Wetland-atmosphere methane exchange in Northeast China: A comparison of permafrost peatland and freshwater wetlands

Agricultural and Forest Meteorology xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Agricultural and Forest Meteorology journal homepage: www.elsevier.com/locate/agrformet

Research Paper

Wetland-atmosphere methane exchange in Northeast China: A comparison of permafrost peatland and freshwater wetlands ⁎

Li Suna, Changchun Songa, , Peter M. Lafleurb, Yuqing Miaoa, Xianwei Wanga, Chao Gonga, Tianhua Qiaoa, Xueyang Yua, Wenwen Tana a b

Key Laboratory of Wetland Ecology and Environment, Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences, Changchun 130102, PR China Department of Geography, Trent University, Peterborough, Ontario, K9J 7B8, Canada

A R T I C L E I N F O

A B S T R A C T

Keywords: Methane exchange Peatland Marsh Permafrost Northeast China

Northeast China contains a large concentration of wetlands, primarily in two prominent types, freshwater marshes on the lowlands plains to the northeast and mountain permafrost peatlands in the north. Both wetlands types are threatened by disturbance, the marshes from agricultural conversion and the peatlands due to climate warming and loss of permafrost. Here we compare two seasons of ecosystem-scale CH4 fluxes measured via eddy covariance for a permafrost peatland in the Da Xing’anling Mountains and a freshwater marsh on the Sanjiang Plain. The objectives were to quantify CH4 fluxes, compare seasonal trends in the flux and determine the dominant environmental and biophysical drivers of the CH4 flux for these two distinct wetland types. CH4 fluxes at the marsh had a strong seasonal trend peaking in mid-summer, while the pattern for the peatland was muted. Maximum instantaneous fluxes were 1.34 μg CH4 m−2 s−1 and 9.5 μg CH4 m−2 s−1 at the peatland and marsh, respectively. Total seasonal CH4 emissions for the peatland, 0.38–1.27 g C-CH4 m−2, were an order of magnitude smaller than those at the marsh, 19.71–21.8 g C-CH4 m−2. Differences between years were small for both wetlands. We used path analysis to examine environmental and biophysical drivers of the flux and found that soil temperature (average soil temperature between 10 cm to 60 cm depths for the peatland and 10 cm depth for the marsh) was most strongly correlated with seasonal CH4 variability for both wetlands. Secondary influences were thaw depth for the peatland and net ecosystem CO2 exchange for the marsh. Given the temperature sensitivity of CH4 flux for both of these wetlands, future climate warming will likely increase CH4 emissions in northeast China, as well, the continued loss of permafrost in the mountain peatlands will likely further contribute to enhanced CH4 emissions.

1. Introduction Wetlands occupy only a small portion of the global terrestrial surface (5–8%), but play an important role in the global carbon cycle (Mitsch and Gosselink, 2007). Wetlands store more than 30% of the world’s soil carbon, with the majority contained in northern peatlands (Bridgham et al., 2006; Mitra et al., 2005; Mitsch et al., 2013; Zedler and Kercher, 2005). These diverse ecosystems sequester carbon dioxide (CO2) from the atmosphere, but are also identified as the largest natural source for atmospheric methane (CH4) (Forster et al., 2007; Lafleur, 2009; Limpens et al., 2008; Roulet, 2000). Because CH4 has a global warming potential 28 times of that of CO2 on a 100 year time horizon and contributes to over 20% of recent global warming (IPCC, 2013) even a modest change in methane sources can change the sign of the greenhouse gas budgets of wetlands. Wetland type (e.g., fen, bog,

⁎

marsh) greatly influences the relative importance of CH4 emissions, the magnitude and seasonal pattern of CH4 exchange varies greatly among wetlands depending upon temperature, vegetation cover and hydrology (Iwata et al., 2015; Koebsch et al., 2015; Nadeau et al., 2013; Sun et al., 2013). In addition, the presence/absence of permafrost in high-latitude wetlands affects CH4 flux (Olefeldt et al., 2013; Turetsky et al., 2014), suggesting that climate warming could greatly impact emission from these ecosystems. Hence, a better understanding of the dynamics of wetland CH4 emissions and the mechanistic controls on ecosystem scale CH4 exchange is urgently needed for improvement and validation of models used for predicting variations in atmospheric CH4 concentration and for the anticipation of ecosystem feedbacks to global change. There have been many studies of CH4 emissions from wetlands, much of this work has been outlined in extensive reviews (see Bartlett and Harriss, 1993; Bridgham et al., 2006, 2013; Lai, 2009; Turetsky

