Wetlands Ecol Manage DOI 10.1007/s11273-015-9421-7

ORIGINAL PAPER

Carbon dioxide exchange in a tropical wet grassland M. L. Meirelles . R. Bracho . E. A. B. Ferreira

Received: 17 February 2014 / Accepted: 8 April 2015 Ó Springer Science+Business Media Dordrecht 2015

Abstract Wet grasslands in the Brazilian Cerrado occupy around 1 % of the Cerrado extension, acting as ecotones connecting well-drained savanna vegetation and gallery forest. Wet grasslands help maintain biodiversity and water balance, and are an important component of the carbon (C) balance, storing an average of 241 MgC h-1 in the top 60 cm soils. However, in addition to increased climactic variability, Brazilian wet grasslands are under pressure from domestic water use and agricultural irrigation. Here, we present 2 years of carbon balance data from a Brazilian wet grassland, which was exposed to below average seasonal and annual precipitation during the first year and excess precipitation during the second year. We used the eddy covariance approach to measure net ecosystem carbon exchange (NEE) and ecosystem respiration (Re). Results indicated that the wet grassland was a carbon sink during both years,

M. L. Meirelles (&)  E. A. B. Ferreira Embrapa Cerrados, BR 020, Km 18, Brasilia, DF CEP 73310-970, Brazil e-mail: [email protected] R. Bracho Department of Biology, University of Florida, Gainesville, FL 32611, USA R. Bracho School of Forest Resources and Conservation, University of Florida, Gainesville, FL 32611, USA

taking up similar amounts each year (-86.5 gC m-2 year-1 in 2005 and -81.3 gC m-2 year-1 in 2006). However, seasonal NEE was variable and driven by fluctuations in the water table depth (P \ 0.05), shifting the balance from a carbon sink to a carbon source from 1 month to the next by as much as 50 gC m-2 when water table dropped below 20 cm depth. The gross ecosystem carbon uptake (GEE) decreased with water table depth, while Re increased due to the exposure of old organic carbon to microbial decomposition. Results indicated that a prolonged drought, which is probable under future projected climatic variability, could lead the ecosystem to be a carbon source and highlight the importance of management strategy in preserving soil water in these ecosystems. Keywords Wet grassland  Net ecosystem CO2 exchange  Eddy covariance  Water table

Introduction Wetlands are within the most important ecosystem on earth, covering between 8.2 and 10.1 9 106 km2, less than 10 % of the earth’s surface area (Lehner and Doll 2004). Globally, wetlands store between 202 and 535 Gt of carbon in their soils (average of 240 Gt C), which is more C than is stored in many other biomes occupying larger areas (Bernal and Mitsch 2008;

123

Wetlands Ecol Manage

Mitra et al. 2005), and about a third the carbon in the atmosphere. Wetlands have the capacity to accumulate C at rates ranging from\0.1 MgC ha-1 year-1 in arctic and subarctic peatlands to 4.80 MgC ha-1 year-1 in tropical wetlands (Mitsch et al. 2010; Ovenden 1990; Saunders et al. 2007); this ability to store large amounts of C is mainly due to low decompositions rates under anoxic conditions (Chimner and Cooper 2003). Globally, wetlands are susceptible to climactic variability, changes in land use, and anthropogenic disturbances, and are under very high pressure for resources extraction and land development (Hooijer et al. 2010; Kansiime et al. 2007; Saunders et al. 2014), all factors that can turn them into strong carbon sources [e.g., C emissions from drained Southeastern Asian forested peatlands contribute the equivalent of 1.3–3.1 % of the C emitted from burning fossil fuel (Hooijer et al. 2010)]. Wetlands act as ecotones between terrestrial and aquatic ecosystems, allowing for a large flow of mass and energy between connecting systems, which in turn lead to a very high biodiversity and provide many other services such as water storage and water quality control, food source, etc. (Mitra et al. 2005; Mitsch et al. 2008; Saunders et al. 2014). Tropical wetlands cover 2 9 106 Km-2 (about 30 % of global wetlands area) (Lehner and Doll 2004; Mitra et al. 2005; Mitsch et al. 2010), and about 2.5 9 104 Km2 of those are wet grasslands in the Brazilian Cerrado (Sano et al. 2008). These wetlands are one of the several ecosystems in the cerrado biome (Oliveira-Filho and Ratter 2002), occurring in relatively large flat areas between gallery forests (water ways) and well-drained ecosystems. The wetlands are a continuous treeless herbaceous stratum on hydromorphic gray waterlogged soils, which are inundated for most of the year but may dry up during the dry season (Ribeiro and Walter 1998; Oliveira-Filho and Ratter 2002). The cerrado biome is highly diverse and is recognized as one of the 25 global biodiversity hot spots (Myers et al. 2000), covering more than 2.4 million km2 and up to 20 % of land surface of Brazil (Furley and Ratter 1988; Sano et al. 2008). Vegetation in the Cerrado is highly variable, driven mainly by soil fertility and ground water regimens (Furley and Ratter 1988; Oliveira-Filho and Ratter 2002). The hydric regimen in this region is characterised by alternating periods of water deficits and excess, which leads to the presence of seasonal grasslands (Oliveira-Filho and Ratter 2002).

