Introduction
Biogeochemical processes in the lake littoral zones, located at the transitional boundary between terrestrial and aquatic ecosystems, are closely linked to their adjacent terrestrial ecosystems through carbon and nutrient delivery. Lakes receive allochthonous carbon and nutrients from their catchment areas via streams, groundwater and surface water inflow, which can greatly increase CO2 and CH4 production in the littoral zones (Huttunen et al. Reference Huttunen, Alm, Liikanen, Juutinen, Larmola, Hammar, Silvola and Martikainen2003a, Hirota et al. Reference Hirota, Senga, Seike, Nohara and Kunii2007). Significant N2O emissions have been measured from N-enriched rivers, estuarine and coastal water, as well as freshwater lakes, reservoirs and wetlands receiving a high nitrogen load (Seitzinger et al. Reference Seitzinger, Kroeze and Styles2000, Huttunen et al. Reference Huttunen, Juutinen, Alm, Larmola, Hammar, Silvola and Martikainen2003b). High emissions of CH4 and N2O have been found from the littoral zones of nutrient-enriched temperate lakes, indicating that littoral zones are potential “hotspots” of N2O and CH4 emissions in the landscape (Huttunen et al. Reference Huttunen, Juutinen, Alm, Larmola, Hammar, Silvola and Martikainen2003b, Bastviken et al. Reference Bastviken, Cole, Pace and Tranvik2004, Wang et al. Reference Wang, Wang, Yin, Wang and Lu2006). However, very few measurements of greenhouse gas emissions have been made in lake littoral zones, especially in the remote Antarctica (Priscu Reference Priscu1997, Seitzinger et al. Reference Seitzinger, Kroeze and Styles2000, Huttunen et al. Reference Huttunen, Juutinen, Alm, Larmola, Hammar, Silvola and Martikainen2003b).
Numerous lakes exist on the Antarctic continent, all of which vary considerably in their physical and biogeochemical properties (Priscu Reference Priscu1997). To our knowledge, most published reports on the biogeochemical cycles of greenhouse gases (GHGs) in Antarctic lakes have been for the permanently ice-covered lakes of the McMurdo Dry Valleys (Vincent et al. Reference Vincent, Downes and Vincent1981, Priscu et al. Reference Priscu, Downes and McKay1996, Priscu Reference Priscu1997, Neumann et al. Reference Neumann, Lyons, Priscu and Donahoe2001). The sources or sinks of GHGs have received little attention from the open lakes of coastal Antarctica when the ice above the lakes completely melts in the summer (Zhu et al. Reference Zhu, Liu, Xu, Huang, Sun, Ma and Sun2010). The algae can affect the GHG emissions from aquatic ecosystems (Weathers Reference Weathers1984). Photosynthesis of bloomy algae can stimulate coupled denitrification by supplying O2 to nitrifiers, and high N2O emissions occurred during green algal bloom (An & Joye Reference An and Joye2001, Wang et al. Reference Wang, Wang, Yin, Wang and Lu2006). In the summer, the well developed algal community usually forms the bulk of biomass in numerous lakes of coastal Antarctica, thus the algal-rich lakes may represent an important N2O and CH4 source and CO2 sink. Furthermore, the period of the ‘midnight sun’ occurs in the summer (about three months). However, effects of algae and sunlight on GHG fluxes have not been studied from the summertime ice-free lakes. In coastal Antarctica, most algal-rich lakes are small and shallow with a high proportion of littoral zones. The neglect of these zones contributes to large uncertainties in GHG budget estimate. Therefore accurate field data on GHG exchange in algal-rich lakes, especially for their littoral zones, are needed to improve understanding of GHG budgets.
During the summers of 2008/09 and 2009/10, N2O, CH4 and CO2 fluxes were investigated from the littoral zone of an algal-rich lake, Lake Daming, on Millor peninsula (unofficial name), East Antarctica. The objectives of this study were: i) to study summertime variations in their fluxes from the littoral zone of Lake Daming, ii) to study the effects of light conditions and other environmental parameters on their fluxes, and iii) to evaluate their budgets and net global warming potential (GWP) from the lake littoral zone.