Corresponding author. E-mail addresses: [email protected] (L. Sun), [email protected] (C. Song).

https://doi.org/10.1016/j.agrformet.2017.11.009 Received 28 November 2016; Received in revised form 10 October 2017; Accepted 6 November 2017 0168-1923/ © 2017 Elsevier B.V. All rights reserved.

Please cite this article as: Sun, L., Agricultural and Forest Meteorology (2017), http://dx.doi.org/10.1016/j.agrformet.2017.11.009

Agricultural and Forest Meteorology xxx (xxxx) xxx–xxx

L. Sun et al.

extends approximately 1.1 km in east-west direction and 170–650 m in north-south direction (Fig. S2a). Although a branch of the Emuer River runs along the southern border of the peatland, its hydrology is mainly controlled by precipitation and there are no visible streams flowing into or out of the peatland. The peat depths varied from 50 to 100 cm. The surface has a hummock-hollow pattern, with mosses (Sphagnum sp., Aulacomnium androgynum, and Polytrichum juniperinum) covering the surface. The dominant vascular vegetation is shrubs (Betula fruticosa, Ledum palustre, Chamaedaphne calyculata, Vaccinium uliginosum, and Rhododendron parvifolium) with a sparse cover of sedge (Eriophorum vaginatum) and grass (Deyeuxia angustifolia). Average shrub height is 60–70 cm. Active layer thickness ranges from 50 to 70 cm in late summer and summer water table depth typically range between −10 and −36 cm. The freshwater marsh is located near the Sanjiang Experimental Station of Wetland Ecology on the Sanjiang Plain (Figs. S1 and S3a, 47°35′10.66″ N, 133°29′57.06″E, 55 m a.s.l.). Approximately 400 m long in the east-west direction and 300–400 m in north-south direction, the marsh is surrounded by shrubs and forests to the south and east and bordered by cropland to the north. The buildings of the Experimental Station are located about 400 m to the east. The marsh is a protected experimental site and has been preserved in its natural state since 1986. This eutrophic wetland is permanently inundated with a varying water depth (normally between 0 and 50 cm). It is a non-peat forming wetland, with a 20–40 cm root mass layer overlying 5–10 cm of humus and clay soil below. Vegetation at the site is dominated by the sedge Carex lasiocarpa and C. pseudocuraica. C. meyeriana and Glyceria spiculosa are also present. Mean canopy height during the growing season is typically about 0.7 m. The climate is temperate continental monsoon, with mean annual temperature of 2.5 °C and July mean temperature of 22 °C. The mean annual precipitation is 558 mm with approximately 80% occurring from May to September. The topography of the marsh is flat and precipitation is the main water source. There are no identified permanent inflows or outflows. The vegetation in the marsh begins to leaf-out in early May with the leaf area index (LAI) close to zero. LAI increases gradually and approaches the maximum of about 2.4 near the end of July or early August. The plants show visible signs of senescence in late August. Senescence accelerates in September and almost no green leaves remain by late September. Although not directly monitored, site visit observations suggest that vegetation phenology at the peatland is similar to that of the marsh. Accordingly, we define the growing season for both sites as May to September.