123

Large amounts of carbon ([240 ton C ha-1) are stored in the soil profile due to low decomposition rates induced by anaerobic conditions (Ferreira 2009). Wet grasslands are also important for water storage, contributing to the sustainability of the region’s waterway, especially during the dry season. However, this ecosystem has a low resilience, and biological processes (photosynthesis and ecosystem respiration) are sensitive to environmental variability. Fluctuations in the precipitation regime and land use changes (drainage and grazing) have caused a reduction in ground water, which leads to colonisation by native tree and shrub species, which in turn shade the herbaceous species, reducing frequency and coverage of species adapted to a higher degree of flooding (Meirelles et al. 2004). Reduction in ground water exposes the accumulated organic carbon to decomposition, shifting the C storage typical of wetlands to overall C loss (Lloyd 2006; Mitsch et al. 2010). Moreover, future climate projections for the region indicate an increase in air temperature and decrease in precipitation by mid and at the end of the 21st century (Christensen 2013; IPCC 2013), which may produce a positive feedback to the atmosphere. Therefore, there is a need to estimate and understand the biophysical factors controlling the C balance in the wet grasslands in the Brazilian Cerrado. Such information is crucial for ecosystem management and for understanding the global carbon cycle. The eddy covariance technique provides almost continuous measurements of net ecosystem carbon exchange (NEE) with the atmosphere (Baldocchi 2003), and in conjunction with measurements of the physical environment and vegetation, allows for a description of the biophysical controls on the ecosystem C balance. NEE measurements in different wetlands around the world indicate that in these ecosystems, the C balance is controlled by hydrodynamics (precipitation and ground water), temperature, and vegetation phenology (Bonneville et al. 2008; Dusek et al. 2009, 2012; Hao et al. 2011; Sonnentag et al. 2010; Zhou et al. 2009). Here, we present 2 years of NEE measurements from a wet grassland ecosystem in the Brazilian Cerrado with a strong seasonal precipitation deficit during the first year and greater than average precipitation in the second year. This provided us with an opportunity to examine how wide variability in the precipitation regimen affects carbon balance in this important ecosystem in the Brazilian

Wetlands Ecol Manage

Cerrado. The objectives of this study were to (1) estimate seasonal and annual C balance and (2) determine environmental factors controlling C exchange between the wet grassland and the atmosphere. We expect the ecosystem to be a carbon source during the precipitation deficit period and to be a carbon sink on an annual basis.

Materials and methods Site description The study was conducted on 20 ha of natural tropical wet grassland in the cerrado biome (Brazilian savanna) in central Brazil, located about 18 km south-west of Brasilia city, Federal District, Brazil (15.55°S, 47.54°W, 1060 m.a.s.l.). The site is seasonally flooded with a continuous 70 cm-tall herbaceous layer dominated by four species, Axonopus comans (42.1 %), Andropogon lateralis (9.29 %), Andropogon bicornis (5.69 %), and Andropogon virgatus (5.13 %) (Rodrigues Munhoz and Felfili 2008). The soil is a gley soil, pH (H2O) 6.1; Al 0.56 cmolc dm-3; Ca ? Mg 0.33 cmolc dm-3; P 3.42 mg l-1; K 55 mg l-1; bulk density 0.5 g cm-3; saturated hydraulic conductivity 13.1 cm h-1. The C pool is up to 241 Mg C h-1 in the top 60 cm (Ferreira 2009). The drier margin of the wet grassland is bordered by an open woody savannah, while the other margin is bordered by a gallery forest. The area is on a 3 % slope through an open woody savanna—wet grassland— gallery forest and is well preserved; the last fire occurred in August 1999 and the area is not under a grazing regime or directly impacted by any other anthropological disturbance. The climate of the region is classified in the Koeppen’s Aw category, which has dry winters (May to September) and rainy summers (October–April), corresponding to the ‘‘dry’’ and ‘‘rainy’’ seasons in the Cerrado region. Mean 30 years annual precipitation is 1553 mm (http://www. bdclima.cnpm.embrapa.br/resultados/balanco.php? UF=df&COD=51). The study was conducted from January 1st 2005 to the end of December 2006. Net ecosystem carbon exchange (NEE) Net ecosystem carbon exchange (NEE) is the balance between two opposing processes: C uptake through

