Lakes are an important component of the climate system. They provide moisture for precipitation, buffer temperature variations, and contribute to regional atmospheric carbon budgets. This article describes an eddy covariance (EC) mesonet on Lake Taihu, a large (area 2400 km2) and shallow (depth 2 m) lake situated in the heavily populated Yangtze River Delta, China. The mesonet consists of five lake sites, representing different biological attributes and wind–wave patterns, and a land site near the lake shore. Common to all the sites are standard EC instruments for measurement of the momentum, sensible heat, water vapor, and CO2 flux. One site is also equipped with laser-based analyzers for precise measurement of the CO2, CH4, and H2O mixing ratios and their isotopic compositions. To the authors' best knowledge, this is the first lake eddy flux mesonet. Early results reveal evidence of biological and pollution controls on the surface–air fluxes of energy, momentum, and greenhouse gases across the lake. The data will be used to address five science questions: 1) Are lake–air parameterizations established for deep lakes applicable to shallow lakes? 2) Why are lake–land breeze circulations less prevalent in the Taihu lake basin than in lake basins in northern latitudes? 3) How do algal blooms alter the lake–atmosphere interactions? 4) Is this eutrophic lake a source or sink of atmospheric CO2? 5) Does the decay of algal and macrophyte biomass contribute significant amounts of CH4 to the atmosphere?

In situ eddy covariance observations reveal unusually large nocturnal CO2 uptake by submerged vegetation in a shallow lake.

MOTIVATION AND SCIENCE QUESTIONS.

Lakes are an important component of the climate system. Even though lakes and reservoirs occupy only about 4% of the global terrestrial surface (Downing et al. 2006), their societal importance is disproportionately large because many large municipalities are located near lake shorelines. The large thermal contrast between a lake and its surrounding land often triggers thermal circulations, having significant impact on air pollution dispersion and transport in the lake catchment. Being important sources of atmospheric moisture, large lakes can enhance storm formation in the downwind area (Samuelsson et al. 2010; Zhao et al. 2012). Except at times of high algal activities (Hari et al. 2008; Balmer and Downing 2011), lake water is usually supersaturated in CO2 with respect to the atmosphere and acts as a source of atmospheric carbon (Cole et al. 1994). Lakes are also sources of atmospheric CH4 (Bastviken et al. 2011) and N2O (Huttunen et al. 2003).

Eddy covariance (EC) is an in situ technique for measuring momentum, heat, water, and greenhouse gas fluxes. It determines the flux continuously and nonintrusively from simultaneous measurements, in the atmospheric surface layer, of turbulent fluctuations in the air velocity and the scalar quantity of interest. The method is a key measurement tool deployed by several large observational networks, such as the Global Flux Network (FLUXNET; Baldocchi et al. 2001), the National Ecological Observatory Network in the United States (Schimel et al. 2007), and the Integrated Carbon Observation System in Europe (www.icos-infrastructure.eu). These networks are playing increasingly important roles in Earth and ecological sciences. Most of the EC sites in these networks are located in upland ecosystems. Owing to logistical difficulties, long-term (>1 yr) EC applications are still rare for lake systems (Rouse et al. 2008; Blanken et al. 2011; Nordbo et al. 2011; Liu et al. 2012). The water equilibrium method and floating chambers, two traditional methods for measuring lake–air gaseous fluxes, are suitable for short field campaigns but are difficult to deploy in uninterrupted long-term operations (St. Louis et al. 2000; Schubert et al. 2012). In addition, they cannot measure energy, momentum, and water fluxes, and yet these fluxes are the primary drivers of the physical state of the atmosphere and the lake system.

Imbedded in the above regional and global networks are a number of small EC clusters. By design, sites in these clusters are positioned in close proximity, usually within tens of kilometers from one another, and are influenced by nearly identical climate conditions. The observed spatial variations reveal gradient effects of land management (Zha et al. 2009; Prescher et al. 2010), ecological succession (Stoy et al. 2008), and natural disturbance (Goulden et al. 2006; Brown et al. 2012). Here we adopt the same research strategy to monitor the temporal and spatial patterns of lake–air fluxes.

In this article, we describe an EC network on Lake Taihu, a large and shallow lake in southeastern China. Globally, 28% of inland lakes are shallow (depth < 5 m) according to the Global Lake Database (www.flake.igb-berlin.de/ep-data.shtml). The network consists of five lake sites, representing different biological attributes, pollution status, and wind–wave patterns, and a land site near the lake shore. Lake Taihu is an environmental hot spot. It is located in the Yangtze River Delta and is the third largest freshwater lake in China. The lake basin is heavily urbanized, with five large municipalities (each with populations greater than one million) situated near the lake shoreline. Frequent algal blooms and air quality problems in recent years have spurred considerable interest among scientists and decision makers in processes governing the lake–atmosphere interactions.

Our study appears to represent the first lake EC network. Our goal is to quantify the lake–air fluxes of energy, momentum, and greenhouse gases across pollution and biological gradients in the lake. The data will be used to address five science questions:

  1. Are lake–air parameterizations established for deep lakes applicable to shallow lakes?

  2. Why are lake–land breeze circulations less prevalent in the Taihu lake basin than in lake basins in northern latitudes?

  3. How do algal blooms alter the lake–atmosphere interactions?

  4. Is this eutrophic lake a source or sink of atmospheric CO2?

  5. Does the decay of algal and macrophyte biomass contribute significant amounts of CH4 to the atmosphere?

The objective of this paper is to provide an overview of this field program. To date, some progress has been made toward answering the first science question, using data from a subset of the eddy flux sites (Deng et al. 2013; Xiao et al. 2013; Wang et al. 2014). A brief summary of the results is given below. In addition, we describe an unusual phenomenon of large nocturnal CO2 uptake by submerged vegetation in the lake.

SITE DESCRIPTION.