et al., 2014). While these syntheses tend to focus on the magnitude and variability of the CH4 flux in time and space and controls on the flux, it is also clear that most of the existing studies have been concentrated in North America and Europe, leaving some regions, such as Asia and the tropics, relatively understudied. China contains a considerable amount of wetland, ∼3.85 × 105 km2 representing about 4% of the surface area (Ma et al., 2012). Yet, only a few whole ecosystem studies of wetland methane flux have taken place in China. One of the greatest concentrations of natural wetlands in China is in the far northeast region (Niu et al., 2009). The dominant wetland types (excluding rice paddies) in this region are freshwater herbaceous wetlands (fens and marshes) with little or no peat accumulation and woody-herbaceous-moss peatlands with peat accumulations ranging from 30 cm to 9 m (Zhao, 1999). The fresh water systems are located toward the northeast and east of the region, the most notable concentration being on the Sanjiang Plain (Niu et al., 2009). Nine national wetland nature reserves, six of which are on the list of Ramsar wetlands of international importance, are on the Sanjiang Plain. Mountain peatlands are located toward the north and east and contain about 7% of the national peatland resources. The majority of these located in the mountainous permafrost zone south of the Russian boarder (Liu, 2005; Ma et al., 2013). Both of these wetland types are under threat. In recent decades marshes on the Sanjiang Plain have been reduced dramatically due to agricultural conversions, with total wetland loss of > 75% since the 1950’s (Dong et al., 2015; Wang et al., 2011). Meanwhile, regional warming after the 1970s has affected all of northeast China, but has particularly impacted the mountain peatlands where more than 35% of the permafrost area has been lost and temperature increases in the coming decades are projected to reduce permafrost area by an additional 28–50% (Chang et al., 2008; Jin et al., 2009). Therefore, understanding the impact of these types of changes on greenhouse gas emission of wetlands is of critical concern for regional and national assessments. In this study we compare two growing seasons of whole ecosystem CH4 fluxes from a permafrost peatland located in the Da Xing’anling Mountains and a freshwater marsh on the Sanjiang Plain. As noted above, these are the two most dominant natural wetland types in northeast China (Niu et al., 2009; Zhao, 1999). Although aspects of the C cycle for both of these wetlands have been studied previously (e.g., Meng et al., 2014; Miao et al., 2012; Song et al., 2009; Sun et al., 2013; Wang et al., 2013), no previous comparison of whole ecosystem methane fluxes from these wetland types has been conducted. The two wetlands differ significantly in their hydrological, soil and vegetation characteristics. Although climatic conditions also differ between the sites, where the more northerly peatland has somewhat cooler temperatures and less precipitation than the marsh, we expected that hydrologic and biophysical differences would be the primary factors regulating differences in wetland-atmosphere CH4 exchange. The overall objectives of the study are to compare whole ecosystem CH4 fluxes from these two wetlands as measured via eddy covariance, to clarify differences in CH4 exchange at diurnal to seasonal time scales, to determine the primary drivers of CH4 flux at each site and to quantify the total seasonal CH4 exchange at the two wetlands.

2.2. Measurements Flux measurement periods were from 1 May to early October in 2012 and 2013 at the marsh and 23 May to early October in 2012 and 8 May to early October in 2013 at the peatland. The closed-path eddy covariance technique was used to measure CH4, CO2 and H2O fluxes at both sites. The system included a three-dimensional ultrasonic anemometer (CSAT-3 Campbell, Scientific, USA), a fast greenhouse gas analyzer (FGGA, Los Gatos Research, Mountain View, CA, USA) and a dry vacuum scroll pump (XDS35i, BOC Edwards, Crawley, UK). Anemometers were installed on the masts at the height of 3.0 m and 2.5 m above the ground at the peatland and marsh, respectively. An inlet tube situated at the same height of the anemometer, with a separation of 15 cm, was used to draw air toward FGGA where CH4, CO2 and H2O concentrations were measured based on off-axis integrated cavity ringdown spectroscopy (Baer et al., 2002). All measurements were taken at a frequency of 10 Hz and the data were stored on dataloggers (CR3000, Campbell, Scientific, USA). The pump used at each site drew the sample air through a 7 m tube (inner diameter 6.4 mm, made of fluorinated ethylene propylene to minimize sorption or desorption) at flow rates about 40 L min−1 into the measuring cell at an operating pressure of approximately 19 kPa.