Materials and methods
Study area
The study area is located on Millor peninsula (69°23′–69°56′S, 76°20′–76°45′E), Larsemann Hills, East Antarctica, with c. 50 km2 of ice-free area. It has a continental Antarctic climate. According to the record from Chinese Zhongshan Station, the annual mean air temperature is about -10°C, and precipitation occurs as snow with a maximum annual water equivalent of 250 mm. There are more than 150 freshwater lakes in the Larsemann Hills ranging from small ephemeral ponds to large water bodies such as Lake Daming where the observation sites were set up. Lake Daming is on the coast of Millor peninsula with an area of c. 0.2 km2, a maximum water depth of 3–4 m and sediments up to 3 m thick. In the water column during the summer, the temperature, salinity, pH and dissolved oxygen gradients are small because the open lake is shallow and well mixed by the wind (Li et al. Reference Li, Wang, Lei, Liang, Chen and Liang1997). Some seabirds, such as skuas and snow petrels inhabit the catchments of the lake. Although this lake is considered oligotrophic, the concentrations of nutrient elements are higher than other lakes in the Larsemann Hills possibly due to seabird activity (Wang & Deprez Reference Wang and Deprez2000). The local benthic algal community dominated by Lynobya, Nostoc, Cosmarium and other Cyanobacteria forms the bulk of biomass in Lake Daming, and the plankton is minor compared to the benthos in terms of lake production (Li et al. Reference Li, Wang, Lei, Liang, Chen and Liang1997). The measurement sites were selected to investigate the variations of GHG fluxes from the littoral zone of Lake Daming. The parallel sites L1 and D1, L2 and D2 were set up from 27 December 2008–15 February 2009. The parallel sites L3 and D3, L4 and D4 were also established from 19–27 February 2010 (Fig. 1).
Millor peninsula, the Larseman Hills of East Antarctica, showing N2O, CH4 and CO2 flux measurement sites at the littoral zone of Lake Daming with the parallel sites L1 and D1, L2 and D2 in summer 2008/09, and the parallel sites L3 and D3, L4 and D4 in summer 2009/10.
Greenhouse gas flux measurements
The N2O, CH4 and CO2 fluxes from algal-rich lake littoral sites were determined during two sampling campaigns (27 December 2008–15 February 2009 and 19–27 February 2010) using a static chamber technique (Huttunen et al. Reference Huttunen, Juutinen, Alm, Larmola, Hammar, Silvola and Martikainen2003b, Zhu et al. Reference Zhu, Liu, Ma, Xu and Sun2008b, Reference Zhu, Liu, Xu, Huang, Sun, Ma and Sun2010). In our study area, the sun remains above the horizon during the summer (about three months, the period of the ‘midnight sun’), whereas it is predominantly dark the rest of the year. To study the effects of sunlight on GHG fluxes, the transparent chambers were used to investigate the fluxes at L1, L2, L3 and L4 under the light conditions, and opaque chambers were used to measure the fluxes at D1, D2, D3 and D4 under the dark conditions. The water depth varied from 0–50 cm in the lake littoral zone during the flux measurement. Open-bottomed acrylic resin cylinder chambers (inner diameter 40 cm × height 45 cm) were placed on the floating collars at the measurement sites for the entire observation period, keeping the lower parts of the collars about 5–10 cm below the water surface. The chambers were directly inserted into the sediments to 5 cm when the water was almost dry. The chambers enclosed an area of about 0.12 m2. Care was taken to avoid the perturbations to water surface, the phytobenthos and sediments during the sampling. The gas samples were taken between 9h00 and 12h00 (local time), and they were done under conditions of relatively light wind. The use of flux collars allows the same spot to be measured repeatedly, and minimizes the site disturbance.
During the flux measurements, headspace gas samples were transferred from the chamber into vacuum vials (17.5 ml) at 0, 15 and 30 min (regressions done on three data points). The fluxes were measured eight times from sites L1, L2, D1 and D2, and nine times from sites L3, L4, D3 and D4 during the observation period. Net ecosystem exchange (NEE) was defined as ecosystem exchange of CO2 under light conditions. Ecosystem respiration rates (ER) were obtained from D1 and D2, D3 and D4 by the same procedure using the dark chambers. Gross photosynthesis (Pg) was calculated as the difference between NEE and dark respiration (Ström & Christensen Reference Ström and Christensen2007). Two plots at each site were used for the flux measurements.
The methods of analysing N2O, CH4 and CO2 concentrations and flux calculation have been described in our previous papers (Zhu et al. Reference Zhu, Liu, Ma, Xu and Sun2008b, Reference Zhu, Liu, Xu, Huang, Sun, Ma and Sun2010). In brief, N2O, CH4 and CO2 concentrations were analysed using gas chromatography equipped with a 63Ni electron capture detector (ECD), a flame ionization detector (FID) and a thermal conductivity detector (TCD) respectively (GC-HP5890 II, USA; Shimadzu GC-12A, Japan; Shimadzu GC-14B, Japan). The net GHG fluxes were calculated using a linear least squares fit to three points in the time series of concentrations with an average chamber temperature for each flux. The least squares regression lines were first visually inspected for abrupt changes in the direction of the flux, resulting from disturbances such as the leakage of the chamber during sampling. The fluxes were usually omitted if the slope of the linear fitting had r 2 < 0.90. For GHG fluxes, positive values indicate net emission to the atmosphere and negative values indicate net uptake from the atmosphere.