The Lake Taihu catchment belongs to four administrative units: Jiangsu Province, Zhejiang Province, Anhui Province, and Shanghai Municipality. To the north and east of the lake are floodplains with an elevation range of 4–6 m and to the west are highlands with elevation of 10–1250 m. The lake itself is situated at 30°5′40″–31°32′58″N, 119°52′32″–120°36′10″E, with a total area of approximately 2400 km2 (Fig. 1). A reference map is given by Xiao et al. (2013). The lake has a mean depth of 1.9 m, with the northern and western portions being deeper (depth 2.5 m) and eastern portion shallower (depth <1.5 m; Qin et al. 2007). The lake bottom has an average elevation of 1.1 m above mean sea level. According to spatial variations of pollution loading (Zhao et al. 2011), vegetation abundance (Liu et al. 2007), and wind–current interaction (Qin et al. 2007), the lake is now divided into seven biophysically distinct zones (Fig. 1; Hu et al. 2011). Zone 1 is a semi-enclosed bay. It has low wind speed and nearly stagnant water, and is eutrophic. Zone 2 is connected to one of the largest outflow waterways, with half of the water covered by floating plants and emergent macrophytes. Receiving large quantities of nutrients from the inflow rivers, Zone 3 is hypereutrophic. Zone 4 in the center of the lake has relatively low photosynthetic activity. Zone 5 is dominated by submerged macrophytes, zone 6 represents a transition between phytoplankton dominance and macrophyte dominance, and zone 7 is characterized by submerged macrophytes and pen fish farming.

Fig. 1.

A Landsat 8 image of Lake Taihu and its surrounding area, taken on 14 Apr 2013. Color composite is 654 red-blue-green (RGB). Red crosses mark locations of the eddy covariance sites. Blue and green lines mark inflow and outflow rivers, respectively. Areas in green are vegetation and areas in purple and brown are cities.

Fig. 1.

A Landsat 8 image of Lake Taihu and its surrounding area, taken on 14 Apr 2013. Color composite is 654 red-blue-green (RGB). Red crosses mark locations of the eddy covariance sites. Blue and green lines mark inflow and outflow rivers, respectively. Areas in green are vegetation and areas in purple and brown are cities.

Table 1 lists a few key climate variables. The climate in the lake catchment is humid-subtropical according to the Köppen climate classification. Water flows into the lake through 27 rivers and canals connected to the west and north shores and outflow water leaves the lake through 22 waterways connected to the east shore. The annual mean flow is about 88 × 108 m3 (Yan et al. 2011). In some of the rivers, flow reversal can occur at times of low water level in the summer. The mean residence time of the lake water is about 350 days (An and Wang 2008).

Table 1.

Climate variables and energy balance components at Lake Taihu. Climate variables are for 1981–2010 measured at the Wuxi surface weather station near the north shore of the lake. Energy flux components are for year 2011 measured at MLW in the lake and are averaged over 24-h period of the day (Fig. 1).

Climate variables and energy balance components at Lake Taihu. Climate variables are for 1981–2010 measured at the Wuxi surface weather station near the north shore of the lake. Energy flux components are for year 2011 measured at MLW in the lake and are averaged over 24-h period of the day (Fig. 1).
Climate variables and energy balance components at Lake Taihu. Climate variables are for 1981–2010 measured at the Wuxi surface weather station near the north shore of the lake. Energy flux components are for year 2011 measured at MLW in the lake and are averaged over 24-h period of the day (Fig. 1).

In recent decades, rapid urbanization and intensification of industry in the lake catchment have caused severe environmental pollution. The lake total nitrogen concentration increased from the pre-1980 level of 0.05 mg L−1 to 1.97 mg L−1 in 2012 (An and Wang 2008; Taihu Office 2013). Frequent algal blooms (see supplementary Fig. ES1; supplementary data are available online at http://dx.doi.org/10.1175/BAMS-D-13-00136.2) have altered the biophysical characteristics of the lake system and have threatened the water supply to several large cities in the lake catchment (Zhang et al. 2010).

Local air quality has also deteriorated. According to the data provided by the Suzhou Meteorological Bureau, the 2012 annual mean PM2.5 concentration in Suzhou City is 43 μg m−3, with the daily mean reaching as high as 150 μg m−3, far exceeding the U.S. Environmental Protection Agency air quality annual limit of 15 μgm−3 and daily limit of 35 μgm−3. Generally, cities at the shoreline of a large lake should have good air quality because of favorable dispersion conditions associated with lake breeze circulations. But this is not the case for Suzhou and other cities near Lake Taihu. Using the detection method described by Sills et al. (2011), we found that the frequency of lake breeze occurrence is about 20% in the summer of 2012, much lower than that reported for the southern Great Lakes region (76% in June–August 2007; Sills et al. 2011). Concurrent energy flux measurements on land and in the lake will provide critical data input to mesoscale models to determine whether the low occurrence frequency is intrinsic to shallow lake systems in subtropical latitudes.

RESEARCH DESIGN.

Eddy covariance.

The field experiment addresses science questions 3–5. Currently, the Lake Taihu Eddy Flux Network consists of five lake sites and one land site (Fig. 1; Table 2). Common to these sites are an EC system consisting of a sonic anemometer/thermometer (model CSAT3, Campbell Scientific Inc., Logan, Utah) and an open-path H2O/CO2 analyzer (Table 2), a four-way net radiometer (model CNR4, Kipp & Zonen B. V., Delft, the Netherlands), an anemometer and wind vane (model 05103; R M Young Company, Traverse City, Michigan), and an air temperature and humidity probe (model HMP155A; Vaisala, Inc., Helsinki, Finland). At the lake sites, water temperature is measured at depths of 0.20, 0.50, 1.00, and 1.50 m, and sediment temperature at the depth of ~0.10 m below the water column (temperature probes model 109-L, Campbell Scientific). At the land site, soil temperature is measured at depths of 0.05, 0.10, and 0.20 m (temperature probes model 109-L, Campbell Scientific) and soil heat flux at the depth of 0.05 and 0.10 m (heat flux plate model HFP01SC; Hukseflux Thermal Sensors B. V., Delft, the Netherlands). At the MLW lake site, the sensors are mounted on small (diameter 30 cm) concrete pillars and connected to A/C power. At the other four lake sites (DPK, BFG, PTS, XLS; see Table 2), the sensors are mounted on an in-lake platform (dimension 3–5 m) and powered by an array of solar panels (Fig. 2). The EC signals are recorded at 10 Hz by a datalogger (model CR 3000, Campbell Scientific), which also performs online flux computation using the block averaging procedure (Lee et al. 2004). All other variables are sampled at 1 Hz by another datalogger (model CR1000, Campbell Scientific).

Fig. 2.

The PTS instrument platform in zone 4 (Fig. 1). Insets are photos of the (left) four-way radiometer and (right) eddy covariance sensors.

Fig. 2.

The PTS instrument platform in zone 4 (Fig. 1). Insets are photos of the (left) four-way radiometer and (right) eddy covariance sensors.