photosynthesis (gross primary production) and C loss via heterotrophic and autotrophic respiration. Here, we measured NEE using the eddy covariance approach with a time resolution of 30 min (Aubinet et al. 2000; Baldocchi 2003). Negative NEE values indicate C uptake from the atmosphere by the wet grassland. The eddy covariance instrumentation comprised a threedimensional sonic anemometer (CSAT3, Campbell Scientific, Logan UT, USA) and open-path infrared gas analyzer (LI-7500, LI-COR Inc., Lincoln, NE, USA) placed at 4 m above the ground on a mast (CM3, Campbell Scientific, Logan UT, USA). High frequency data were collected at 20 Hz on a data logger (CR5000, Campbell Scientific, Logan UT, USA). The eddy3 program developed by the micrometeorology group of Alterra (Research Institute of Wageningen University and Research Centre, Wageningen, The Netherlands), calculated the flux every 30 min. Fluxes were corrected for high frequency fluctuations, sensors separation, and misalignment of sensors with respect to the local streamline (Aubinet et al. 2000; Moncrieff et al. 1997). Density corrections due to water vapor and heat fluxes were also applied (Webb et al. 1980). Data screening was applied to eliminate half-hourly fluxes resulting from incomplete half hour measurements during the system calibration or maintenance, rainfall during the 30 min measurements, when the canopy was decoupled from the external atmospheric conditions as defined by the friction velocity, u* (0.12 ms-1) (Goulden et al. 1996), and excessive variation from the half-hourly means based on the analysis of standard deviations of u, v, and w wind statistics and CO2 fluxes (where u and v are the orthogonal horizontal and w the vertical wind speed components). Data were quality controlled by inspection of the steady state and turbulent conditions around the tower (Foken and Wichura 1996). On average, 62 % of the data that remained after filtering was applied. The fetch from the tower extended more than 300 m in all directions. Simulation of the foot print (Kormann and Meixner 2001), indicated that [75 % of the fluxes were generated from the first 150 m of the tower footprint under turbulent conditions. Meteorological measurements Standard meteorological measurements were collected at 4 m from a mast (CM3, Campbell Scientific, Logan UT, USA) adjacent to the eddy tower, including photosynthetically active radiation (PAR) (LI-190SA,

123

Wetlands Ecol Manage

LI-COR, Lincoln, NE, USA), global radiation (CM3, Kipp & Zonen, Delf, The Netherlands), net radiation (Q7, REBS, Bellevue, WA, USA), relative humidity (RH) and air temperature (Ta) (HMP-45C,Vaisala, Helsinki, Finland), precipitation (TE525 tipping bucket gage, Texas Electronic, Dallas, TX, USA), wind speed (014A, Met One), wind direction (024A, Met One), soil heat flux (HFT3, Campbell Scientific), and soil temperature (TCAV, Campbell Scientific, Logan UT, USA). All signals from these instruments were logged by a data logger (model CR23X, Campbell Scientific) at a scan rate of 1 min, and recorded as halfhourly mean values. Soil heat flux (G) was obtained from the mean value of two soil heat flux plates, corrected for heat storage above the plates. Distance to the water table (water table depth) was monitored at nine equally spaced locations along a transect that ran from the edge of the open woody savanna, across the wet grassland, and to the edge of the gallery forest. At each measurement point a 2 m PVC pipe was inserted in the ground and distance to the water table was measured monthly using a standard meter stick. Data gap filling strategies Gaps in the flux data were filled by separating day (PAR [ 10 l mols m-2 s-1) and night time conditions using monthly parameters obtained by fitting half hour NEE to PAR using a non-rectangular hyperbola (Falge et al. 2001) in the form NEE =

aPAR þRd 1ðPAR=2200ÞþðaPAR=NEEopt Þ

ð1Þ

where a (lmol CO2 m-2 s-1/lmol quantum m-2 s-1) is the ecosystem apparent quantum yield, NEEopt (lmol CO2 m-2 s-1) is the optimum rate of CO2 exchange at maximum observed PAR (2200 lmol m-2 s-1), and Rd (lmol CO2 m-2 s-1) is the ecosystem dark respiration (NEE at PAR = 0). Night time (PAR \ 10 l mols m-2 s-1) gaps in the data were filled using yearly parameters obtained by relating half hour night time NEE to air temperature in the form NEEnight ¼ a exp

bTa

ð2Þ

where a and b are regression coefficients and Ta (°C) is the half hour air temperature. Parameters for day and night time conditions were estimated (P \ 0.05) using the nonlinear procedure in SAS (SAS Institute Inc., Cary NC, USA).

123

Half hour GEE (gross ecosystem carbon exchange) was estimated as the difference between NEE and ecosystem respiration (Re): GEE ¼ NEE  Re

ð3Þ

where Re was estimated from annual night time parameters obtained using Eq. 2. Monthly and annual NEE is the sum of measured and gap-filled NEE.