2. Material and methods 2.1. Study sites The permafrost peatland is located in the Da Xing’anling Mountains of Heilongjiang province (52°56′32.40″ N, 122°51′23.26″ E, 473 m a.s.l.) south of the Russian boarder (see Supplemental Materials, Fig. S1). It is within the Eurasian zone of continuous permafrost. Climate of the region is cold temperate monsoon, with a mean annual temperature of −3.9 °C, mean July temperature of 18.4 °C and mean annual precipitation is 452 mm. The site is pristine, with no disturbance from human activities. Surrounded by shrubs and forests, the peatland 2

Agricultural and Forest Meteorology xxx (xxxx) xxx–xxx

L. Sun et al.

The air was passed through a filter with a pore size of 10 μm to prevent dust and insects from entering the system and through two 2 μm Swagelok filters (one internal and one external) before entering the measuring cell. The 10 μm filters were replaced every 5–10 days and 2 μm external Swagelok filter was cleaned about once per month to maintain clean optics and avoid the restriction of inflow. We changed the bypass valve of each FGGA from internal to external placement on the instrument to facilitate manual adjustment of optimal cell operating pressure. Because the pump and the gas analyzer had a high power requirement, the EC systems at both sites ran on AC power supply during the measurement period. An uninterruptible power supply was added to each site to stabilize the starting voltage of the pump. Fetch around the EC mast at the peatland ranged from at least 150 m (parts of southwest and northwest) to approximately 700 m toward the east (Fig. S2). At the marsh, fetch around the EC mast ranged from at least 100 m (part of northwest) to approximately 400 m toward the southeast (Fig. S3).

with average data derived from compiling the 30 min ensemble average trend of five days prior to and following the gap. Before further analysis, we examined the availability of acceptable non-gap-filled halfhourly fluxes on a daily time scale and found that the number of days with more than 24 and 12 available half-hourly CH4 fluxes were greater than 66% and 91%, respectively, of the total growing season days at both sites. On the premise of assuring data quality, we chose to retain the greatest number of fluxes possible at the daily time scale to improve the validity of statistical inference. Therefore, days with more than 12 half-hourly CH4 fluxes were included in the following statistical analysis. 2.4. Micrometeorological and auxiliary measurements Micrometeorological variables were measured by the automatic weather stations located within 100 m and 150 m of the EC systems at the peatland and marsh site, respectively, and included photosynthetically active radiation (PAR), long- and short-wave radiation, barometric pressure (P), air temperature (Ta), soil temperature (Ts) at 10, 15, 20, 40 and 60 cm depth, relative humidity, and precipitation. With the exception of soil temperatures in the peatland which were recorded every 15 min, all the other meteorological data were recorded every 1 min by data loggers. Water level (WL) at the marsh was measured by pressure transducers (Odyssey, Dataflow Systems Pty Ltd., Christchurch, New Zealand) with a sampling frequency of 2 h. WL and thaw depth (TD) at the peatland site were measured manually one or two times per week during the study. WL was measured in five fixed wells located around the flux tower and these data were averaged for the proceeding analyses. Thaw depth was measured along 12 lines radiating out from the tower at approximately 30° intervals. Six measurements were made at distances between 50 and 100 m from the EC mast on each line and all data (72 measurements in total) were averaged for the analysis. Thaw depth was measured by inserting a steel rod into the soil until resistance was met and the distance from the bottom of the probe to the top of the moss layer was recorded. Long term meteorological data (air temperature and precipitation) were obtained from an automatic weather station located near the marsh site and the Mohe Meteorological Station about 20 km away from the peatland site.