Table 2.

A list of eddy covariance sites.

A list of eddy covariance sites.
A list of eddy covariance sites.

The prevailing wind is southeast in the summer and northeast in the winter. Wind speed varies among the lake sites. Having more open fetch, PTS and DPK are windier than the other sites. The 10-m mean wind speed for the period July–September 2013 is 3.6 m s−1 at MLW, 4.8 m s−1 at DPK, 5.0 at m s−1 BFG, 4.7 m s−1 at XLS, and 5.0 m s−1 at PTS. A footprint analysis for MLW, DPK, and BFG is given by Wang et al. (2014).

Two lake sites (MLW and BFG) have enhanced measurement capacity. MLW is a supersite located in the Meiliangwan Bay. A 250-m-long boardwalk provides easy access to the instruments. A small raised instrument shed at the end of the boardwalk houses two laser-based analyzers for simultaneous measurement of CO2, CH4, and H2O mixing ratios (Model G1301, Picarro Inc., Santa Clara, California) and for measurement of the D and 18O isotopic compositions of water vapor (model DLT-100, Los Gatos Research, Mountain View, California). These analyzers are configured in gradient mode, switching every 30–60 s between air samples drawn from the heights of 1.1 and 3.5 m above the water surface. The fluxes of H2O, CO2, and CH4 and the isotopic compositions of lake evaporation are determined with the gradient-diffusion method. Water samples are collected daily at midday from the 20-cm depth for analysis of pH, D, and 18O isotopic compositions and dissolved CO2, CH4, and N2O concentrations. Water chemical parameters (pH, turbidity, dissolved oxygen, chlorophyll, blue-green algae phycocyanin, conductivity, salinity, oxidation reduction potential) are recorded every 15 min by an early warming monitoring system (model GuardianBlue, HACH Company, Loveland, Colorado) and a multi-parameter water quality analyzer (model YSI6600, YSI Inc., Yellow Spring, Ohio).

At the BFG site, the dissolved oxygen content, oxidation reduction potential, water temperature, conductivity, and pH are monitored continuously (model Professional Plus, YSI Inc.). A T-chain (PME Inc., Vista, California) provides another set of water temperature profile measurements. A second eddy covariance system, consisting of a sonic anemometer/thermometer (model CSAT3, Campbell Scientific) and an open-path CH4 analyzer (model LI-7700, LI-COR, Inc., Lincoln, Nebraska), measures the CH4 flux. Submerged macrophyte vegetation is abundant in the eddy flux footprint of this site (peak biomass density 1.4 kg m−2; Liu et al. 2007). These additional measurements will help us to better understand its ecophysiological processes.

All the micrometeorological variables are averaged to half-hourly intervals. The eddy flux and micro-meteorological data are retrieved weekly via a cellular communication device (model H7118 GPRS/EDGE DTU, Hongdian Technologies Corporation, Shenzhen, China) and inspected for quality assurance. The 10-Hz EC data are retrieved upon biweekly to monthly site visits. Coordinate rotation into the natural wind system is performed on the EC statistics (Lee et al. 2004) and density corrections are applied to the CO2, CH4, and H2O fluxes (Webb et al. 1980).

Lake surveys.

To gain a more detailed understanding of spatial variations of the greenhouse gas fluxes, we supplement the eddy covariance measurement with periodic spatial surveys of the concentrations of CO2, CH4, and N2O dissolved in the lake water. At monthly intervals, water samples are collected at the 20-cm depth from 17 points across the northern half of the lake. Every three months, water samples are collected from a network of 29 points across the whole lake. The half-lake survey is completed in one day and the whole-lake survey is completed in two consecutive days. The water samples are sealed in 300-mL glass bottles without air space and are analyzed within 24–48 h after collection for the dissolved CO2, CH4, and N2O concentrations using the equilibrium method. Pure N2 gas is injected into the glass bottle, creating a 100-mL headspace. The displaced water is analyzed for alkalinity, pH, and water isotopic compositions. The remaining 200-mL water is then vigorously mixed with the N2 gas for 5 min to promote liquid–gas equilibrium. The gaseous concentrations of CO2, CH4, and N2O are analyzed on a gas chromatography (model 6890N, Agilent Technologies Inc., Loveland, Colorado) and subsequently converted to the dissolved concentration using the Henry law. The water–air flux F is calculated from the bulk diffusion model,

 
formula

where Cs is the CO2, CH4, or N2O concentration in equilibrium with ambient air, Cb is the concentration in the bulk water below the interfacial layer, and k is an exchange coefficient (e.g., Cole et al. 1994).

The D and 18O compositions of the water samples are analyzed with isotopic ratio infrared spectroscopy (Model DLT-100, Los Gatos Research). Each sample is measured three times against a working standard traceable to the Vienna Standard Mean Ocean Water (VSMOW) scale. The isotope compositions of the lake water are useful tracer of water currents and provide constraints on the lake evaporative flux.

Modeling lake–air interactions.

The modeling component of the project is concerned with science questions 1 and 2. The experimental data are used to improve parameterization of the lake–air interactions. We will test the bulk parameterizations for momentum, sensible heat, and latent heat transfers between the water surface and the atmosphere (Garratt 1992), models of varying complexity for lake evaporation (Brutsaert 1982), and the National Center for Atmospheric Research's (NCAR's) 1D diffusion model of heat transfer in lakes (Subin et al. 2012). Another task is to investigate the dynamics of lake–land circulations by embedding these parameterizations in the mesocale Weather Research and Forecasting (WRF) Model.

EARLY RESULTS.

Radiation fluxes.

Figure 3 presents an example of the time series of the surface radiation components and the flux variables. The observation was made in June 2013 over a 10-day period with variable cloudiness. The 24-h mean values for a typical clear-sky day [day of year (DOY) 172 or 21 June 2013] is provided in supplementary Table ES1. The six EC sites track each other very well in terms of the incoming solar radiation (K) and the incoming longwave radiation (L) despite the large spatial separation among the sites (up to 45 km; Fig. 1), indicating uniform sky conditions across the lake and lack of persistent local disturbance to cloud formation. Among the five lake sites, there are little spatial variations in the outgoing longwave radiation (L) except on the clear day (DOY 154) when the BFG site—a site with submerged macrophytes—has a slightly higher value owning to warmer surface water (bottom panel on the right, Fig. 3). The clear-sky albedo shows small but measurable differences among the lake sites: the two eutrophic sites (MLW and DPK) have lower values than the other two cleaner sites (XLS and PTS; see supplementary Table ES1). The macrophyte site (BFG) also has low albedo.