Results Environmental conditions The long-term (30 years, 1965–2005) mean annual precipitation was 1553 mm; annual precipitation in 2005 was below the long-term average (1391 mm), and in 2006 it was above average (1829 mm). However, monthly precipitation deviated from the long-term average (Fig. 1a), e.g., a precipitation deficit of 228 mm accumulated form April to October 2005 (Fig. 1a), followed by precipitation of more than 219 mm above average in the following 2 months. Water table (Fig. 1a), decreased after July and reached the deepest value (71 cm) in October 2005. With the excess precipitation in the following 2 months, the water table recovered and stayed relatively close to the soil surface for the remainder of the study. The yearly maximum mean monthly Ta occurred in October (30.2 °C) 2005 and September (28.8 °C) 2006, and the minimum mean monthly Ta occurred in July each year (Fig. 1b), when minimum values of 6.9 and 6.8 °C were recorded in 2005 and 2006, respectively. Mean monthly maximum (9 am–4 pm) air vapor pressure deficit (VPD, Fig. 1b) occurred at the end of the dry season, reaching 2.78 kPa in October 2005 and 2.14 kPa in September 2006. Total incoming radiation was very similar each year with 6.62 GJ m-2 year-1 and 6.66 GJ m-2 year-1 for 2005 and 2006, respectively, with small monthly fluctuations due to cloud cover. Seasonal carbon exchange The wet grassland acted as a C sink during the 2 years of measurements accumulating a total of -167.8 gC m-2. Similar amounts of C were taken up each year, -86.5 gC m-2 year-1 in 2005 and -81.3 gC m-2 year-1 in 2006, although gross

Wetlands Ecol Manage 400

20

(a)

0 300 -20 200 -40

WTD (cm)

Precipitation (mm)

Fig. 1 a Monthly precipitation (dark bars), and average long-term (30 years) monthly precipitation (gray bars) and average monthly water table depth (closed circles); b monthly average air temperature (open circles) and average maximum (9 am–4 pm) air vapor pressure deficit (VPD, KPa; dark circles) in a Brazilian wet grassland

100 -60

0 24

-80

(b)

2.8

22

2.0 18

VPD (KPa)

20

o

Tair ( C)

2.4

1.6 16 1.2 14 Jan/05

Apr/05

Jul/05

ecosystem carbon exchange (GEE) and ecosystem respiration (Re) were about 100 gC m-2 lower in 2006 than in 2005 and seasonal carbon exchange varied from 1 year to the next. In 2005, GEE was 1509 gC m-2 year-1 and in 2006 it was 1409 gC m-2 year-1, while Re was 1423 gC m-2 year-1 in 2005 and 1327 gC m-2 year-1 in 2006. (Table 1) The ecosystem acted as a sustained C sink from January to July–August each year (Figs. 2a, 3), accumulating up to -185 and -117 gC m-2 in 2005 and 2006, respectively. GEE reached above -4 gC m-2 day-1, dominating over Re (Fig. 2b). NEE reached a maximum around -2 gC m-2 d-1 in June–July of each year. Maximum monthly NEE was -38.8 gC m-2 month-1 in July 2005 and -49.7 gC m-2 month-1 in June 2006 (Fig. 3). However, the ecosystem became a carbon source from September to December in 2005 and September to October 2006 (Figs. 2a, 3), losing 98.6 gC m-2 (53 % of C previously accumulated) to the atmosphere in the last 4 months of 2005 and 36.2 gC m-2 for the same period in 2006, as Re

Oct/05

Jan/06

Apr/06

Jul/06

Oct/06

became the dominant carbon flux reaching [?4 gC m-2 d-1 while GEE decreased to the minimal daily below -3 gC m-2 d-1 (Fig. 2b). NEE reached a maximum daily of ?2 gC m-2 d-1 in November 2005 and earlier in 2006. Maximum monthly C release was ?33.85 gC m-2 month-1 in November 2005 and 24.77 gC m-2 month-1 in September 2006 (Fig. 3). Controls over carbon exchange Monthly NEE was driven by distance to the water table (r2 = 0.33, P \ 0.01, REG procedure SAS, 9.2). The ecosystem became a strong C source in 2005 when the water table dropped below 20 cm. Ecosystem respiration (Re) was driven by air temperature showing a Q10 of 1.65. Parameters obtained from the light response curves were also used to evaluate seasonal environmental constraints on net ecosystem carbon exchange. Optimum NEE (NEEopt) represents the maximum C uptake capacity when light is not

123

Wetlands Ecol Manage Table 1 Monthly carbon fluxes (gC m-2 month-1), GEE, gross ecosystem production; Re, ecosystem respiration; NEE, net ecosystem carbon exchange; and distance to the water table (DTW, cm) in a Brazilian wet grassland Month-year

