (x)Best Fit Equationr2Fpnlakes
  • a

    Excluding outlying value from Lake Chad.

  • b

    Excluding outlying value from Lake Malawi.

  • c

    Excluding outlying value from Lake Winnipeg.

GlobalPP Latitudey = 1080 e−0.0612x0.7188.8<0.000139
   Annual insolationy = 0.322 e0.0015x0.7188.8<0.000139
   MAWTy = 18.6 e0.137x0.7270.5<0.000129
 BE Latitudey = 0.728 e0.0505x0.5227.6<0.000127
   PPy = 204 x−0.8000.6445.2<0.000127
   MAWTy = 18.4 e−0.105x0.4416.40.000623
 pCO2a Latitudey = 7.14 x + 3980.2912.10.00231
   PPy = −114 ln(x) + 12200.4220.40.000130
   MAWTy = −17.6 x + 9430.4218.00.000327
 FCO2a Latitudey = 1.29 x + 7.290.3012.20.001631
   PPy = −0.047 x + 72.20.248.70.006430
   MAWTy = −2.78 x + 97.90.3111.30.002527
RegionalPPtropicszavgy = 2508 x−0.3370.5310.00.0111a
  temperateW:LAy = 26.3 x + 24.80.519.30.0111
  highzavgy = 123 e−0.0442x0.7425.10.00111
 BEtropicszavgy = 0.0055 x + 0.6210.8124.80.0028b
  temperateLSRy = 57.0 x + 0.10.95119.3<0.00018c
  temperate-highzavgy = 158.5 x0.6990.244.50.0516

3. Environmental Controls on Carbon Fluxes in Large Lakes

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Assembly of the Data Set
  5. 3. Environmental Controls on Carbon Fluxes in Large Lakes
  6. 4. Human Impacts: Eutrophication and Carbon Burial
  7. 5. Patterns at the Global Scale
  8. 6. Summary and Conclusions
  9. Acknowledgments
  10. References
  11. Supporting Information

3.1. Characteristics of Lakes in the Database

[23] Thirteen of the 41 lakes represented in the database fall within tropical latitudes (0–20°), fifteen in temperate latitudes (35–53°), and thirteen in higher latitudes (57–66°) (auxiliary material Table S1). Salinity in the lakes ranges from 0 to 50 p.p.t., but 36 of the 41 lakes are typified by salinities between 0 and 1.5 p.p.t. and thus constitute truly freshwater systems. The five lakes with higher salinity are, in order of increasing salinity: Turkana, Issyk Kul, Caspian, Van, and Great Salt (range: 2.7–49.8 p.p.t.). The 41 lakes included in this study cumulatively represent 71% of the total surface area and 67% of the total volume of freshwater lakes in the world, as well as 57% of the area and 78% of the volume of saline lakes globally [Herdendorf, 1982].

[24] Individual lakes within the database span a range of basin morphology characteristics. Although the size cutoff for inclusion of lakes in this study is 500 km2, the majority of lakes included are substantially larger than this. Four lakes fall in the range of 500–1000 km2, 21 between 1000 and 10,000 km2, fifteen between 10,000 and 100,000 km2, and one >100,000 km2 (auxiliary material Table S1). Table S1 summarizes the range of volumes, surface elevations and average depths of the lake basins.

3.2. Primary Production

[25] Primary production (PP) in large lakes ranges from less than 10 g C m−2 yr−1 at high latitudes to nearly 1900 g C m−2 yr−1 in the tropics (Figure 2 and auxiliary material Table S2). The log of PP shows a statistically significant negative relationship with latitude at the global scale. Annual primary production also exhibits a positive exponential relationship with other variables correlated with latitude, such as annual insolation and mean annual water temperature (MAWT) (Figures 2a–2c and Table 2). Latitude fundamentally determines the availability of light to provide energy for photosynthesis. The decrease in annual insolation with latitude across the large lakes in this data set is best fit with a second-order polynomial curve (r2 = 0.91, not shown). Mean annual water temperature, which can limit productivity at the lower end of the observed range, is linearly correlated with both latitude (r2 = 0.89, not shown) and insolation (r2 = 0.78, not shown). Nutrient influx to lakes should also be on the order of 100 times higher in the tropics than at the poles as a result of the intense weathering that occurs under tropical precipitation regimes [e.g., Schindler, 1978]. None of the environmental controls on PP, with the possible exception of nutrient weathering for which we do not have direct data, decreases exponentially with latitude. However, the exponential relationships shown between PP and latitude, insolation, and MAWT in Figure 2 and Table 2 provide 11–23% more explanatory power than linear relationships because they accommodate the high variability in PP among tropical lakes and the low PP values in high-latitude lakes without yielding negative PP values. Thus we do not attempt with the current data to differentiate the relative strength of the three latitude-controlled variables, insolation, nutrient delivery, and water temperature, in controlling lacustrine primary production at the global scale. Finally, we note that the errors in measured PP rates associated with the 14C uptake method probably also introduce some error into the relationship between latitude and PP rates.

image

Figure 2. Global-scale relationships between annual primary production and environmental variables. Environmental variables are (a) latitude, (b) incident solar radiation calculated for each lake, and (c) mean annual water temperature (MAWT). Tropical lakes are indicated with circles, temperate lakes with squares, and high-latitude lakes with triangles. Regional-scale relationships between annual primary production and basin morphometry variables observed within latitude zones: (d) tropical lakes and average depth (zavg), (e) temperate lakes and watershed to lake area ratio, and (f) high-latitude lakes and zavg. Note that some axes are shown in linear scale, others in log scale. Statistical outliers excluded from regressions are indicated by gray symbols and identified in Table 2.