Environmental variables
The meteorological data were from the weather station of Zhongshan Station on Millor peninsula. Chamber temperature was measured with a thermometer inserted into the chambers. The 5 cm, 10 cm and 15 cm sediment temperatures were determined with ground thermometers inserted into the corresponding depths. Water temperature was determined using a thermometer, and water table was monitored using a ruler adjacent to each chamber. The lake sediment samples (0–15 cm) were collected from the locations adjacent to the GHG flux sites, and stored at 4°C until analysed. The pH was determined in distilled water and 1 M KCl solution (sediment: solution ratio = 1:3). Total organic carbon (TOC), total nitrogen (TN) and total sulfur (TS) contents were determined using vario MACRO CHNS analyser (Zhu et al. Reference Zhu, Liu, Xu, Huang, Sun, Ma and Sun2010). NH4 +-N and NO3 --N were determined using the colorimetric indophenol blue method and Griess-Ilosvay colorimetric method (Keeney & Nelson Reference Keeney and Nelson1982).
Global warming potential calculation
Calculation of the GWP allows an accurate comparison to be made of the radiative forcing of different GHGs relative to the reference gas (CO2). This study refers to the combined GWP of N2O and CH4 emissions from the investigated sites, and NEE was taken into account for the calculation of net GWP. Based on a 100-year time frame, the GWP of CH4 and N2O are, respectively, 25 and 298 times higher than that of CO2 (IPCC 2007). The GWP were calculated by taking mass factors, 1 for CO2, 25 for CH4, and 298 for N2O.
Statistical analyses
All statistical analyses were done using SPSS 12.0.1 and Microsoft Excel for Window 2003. The mean GHG fluxes and standard errors were calculated for each flux measurement at the observation sites. The correlations between GHG fluxes and environmental parameters were tested using Pearson correlation. The factor tested and the relationship were considered statistically significant where P < 0.05. Differences in environmental parameters and mean GHG fluxes between the sites were tested with Student's t-test at P = 0.05.
Results
Climatic characteristics and lake sediment properties
The daily air temperatures ranged from -4.5 to 5.7°C and from -8 to -2.8°C in the summers of 2008/09 and 2009/10 respectively and slightly decreased with date. Corresponding to air temperature variations, the 0–15 cm sediment temperatures ranged from -1.1 to 8.8°C (summer 2008/09) and from -3.1 to -1.8°C (summer 2009/10). The daily sunlight time (i.e. the length of daily time in which the direct solar irradiance is ≥120 W m-2) varied between 0.4 and 18.3 h and between 0 and 7.2 h respectively. Water depths were almost stable with date except 20 February corresponding to a high temperature in summer 2009/10 (Fig. 2). The lake sediment was acidic with the mean pH of about 4.5. The mean TOC, TN, TS contents and C/N ratio were 4.10%, 0.07%, 0.12% and 61.5 respectively. The concentrations of NH4 +-N and NO3 --N were considerably high with the ranges of 26.48–61.35 μg g-1 and 15.48–18.45 μg g-1 respectively (Table I).
Meteorological characteristics in the study area during the summers of 2008/09 and 2009/10. The data for air temperature and sunlight time were obtained from Chinese Zhongshan Station. The water temperatures during the summer of 2009/10 were not measured in the field.
Chemical properties of sediment of Lake Daming, Larsemann Hills, Antarctica.
Note: TOC, TN, TS and C/N indicate total organic carbon, total nitrogen, total sulfur and the ratio of carbon and nitrogen respectively.
N2O fluxes
In summer 2008/09, the mean fluxes varied from 0.14–1.61 μmol N2O m-2 h-1 at L1 and from -0.24–1.44 μmol N2O m-2 h-1 at L2 under light conditions (Fig. 3). Almost all of the fluxes exceeded 0.3 μmol N2O m-2 h-1, and on occasions they amounted to 1.5 μmol N2O m-2 h-1. However, under dark conditions D1 and D2 generally released less N2O compared to L1 and L2 (Table II). Almost half of fluxes at D1 and D2 were below or close to zero, suggesting that significant N2O uptake occurred there. In summer 2009/10, extremely high N2O emissions occurred at L3 and L4 with the ranges of 0.53–15.43 μmol N2Om-2h-1 and 0.16–8.76 μmol N2Om-2h-1 respectively, under light conditions (Fig. 3), whereas D3 and D4 showed much lower N2O fluxes under dark conditions. The summertime period per year when the sun remains above the horizon (the period of the ‘midnight sun’) is about 90 days long. Summertime N2O budgets were estimated to be 0.56 and 10.21 mmol N2O m-2 from the littoral zone of Lake Daming in the summers of 2008/09 and 2009/10 respectively, indicating that the lake littoral zone was a significant source of atmospheric N2O.