GEE

Re

NEE

WTD

Jan-05

-155.533

130.878

-24.653

-16.88

Feb-05

-143.834

118.028

-25.805

-2.80

Mar-05

-138.749

127.955

-10.794

-0.10

Apr-05

-139.846

118.547

-21.298

2.10

May-05

-127.613

108.620

-18.992

1.00

Jun-05 Jul-05

-117.462 -139.42

100.403 100.587

-17.059 -38.831

0.70 -0.40

Aug-05

-138.216

110.483

-27.732

-12.20

Sept-05

-95.722

121.385

25.662

-25.90

Oct-05

-110.228

136.879

26.651

-70.80

Nov-05

-90.725

124.578

33.853

-46.10

Dec-05

-112.220

124.741

12.521

-9.00

Jan-06

-138.493

130.313

-8.178

-1.80

Feb-06

-125.864

114.837

-11.026

-0.80

Mar-06

-128.031

119.328

-8.703

0.50

Apr-06

-125.400

116.804

-8.596

2.40

May-06

-125.736

92.878

-32.857

3.00

Jun-06

-134.125

84.394

-49.730

2.90

Jul-06

-91.4254

93.252

1.827

2.40

Aug-06

-98.6954

98.484

-0.210

1.20

Sept-06

-89.4201

114.194

24.774

-8.30

Oct-06 Nov-06

-100.648 -126.336

119.167 117.028

18.520 -9.306

-2.80 3.30

Dec-06

-124.908

127.125

2.217

3.90

limiting. Highest NEEopt occurred in July and August in 2005 (-16.19 and -16.41 lmol CO2 m-2 s-1, respectively) and in June 2006 (-21.26 lmol CO2 m-2 s-1), decreasing to minimum of -7.3 lmol CO2 m-2 s-1 in September 2005 and -5.4 lmol CO2 m-2 s-1 in August 2006 (Fig. 4a). Monthly NEEopt was explained by air vapor pressure deficit (VPD) (r2 = 0.22, P = 0.02). Highest NEEopt was reached at VPD between 1.0 and 1.5 kPa and reached minimum when air VPD exceeded 2.0 kPa. Apparent quantum yield (a), the maximum light use efficiency for C uptake (Fig. 4b) showed a seasonal trend and was explained by Ta (r2 = 0.47, P \ 0.01). Quantum yield reached minimum values at maximum Ta. The ecosystem dark respiration (Rd) (Fig. 4c), followed air temperature trend, similar to Re.

123

Discussion This tropical wet grassland assimilated a total of -167.8 gC m-2 over a 2-year period; however, the C uptake was relatively similar between 2 years with differing precipitation regimes (-86.5 gC m-2 year-1 in 2005 and -81.3 gC m-2 year-1in 2006). Net C fluxes from different wetlands around the world are highly variable, ranging from net C sources to a carbon sink greater than -500 gC m-2 year-1 in tropical wetlands. The large variation in C balance is driven by the ecosystem type, location, management history, and the approach used to estimate the annual carbon balance (Bonneville et al. 2008; Dusek et al. 2009; Dusek et al. 2012; Hao et al. 2011; Hirota et al. 2006; Lund et al. 2010; Mitsch et al. 2010; Sonnentag et al. 2010; Suzuki et al. 1999; Zhou et al. 2009). When compared with other ecosystems in the Cerrado, the wet grassland is in the range of measured and modeled NEE (Rocha et al. 2002; von Randow et al. 2013). In general, tropical wetlands can be more productive than those located at higher latitudes, however; respiration and decomposition rates are also higher in tropical wetlands driven by higher temperatures when compared with high latitude wetlands (Mitsch et al. 2010). For example, the Brazilian wet grassland in this study had an annual GEE above 1400 gC m-2 year-1 and ecosystem respiration represented about 94 % of gross carbon uptake, leading to an annual C balance below 90 gC m-2, while a temperate cattail wetland had a net yearly carbon uptake of[260 gC m-2 and a GEE of 831 gC m-2 year-1, but Re represented only 68 % of total carbon uptake (Bonneville et al. 2008). On a seasonal basis, the highest NEE occurred when Re, driven by low Ta, decreased and GEE was still high, indicating a higher sensitivity of Re to temperature than GEE. However, NEE in this wet grassland was very sensitive to changes in water availability; it shifted direction by more than 50 gC m-2 from first month to the next in 2005 and the ecosystem moved from a net carbon sink of -27.7 gC m-2 in August to a net carbon source of 25.6 gC m-2 in September (Table 1), and remained as a sustained carbon source during the four consecutive months as the water table stayed below 20 cm from September to November. Gross ecosystem carbon uptake decreased by more than 40 gC m-2 from August to September and continued at a low rate for the rest of the year, as grasses were probably under strong water stress since most of the roots were unable