Download figure to PowerPoint

[26] Within latitude zones, a few additional correlations were observed between basin morphometry variables and annual PP (Figures 2d−2f and Table 2). In the tropics, lake average depth (zavg, calculated as the inverse of the surface area to volume ratio where depth data were lacking) exhibits a negative power relationship with annual PP. At high latitudes, PP declines exponentially with zavg. Average depth is linked to water column mixing and nutrient availability in the euphotic zone. Watershed to lake area (W:LA) ratios are positively correlated to PP rates in temperate latitudes. W:LA should scale with nutrient inputs to a basin. Thus, within regions, basin morphometric variables linked to nutrient input and availability are correlated with primary production rates.

[27] Despite methodological differences in PP measurements among study lakes, this data synthesis shows that PP varies as a function of latitude in large lakes as it does in small lakes and reservoirs [cf. Brylinsky and Mann, 1973]. The latitude-related variables of water temperature, insolation, and probably also nutrient input control primary production in large lakes at global and regional scales [cf. Schindler, 1978].

3.3. Organic Carbon Burial in Sediments

[28] Burial rates of organic carbon in large lakes range between about 1 and 20 g C m−2 yr−1 (auxiliary material Table S3). At the global scale, we found no significant relationships between organic carbon mass accumulation rates and either latitude or primary production (r2 values of 0.12 and 0.01, respectively, not shown). After normalizing annual MAROC to PP rates by determining the burial efficiency (BE = PP/MAROC, expressed as a percent), it becomes clear that BE, ranging between about 0.3 and 60% in most lakes, is correlated with latitude, PP, and MAWT (Figures 3a–3c and Table 2). Burial efficiency increases exponentially with latitude, exceeding 100% in one high-latitude lake, reflecting either poor data quality or relatively substantial burial of terrigenous organic matter, and shows a negative power relationship with PP. The exponential decline of BE with increasing mean annual water temperature suggests that colder water facilitates higher burial efficiency by limiting remineralization rates.

image

Figure 3. Global-scale relationships between organic carbon burial efficiency (% of annual PP buried) and environmental variables: (a) latitude, (b) annual primary production, and (c) mean annual water temperature (MAWT). Tropical lakes are indicated with circles, temperate lakes with squares, and high-latitude lakes with triangles. Note that one data point lies off the chart at the location indicated by the arrow in Figures 3a-3c (Lake Vänern, burial efficiency = 190%). Regional-scale relationships between OC burial efficiency and environmental variables within latitude zones: (d) tropical, (e) temperate, and (f) temperate and high latitudes. LSR is the linear sedimentation rate. Note that some axes are shown in linear scale, others in log scale. Statistical outliers excluded from regressions are indicated by gray symbols and identified in Table 2.

Download figure to PowerPoint

[29] At regional scales, burial efficiency is correlated to basin morphometry and other variables that may affect oxygen exposure time for sinking particules (Figures 3d–3f and Table 2). In the tropics, burial efficiency increases with average depth. Permanently stratified deep lakes in the tropics have anoxic hypolimnia, which reduces the potential for oxidation as well as resuspension and reworking of particulate organic carbon once it sinks below the mixed layer. Indeed, burial efficiency in the tropics appears to be favored by a deeper water column. In contrast, burial efficiency has the opposite relationship with average depth at temperate and high latitudes (Figure 3f), with shallower water columns correlated with greater burial efficiency. The water columns of large lakes at high latitudes are generally well oxygenated to the bottom, providing more opportunity for oxic bacterial breakdown of organic particles as they sink through the water column. We reason that particulate organic matter settles to the lake floor faster in a shallow lake than in a deep lake before being buried to the depth of sedimentary anoxia, thereby enhancing its chance for preservation. Along the same lines, in temperate lakes, linear sedimentation rates correlate with burial efficiency, supporting the idea that hastier removal from oxic conditions enhances OC preservation in lake sediments. Without data on oxygen penetration depth in lake sediments, it is not possible to calculate oxygen exposure times to test this idea [cf. Hartnett et al., 1998]. However, large temperate lakes generally overturn at least once per year and maintain permanently oxic water columns, so it is reasonable to expect that linear sedimentation rates would correlate with average time to burial below the diagenetically active zone at the sediment surface.