Temporal variation of N2O fluxes from the observation sites under light and dark conditions during the summers of 2008/09 and 2009/10. Note: L1 and D1, L2 and D2 were the parallel sites respectively in summer 2008/09. L3 and D3, L4 and D4 were parallel sites respectively in summer 2009/10.
Summary of N2O fluxes, CH4 fluxes, net CO2 fluxes, respiration rates, photosynthetic rates and global warming potential (GWP) for the 100 year time horizon in the littoral zones of Lake Daming, East Antarctica during the summers of 2008/09 and 2009/10.
Note: The mean fluxes with the different superscript letter (a, b or c) indicate significant difference at P < 0.05 between the observation sites. The data from all the sites under the light and dark conditions were analysed to calculate the GWP. Based on a 100-year time frame, the GWP of CH4 and N2O are, respectively, 25 and 298 times higher than that of CO2 (IPCC 2007). The GWP were calculated by taking mass factors, 1 for CO2, 25 for CH4, and 298 for N2O. NEE = net ecosystem exchange.
Overall, there were significant differences (P < 0.01) in N2O fluxes under light and dark conditions, and the mean flux under the light (4.73 ± 0.55 μmol N2O m-2 h-1) was one order of magnitude higher than that under the dark (0.91 ± 0.23 μmol N2O m-2 h-1), indicating sunlight could stimulate N2O emissions from the algal-rich littoral zone (Fig. 4a). Additionally, the variation of N2O fluxes corresponded to sediment temperature (r = 0.47, P < 0.01). The N2O fluxes showed no significant positive correlation with NEE and algal photosynthesis rate (Pg), but they significantly positively correlated with ER (P < 0.01) (Table III).
Comparison between mean a. N2O and b. CH4 fluxes from the observation sites under light and dark conditions. The box graph showed the differences of N2O and CH4 fluxes under light and dark conditions. The flux data under light conditions were obtained from the sites L1, L2, L3 and L4, and the data under dark conditions from the sites D1, D2, D3 and D4.
Correlation between summertime N2O, CH4 and CO2 fluxes and environmental variables at the observation sites in the littoral zone of Lake Daming.
Note: Only statistically significant correlation coefficients were shown in this table. * and ** indicate correlation significant at the 0.05 and 0.01 level respectively. NEE = net ecosystem exchange, ER = ecosystem respiration rate, Pg = photosynthesis rate, CT = chamber temperature, WD = water depth, AT = air temperature, SL = sunlight time, ST = sediment temperature. A digital corresponding to a relationship means that all the data during two summers were analysed together, while two digitals mean that the data during two summers were analysed respectively because of poor correlations.
CH4 fluxes
The CH4 emissions under light conditions showed a consistent variation at L1 and L2 in summer 2008/09, and at L3 and L4 in summer 2009/10. Significantly enhanced CH4 emissions occurred at D1 and D2 under dark conditions (Fig. 5). Overall, there was a significant difference (P < 0.05) in CH4 fluxes under light and dark conditions, and the mean flux under the light (2.78 ± 0.85 μmol CH4 m-2 h-1) was significantly lower than that under the dark (17.80 ± 3.32 μmol CH4 m-2 h-1) (Fig. 4b). The accumulated CH4 emissions from the littoral zone were 13.01 and 2.11 mmol CH4 m-2 respectively during the period of the ‘midnight sun’ in the summers of 2008/09 and 2009/10.
Temporal variation of CH4 fluxes from the observation sites under light and dark conditions during the summers of 2008/09 and 2009/10. Note: L1 and D1, L2 and D2 were the parallel sites respectively in summer 2008/09. L3 and D3, L4 and D4 are parallel sites respectively in summer 2009/10.
A significant positive correlation (r = 0.32, P < 0.01) was found between CH4 fluxes and Pg (Table III), and the fluxes also showed a significant positive correlation with water temperature (r = 0.37, P < 0.01), air temperature (r = 0.30, P = 0.04) and sediment temperature (r = 0.35, P < 0.01), indicating that light conditions and the lake thermal regime had important effects on CH4 emissions from the lake littoral zone. In addition, we observed a positive but not statistically significant correlation between CH4 flux and water depth at the observation sites, suggesting that water depth might be an important, but not key factor affecting CH4 emissions.
CO2 fluxes
The mean CO2 fluxes (NEE) varied between a weak sink and a weak source (-0.37–0.13 mmol CO2 m-2 h-1), and there were no significant differences between sites L1 and L2 (Table II). Overall the fluxes were negative over the observation period, indicating that lake littoral zone was a net CO2 sink. Respiration rate varied from 0.47–2.90 mmol CO2 m-2 h-1 in summer 2008/09 and from 0.66–1.71 mmol CO2 m-2 h-1 in summer 2009/10, whereas Pg varied from -0.15 to -1.30 mmol CO2 m-2 h-1, and from -0.33 to -2.63 mmol CO2 m-2 h-1 respectively (Fig. 6). The CO2 budgets in the lake littoral zone were estimated to be -2.50 and -386.36 mmol CO2 m-2 respectively over the summers of 2008/09 and 2009/10.