Wetlands Ecol Manage 3

(a) 2 1

-2

-1

NEE (gC m d )

Fig. 2 Daily carbon fluxes (gC m-2 day-1) a net ecosystem carbon exchange (NEE), and b gross ecosystem carbon exchange (GEE, open circles), ecosystem respiration (Re, closed circles) measured on a Brazilian wet grassland for two consecutive years

0 -1 -2 -3 7

-7

(b)

5

-4

4

-3

3

-2

2

-2

-1 Jan/05

1 May/05

40

20

0

-2

-1

NEE (gC m month )

-1

-5

Re (gC m d )

6

Daily NEP

-2

-1

GEE (gC m d )

-6

-20

-40

-60 Jan/05 May/05 Sep/05 Jan/06 May/06 Sep/06 Jan/07

Fig. 3 Monthly net ecosystem carbon exchange (NEE, gC m-2 month-1) measured on a Brazilian wet grassland for two consecutive years

to reach the water-saturated soil. Grasses in general, and especially tropical grasses, develop more than 80 % of their root biomass in the top 25 cm soil (Jackson et al. 1996). This ecosystem was still a C

Sep/05

Jan/06

May/06

Sep/06

Jan/07

source at the end of 2005 even though the water table recovered in December, and was a small C sink in the following 4 months in 2006 as compared with 2005 (Table 1). It is likely that water stress in 2005 induced early senescence, which reduced green tissues that took few months to grow back in 2006. Moreover, Re increased with Ta but was also enhanced by increased substrate availability to microbial decomposition when top soil with high organic C was exposed as water table receded. Indeed, Re was higher during the months when the water table was below the 20 cm depth in 2005 than in 2006, when water table was near to or at the soil surface. The maximum ecosystem C uptake capacity when light is not limiting (NEEopt) was considerably reduced at highest air vapor pressure deficit, which coincided with periods of low precipitation; this may indicate stomatal control on C exchange at high vapor pressure. In general, the direction and magnitude of C fluxes in wetlands around the world are driven by two main

123

NEEopt (µmol CO2 m s )

Wetlands Ecol Manage

-2

-1

-4

hundreds, even thousands of years in rich organic soils (Mack et al. 2011). This ecosystem is already going under degradation and local disturbances, and changes in land use have reduced water table below the surface (Meirelles et al. 2004). Under this ongoing land use changes and the projected climate change scenarios of reduced precipitation and increased air temperature for the region (Christensen 2013; IPCC 2013), most of the C accumulated in soils of this wet grassland most likely will be fed back to the atmosphere.

(a)

-8 -12 -16 -20 -24 0.00

(b) -0.01 -0.02 -0.03

6

-2

-1

Rd (µmol CO2 m s )

-0.04

5

(c) Date vs alpha

4 3 2 1 0 Jan/05 May/05 Sep/05

Jan/06 May/06 Sep/06

Jan/07

Fig. 4 Monthly parameters of the light response curve derived from Eq. 1, a optimum net ecosystem exchange (NEEopt), b ecosystem apparent quantum yield (a, lmol CO2 m-2 s-1/ lmol quantum m-2 s-1), and c dark respiration (Rd) in a Brazilian wet grassland for two consecutive years

environmental drivers: the hydrologic and temperature regimes, and their interactions with biological variables and processes such as the size of the photosynthetic apparatus (leaf area index, LAI), photosynthetic rates, autotrophic respiration, and soil organic matter decomposition (SOM). The Brazilian wet grassland behaved as a C sink during the 2 years of measurements; however, this ecosystem is very sensitive to fluctuations in the hydric regime, which has an impact on the major ecosystem C fluxes (GEE and Re), affecting direction and magnitude of net C exchange. Environmental fluctuations like consecutive drought years or any change in land use like drainage or increased water use for agricultural or domestic purposes from this wetland will reduce the ecosystem capacity to take up carbon (GEE) by reducing the photosynthetic apparatus. At the same time, large amounts of old carbon are exposed to decomposition or fire and most likely, large amounts of carbon will be released to the atmosphere. A single fire event can release C accumulated over

123

Acknowledgments This study was supported by the Agricultural Technology Development Project for Brazil (PRODETAB), negotiated between the Brazilian government and the World Bank. The authors thank the computational support of Dr. Celso von Randow of the National Institute for Space Research (INPE-Brazil), and the fieldwork assistance provided by the technical experts Edim Borges Vieira, Nelson de Oliveira Pais, and Valdeci de Matos Lima of the Brazilian Agricultural Research Corporation (EMBRAPA). R. Bracho was supported by the U.S. Department of Energy Office of Science Award Number DE-FG02-99ER62848 while working on this manuscript. The authors thank Elizabeth Weeb for proofreading the manuscript.