Fig. 3.

Time series of micrometeorological and flux variables, DOY 152–162 (1–10 Jun 2013). Ticks on the horizontal axis mark the 00:00 h of the day. Here Rn is net radiation (W m−2), L is incoming longwave radiation (W m−2), L is outgoing longwave radiation, K is incoming shortwave radiation (W m−2), K is reflected shortwave radiation (W m−2), u* is friction velocity (m s−1), Fc is CO2 flux (μmol m−2 s−1), λE is latent heat flux (W m−2), H is sensible heat flux (W m−2), and Ts is surface temperature (°C).

Fig. 3.

Time series of micrometeorological and flux variables, DOY 152–162 (1–10 Jun 2013). Ticks on the horizontal axis mark the 00:00 h of the day. Here Rn is net radiation (W m−2), L is incoming longwave radiation (W m−2), L is outgoing longwave radiation, K is incoming shortwave radiation (W m−2), K is reflected shortwave radiation (W m−2), u* is friction velocity (m s−1), Fc is CO2 flux (μmol m−2 s−1), λE is latent heat flux (W m−2), H is sensible heat flux (W m−2), and Ts is surface temperature (°C).

The surface water temperature is surprisingly uniform across the lake, varying by less than 1.5°C among these sites over 90% of the observations shown in Fig. 3. A notable exception occurred at midday of DOY 154, when the macrophye site (BFG) was 5.5°C warmer than the rest of the sites. In a sensitivity analysis using the NCAR's lake model, Deng et al. (2013) found that the lake surface temperature is not sensitive to water pollution status or wind speed.

The reflected shortwave radiation (K) and the outgoing longwave radiation (L) show large land–lake contrasts. On DOY 172, the albedo of the land site (DS) is 0.09–0.11 greater than the lake sites. The outgoing longwave radiation at DS is higher in the day and lower at night than the lake sites. The land surface at DS is warmer in the daytime and cooler at night than the water surface. Generally, the daily mean net radiation is lower at DS than at the lake sites. For example, on DOY174, the net radiation at DS is 174.6 W m−2 and is 30–40 W m−2 lower than at the lake sites.

Sensible and latent heat fluxes.

The time series plot reveals several features typical of the sensible (H) and latent heat flux (λE) at this lake. Owing to the heat storage in the water column, the fluxes at the lake sites with open fetch (DPK, BFG, XLS, and PTS) show little diurnal variations. At the MLW site near the north shore, the land influence is evidenced by the large H during a few midday periods (DOY 155–157) when wind blew from the shore. Unlike at Lake Superior where the atmospheric surface layer is generally stable and H is negative during the warm season (Blanken et al. 2011), positive H occurred for about 75% of the observations shown in Fig. 3. The half-hourly H and λE time series at the sites with open fetch track each reasonably well. In Lake Kinneret in a semi-arid climate, spatial variations in evaporation can be as large as 100% on the daily basis (Assouline and Mahrer 1996).

On the monthly time scale, the lake evaporation is controlled primarily by net radiation (Fig. 4a; Wang et al. 2014). When plotted against the available energy, the data collapse onto a single line represented by the Priestley–Taylor evaporation model with the original α coefficient of 1.26, regardless of measurement location in the lake (Fig. 4b). Similar agreement has been reported for a tropical reservoir (dos Reis and Dias 1998). Here we have removed the bias errors of energy imbalance by forcing energy balance closure according to the procedure described by Blanken et al. (1997) and Twine et al. (2000). The data from the land site also follow the Priestley–Taylor model prediction, but with a modified coefficient of 1.0.

Fig. 4.

Monthly latent heat flux λE as a function of (top) net radiation Rn and (bottom) available energy RnG, where G is heat storage in soil or water, Δ is the slope of the saturation vapor pressure curve, and γ is the psychrometric constant.

Fig. 4.

Monthly latent heat flux λE as a function of (top) net radiation Rn and (bottom) available energy RnG, where G is heat storage in soil or water, Δ is the slope of the saturation vapor pressure curve, and γ is the psychrometric constant.

Momentum exchange.

Unsurprisingly, the friction velocity u* at the land site (DS) is much higher than at the open-water lake sites (DPK, BFG, XLS, PTS). For the data period shown in Fig. 3, the difference is about a factor of 2. The MLW site has higher friction velocity than the other lake sites, once again a result of land influences. In lake zones sheltered by land vegetation, momentum transfer into the water is usually reduced (Hondzo and Stefan 1993), so the enhanced friction velocity at MLW, measured at the height of 3.5 m above the water, is perhaps more indicative of a wake effect of the land than the true momentum flux from the atmosphere to the water (Markfort et al. 2010). During periods with open fetch (wind direction 200° to 315°; Fig. 1), however, the MLW drag coefficient is not much different from that at DPK. The drag coefficient is a measure of efficiency of momentum exchange between the lake water and the atmosphere. Under neutral stability it is given as

 
formula

where u* is friction velocity and u10 is wind speed at the 10-m height. According to Xiao et al. (2013), the open-fetch drag coefficient at the wind speed of u10 = 5 m s−1 is 2.0 (±0.4) × 10−3 at DPK, 1.9 (±0.3) × 10−3 at MLW, and 1.1 (±0.4) × 10−3 at BFG. The BFG site has much lower CD10N because the presence of submerged macrophytes has the effect of reducing the roughness of the water surface. Using the same data analysis method of Xiao et al. (2013), we found that the CD10N values at the other two newer lake sites (XLS and PTS) are also higher than that at BFG (1.3 ± 0.4 × 10−3 for u10 = 5 m s−1). For comparison, the standard drag coefficient model, established from observations in the marine environment, predicts a CD10N value of 1.0 × 10−3 at u10 = 5 m s−1 (Garratt 1992).

Carbon dioxide flux.