[30] In marine settings, less than 0.5% of primary production is ultimately preserved in sediments [Hedges and Keil, 1995]. In contrast, burial efficiencies calculated here in the same way ranged from 0.3 to 60% (mean = 9 ± 13%, median = 3, excluding one outlier) and increased with latitude. A careful analysis of new production, sedimentation, and early diagenesis in Lake Baikal [Müller et al., 2005] used in combination with the production numbers in Table 2, suggests that burial efficiency in most large lakes is of the same approximate magnitude and controlled by the same factors as in the oceans [cf. Hedges and Keil, 1995]. A study on OC accumulation rates and early diagenesis in Lake Superior also supports this conclusion [Johnson et al., 1982]. Both colder mean annual surface water temperature and factors that may reduce oxygen exposure times, high sedimentation rates and favorable average depths (high at low latitudes and vice versa), appear to enhance organic matter preservation in large lake sediments [cf. Hartnett et al., 1998].

3.4. Carbon Dioxide Concentrations and Lake-Atmosphere Exchange

[31] Partial pressures of CO2 in the surface waters of large lakes range between 0 and ∼27,600 μatm, with an average across latitudes of 850 μatm (auxiliary material Table S4). Fluxes of CO2 between large lakes and the atmosphere range from a small net invasion of −29 g C m−2 yr−1 to a large net evasion of 718 g C m−2 yr−1, with an overall average of 72 g C m−2 yr−1. Both pCO2 values and CO2 lake-atmosphere exchange fluxes (FCO2), primarily from lake to atmosphere, increase with latitude and decrease with increases in both PP and MAWT (Figure 4 and Table 2). Higher PP rates draw down CO2 concentrations in surface waters. CO2 is more soluble in colder water, so the higher saturation levels seen at lower temperature are not attributable to water temperature, as warming of surface water alone would increase CO2 saturation in the absence of biological processes. Thus, on the basis of CO2 concentrations, we see a gradient ranging from essentially autotrophic lakes at tropical latitudes to increasingly heterotrophic lakes at higher latitudes. It is not surprising that the observed pCO2 values for most tropical lakes are somewhat supersaturated (Figure 4), given the greater likelihood of measuring supersaturated CO2 concentrations to detecting a surplus of O2 during a random sampling and thus falsely concluding that conditions are heterotrophic [cf. Karl et al., 2003]. With few exceptions (4 of 13 lakes), all tropical lakes had at least one undersaturated CO2 value (10 of 109 values, or 9%).

image

Figure 4. Global-scale relationships between CO2 concentrations (pCO2), lake-atmosphere CO2 exchange fluxes (FCO2), and environmental variables. Environmental variables are (a, d) latitude, (b, e) annual primary production (PP), and (c, f) mean annual water temperature. Positive FCO2 values correspond to net evasion of CO2 from the lake to the atmosphere, negative FCO2 reflect net invasion, and zero value indicates no net gas exchange. Tropical lakes are indicated with circles, temperate lakes with squares, and high-latitude lakes with triangles.

Download figure to PowerPoint

[32] To further illustrate sampling bias in favor of CO2 supersaturation, a total of 462 individual sampling events in oligotrophic Lake Superior during the 1968 and 1969 growing seasons yielded only five undersaturated pCO2 values in surface open water (1%). In contrast, an in situ pCO2 probe (SAMI CO2 sensor, Sunburst Instruments) experimentally deployed in Lake Superior's western arm revealed pCO2 levels ranging from 110–510 μatm during high-frequency measurements between June and September 2001, including a 2-week period of CO2 undersaturation not linked to cooler water temperatures (M. Baehr et al., unpublished data, 2001). This observation of relatively prolonged CO2 undersaturation is consistent with oxygen stable isotope results indicating that production rates exceed respiration rates in the epilimnion for part of the thermally stratified period in Lake Superior [Russ et al., 2004]. Nonetheless, Lake Superior is clearly net heterotrophic on an annual timescale [e.g., Urban et al., 2005].

[33] The dominant controls on gas exchange in lake environments include surface water pCO2 levels, wind speed, and penetrative convection [e.g., Cole and Caraco, 1998; MacIntyre et al., 2002; Sobek et al., 2005]. The quality and sampling density of published data available for estimating lake-atmosphere fluxes was generally the weakest of those presented in this study, because most of the data were not collected with the intention of estimating gas exchange fluxes. Thus the flux estimates given here are inherently rudimentary. The pCO2 calculations are very sensitive to errors in measured pH values, which may be considerable in freshwater [cf. French et al., 2002] and subject to changing methods through time. Moreover, most wind speed values used to calculate fluxes were averaged over monthly or longer time periods, which does not reflect the substantial gas exchange associated with storms. Additionally, we did not attempt to incorporate a term for gas exchange related to rainfall, which can be significant [Ho et al., 1997, 2000]. These factors suggest that our gas exchange estimates may be underestimated.

[34] For comparison, Weiler [1974, 1975] estimated significantly higher gas evasion fluxes from lakes Erie and Ontario (400–500 g C m−2 yr−1) than those presented here. Weiler's evasion rates are double to triple the basin-wide estimates of primary production in these mesotrophic-eutrophic lakes, and on the basis of our estimates, appear to be overestimates. On the other hand, Urban et al. [2005] estimate that 3 Tg C yr−1 (1 Tg = 1012 g) (∼37 g C m−2 yr−1) is outgassed as CO2 from Lake Superior each year, which is only 25% of our estimated net annual evasion. The discrepancy could result from data quality issues in this study or to the Urban et al. study sites not being representative of the lake as a whole. Combined, these results suggest that our estimates are in the right ballpark.