Temporal variation of net CO2 flux (net ecosystem exchange (NEE)), respiration rate (ER) and photosynthesis rate (Pg) from the observation sites during the summers of 2008/09 and 2009/10. Note: NEE was obtained from the transparent chambers at the sites L1 and L2, L3 and L4, and ER was obtained from the opaque chambers at the sites D1 and D2, D3 and D4.
Sunlight time showed a significant negative correlation with NEE (r = -0.45, P < 0.05) and Pg (r = -0.66, P < 0.01) (Table III). Therefore light conditions and the photosynthesis of algae were predominant factors affecting NEE. Ecosystem respiration rates corresponded to the variations in air or sediment temperature, and they almost simultaneously reached a maximum (Figs 2 & 6), suggesting that ER might be limited by the temperature in the lake littoral zone.
Global warming potential
The littoral zone sites of Lake Daming in summer 2009/10 (54.46 mgCO2eq m-2 h-1) had ten times higher radiative forcing than those in summer 2008/09 (5.73 mgCO2eq m-2 h-1) due to much higher N2O emissions (Table II). The combined GWP of N2O and CH4 fluxes completely counteracted and surpassed CO2 uptake by algal photosynthesis, with 2.41 mgCO2eq m-2 h-1 for CH4 fluxes in 2008/09 and 0.14 mgCO2eq m-2 h-1 in 2009/10, and with 3.37 mgCO2eq m-2 h-1 for N2O fluxes in 2008/09 and 61.81 mgCO2eq m-2 h-1 in 2009/10. Therefore a high GWP-positive of N2O and CH4 emissions might convert an algal-rich lake site with net CO2 uptake into a net radiative forcing source in Antarctica.
Discussion
N2O fluxes from lake littoral zone
Aerobic nitrification and anaerobic denitrification are the major processes producing N2O in the soils/sediments, whereas N2O can be consumed by denitrification under highly anoxic conditions (Davidson & Schimel Reference Davidson and Schimel1995, Chapuis-Lardy et al. Reference Chapuis-Lardy, Wrage, Metay, Chotte and Bernoux2007). In this study, enhanced N2O emissions occurred at the observation sites under light conditions, indicating that sunlight might stimulate N2O production from algal-rich lake littoral zones in East Antarctica. Recently, Stewart et al. (Reference Stewart, Brummell, Farrell and Siciliano2012) found that increasing efflux of N2O from some plant communities in high Arctic polar deserts occurred with increasing soil moisture under light conditions, while increasing consumption of N2O occurred under dark conditions, which is similar to our results. Some studies suggest N2O may be produced by nitrification in plants (Smart & Bloom Reference Smart and Bloom2001), and/or it may be taken up from the soil by roots and then released to the atmosphere via transpiration (Pihlatie et al. Reference Pihlatie, Ambus, Rinne, Pilegaard and Vesala2005). Light-dependent plant internal N2O production during N-assimilation has been hypothesized to occur within the aboveground biomass of some plants (Stewart et al. Reference Stewart, Brummell, Farrell and Siciliano2012). Many studies showed that the presence of abundant algae might have significant effects on lake N2O production because photosynthesis of algae can stimulate coupled denitrification by supplying O2 to nitrifiers, thereby stimulating nitrification and indirectly providing NO3 - to denitrifiers, and thus higher N2O fluxes occurred during serious algae blooms (Weathers Reference Weathers1984, An & Joye Reference An and Joye2001, Wang et al. Reference Wang, Wang, Yin, Wang and Lu2006). Jørgensen et al. (Reference Jørgensen, Struwe and Elberling2011) observed a significant positive correlation between O2 concentration in the rooting zone and incoming light intensity. In our study area, the period of the ‘midnight sun’ occurs in the short summer, when light conditions and high temperature favour the algal photosynthesis. Increased N2O production might occur in the algae and the sediments of Lake Daming through enhanced nitrification and coupled denitrification due to the O2 release during the algal photosynthesis. Therefore, higher N2O emissions occurred in lake littoral zone under light conditions than under dark conditions.