The most interesting feature regarding the CO2 flux (Fc) data is the negative flux, or CO2 diffusion into the water, in darkness (DOY 154–157; Fig. 3). These nocturnal uptake events usually persist through the whole evening and have been observed at all the lake sites. They tend to occur in the night following a daylight period of strong solar radiation, although not every high solar radiation day is followed with an uptake event, and are rarely seen under overcast conditions. The nocturnal uptake occurs more frequently at the macrophyte site (BFG), where the plants are totally submerged in water, than at the other lake sites. In the calendar year 2012, there were 106 night periods with persistent negative flux at BFG and only 76 at DPK. Our results indicate that the macrophyte (Potamogeton malaianus and Hydrilla verticillata) and phytoplankton species (Microcystis and Navicula) in Lake Taihu have the capacity to deploy Crassulacean acid metabolism (CAM) for carbon capture (e.g., Maberly and Madsen 2002).

The most negative half-hourly flux (–24.6 μmol m−2 s−1) was recorded at BFG on the evening of 30 July (DOY 212; Fig. 5). The flux is too large to be explained by measurement uncertainties; furthermore, the large and persistent nocturnal negative flux caused the CO2 concentration in the surface air at BFG to reach minima at sunrise as opposed to minima in midafternoon at the site near land (MLW). For comparison, the midday photosynthetic flux of coniferous forests on land is typically −15 μmol m−2 s−1 during the growing season (e.g., Goulden et al. 2006), and the nocturnal uptake flux of a land CAM plant is about −10 μmol m−2 s−1 following a period of exposure to high photosynthetically active radiation (Nobel and Hartsock 1983). In this latter study, the flux is expressed at the leaf scale as flux density per unit leaf area. In the only published ecosystem-scale carbon flux study involving CAM plants (pineapple), José et al. (2007) reported a nighttime uptake of −1.1 μmol m−2 s−1.

Fig. 5.

(top) Half-hourly CO2 flux at BFG. (bottom) Half-hourly time series for a 10-day period (23 Jul–1 Aug 2012). (a) Incoming shortwave radiation at BFG, (b) CO2 mixing ratio at BFG (solid line) and at MLW (dashed line), (c) CO2 flux (black solid line) and 0.2–1.0-m water temperature difference ΔTw (1.0-m temperature minus 0.2-m temperature; red dashed line) at BFG, and (d) water (purple dashed line, 0.2 cm; solid black line, 0.5 m; blue dashed line, 1 m; green dashed line, 1.5 m) and sediment temperature (gray line) at BFG.

Fig. 5.

(top) Half-hourly CO2 flux at BFG. (bottom) Half-hourly time series for a 10-day period (23 Jul–1 Aug 2012). (a) Incoming shortwave radiation at BFG, (b) CO2 mixing ratio at BFG (solid line) and at MLW (dashed line), (c) CO2 flux (black solid line) and 0.2–1.0-m water temperature difference ΔTw (1.0-m temperature minus 0.2-m temperature; red dashed line) at BFG, and (d) water (purple dashed line, 0.2 cm; solid black line, 0.5 m; blue dashed line, 1 m; green dashed line, 1.5 m) and sediment temperature (gray line) at BFG.

The growth of submerged aquatic plants is thought to be diffusion limited (Raven 1970; Smith and Walker 1980; Maberly and Madsen 2002). The molecular diffusivity of CO2 in water is four orders of magnitude smaller than in air. The leaf boundary layer resistance of plants submerged in stirred water is typically 60–100 s cm−1 (Black et al. 1981), whereas land plants have a resistance on the order of 0.1–1 s cm−1 (Campbell 1977). The slow diffusion is also manifested in the small water–air exchange coefficient k. A typical k value for inland lakes is ~0.5 m day−1 (Read et al. 2012), which would cap the CO2 uptake at no more than ~0.1 μmol m−2 s−1 according to Eq. (1). The negative Fc values in this study exceed the diffusion limit by two orders of magnitude, calling into question the applicability of the diffusion limitation paradigm at this shallow lake.

The uptake events start precisely at the time when ΔTw switches from being negative to being positive in the late afternoon and end when ΔTw turns negative again the next morning (Fig. 5). Here ΔTw is the difference in water temperature between the 100- and the 20-cm depth. At this lake, water temperature is almost always greater than 4°C, so a positive ΔTw is indicative of convective instability of the water column. It appears that the bulk diffusion model [Eq. (1)] needs serious revision under conditions of convective instability and in the presence of an active sink in the water column.

The CAM of aquatic plants is hypothesized to be an adaptation to diffusion limitation (Maberly and Madsen 2002). Our alternative hypothesis is that in shallow lakes, diffusion is not a limiting factor during most nights because of convective instability in the water column, and CAM is a physiological mechanism adapted to the high availability of the CO2 resource at night.

Spatial patterns from lake surveys.

Figure 6 presents an example of the data collected during a whole-lake survey. This and other surveys reveal spatially coherent patterns in the dissolved CO2, CH4, and N2O concentrations and in the water 18O/16O isotopic ratio. During this survey, which took place on 13–14 May 2013 between 0600 and 1630 local time, the mean lake water temperature is 22.1°C. For reference, the gas concentration in equilibrium with the ambient air is 16.3 μmol L−1 (CO2), 2.9 nmol L−1 (CH4), and 9.4 nmol L−1 (N2O). The mean (±1 standard deviation) of the dissolved CO2 concentration is 36.4 (±29.5) μmol L−1. Of the 29 sampling locations, 27 are supersaturated, having CO2 concentration values greater than the equilibrium value. The highest values were observed in zone 3 (Fig. 1), indicating the dominant role of river carbon import.

Fig. 6.

Spatial distributions of dissolved (a) CO2, (b) CH4, and (c) N2O concentration and (d) the 18O composition of the lake water, 13–14 May 2013.

Fig. 6.

Spatial distributions of dissolved (a) CO2, (b) CH4, and (c) N2O concentration and (d) the 18O composition of the lake water, 13–14 May 2013.

The mean N2O concentration is 13.3 ± 9.4 nmol L−1, with the majority of the sampling locations (22) being supersaturated with respect to the atmosphere. The N2O concentration shows a northwest to southeast gradient with the highest values observed in zone 3 (Fig. 6c). This pattern is similar to that of CO2 (Fig. 6a) and also resembles the spatial variations of the total N concentration reported for the lake (Yan et al. 2011).

In contrast to CO2 and N2O, the highest CH4 concentrations were observed in zones 5 and 7 dominated by macrophyte habitats (Fig. 6b). We interpret this as evidence that the main source of CH4 in this lake is the organic carbon from the primary production of macrophyte plants. All the sampling sites are supersaturated in CH4 with respect to the atmosphere, with the mean concentration of 94.2 ± 120.1 nmol L−1. Among the three gas species measured, CH4 shows the largest spatial variations: the ratio of the maximum to minimum concentration is 61. For comparison, the maximum-to-minimum concentration ratio is 17 for CO2 and 7 for N2O.