4. Human Impacts: Eutrophication and Carbon Burial

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Assembly of the Data Set
  5. 3. Environmental Controls on Carbon Fluxes in Large Lakes
  6. 4. Human Impacts: Eutrophication and Carbon Burial
  7. 5. Patterns at the Global Scale
  8. 6. Summary and Conclusions
  9. Acknowledgments
  10. References
  11. Supporting Information

[35] Several of the lakes in this study have undergone varying degrees of cultural eutrophication during the 20th century: Victoria, Managua, Tahoe, Erie, Ontario, Constance, Mälaren, and Ladoga. Primary production rate measurements are available for at least two distinct time periods with respect to basin-wide productivity levels for all of these lakes except Managua and Mälaren (Figure 5a and auxiliary material Table S2). Lakes Victoria, Tahoe, and Ladoga had either preeutrophication PP measurements or reasonable estimates thereof in the literature to compare to posteutrophication values; these lakes experienced 88–500% increases in PP over timescales of a few to several decades, with percent of increase scaling inversely with lake surface area. Only postreoligotrophication values were available for lakes Erie, Ontario, and Constance. The apparent increases in PP associated with eutrophication in these lakes, relative to the postreoligotrophication rates, ranged from 31% (Constance) to 135% (Erie). It is possible that the postreoligotrophication and preeutrophication PP rates were different or that not all studies captured PP rates at peak eutrophic conditions.

image

Figure 5. Changes in (a) primary production, (b) OC mass accumulation rates, and (c) burial efficiency associated with eutrophication in lakes Victoria (V), Tahoe (T), Erie (E), Ontario (O), Constance (C), and Ladoga (L). Historical (or postreoligotrophication, indicated with asterisks) values are in black and posteutrophication values are shown in white bars.

Download figure to PowerPoint

[36] Annual carbon burial rates also changed with lake eutrophication. For lakes that had both pre- and post-eutrophication OC mass accumulation rate data reported (Victoria, Erie, Ontario, Constance), MAROC values increased between 52 and 304%, although the latter value may not account for early sediment diagenesis, and may also include some error based on averaging MAROC values from different studies for the pre- and post-eutrophication values (Figure 5b). Excluding the highest value, the range is 52–248%. Burial efficiencies did not necessarily increase with eutrophication, with change relative to preeutrophication efficiencies ranging from −24% to +166% (Figure 5c). In any case, whether or not annual OC burial rates and efficiencies increase with eutrophication, the total percentage of organic carbon buried in a basin remains quite small (for this comparison: 7 ± 10% preeutrophication and 9 ± 11% under eutrophic conditions). The vast majority of carbon fixed within the basin remains available for consumption, respiration, and gas exchange. However, this would not necessarily result in a net increase of CO2 emitted from the basin, but may rather appear as an increase in the turnover rate of atmospheric CO2 through primary production.

5. Patterns at the Global Scale

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Assembly of the Data Set
  5. 3. Environmental Controls on Carbon Fluxes in Large Lakes
  6. 4. Human Impacts: Eutrophication and Carbon Burial
  7. 5. Patterns at the Global Scale
  8. 6. Summary and Conclusions
  9. Acknowledgments
  10. References
  11. Supporting Information

5.1. Estimated Global Carbon Fluxes Through Large Lakes

[37] We estimated the magnitude of global carbon fluxes through production, burial, and gas exchange in large lakes on the basis of the relationships observed here between carbon fluxes and latitude, specifically PP, BE, and FCO2 versus latitude (Table 3). Herdendorf [1982] supplied the cumulative surface area of large lakes, also defined as >500 km2 in his study, across latitudes. Using PP, BE, and FCO2 values for the midpoint of each 10° latitude band, with northern and southern latitude areas summed, we calculated average flux rates for each term using the relationships in Table 2, then scaled up by area in each zone before summing to obtain the estimated global total. We excluded the three lakes >500 km2 at latitudes higher than 70° from Herdendorf, so as not to extrapolate outside the range of our data coverage.

Table 3. Global Estimates of Annual Production, Burial, and Gas Exchange Fluxes in All Large Lakes
Latitude, degnlakesaTotal Area,a km2Average PP,b g C m−2 yr−1Average BE,b %Average FCO2,b g C m−2 yr−1Annual PP, Tg C yr−1Annual MAROC, Tg C yr−1Annual FCO2, Tg C yr−1%PP Recycled in Surface Waters
  • a

    Data on area and number of lakes in each 10° latitude band from Herdendorf [1982].

  • b

    Average values based on middle value in latitude range and correlations with latitude in Table 2.