References Aubinet M et al (2000) Estimates of the annual net carbon and water exchange of forests: the EUROFLUX methodology. Adv Ecol Res 30(30):113–175 Baldocchi DD (2003) Assessing the eddy covariance technique for evaluating carbon dioxide exchange rates of ecosystems: past, present and future. Glob Change Biol 9(4):479–492 Bernal B, Mitsch WJ (2008) A comparison of soil carbon pools and profiles in wetlands in Costa Rica and Ohio. Ecol Eng 34(4):311–323 Bonneville M-C, Strachan IB, Humphreys ER, Roulet NT (2008) Net ecosystem CO2 exchange in a temperate cattail marsh in relation to biophysical properties. Agric For Meteorol 148(1):69–81 Chimner RA, Cooper DJ (2003) Influence of water table levels on CO2 emissions in a Colorado subalpine fen: an in situ microcosm study. Soil Biol Biochem 35(3):345–351 Christensen JH, Krishna Kumar K, Aldrian E, An S-I, Cavalcanti IFA, de Castro M, Dong W, Goswami P, Hall A, Kanyanga JK, Kitoh A, Kossin J, Lau N-C, Renwick J, Stephenson DB, Xie S-P, Zhou T (2013) Climate phenomena and their relevance for future regional climate change. In: Stocker TF, Qin D, Plattner G-K, Tignor M, Allen SK, Boschung J, Nauels A, Xia Y, Bex V, Midgley PM (eds) Climate change 2013: the physical science basis. Contribution of working Group I to the fifth assessment report of the intergovernmental panel on climate change. Cambridge University Press, Cambridge Dusek J et al (2009) Influence of summer flood on the net ecosystem exchange of CO2 in a temperate sedge-grass marsh. Agric For Meteorol 149(9):1524–1530

Wetlands Ecol Manage Dusek J, Cizkova H, Stellner S, Czerny R, Kvet J (2012) Fluctuating water table affects gross ecosystem production and gross radiation use efficiency in a sedge-grass marsh. Hydrobiologia 692(1):57–66 Falge E et al (2001) Gap filling strategies for defensible annual sums of net ecosystem exchange. Agric For Meteorol 107(1):43–69 Ferreira EAB, Meirelles ML, Franco AC, Santos JLM (2009) Influeˆncia de variac¸o˜es ambientais sobre o fluxo de CO2 em ´ mido no Cerrado (Influence of solos sob Campo Limpo U environmental changes on the CO2 flow in wet grassland Cerrado soil). In: Faleiro FG, Farias Neto AL (eds) Menc¸o˜es honrosas: IX Simpo´sio Nacional do Cerrado e II Simpo´sio Internacional de Savanas Tropicais. Embrapa Cerrados, DF, Brasil, pp 238–242 Foken T, Wichura B (1996) Tools for quality assessment of surface-based flux measurements. Agric For Meteorol 78(1–2):83–105 Furley PA, Ratter JA (1988) Soil resources and plant-communities of the central Brazilian cerrado and their development. J Biogeogr 15(1):97–108 Goulden ML, Munger JW, Fan SM, Daube BC, Wofsy SC (1996) Exchange of carbon dioxide by a deciduous forest: response to interannual climate variability. Science 271(5255): 1576–1578 Hao YB et al (2011) Predominance of precipitation and temperature controls on ecosystem CO2 exchange in zoige alpine wetlands of southwest China. Wetlands 31(2):413–422 Hirota M et al (2006) Carbon dioxide dynamics and controls in a deep-water wetland on the Qinghai-Tibetan Plateau. Ecosystems 9(4):673–688 Hooijer A et al (2010) Current and future CO2 emissions from drained peatlands in Southeast Asia. Biogeosciences 7(5): 1505–1514 IPCC (2013) Annex I: atlas of global and regional climate projections [van Oldenborgh GJ, Collins M, Arblaster J, Christensen JH, Marotzke J, Power SB, Rummukainen M, Zhou T (eds)]. In: Climate change 2013: the physical science basis. Contribution of working Group I to the fifth assessment report of the intergovernmental panel on climate change [Stocker TF, Qin D, Plattner G-K, Tignor M, Allen SK, Boschung J, Nauels A, Xia Y, Bex V, Midgley PM (eds)]. Cambridge University Press, Cambridge, UK and New York, NY, USA Jackson RB et al (1996) A global analysis of root distributions for terrestrial biomes. Oecologia 108(3):389–411 Kansiime F, Saunders MJ, Loiselle SA (2007) Functioning and dynamics of wetland vegetation of Lake Victoria: an overview. Wetlands Ecol Manage 15(6):443–451 Kormann R, Meixner FX (2001) An analytical footprint model for non-neutral stratification. Bound-Layer Meteorol 99(2):207–224 Lehner B, Doll P (2004) Development and validation of a global database of lakes, reservoirs and wetlands. J Hydrol 296(1–4):1–22 Lloyd CR (2006) Annual carbon balance of a managed wetland meadow in the Somerset Levels, UK. Agricultural and Forest Meteorology 138(1–4):168–179 Lund M et al (2010) Variability in exchange of CO2 across 12 northern peatland and tundra sites. Glob Change Biol 16(9):2436–2448