The 18O composition of the lake water becomes progressively enriched from north to south (Fig. 6d). The mean value for the whole lake is −3.71 ± 0.52 per mil (VSMOW scale) and is closer to the 18O composition of the outflows (–3.88 ± 0.87 per mil) than that of the inflows (–5.08 ± 0.86 per mil), both of which were measured on 21–22 May 2013. The enrichment, caused by the kinetic and equilibrium fractionation during evaporation, can be viewed as a proxy for “water age” or the time elapsed since the water enters the lake via the inflow rivers and precipitation. It appears that at this time of the year, the lake water moved predominantly in the north–south direction.

Modeling lake–air interactions.

The data collected at Lake Taihu have been used to improve the lake model in NCAR's Community Land Model (CLM) system. In the latest version of the CLM lake model (Subin et al. 2012), the water column is divided into 10 layers, and heat diffusion between the adjacent layers is described by a one-dimensional diffusion equation. Tests of the model for deep lakes reveal that the eddy thermal diffusivity given by the model is 1–2 orders of magnitude too low (Martynov et al. 2010; Subin et al. 2012). In contrast, we found that for Lake Taihu the diffusivity should be adjusted downward by 92% in order for the model to reproduce successfully the observed diurnal variations in the lake surface temperature (Deng et al. 2013). The small eddy diffusivity suggests that unresolved vertical eddy motions are weak or absent at this large and shallow lake.

The tuned lake model has now been successfully coupled to the WRF Model. In the default version of WRF V3.3.1, the lake surface temperature is set to the sea surface temperature at the same latitude and the surface sensible and latent heat fluxes are computed with a bulk scheme. The tuned lake model shows significant improvement over the WRF default in the predictions of the surface state and flux variables (see Fig. ES2 and Table ES2 in the online supplement). These improvements should lead to more realistic simulations of the lake breeze circulation.

SUMMARY.

The experimental objective of this field program is to quantify the lake–air fluxes of energy, momentum, and greenhouse gases across pollution and biological gradients in the lake. The data collected so far indicate that (i) the radiation balance components and surface temperature show very small variations across the lake, (ii) the momentum exchange is reduced at the habitat of submerged macrophytes, (iii) there is evidence of CAM for carbon capture, and (iv) there are large spatial gradients of the concentrations of CO2, CH4, and N2O dissolved in the lake water. The ongoing field experiment, data analysis, and model development will attempt to answer the five science questions outlined in section 1.

ACKNOWLEDGMENTS

This research was supported by the Ministry of Education of China (Grant PCSIRT), the Priority Academic Program Development of Jiangsu Higher Education Institutions (Grant PAPD), the National Science Foundation of China (Grant 41275024), and the Natural Science Foundation of Jiangsu Province, China (Grant BK2010572). We thank Prof. Peter Raymond of Yale University and Prof. Shuqing An of Nanjing University for helpful discussions and Dr. David Fitzjarrald and two anonymous reviewers for their constructive comments on this paper.