While the difference between light and dark N2O flux may be alga dependent, our results also suggest that additional factors, such as microbial activity, water depth, and sediment properties, may also affect N2O emissions from lake littoral zone. The N2O fluxes showed no significant positive correlation with NEE and Pg, but they significantly positively correlated with ER (P < 0.01) (Table III). Stewart et al. (Reference Stewart, Brummell, Farrell and Siciliano2012) proposed that the influence of light on N2O flux might result from short-term effects of resource competition between vegetation and soil microbes in response to light-driven changes in O2 availability. Ecosystem respiration rate is an indirect comprehensive indicator for the activity strength of the algae and microbes in the sediments while Pg and NEE are predominantly affected by algal activity (Davidson & Schimel Reference Davidson and Schimel1995). The N2O production in the lake littoral zone might be related with the activity of both the algae and some microbes, such as nitrifiers and denitrifiers in the sediments (An & Joye Reference An and Joye2001, Chapuis-Lardy et al. Reference Chapuis-Lardy, Wrage, Metay, Chotte and Bernoux2007). Therefore, N2O fluxes showed a significant positive correlation with ER instead of Pg and NEE. Water depth might impact N2O fluxes from the littoral zones (Zhu et al. Reference Zhu, Liu, Ma, Xu and Sun2008b, Fromin et al. Reference Fromin, Pinay and Montuelle2010, Liu et al. Reference Liu, Zhu, Ma, Xu, Luo, Huang and Sun2011). In our study area N2O fluxes were negatively correlated with water depth although this relation was not statistically significant. The sites with low water depth exposed more sediment to air, and sediment drying accelerated the release of bio-available nutrients from organic matter (Davidson & Schimel Reference Davidson and Schimel1995, Fromin et al. Reference Fromin, Pinay and Montuelle2010) Furthermore, these sites showed high NH4 + and NO3 - concentrations (Table I), which could increase the proportion of N2O production in either nitrification or denitrification.
CH4 fluxes from lake littoral zone
Sunlight could greatly decrease CH4 emissions from the lake littoral zone due to O2 release by the algal photosynthesis, and either CH4 was effectively oxidized or its production was inhibited by O2 availability (Frenzel & Karofeld Reference Frenzel and Karofeld2000). On the other hand, sunlight can impact algal production, and high algal production in the summer could produce new autochthonous organic matter, which in turn could supply the extra substrates for CH4 production especially in shallow, highly productive waters (Huttunen et al. Reference Huttunen, Alm, Liikanen, Juutinen, Larmola, Hammar, Silvola and Martikainen2003a). A positive correlation was found between CH4 fluxes and Pg (Table III), but the estimated NEE in this study area (Fig. 6) was much lower than those in some arctic, temperate and tropical lakes, and boreal wetlands with aquatic plants, such as fringing zones of Lake Nakaumi in Japan (-7.27 to -0.52 mmol CO2 m-2 h-1, Hirota et al. Reference Hirota, Senga, Seike, Nohara and Kunii2007), Mackenzie Delta lakes in the Arctic landscape (-1.50 to -0.28 mmol CO2 m-2 h-1, Tank et al. Reference Tank, Lesack and Hesslein2009), and tundra wetlands in Alaska (-8.33 to -1.67 mmol CO2 m-2 h-1, Harazono et al. Reference Harazono, Mano, Miyata, Zulueta and Oechel2003), suggesting that CH4 emission from the lake littoral zone was insignificant through the supply of extra substrates due to low NEE. Therefore algae might predominantly produce oxygen via the photosynthesis to decrease CH4 production or increase its oxidation in the littoral zone of Lake Daming.
The CH4 fluxes showed a significant positive correlation with water temperature (r = 0.37, P < 0.01), air temperature (r = 0.30, P = 0.04) and sediment temperature (r = 0.35, P < 0.01), indicating that CH4 emissions were related to the lake thermal regime (Table III). The increase in the temperature might stimulate the activity of methanogenic bacteria, and thus contribute to higher CH4 emissions (Davidson & Schimel Reference Davidson and Schimel1995). Water depth is also generally considered to be an important factor influencing CH4 production and oxidation (Frenzel & Karofeld Reference Frenzel and Karofeld2000). In this study, we observed a positive but not statistically significant correlation between CH4 flux and water depth at the observation sites, suggesting that water depth might be an important, but not key factor affecting CH4 production and emission from lake littoral zones.
CO2 fluxes from lake littoral zone
The CO2 flux dynamics we observed are probably linked to the processes of photosynthesis and sediment and algal respiration. Photosynthesis tends to dominate under light conditions, and algal respiration, microbial respiration and microbial decomposition of organic matter under dark conditions (Kuzyakov & Gavrichkova Reference Kuzyakov and Gavrichkova2010, Jørgensen et al. Reference Jørgensen, Struwe and Elberling2011). In this study, ER showed a relatively small fluctuation, whereas the variability in NEE highly corresponded to Pg, indicating that photosynthesis rate was the primary explanatory variable for NEE at the observation sites, and that algal activity might play a more important role in net CO2 flux rather than the bacterial respiration and the mineralization of organic matter. A similar pattern of CO2 flux and Pg has been found in other global lakes, including Japanese fringing zones of Lake Nakaumi (Hirota et al. Reference Hirota, Senga, Seike, Nohara and Kunii2007), Mackenzie Delta lakes in the Arctic (Tank et al. Reference Tank, Lesack and Hesslein2009), and some other Antarctic lakes such as Lake Tuanjie and Lake Mochou (Zhu et al. Reference Zhu, Liu, Xu, Huang, Sun, Ma and Sun2010). In addition, photosynthetically active radiation (PAR) was the most important abiotic factor for NEE or Pg, and they generally showed a negative correlation (Harazono et al. Reference Harazono, Mano, Miyata, Zulueta and Oechel2003, Kuzyakov & Gavrichkova Reference Kuzyakov and Gavrichkova2010). Sunlight time (SL) was as an indirect proxy for light intensity and PAR, thus SL showed a significant negative correlation with NEE (r = -0.45, P < 0.05) and Pg (r = -0.66, P < 0.01) (Table III), indicating that light conditions and the algal photosynthesis were predominant factors affecting NEE (Zhu et al. Reference Zhu, Liu, Xu, Huang, Sun, Ma and Sun2010).