2.3. Data processing Processing of high-frequency EC data was performed with EddyPro4.1 (LI-COR, Lincoln, Nebraska, USA). Raw data were filtered for spikes and linear detrending was used. Double coordinate rotations were performed to align the mean vertical velocity measurements normal to the mean wind streamlines before scalar flux calculations. Using the covariance maximization method (relative to the vertical velocity or temperature), time lag was determined for each half-hourly period. Half-hourly fluxes of CH4, CO2 and H2O were calculated as the mean covariance of vertical wind velocity and scalar fluctuations. The WPL correction for density fluctuations arising from variations in water vapor was applied according to Ibrom et al. (2007b). The low-pass filtering effects were assessed and corrected using the method of Ibrom et al. (2007a) based on in situ determination of water vapor attenuation and on a model for the corresponding spectral correction factor. Quality control criteria according to Mauder and Foken (2004) were used to reject bad data. Additionally, data were excluded when the pump stopped working due to maintenance or high temperature in summer and when the sonic anemometer signal was degraded by heavy rain. The data collected during weak turbulence were removed from further analyses by filtering all half hour fluxes with friction velocity (u*) below 0.1 m s−1. The threshold was determined by analyzing the relationship between u* and CH4 flux during the growing seasons (Long et al., 2010). The footprint contribution for each half-hourly flux was examined by using the one dimensional (crosswind-integrated) estimation method of Kljun et al. (2004). In the micrometeorological conditions (u* < 0.2 m s−2, stability parameter zm/L < −200 or zm/L > 1; zm, measurement height and L, Obukhov length) when the results of Kljun et al. (2004) may not be approximate, the model of Kormann and Meixner (2001) was used. The peak contributing distances were on average 20 m from the EC mast at the peatland and 24 m at the marsh. To ensure that 90% fetch were within the wetland boundary, careful field investigations and footprint analysis were conducted. 4.8% and 9.3% of the data from peatland and marsh, respectively, were excluded due to the footprints extending beyond the area of interest. CH4 flux was calculated by adding the rate of CH4 storage change (S) to the turbulent flux. S was estimated from the changes in the average CH4 concentrations at the sensor height over the 30-min intervals assuming that the CH4 concentrations were representative for the entire air column below the sensors. In this study, positive CH4 and CO2 fluxes indicate net flux from the ecosystem to the atmosphere and negative from the atmosphere to the ecosystem. After all filtering and data exclusion 39% and 37% of the data were removed at the peatland and marsh, respectively. Since clear relationships between half-hourly CH4 fluxes and other environmental or biophysical variables were not found for the peatland, data gaps less than 2.5 h were filled by linear interpolation and longer gaps were filled

2.5. Analysis Daily values of meteorological variables were calculated based on the 1 min observations for each day. Daily thaw depth and water level data of the peatland were derived by linearly interpolating between the weekly measurements, with constant values assigned for the day. CH4 flux data were divided into three separate time periods based on the stage of plant growth and general physical condition of the wetlands: early growing season (early May–June 15), peak growing season (June 16–August 20) and late growing season (August 21 to the end of September). Diurnal variations of CH4 flux were analyzed during the three periods. To investigate the regulation of CH4 at seasonal time scale, we examined the relationship between daily CH4 and environmental and biophysical factors by stepwise regression where a variance inflation factor (VIF) greater than 5 was used to indicate severe potential multicollinearity problems (Rogerson, 2001; Ström and Christensen, 2007). We select Ta, Ts (at 10, 15, 20, 40 and 60 cm depths and the average of all depths), PAR, u*, P, TD (only at the peatland site) and WL as the targeted environmental variables and net ecosystem CO2 exchange (NEE) as the biophysical variable. The stepwise regression was used for preliminary screening of variables. CH4 flux data were logarithmic transformed to reduce the variation of the original data and stabilize the variance of errors before the regression analysis. Then the selected variables from stepwise regression were used to construct the hypothesized model using path analysis to examine the relative importance 3

Agricultural and Forest Meteorology xxx (xxxx) xxx–xxx

L. Sun et al.

precipitation in September (179.7 mm). Growing season precipitation in 2013 was less (495.6 mm), but still greater than the long-term average. Since precipitation was the main water source for the two wetlands, water level generally increased with rainfall events and decreased by evapotranspiration during fair weather periods (Fig. 1e and f). Water level for the peatland was always below the surface with averages of −18.5 cm and −14.1 cm in 2012 and 2013, respectively. In contrast, water level was always above the surface at the marsh and was on average 8 cm higher in 2012 than in 2013, corresponding to greater precipitation in 2012. The soil at the peatland began to thaw at the end of April. Thaw depth increased gradually from about 10 cm at the beginning of May and approached maximum in mid-September (Fig. 1e). Maximum thaw depth was 71.5 cm in 2012 and 68.5 cm in 2013.