REFERENCES

REFERENCES
An
,
S.
, and
R.
Wang
,
2008
:
Human-induced drivers of the development of Lake Taihu
.
Lectures on China's Environment
,
X.
Lee
,
Ed.
,
Yale School of Forestry and Environmental Studies Publication Series
,
Vol. 20
,
151
165
.
Assouline
,
S.
, and
Y.
Mahrer
,
1996
:
Spatial and temporal variability in microclimate and evaporation over Lake Kinneret: Experimental evaluation
.
J. Appl. Meteor.
,
35
,
1076
1084
,
doi:10.1175/1520-0450(1996)035<1076:SATVIM>2.0.CO;2
.
Baldocchi
,
D.
, and
Coauthors
,
2001
:
FLUXNET: A new tool to study the temporal and spatial variability of ecosystem-scale carbon dioxide, water vapor, and energy flux densities
.
Bull. Amer. Meteor. Soc.
,
82
,
2415
2434
,
doi:10.1175/1520-0477(2001)082<2415:FANTTS>2.3.CO;2
.
Balmer
,
M. B.
, and
J. A.
Downing
,
2011
:
Carbon dioxide concentrations in eutrophic lakes: Undersaturation implies atmospheric uptake
.
Inland Waters
,
1
,
125
132
,
doi:10.5268/IW-1.2.366
.
Bastviken
,
D.
,
L. J.
Tranvik
,
J. A.
Downing
,
P. M.
Crill
, and
A.
Enrich-Prast
,
2011
:
Freshwater methane emissions offset the continental carbon sink
.
Science
,
331
,
50
,
doi:10.1126/science.1196808
.
Black
,
M. A.
,
S. C.
Maberly
, and
D. H. N.
Spence
,
1981
:
Resistances to carbon dioxide fixation in four submerged freshwater macrophytes
.
New Phytol.
,
89
,
557
568
,
doi:10.1111/j.1469-8137.1981.tb02335.x
.
Blanken
,
P. D.
, and
Coauthors
,
1997
:
Energy balance and canopy conductance of a boreal aspen forest: Partitioning overstory and understory components
.
J. Geophys. Res.
,
102
,
28 915
28 927
,
doi:10.1029/97JD00193
.
Blanken
,
P. D.
, and
Coauthors
,
2011
:
Evaporation from Lake Superior: 1. Physical controls and processes
.
J. Great Lakes Res.
,
37
,
707
716
Brown
,
M. G.
, and
Coauthors
,
2012
:
The carbon balance of two lodgepole pine stands recovering from mountain pine beetle attack in British Columbia
.
Agric. For. Meteor.
,
153
,
82
93
,
doi:10.1016/j.agrformet.2011.07.010
.
Brutsaert
,
W. H.
,
1982
:
Evaporation into the Atmosphere
.
D.
Reidel
,
299
pp
.
Campbell
,
G. S.
,
1977
:
Introduction to Environmental Biophysics
.
Springer
,
159
pp
.
Cole
,
J. J.
, and
N. F.
Caraco
,
G. W.
Kling
, and
T. K.
Kratz
,
1994
:
Carbon dioxide supersaturation in the surface waters of lakes
.
Science
,
265
,
1568
1570
.
Deng
,
B.
,
S.
Liu
,
W.
Xiao
,
W.
Wang
,
J.
Jin
, and
X.
Lee
,
2013
:
Evaluation of the CLM4 lake model at a large and shallow freshwater lake
.
J. Hydrometeor.
,
14
,
636
649
,
doi:10.1175/JHM-D-12-067.1
.
dos Reis
,
R. J.
, and
N. L.
Dias
,
1998
:
Multi-season lake evaporation: Energy-budget estimates and CRLE model assessment with limited meteorological observations
.
J. Hydrol.
,
208
,
135
147
,
doi:10.1016/S0022-1694(98)00160-7
.
Downing
,
J. A.
, and
Coauthors
,
2006
:
The global abundance and size distribution of lakes, ponds, and impoundments
.
Limnol. Oceanogr.
,
51
,
2388
2397
,
doi:10.4319/lo.2006.51.5.2388
.
Garratt
,
J. R.
,
1992
:
The Atmospheric Boundary Layer
.
Cambridge University Press
,
316
pp
.
Goulden
,
M. L.
, and
Coauthors
,
2006
:
An eddy covariance mesonet to measure the effect of forest age on land-atmosphere exchange
.
Global Change Biol.
,
12
,
2146
2162
,
doi:10.1111/j.1365-2486.2006.01251.x
.
Hari
,
P.
,
J.
Pumpanen
,
J.
Huotari
,
P.
Kolari
,
J.
Grace
,
T.
Vesala
, and
A.
Ojala
,
2008
:
High-frequency measurements of productivity of planktonic algae using rugged nondispersive infrared carbon dioxide probes
.
Limnol. Oceanogr. Methods
,
6
,
347
354
,
doi:10.4319/lom.2008.6.347
.
Hondzo
,
M.
, and
H. G.
Stefan
,
1993
:
Lake water temperature simulation model
.
J. Hydraul. Eng.
,
119
,
1251
1273
,
doi:10.1061/(ASCE)0733-9429(1993)119:11(1251)
.
Hu
,
W.
,
S. E.
Jørgensenb
,
F.
Zhang
,
Y.
Chen
,
Z.
Hu
, and
L.
Yang
,
2011
:
A model on the carbon cycling in Lake Taihu, China
.
Ecol. Modell.
,
222
,
2973
2991
,
doi:10.1016/j.ecolmodel.2011.04.018
.
Huttunen
,
J. T.
, and
Coauthors
,
2003
:
Fluxes of methane, carbon dioxide and nitrous oxide in boreal lakes and potential anthropogenic effects on the aquatic greenhouse gas emissions
.
Chemosphere
,
52
,
609
621
,
doi:10.1016/S0045-6535(03)00243-1
.
José
,
J. S.
,
R.
Montes
, and
N.
Nikonova
,
2007
:
Seasonal patterns of carbon dioxide, water vapour and energy fluxes in pineapple
.
Agric. For. Meteor.
,
147
,
16
34
,
doi:10.1016/j.agrformet.2007.06.003
.
Lee
,
X.
,
W.
Massman
, and
B.
Law
,
2004
:
Handbook of Micrometeorology: A Guide for Surface Flux Measurement and Analysis
.
Kluwer Academic
,
250
pp
.
Liu
,
H.
,
Q.
Zhang
, and
G.
Dowler
,
2012
:
Environmental controls on the surface energy budget over a large southern inland water in the United States: An analysis of one-year eddy covariance flux data
.
J. Hydrometeor.
,
13
,
1893
1910
,
doi:10.1175/JHM-D-12-020.1
.
Liu
,
W.
,
W.
Hu
,
Y.
Chen
,
X.
Gu
,
Z.
Hu
,
Y.
Chen
, and
J.
Ji
,
2007
:
Temporal and spatial variation of aquatic macrophytes in West Taihu Lake
.
Acta Ecol. Sin.
,
27
,
159
170
.
Maberly
,
S. C.
, and
T. V.
Madsen
,
2002
:
Freshwater angiosperm carbon concentrating mechanisms: Processes and patterns
.
Funct. Plant Biol.
,
29
,
393
405
,
doi:10.1071/PP01187
.
Markfort
,
C. D.
,
A. L. S.
Perez
,
J. W.
Thill
,
D. A.
Jaster
,
F.
Porté-Agel
, and
H. G.
Stefan
,
2010
:
Wind sheltering of a lake by a tree canopy or bluff topography
.
Water Resour. Res.
,
46
,
W03530
,
doi:10.1029/2009WR007759
.
Martynov
,
A.
,
L.
Sushama
, and
R.
Laprise
,
2010
:
Simulation of temperate freezing lakes by one-dimensional lake models: Performance assessment for interactive coupling with regional climate models
.
Boreal Environ. Res.
,
15
,
143
164
.
Nobel
,
P. S.
, and