0–1019151,1607950.9141201.12.197
10–2018123,8704311.627530.833.392
20–30929,3792342.6406.90.181.280
30–403774,9881274.3529.50.413.955
40–5051768,279697.165533.750−13
50–6062159,1773711.7785.90.6912−115
60–7054142,9342019.4912.90.5613−368
Total2501,449,7872517.4786−23

[38] We estimate the total annual PP flux in large lake basins to be 250 Tg C yr−1 and losses of organic carbon in the forms of MAROC and CO2 evasion to be on the order of 7 and 90 Tg C yr−1, respectively. Dean and Gorham [1998] estimated OC burial rates of 11 Tg C yr−1 in large lakes and inland seas (both of which are included in our estimate), although they use 5000 km2 as their size cutoff for large lakes. Our results are broadly consistent with the Dean and Gorham [1998] estimate, and the small discrepancy appears to result from the uncertainty in the total global surface area of large lakes. Our estimates suggest that evasion accounts for an order of magnitude greater loss than carbon burial from large lake ecosystems and that carbon losses through both pathways total nearly 100 Tg C yr−1 or 40% of global large lake PP (Table 3).

[39] According to our calculations (Table 3), PP exceeds burial plus evasion 40-fold near the equator, indicating that 97% of autochthonous production appears to be recycled within these warm aquatic ecosystems at any given time rather than lost from the system through outgassing or burial. At the other extreme, it appears that autochthonous PP can only account for 22% of the CO2 evaded from lakes at 60–70° latitudes, suggesting that heterotrophy and reliance on terrestrial subsidies increase with latitude in large lakes.

5.2. Storage of Organic Carbon in Terrestrial Versus Aquatic Environments

[40] Large differences may exist between the total amount of organic carbon stored in lake sediments and in terrestrial biomes within their watersheds. The magnitude of total carbon storage in a lake basin relative to the soils and biota in its watershed appears to depend more on the geological origin of the basin than latitude, although annual rates of sedimentary OC accumulation do not differ significantly among basin types (auxiliary material Tables S1 and S3). In each of four large tectonic lake basins, orders of magnitude more organic carbon are stored in lake sediments than in their watersheds. Ranging in latitude from Lake Tanganyika (3°S–8°S) to Lake Baikal (51°N–56°N), the sediments of these tectonic lakes contain 101.7 to 103.6 times as much organic carbon as is stored in the soils and aboveground and belowground vegetation of their watersheds (Figure 6 and Table 1). These results provide a window into the importance of large lake basins in sequestering organic carbon over longer geological time periods. By contrast, large lakes of glacial origin have been accumulating carbon for only roughly the past 10,000 years. Sediment thickness underlying these basins is orders of magnitude lower than what lies buried beneath lakes of tectonic origin. Consequently, the amount of organic carbon stored in several of the Laurentian Great Lakes ranges from 10−0.7 to 100.2 times the drainage basin inventory.

image

Figure 6. Order of magnitude differences between total organic carbon content of lake sediments and watershed soils and biota for tectonic (black) and glacial (gray) large-lake basins. Lakes included are Tanganyika (T), Malawi (M), Biwa (Bi), Erie (E), Ontario (O), Michigan (Mi), Superior (S), and Baikal (Ba).

Download figure to PowerPoint

6. Summary and Conclusions

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Assembly of the Data Set
  5. 3. Environmental Controls on Carbon Fluxes in Large Lakes
  6. 4. Human Impacts: Eutrophication and Carbon Burial
  7. 5. Patterns at the Global Scale
  8. 6. Summary and Conclusions
  9. Acknowledgments
  10. References
  11. Supporting Information

[41] We have presented our current best estimates of organic carbon production, burial, and CO2 exchange fluxes in large lakes of the world. Production appears to be controlled primarily by factors associated with latitude (insolation, water temperature, and watershed weathering/nutrient delivery) at the global scale and by factors influencing nutrient cycle dynamics at broad regional scales. Burial efficiency in large lakes is most correlated with water temperature and factors that may reflect oxygen exposure time. Although CO2 concentration and flux estimates are less robust, a trend is suggested from autotrophic to increasingly heterotrophic conditions with increasing latitude and decreasing mean annual water temperature.

[42] We estimate that the global large-lake fluxes of carbon through production, evasion, and burial are on the order of 250, 90, and 7 Tg C yr−1, respectively. Thus roughly an order of magnitude more carbon is lost to the atmosphere than to sediment burial annually in large lakes. The remaining amount of organic carbon from lake PP available to support the food web decreases with latitude from 80 to 100% in the tropics to negative rates (−110 to −370%) at the highest latitudes, implying that the importance of allochthonous carbon subsidies increases dramatically with latitude. These data constitute a late 20th century benchmark against which to evaluate future rate studies both on an individual lake basis and relative to global trends.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Assembly of the Data Set
  5. 3. Environmental Controls on Carbon Fluxes in Large Lakes
  6. 4. Human Impacts: Eutrophication and Carbon Burial
  7. 5. Patterns at the Global Scale
  8. 6. Summary and Conclusions
  9. Acknowledgments
  10. References
  11. Supporting Information