of variables controlling CH4 exchange. Path analysis is an extension of multiple regression (Bassow and Bazzaz, 1998; Iwata et al., 2015; Schemske and Horvitz, 1988). It is used to disassemble the correlation coefficient between dependent and independent variables and provide estimates of the magnitude and significance of hypothesized connections among variables. It is particular useful when independent variables have a causal or correlated relationship. To estimate the path coefficients for the effects of each independent variable on CH4 fluxes, all the variables were first standardised by subtracting the mean of each variable and dividing by the standard deviation of each variable (Schemske and Horvitz, 1988). The adequacy of the model by path analysis was determined using a Goodness-of-fit Index (GFI). When the GFI was greater than 0.9, the model was considered adequate. The stepwise regression and path analysis were performed by SPSS 18.0 software (SPSS Inc., Chicago II, USA) and Amos 17.0 (Analysis of Moment Structure, SPSS Inc., Chicago, IL, USA), respectively.

3.2. Diurnal variations in CH4 fluxes Short periods (five to seven days) of consecutive sunny days were selected in each of the different periods of the growing seasons to examine the diurnal cycles of CH4 fluxes (only non-gap-filled data were used for this analysis). CH4 exchange at the peatland showed no discernible daily pattern during any period for either of the two growing seasons (Figs. S4a and S4b). CH4 fluxes at the marsh also did not show a discernible diurnal variation except for the peak growing season of 2013 when a weak diurnal pattern was observed with average daytime emission 12% higher than that of the nighttime (Fig. S4d). Nighttime data generally showed relatively larger variability than the daytime data at both sites because fewer data were available because of high rejection rates from u* and footprint filters.

3. Results 3.1. Environmental conditions There was little difference in growing season temperature between years at either site (Fig. 1a and b). For the peatland, the average air temperature was 0.6 °C and 0.5 °C higher, in 2012 and 2013 respectively, than the long-term (1981–2013) average of 13.3 °C. Soil temperature at 10 cm depth (Ts10) was less variable, but followed the seasonal trend as Ta (Fig. 1a). Mean Ts10 for the peatland was 6.0 °C in 2012 and 5.5 °C in 2013. For the marsh, the average air temperature was 0.7 °C and 0.8 °C higher, in 2012 and 2013 respectively, than the long-term (1991–2013) average of 17.6 °C. Ts10 followed a similar pattern Ta, but averaged 4.5–4.8 °C lower (Fig. 1b). Precipitation during the two growing seasons tended to be greater than the long-term average for both sites (Fig. 1c and d). At the peatland, total growing season precipitation in 2012 (449.5 mm) and 2013 (536.5 mm) was 83.6 mm and 170.6 mm higher, respectively, than the long-term mean. The relatively high precipitation in 2013 was mainly caused by high precipitation in May (115.6 mm), which was 81.2 mm (236%) higher than the long-term average of 34.4 mm. At the marsh, growing season precipitation in 2012 was 135.8 mm higher than the long-term average of 460.9 mm, mainly caused by anomalously high

3.3. Seasonal variability in CH4 fluxes There were distinct differences in seasonal trends and magnitude in CH4 flux between the two wetlands (Fig. 2). Overall, instantaneous CH4 fluxes were an order of magnitude smaller at the peatland site, ranging from −0.5 to 0.87 μg CH4 m−2 s−1 in 2012 and from −0.3 to 1.34 μg CH4 m−2 s−1 in 2013. There was little seasonal trend in 2012, but in 2013 maximum hourly fluxes increased throughout the study period until mid-September, then decreased modestly (Fig. 2a and b). At the marsh, instantaneous CH4 fluxes ranged from uptake of −1.3 μg CH4

Fig. 1. Time series of the daily micrometeorological variables at the peatland and marsh during the 2012 and 2013 growing seasons, including (a, b) air temperature (Ta, grey empty circles), soil temperature at 10 cm depth (Ts10, grey filled circles) and long-term (1981–2013 at the peatland and 1991–2013 at the marsh) average air temperature (black lines), (c, d) precipitation (black bars), cumulative precipitation (black lines) and long-term average of cumulative precipitation (grey lines) and (e, f) water level (black lines) and thaw depth (grey lines).