Comparisons with other measurements and regional implications
The mean N2O fluxes in the littoral zones of Lake Daming were one to two orders of magnitude higher than those from Lake Mochou (0.11 ± 0.24 μmol N2O m-2 h-1) and Lake Tuanjie (0.09 ± 0.10 μmol N2O m-2 h-1) in East Antarctica (Liu et al. Reference Liu, Zhu, Ma, Xu, Luo, Huang and Sun2011). Our results were also much higher than the fluxes from some wetlands and aquatic ecosystems in the boreal regions (-0.04–0.13 μmol N2O m-2 h-1 in Reservoir Lokka; -0.04–0.01 μmol N2O m-2 h-1 in Lake Kevätön; Huttunen et al. Reference Huttunen, Alm, Liikanen, Juutinen, Larmola, Hammar, Silvola and Martikainen2003a). The N2O emissions from this study were comparable with those from ornithogenic soils in Maritime Antarctica (0.33–8.06 μmol N2O m-2 h-1; Zhu et al. Reference Zhu, Liu, Xu, Ma, Zhao and Sun2008a), lakeshore soils in Antarctic Garwood Valley (0.94–4.71 μmol N2O m-2 h-1; Gregorich et al. Reference Gregorich, Hopkins, Elberling, Sparrow, Novis, Greenfield and Rochette2006), and tropical hydroelectric reservoirs in French Guiana (2.08 ± 0.94 μmol N2O m-2 h-1; Guérin et al. Reference Guérin, Abril, Tremblay and Delmas2008). Therefore, the sparse data in this study hint at the potential importance of lake littoral zones in the total local atmospheric N2O load in East Antarctica.
The mean CH4 fluxes in the littoral zones of Lake Daming were substantially lower than the global mean fluxes of lakes (0.11–0.18 mmol CH4 m-2 h-1), plant mediated mean fluxes of lakes (0.52 mmol CH4 m-2 h-1), and the fluxes from some temperate wetlands (-18.25–1.13 mmol CH4 m-2 h-1) (Bastviken et al. Reference Bastviken, Cole, Pace and Tranvik2004, Walter et al. Reference Walter, Vas, Brosius, Chapin, Zimov and Zhuang2010). However, our results were comparable to those from boreal wetlands (-8.33–14.33 μmol CH4 m-2 h-1) (Bartlett et al. Reference Bartlett, Crill, Sass, Harriss and Dise1992). Furthermore, these fluxes were similar in magnitude to those from Lake Mochou (9.11 μmol CH4 m-2 h-1), Lake Tuanjie (6.77 μmol CH4 m-2 h-1) (Zhu et al. Reference Zhu, Liu, Xu, Huang, Sun, Ma and Sun2010), and the wetlands (-12.2–36.7 μmol CH4 m-2 h-1) in East Antarctica (Zhu et al. Reference Zhu, Liu, Ma, Xu and Sun2008b), but they were lower than those from lakeshore soils (30–900 μmol CH4 m-2 h-1) in Antarctic Garwood Valley (Gregorich et al. Reference Gregorich, Hopkins, Elberling, Sparrow, Novis, Greenfield and Rochette2006). The results of this study indicate that the shallow lake littoral zones might be a weak CH4 emitter in Antarctica.