[43] We thank Paul Baker, Bruce Finney, Doug Ricketts, and Jim Russell for sharing data and wisdom about carbon fluxes for several lakes included in this study, as well as Matt Baehr, Jon Cole, Gordon Holtgrieve, Ernie Lewis, Jim Murray, Paul Quay, and Joe Werne for helpful discussions. S. R. A. is particularly grateful to the NOAA Climate and Global Change Postdoctoral Fellowship Program for making this research possible and to Jim McManus for support and feedback throughout this project, as well as insightful comments on an earlier draft of the manuscript. The authors thank two anonymous reviewers and the Associate Editor for comments that significantly improved the manuscript.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Assembly of the Data Set
  5. 3. Environmental Controls on Carbon Fluxes in Large Lakes
  6. 4. Human Impacts: Eutrophication and Carbon Burial
  7. 5. Patterns at the Global Scale
  8. 6. Summary and Conclusions
  9. Acknowledgments
  10. References
  11. Supporting Information
  • Berberan-Santos, M. N., et al. (1997), On the barometric formula, Am. J. Phys., 65, 404412.
  • Biddanda, B. A., and J. B. Cotner (2002), Love handles in aquatic ecosystems: The role of dissolved organic carbon drawdown, resuspended sediments, and terrigenous inputs in the carbon balance of Lake Michigan, Ecosystems, 5, 431445.
  • Bohacs, K. M., et al. (2000), Lake-basin type, source potential, and hydrocarbon character: An integrated sequence-stratigraphic–geochemical framework, in Lake Basins Through Space and Time, edited by E. Gierlowski-Kordesch, and K. Kelts, pp. 334, Am. Assoc. of Pet. Geol., Tulsa, Okla.
  • Bonan, G. B. (1995), Sensitivity of a GCM simulation to inclusion of inland water surfaces, J. Clim., 8, 26912708.
  • Brylinsky, M., and K. H. Mann (1973), An analysis of factors governing productivity in lakes and reservoirs, Limnol. Oceanogr., 18, 114.
  • Cole, J. J., and N. F. Caraco (1998), Atmospheric exchange of carbon dioxide in a low-wind oligotrophic lake measured by the addition of SF6, Limnol. Oceanogr., 43, 647656.
  • Cole, J. J., and N. F. Caraco (2001), Carbon in catchments: Connecting terrestrial carbon losses with aquatic metabolism, Mar. Freshwater Res., 52, 101110.
  • Cole, J. J., et al. (1994), Carbon dioxide supersaturation in the surface waters of lakes, Science, 265, 15681570.
  • Colman, S. M., and D. S. Foster (1994), A sediment budget for southern Lake Michigan: Source and sink models for different time intervals, J. Great Lakes Res., 20, 215228.
  • Dean, W. E., and E. Gorham (1998), Magnitude and significance of carbon burial in lakes, reservoirs, and peatlands, Geology, 26, 535538.
  • Einsele, G., et al. (2001), Atmospheric carbon burial in modern lake basins and its significance for the global carbon budget, Global Planet. Change, 30, 167195.
  • Falkowski, P. G., et al. (2003), Phytoplankton and their role in primary, new, and export production, in Ocean Biogeochemistry: The Role of the Ocean Carbon Cycle in Global Change, edited by M. J. R. Fasham, pp. 99121, Springer, New York.
  • French, C. R., et al. (2002), Spectrophotometric pH measurements of freshwater, Anal. Chim. Acta, 453, 1320.
  • Grande, K. D., et al. (1989), Primary production in the North Pacific gyre: A comparison of rates determined by the 14C, O2 concentration and 18O methods, Deep Sea Res., 36, 16211634.
  • Hartnett, H. E., et al. (1998), Influence of oxygen exposure time on organic carbon preservation in continental margin sediments, Nature, 391, 572574.
  • Hedges, J. I., and R. G. Keil (1995), Sedimentary organic matter preservation: An assessment and speculative synthesis, Mar. Chem., 49, 81115.
  • Herdendorf, C. E. (1982), Large lakes of the world, J. Great Lakes Res., 8, 379412.
  • Ho, D. T., et al. (1997), The effect of rain on air-water gas exchange, Tellus, Ser. B, 49, 149158.
  • Ho, D. T., W. E. Asher, L. F. Bliven, P. Schlosser, and E. L. Gordan (2000), On mechanisms of rain-induced air-water gas exchange, J. Geophys. Res., 105, 24,02524,057.
  • Horie, S. (1981), On the significance of paleolimnological studies of ancient lakes—Lake Biwa and other relict lakes, Verh. Int. Verein. Limnol., 21, 1344.
  • Jähne, B., K. O. Münnich, R. Bösinger, A. Dutzi, W. Huber, and P. Libner (1987), On the parameters influencing air-water gas exchange, J. Geophys. Res., 92, 19371949.
  • Jassby, A. D., et al. (2001), Lake Tahoe: Diagnosis and rehabilitation of a large mountain lake, in The Great Lakes of the World (GLOW): Food-Web, Health and Integrity, edited by M. Munawar, and R. E. Hecky, pp. 431454, Backhuys, Leiden, Netherlands.
  • Johnson, T. C. (1980), Sediment redistribution by waves in lakes, reservoirs and embayments, in Proceedings of the Symposium on Surface Water Impoundments, edited by H. Stefan, pp. 13071317, Soc. of Civ. Eng., New York.
  • Johnson, T. C., et al. (1982), Total organic carbon in Lake Superior sediments: Comparisons with hemipelagic and pelagic marine environments, Limnol. Oceanogr., 27, 481491.
  • Juranek, L. W., and P. D. Quay (2005), In vitro and in situ gross primary and net community production in the North Pacific Subtropical Gyre using labeled and natural abundance isotopes of dissolved O2, Global Biogeochem. Cycles, 19, GB3009, doi:10.1029/2004GB002384.
  • Karl, D. M., et al. (2003), Metabolic balance of the open sea, Nature, 426, 32.
  • Kemp, A. L. W., et al. (1978), Sedimentation rates and a sediment budget for Lake Superior, J. Great Lakes Res., 4, 276287.
  • Kling, G. W., et al. (1991), Arctic lakes and streams as gas conduits to the atmosphere: Implications for tundra carbon budgets, Science, 251, 298301.
  • Kling, G. W., et al. (1992), The flux of CO2 and CH4 from lakes and rivers in arctic Alaska, Hydrobiologia, 240, 2336.
  • Kortelainen, P., et al. (2004), A large carbon pool and small sink in boreal Holocene lake sediments, Global Change Biol., 10, 16481653.
  • Kozhova, O. M., and L. R. Izmest'eva (1998), Lake Baikal: Evolution and Biodiversity, Backhuys, Leiden, Netherlands.