4

Agricultural and Forest Meteorology xxx (xxxx) xxx–xxx

L. Sun et al.

Fig. 2. Time series of half-hourly and daily methane fluxes at the peatland (a, b) and marsh (c, d) sites during the two growing seasons. Half-hourly data were non-gap filled and daily data were integrated from gap-filled half-hourly data. Note the difference in y-axis scales between the upper and lower panels.

m−2 s−1 to emission of 9.5 μg CH4 m−2 s−1 in both years. Fluxes were small at the beginning and end of the measurement period and reached a broad peak during June, July and August (Fig. 2c and d). Daily CH4 fluxes at the two wetland sites more clearly showed the seasonal patterns. At the peatland, fluxes in May and most of June were small (< 4.0 mg C-CH4 m−2 d−1) when soil temperature was low and thaw depth < 23.5 cm (Figs. 2 a, b, 1 a, and e), but fluxes increased in late-June and July as soil temperature and thaw depth increased. In 2012 this step increase resulted in a flat response, with daily maximum fluxes plateauing near 6.0 mg C-CH4 m−2 d−1. The response was more dramatic in 2013, when daily maximum flux increased steadily and reached a maximum of 18.1 mg C-CH4 m−2 d−1 at the end of August and then declined steadily thereafter (Fig. 2a and b). CH4 emissions at the marsh showed much stronger seasonality and followed closely the seasonal dynamics of Ts10 (Figs. 2 c, d and 1 b). CH4 fluxes increased rapidly following the increase in soil temperature in the early stage of the growing season and peaked around 218.3 mg C-CH4 m−2 d−1 on 24 June in 2012 and 278.2 mg C-CH4 m−2 d−1 on 26 June in 2013. Fluxes declined in the latter part of the measurement period as soil temperatures decreased. Seasonal cumulative CH4 emission at the peatland was 0.38 g C-CH4 m−2 and 1.27 g C-CH4 m−2 in 2012 and 2013, respectively (gap-filled data were used to calculate these totals). In contrast, cumulative CH4 emissions at the marsh were much larger, totalling 19.7 g C-CH4 m−2 and 21.8 g C-CH4 m−2 in 2012 and 2013, respectively.

Table 1 Variable selection and collinearity statistics by stepwise regression results between CH4 fluxes and the environmental and biophysical variables. Peatland

2012

2013

p

Variables entered

VIF

Tsaverage TD

2.19 2.19

*

2.55 3.17 2.43 1.95

*

Tsaverage TD WL Ta

*

* *

0.02

Marsh

p

Variables entered

VIF

Ts10 NEE PAR

1.12 2.78 2.68

*

Ts10 NEE PAR WL

1.64 2.55 1.60 1.45

*

* *

* * *

Tsaverage, average soil temperature at depths of 10, 15, 20, 40 and 60 cm; TD, thaw depth; WL, water level; Ta, air temperature; Ts10, soil temperature at 10 cm depth; NEE, net ecosystem CO2 exchange; PAR, photosynthetically active radiation; VIF, variance inflation factor. * Stands for significant at p < 0.01.

was explored through the path analysis (Figs. 3 & 4). For the peatland ecosystem, Tsaverage (path coefficient 0.60–0.65, p < 0.01) and TD (path coefficient 0.52–0.55, p < 0.01) were the dominant factors directly related to CH4 during the growing seasons (Fig. 3). Tsaverage was also indirectly related to CH4 through TD. The direct effect from WL was either not significant (p = 0.36 in 2012) or weak (path coefficient 0.1 in 2013, p < 0.01) and the indirect effects from WL on CH4 were slightly higher (0.12 in 2012 and 0.18 in 2013). The total indirect effects from Ta were negative (−0.05 in 2012 and −0.16 in 2013), which could be attributed mainly to the hysteresis response between TD and Ta at the seasonal time scale. For the marsh ecosystem, path analysis showed the dominant variables directly related to CH4 variability were Ts10 (path coefficient 0.73–0.81, p < 0.01) and NEE (path coefficient −0.40 to −0.44, p < 0.01) (Fig. 4). Since negative NEE indicates a net flux from the atmosphere to the marsh, CH4 emission increased with higher carbon dioxide uptake. The direct effect from WL was marginal (path coefficient −0.10 in 2012, p = 0.07) or weak (path coefficients −0.12 in 2013, p < 0.01). WL had a negative indirect effect on CH4 emission through Ts10, probably because the standing water in the marsh buffered short-term impacts of the ambient temperature variations, as well as the relatively low air temperatures occurring on rainy days when WL increased. The effects of PAR were complex. In both years PAR had a significant direct effect (−0.30 and −0.15) on CH4 flux, possibly