Mack MC et al (2011) Carbon loss from an unprecedented Arctic tundra wildfire. Nature 475(7357):489–492 Meirelles ML, Guaimara˜es AJM, Oliveira RC, Glein MA, Ribeiro FF (2004) Impactos sobre o estrato herba´ceo de ´ reas U ´ midas do Cerrado (Impacts on the herbaceous layer A of the Cerrado Wetlands). In: Aguiar LM, Camargo AJA (eds) Cerrado: ecologia e caracterizac¸a˜o. Embrapa Cerrados, Brasilia, pp 41–68 Mitra S, Wassmann R, Vlek PLG (2005) An appraisal of global wetland area and its organic carbon stock. Curr Sci 88(1): 25–35 Mitsch WJ et al (2008) Tropical wetlands for climate change research, water quality management and conservation education on a university campus in Costa Rica. Ecol Eng 34(4):276–288 Mitsch WJ et al (2010) Tropical wetlands: seasonal hydrologic pulsing, carbon sequestration, and methane emissions. Wetlands Ecol Manag 18(5):573–586 Moncrieff JB et al (1997) A system to measure surface fluxes of momentum, sensible heat, water vapour and carbon dioxide. J Hydrol (Amsterdam) 188–189(1/4):589–611 Myers N, Mittermeier RA, Mittermeier CG, da Fonseca GAB, Kent J (2000) Biodiversity hotspots for conservation priorities. Nature 403(6772):853–858 Oliveira-Filho AT, Ratter JA (2002) Vegetation physiognomies and woody flora of the Cerrado biome. In: Oliveira PS, Marquis RJ (eds) The Cerrado of Brazil: ecology and natural history of a neotropical savanna. Columbia University Press, New York, pp 91–120. ISBN 0-23112043-5 Ovenden L (1990) Peat accumulation in northern wetlands. Quatern Res 33(3):377–386 Ribeiro JF, Walter BMT (1998) Fitofisionomias do Bioma Cerrado. In: Sano SM, Almeida SP (eds) Cerrado: ambiente e flora. Embrapa Cerrado, Brası´lia, pp 89–166 Rocha HRD et al (2002) Measurements of CO2 exchange over a woodland savanna (Cerrado Sensu stricto) in southeast Brasil. Biota Neotropica 2(1):1–11 Rodrigues Munhoz CB, Felfili JM (2008) Phytosociology of the herb-subshrub layer of a moist grassland community in Central Brazil. Acta Botanica Brasilica 22(4):905–913 Sano EE, Rosa R, Brito JLS, Ferreira LG (2008) Mapping the Cerrado’s vegetation cover. Boletim de Pesquisa e Desenvolvimento—Embrapa Cerrados (205):60 Saunders MJ, Jones MB, Kansiime F (2007) Carbon and water cycles in tropical papyrus wetlands. Wetlands Ecol Manag 15(6):489–498 Saunders MJ, Kansiime F, Jones MB (2014) Reviewing the carbon cycle dynamics and carbon sequestration potential of Cyperus papyrus L. wetlands in tropical Africa. Wetlands Ecol Manag 22(2):143–155 Sonnentag O, van der Kamp G, Barr AG, Chen JM (2010) On the relationship between water table depth and water vapor and carbon dioxide fluxes in a minerotrophic fen. Glob Change Biol 16(6):1762–1776 Suzuki S, Ishida T, Nagano T, Waijaroen S (1999) Influences of deforestation on carbon balance in a natural tropical peat swamp forest in Thailand. Environ Control Biol 37(2): 115–128 von Randow C et al (2013) Inter-annual variability of carbon and water fluxes in Amazonian forest, Cerrado and pasture

123

Wetlands Ecol Manage sites, as simulated by terrestrial biosphere models. Agric For Meteorol 182:145–155 Webb EK, Pearman GI, Leuning R (1980) Correction of flux measurements for density effects due to heat and watervapor transfer. Q J R Meteorol Soc 106(447):85–100

123

Zhou L, Zhou G, Jia Q (2009) Annual cycle of CO2 exchange over a reed (Phragmites australis) wetland in Northeast China. Aquat Bot 91(2):91–98