  • Lewis, E., and D. Wallace (1998), Program developed for CO2 system calculations, report, Carbon Dioxide Inf. Anal. Cent., Oak Ridge Natl. Lab., Oak Ridge, Tenn.
  • Luz, B., et al. (2002), Evaluation of community respiratory mechanisms with oxygen isotopes: A case study in Lake Kinneret, Limnol. Oceanogr., 47, 3342.
  • MacIntyre, S., et al. (2002), The critical importance of buoyancy flux for gas flux across the air-water interface, in Gas Transfer at Water Surfaces, Geophys. Monogr. Ser., vol. 127, edited by M. A. Donelan et al., pp. 135139, AGU, Washington, D. C.
  • Marra, J. (2002), Approaches to the measurement of plankton production, in Phytoplankton Productivity and Carbon Assimilation in Marine and Freshwater Ecosystems, edited by P. J. Williams et al., pp. 78108, Blackwell, Malden, Mass.
  • Martens, K., et al. (Eds.) (1994), Speciation in Ancient Lakes, Schweizerbart'sche Verlagsbuchhandlung, Stuttgart, Germany.
  • McFadden, M. A., et al. (2004), Paleoproductivity of eastern Lake Ontario over the past 10,000 years, Limnol. Oceanogr., 49, 15701581.
  • McManus, J., et al. (2003), Hypolimnetic oxidation rates in Lake Superior: Role of dissolved organic material on the lake's carbon budget, Limnol. Oceanogr., 48, 16241632.
  • Millero, F. J. (1979), The thermodynamics of the carbonate system in seawater, Geochim. Cosmochim. Acta, 43, 16511661.
  • Molsa, H., et al. (1999), Fisheries research toward resource management on Lake Tanganyika, Hydrobiologia, 407, 124.
  • Müller, B., et al. (2005), Internal carbon and nutrient cycling in Lake Baikal: Sedimentation, upwelling, and early diagenesis, Global Planet. Change, 46, 101124.
  • O'Reilly, C. M., et al. (2003), Climate change decreases aquatic ecosystem productivity of Lake Tanganyika, Africa, Nature, 424, 766768.
  • O'Reilly, C. M., et al. (2004), Reply to “Climate change effect on Lake Tanganyika?,”, Nature, 430, doi:10.1038/nature02737.
  • Patterson, G., et al. (2000), Effect of hydrological cycles on planktonic primary production in Lake Malawi/Niassa, in Ancient Lakes: Biodiversity, Ecology and Evolution, edited by A. Rossiter, and H. Kawanabe, pp. 421430, Academic, London.
  • Petersen, R. R. (1975), A paleolimnological study of the eutrophication of Lake Erie, Verh. Int. Verein. Limnol., 19, 22742283.
  • Peterson, B. J. (1980), Aquatic primary productivity and the 14C-CO2 method: A history of the productivity problem, Annu. Rev. Ecol. Syst., 11, 359385.
  • Ramlal, P. S., et al. (2003), Sources and fluxes of organic carbon in Lake Malawi/Nyasa, J. Great Lakes Res., 29, Suppl. 2, 107120.
  • Rea, D. K., et al. (1980), Southern Lake Michigan sediments: Changes in accumulation rate, mineralogy, and organic content, J. Great Lakes Res., 6, 321330.
  • Rosendahl, B. R. (Ed.) (1988), Seismic Atlas of Lake Tanganyika, East Africa, Dep. of Geol., Duke Univ., Durham, N. C.
  • Rowan, D. J., et al. (1992), Estimating the mud deposition boundary depth in lakes from wave theory, Can. J. Earth Sci., 49, 24902497.
  • Roy, R. N., et al. (1993), The dissociation constants of carbonic acid in seawater at salinities 5 to 45 and temperatures 0 to 45°C, Mar. Chem., 44, 249267.
  • Roy, R. N., et al. (1994), Erratum, Mar. Chem., 45, 337.
  • Roy, R. N., et al. (1996), Erratum, Mar. Chem., 52, 183.
  • Russ, M. E., N. E. Ostrom, H. Gandhi, P. H. Ostrom, and N. R. Urban (2004), Temporal and spatial variations in R:P ratios in Lake Superior, an oligotrophic freshwater environment, J. Geophys. Res., 109, C10S12, doi:10.1029/2003JC001890.
  • Sarch, M. T., and C. M. Birkett (2000), Fishing and farming at Lake Chad: Responses to lake level fluctuations, Geogr. J., 166, 156172.
  • Sarvala, J., et al. (1999), Trophic structure of Lake Tanganyika: Carbon flows in the pelagic food web, Hydrobiologia, 407, 149173.
  • Schindler, D. W. (1978), Factors regulating phytoplankton production and standing crop in the world's freshwaters, Limnol. Oceanogr., 23, 478486.
  • Scholz, C. A. (Ed.) (1989), Seismic Atlas of Lake Malawi (Nyasa), East Africa, Dep. of Geol., Duke Univ., Durham, North Carolina.
  • Scholz, C. A., et al. (1998), Comparative sequence stratigraphy of low-latitude versus high-latitude lacustrine rift basins: Seismic data examples from the East African and Baikal rifts, Palaeogeogr. Palaeoclimatol. Palaeoecol., 140, 401420.
  • Silliman, J. E., et al. (1996), Record of postglacial organic matter delivery and burial in sediments of Lake Ontario, Org. Geochem., 24, 463472.
  • 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.
  • Stallard, R. F. (1998), Terrestrial sedimentation and the carbon cycle: Coupling weathering and erosion to carbon burial, Global Biogeochem. Cycles, 12, 231257.
  • Straškrábová, V., et al. (2005), Primary production and microbial activity in the euphotic zone of Lake Baikal (Southern Basin) during late winter, Global Planet. Change, 46, 5773.
  • Urban, N. R., M. T. Auer, S. A. Green, X. Lu, D. S. Apul, K. D. Powell, and L. Bub (2005), Carbon cycling in Lake Superior, J. Geophys. Res., 110, C06S90, doi:10.1029/2003JC002230.
  • Van Alstine, J. (2006), A high resolution study of the spatial and temporal variability of natural and anthropogenic compounds in offshore Lake Superior sediments, thesis, 125 pp., Univ. of Minn. at Duluth, Duluth.
  • Verburg, P., et al. (2003), Ecological consequences of a century of warming in Lake Tanganyika, Science, 301, 505507.
  • Wanninkhof, R. (1992), Relationship between wind speed and gas exchange over the ocean, J. Geophys. Res., 97, 73737382.
  • Weiler, R. R. (1974), Exchange of carbon dioxide between the atmosphere and Lake Ontario, J. Fish. Res. Board Can., 31, 329332.
  • Weiler, R. R. (1975), Carbon dioxide exchange and productivity in Lake Erie and Lake Ontario, Verh. Int. Verein. Limnol., 19, 694704.
  • Weiss, R. F. (1974), Carbon dioxide in water and seawater: The solubility of a non-ideal gas, Mar. Chem., 2, 203215.