T. L.
Hartsock
,
1983
:
Relationships between photosynthetically active radiation, nocturnal acid accumulation, and CO2 uptake for a Crassulacean acid metabolism plant, Opunita ficus-indica
.
Plant Physiol.
,
71
,
71
75
,
doi:10.1104/pp.71.1.71
.
Nordbo
,
A.
,
S .
Launiainen
,
I.
Mammarella
,
M.
Leppäranta
,
J.
Huotari
,
A.
Ojala
, and
T.
Vesala
,
2011
:
Long-term energy flux measurements and energy balance over a small boreal lake using eddy covariance technique
.
J. Geophys. Res.
,
116
,
D02119
,
doi:10.1029/2010JD014542
.
Prescher
,
A. K.
,
T.
Grunwald
, and
C.
Bernhofer
,
2010
:
Land use regulates carbon budgets in eastern Germany: From NEE to NBP
.
Agric. For. Meteor.
,
150
,
1016
1025
,
doi:10.1016/j.agrformet.2010.03.008
.
Qin
,
Q.
,
P.
Xu
,
Q.
Wu
,
L.
Luo
, and
Y.
Zhang
,
2007
:
Environmental issues of Lake Taihu, China
.
Hydrobiologia
,
581
,
3
14
,
doi:10.1007/s10750-006-0521-5
.
Raven
,
J. A.
,
1970
:
Exogenous inorganic carbon sources in plant photosynthesis
.
Biol. Rev. Camb. Philos. Soc.
,
45
,
167
221
,
doi:10.1111/j.1469-185X.1970.tb01629.x
.
Read
,
J. S.
, and
Coauthors
,
2012
:
Lake-size dependency of wind shear and convection as controls on gas exchange
.
Geophys. Res. Lett.
,
39
,
L09405
,
doi:10.1029/2012GL051886
.
Rouse
,
W. R.
,
P. D.
Blanken
,
N.
Bussières
,
C. J.
Oswald
,
W. M.
Schertzer
,
C.
Spence
, and
A. E.
Walker
,
2008
:
Investigation of the thermal and energy balance regimes of Great Slave and Great Bear Lakes
.
J. Hydrometeor.
,
9
,
1318
1333
,
doi:10.1175/2008JHM977.1
.
Samuelsson
,
P.
,
E.
Kourzeneva
, and
D.
Mironov
,
2010
:
The impact of lakes on the European climate as simulated by a regional climate model
.
Boreal Environ. Res.
,
15
,
113
129
.
Schimel
,
D.
,
W.
Hargrove
,
F.
Hoffman
, and
J.
MacMahon
,
2007
:
NEON: A hierarchically designed national ecological network
.
Front. Ecol. Environ
,
5
,
59
,
doi:10.1890/1540-9295(2007)5 [59:NAHDNE]2.0.CO;2
.
Schubert
,
C. J.
,
T.
Diem
, and
W.
Eugster
,
2012
:
Methane emissions from a small wind shielded lake determined by eddy covariance, flux chambers, anchored funnels, and boundary model calculations: A comparison
.
Environ. Sci. Technol.
,
46
,
4515
4522
,
doi:10.1021/es203465x
.
Sills
,
D. M. L.
,
J. R.
Brook
,
I.
Levy
,
P. A.
Makar
,
J.
Zhang
, and
P. A.
Taylor
,
2011
:
Lake breezes in the southern Great Lakes region and their inf luence during BAQS-Met 2007
.
Atmos. Chem. Phys.
,
11
,
7955
7973
,
doi:10.5194/acp-11-7955-2011
.
Smith
,
F. A.
, and
N. A.
Walker
,
1980
:
Photosynthesis by aquatic plants: Effects of unstirred layers in relation to assimilation of CO2 and HCO3 and to carbon isotope discrimination
.
New Phytol.
,
86
,
245
259
,
doi:10.1111/j.1469-8137.1980.tb00785.x
.
St. Louis
,
V. L.
,
C. A.
Kelly
,
É.
Duchemin
,
J. W. M.
Rudd
, and
D. M.
Rosenberg
,
2000
:
Reservoir surfaces as sources of greenhouse gases to the atmosphere: A global estimate
.
Bioscience
,
50
,
766
775
,
doi:10.1641/0006–3568(2000)050[0766:RSASOG]2.0.CO;2.
Stoy
,
P. C.
,
G. G.
Katul
,
M. B. S.
Siqueira
,
J. Y.
Juang
,
K. A.
Novick
,
H. R.
McCarthy
,
A. C.
Oishi
, and
R.
Oren
,
2008
:
Role of vegetation in determining carbon sequestration along ecological succession in the southeastern United States
.
Global Change Biol.
,
14
,
1409
1427
,
doi:10.1111/j.1365-2486.2008.01587.x
.
Subin
,
Z. M.
,
W. J.
Riley
, and
D. V.
Mironov
,
2012
:
An improved lake model for climate simulations: Model structure, evaluation, and sensitivity analyses in CESM1
.
J. Adv. Model. Earth Syst.
,
4
,
M02001
,
doi:10.1029/2011MS000072
.
Taihu Office
,
2012
:
The 2013 Health Report of Lake Taihu
.
Taihu Bureau, Ministry of Water Resources of China, Shanghai Office
,
16
pp
.
Twine
,
T. E.
, and
Coauthors
,
2000
:
Correcting eddycovariance flux underestimates over a grassland
.
Agric. For. Meteor.
,
103
,
279
300
,
doi:10.1016/S0168-1923(00)00123-4
.
Wang
,
W.
, and
Coauthors
,
2014
:
Temporal and spatial variations in radiation and energy balance across a large freshwater lake in China
.
J. Hydrol.
,
511
,
811
824
,
doi:10.1016/j.jhydrol.2014.02.012
.
Webb
,
E. K.
,
G. I.
Pearman
, and
R.
Leuning
,
1980
:
Correction of flux measurements for density effects due to heat and water vapor transfer
.
Quart. J. Roy. Meteor. Soc.
,
106
,
85
100
,
doi:10.1002/qj.49710644707
.
Xiao
,
W.
,
S.
Liu
,
W.
Wang
,
D.
Yang
,
J.
Xu
,
C.
Cao
,
H.
Li
, and
X.
Lee
,
2013
:
Transfer coefficients of momentum, heat and water vapour in the atmospheric surface layer of a large freshwater lake
.
Bound.-Layer Meteor.
,
148
,
479
494
,
doi:10.1007/s10546-013-9827-9.
Yan
,
S.
,
H.
Yu
,
L.
Zhang
,
J.
Xu
, and
Z.
Wang
,
2011
:
Water quantity and pollutant fluxes of inflow and outflow rivers of Lake Taihu 2009
.
J. Lake Sci.
,
23
,
855
862
.
Zha
,
T.
, and
Coauthors
,
2009
:
Carbon sequestration in boreal jack pine stands following harvesting
.
Global Change Biol.
,
15
,
1475
1487
,
doi:10.1111/j.1365-2486.2008.01817.x.
Zhang
,
X.
, and
Coauthors
,
2010
:
The 2007 water crisis in Wuxi, China: Analysis of the origin
.
J. Hazard. Mater.
,
182
,
130
135
,
doi:10.1016/j.jhazmat.2010.06.006
.
Zhao
,
G.
,
J.
Gao
,
P.
Tian
,
K.
Tian
, and
G.
Ni
,
2011
:
Spatial–temporal characteristics of surface water quality in the Taihu Basin, China
.
Environ. Earth Sci.
,
64
,
809
819
,
doi:10.1007/s12665-011-0902-6
.
Zhao
,
L.
,
J.
Jin
,
S. Y.
Wang
, and
M. B.
Ek
,
2012
:
Integration of remote-sensing data with WRF to improve lake-effect precipitation simulations over the Great Lakes region
.
J. Geophys. Res.
,
117
,
D09102
,
doi:10.1029/2011JD016979
.

Footnotes

A supplement to this article is available online (10.1175/BAMS-D-13-00136.2)

Supplemental Material