The mean CO2 fluxes from the littoral zone of Lake Daming agreed with those from terrestrial populations of Nostoc commune (Cyanobacteria) in an Antarctic dry valley (Novis et al. Reference Novis, Whitehead, Gregorich, Hunt, Sparrow, Hopkins, Elberling and Greenfield2007; -0.14 to -0.20 mmol CO2 m-2 h-1) and those from studies of benthic communities at 10 m depth in Lake Hoare (Moorhead et al. Reference Moorhead, Schmeling and Hawes2005; c. -0.15 mmol CO2 m-2 h-1), and were well within the range of values for benthic algae in Antarctic lakes summarized by Moorhead et al. (Reference Moorhead, Davis and Wharton1997) (0 to -13.32 mmol CO2 m-2 h-1). However, the results of this study were one to two orders of magnitude lower than those observed in Lake Mochou (-1.61 mmol CO2 m-2 h-1) and Lake Tuanjie (-0.84 mmol CO2 m-2 h-1) of East Antarctica in summer 2007/08 (Zhu et al. Reference Zhu, Liu, Xu, Huang, Sun, Ma and Sun2010). They were also substantially lower than those from some temperate and Arctic lakes, such as Lake Nakaumi in Japan (-7.27 to -0.52 mmol CO2 m-2 h-1, Hirota et al. Reference Hirota, Senga, Seike, Nohara and Kunii2007) and Mackenzie Delta lakes in the Arctic landscape (-1.50 to -0.28 mmol CO2 m-2 h-1, Tank et al. Reference Tank, Lesack and Hesslein2009). Our results revealed that the littoral zone of Lake Daming was a weak CO2 sink in accordance with some reports in the lakes of the McMurdo Dry Valleys, Antarctica (Neumann et al. Reference Neumann, Lyons, Priscu and Donahoe2001).
Uncertainty analysis of GHG flux
This study was limited by several uncertainties. Firstly, the uncertainties of GHG flux quantification may be caused by the error associated with chamber measurement. This may be due to: i) lack of spatial representation, which might result from a limited number of sampling points, effects of the spatial variability in the algal coverage, community compositions and sediment properties of the lake littoral zone, and inadequate measurements available for the observation period; ii) gradient problems, such as build-up/rapid reduction of GHG concentrations in the chamber; and iii) changes in temperature during translucent chamber enclosure (Zheng et al. Reference Zheng, Xie, Liu, Zhou, Yao, Wang, Wang, Yang, Zhu, Huang and Butterbach-Bah2008). These factors could have affected the flux measurements in this study.
Secondly, our observation provided short summertime GHG fluxes from the ice-free lake littoral zones, and furthermore, the fluxes were measured only between 9h00 and 12h00. It was possible that the sun went down in the afternoon or at night, and heat exchange could modify the turbulence in lake water and thus GHG fluxes. These fluxes are still not known outside of the measured period, and it is difficult to evaluate the importance of summertime GHG fluxes from the littoral zone of Lake Daming in net annual budget.
Thirdly, these data may represent a minimum or conservative estimate of GHG fluxes. This is because it is possible that our sampling regime did not capture all the episodic, high magnitude events, such as the ebullition, which usually happen in a short time period. Generally, the ebullition is an important component of whole-lake CH4 emissions, and if accounted for, could significantly increase estimates of regional lake emissions. Furthermore, the area of the lake with shallow water generally has a high probability of ebullition (Bastviken et al. Reference Bastviken, Cole, Pace and Tranvik2004, Walter et al. Reference Walter, Vas, Brosius, Chapin, Zimov and Zhuang2010). Water depth was generally below 1 m in the littoral zone of Lake Daming, and thus the ebullition might be very important in CH4 flux estimate, which was not likely included in 30-minute chamber trials. Therefore GHG fluxes, especially CH4 emissions, could be underestimated from the littoral zone of Lake Daming.
Conclusions
To our knowledge, few GHG measurements have simultaneously been conducted in the littoral zone of an ice-free Antarctic lake during the summer. In this study, we investigated the temporal variations of GHG fluxes and environmental variables affecting the fluxes from the littoral zone of Lake Daming, East Antarctica, during the summers of 2008/09 and 2009/10. Overall the lake littoral zone was a strong N2O emitter, weak CH4 emitter, and weak CO2 sink. There were significant differences (P < 0.05) in CH4 and N2O fluxes under the light and dark conditions, and the sunlight could decrease CH4 emissions, but greatly stimulate N2O emissions due to the algal photosynthesis. CH4 emissions were significantly affected by the lake thermal regime. The sunlight was the predominant factor affecting NEE and Pg. In the summer, the combined GWP of N2O and CH4 fluxes completely counteracted and surpassed CO2 uptake by the algal photosynthesis in the littoral zone of Lake Daming. High GWP-positive of N2O and CH4 emissions might convert an algal-rich lake site with net CO2 uptake into a net radiative forcing source in coastal Antarctica.
Acknowledgements
This work was supported by the National Natural Science Foundation of China (Grant No. 41176171, 41076124) and Specialized Research Fund for the Doctoral Program of Higher Education (Grant No. 20123402110026). We thank Dr Yuhan Luo and Dr Jin Huang for their help in field sampling. We also thank the Polar Office of National Ocean Bureau of China and the members of the 25th and 26th Chinese Antarctic Research Expedition for their support. We are also grateful to the anonymous reviewers and editor for their helpful revision and comments on a previous version of this paper.