Supporting Information

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Assembly of the Data Set
  5. 3. Environmental Controls on Carbon Fluxes in Large Lakes
  6. 4. Human Impacts: Eutrophication and Carbon Burial
  7. 5. Patterns at the Global Scale
  8. 6. Summary and Conclusions
  9. Acknowledgments
  10. References
  11. Supporting Information

Auxiliary material for this article contains the data and references used in the synthesis of production, organic carbon burial, and carbon dioxide evasion for large lakes of the world.

Auxiliary material files may require downloading to a local drive depending on platform, browser, configuration, and size. To open auxiliary materials in a browser, click on the label. To download, Right-click and select “Save Target As…” (PC) or CTRL-click and select “Download Link to Disk” (Mac).

See Plugins for a list of applications and supported file formats.

Additional file information is provided in the readme.txt.

FilenameFormatSizeDescription
gbc1393-sup-0001-readme.txtplain text document7Kreadme.txt
gbc1393-sup-0002-ts01.txtplain text document3KTable S1. Geographic, morphometric, and environmental data on lakes included in this study from a variety of sources.
gbc1393-sup-0003-ts01.xlsapplication/excel60KTable S1. Geographic, morphometric, and environmental data on lakes included in this study from a variety of sources.
gbc1393-sup-0004-ts02.txtplain text document4KTable S2. Primary production data used in this study, including information on sources, methods, the number of measurements available, and the year(s) in which they occurred.
gbc1393-sup-0005-ts03.txtplain text document2KTable S3. Estimated annual organic carbon mass accumulation rates (MAROC) and burial efficiency in lake basins.
gbc1393-sup-0006-ts04.txtplain text document3KTable S4. Estimated pCO2 values and rudimentary flux estimates.
gbc1393-sup-0007-txts01.txtplain text document23KText S1. A complete list of all references from which data contained in Tables S1–S4 were gathered.
gbc1393-sup-0008-t01.txtplain text document1KTab-delimited Table 1.
gbc1393-sup-0009-t02.txtplain text document1KTab-delimited Table 2.
gbc1393-sup-0010-t03.txtplain text document1KTab-delimited Table 3.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.