3.4. Controls on CH4 exchange Stepwise regression analysis based on daily values showed that of all the environmental and biophysical variables Tsaverage (average soil temperature at 10, 15, 20, 40 and 60 cm depths), TD, WL and Ta were significantly related to the season fluctuation of CH4 exchange at the peatland (Table 1). There was no severe multicollinearity (VIF < 3) for most variables. The selection of Tsaverage over any single depth soil temperature was confirmed by comparing regression statistics, coefficient of determination (R2) and root mean square error percentage (RMSE%), where in 2013, R2 and RMSE% using Tsaverage were 0.63 and 0.42, respectively, while R2 and RMSE% using single depth Ts varied from 0.43–0.59 and 0.44–0.6, respectively. Statistics for 2012 were similar. In contrast, Ts10, NEE, PAR and WL were the main variables selected for the marsh (Table 1). In this case, Ts10 outperformed all other individual Ts depths and Tsaverage. The relative causality of variables influencing CH4 flux variability 5

Agricultural and Forest Meteorology xxx (xxxx) xxx–xxx

L. Sun et al.

and modelled values from the exponential function shown in Fig. 5a. After the influence of soil temperature was removed in the period before the peak of Tsaverage, TDre only explained 5% to 7% of the CH4re variability (Fig. 5b), which reaffirmed Ts as the dominant control on CH4 flux during this period. However, in the remaining study period, i.e., after Tsaverage peaked around mid-August, TD explained 67%–68% of variability in CH4re daily flux (Fig. 5c), highlighting the importance of TD on CH4 in the late growing season. For the marsh ecosystem, Ts10 explained the greatest amount of the variability in CH4 flux. Non-linear regression produced similar exponential functions for both growing seasons, with explained variance in daily CH4 flux > 85% (Fig. 6a). NEE was the second most important factor, but was also influenced by temperature. To account for the influence of temperature on the NEE-CH4 relationship, a linear regression between NEE and Ta was conducted (not shown). The residuals of NEE and CH4 (NEEre and CH4re respectively) were calculated as the difference between measured and modelled values from the temperature relationships. We regressed CH4re on NEEre and found that NEEre explained 15% to 17% of the variation in CH4re (Fig. 6b). This relationship suggested that plant activity played an important secondary role in CH4 emissions at this site.

Fig. 3. Path diagrams for the peatland illustrating the effects of environmental parameters on CH4 flux during the 2012 (top) and 2013 (bottom) growing seasons. Environmental variables include: air temperature (Ta), average soil temperature at different depths (Tsaverage), thaw depth (TD) and water level (WL). Solid arrows represent positive correlations, and dashed ones are negative correlations. Numbers beside the arrows are standardized path coefficients (* indicates p < 0.01). Direct effects are indicated by the path coefficients. Indirect effects (not shown) are computed as the product of path coefficients linking an environmental variable to CH4 flux. All data used in the analyses were daily mean values. Total effects are the sum of indirect and direct effects.

4. Discussion 4.1. Dynamics of CH4 exchange We have shown that there are significant differences in magnitude and seasonal pattern of CH4 emissions from these two wetlands. These results support findings, as previous reviews have highlighted such differences between wetland types (Bartlett and Harriss, 1993; Mitsch et al., 2013; Turetsky et al., 2014), where wetland type with its attendant vegetation, hydrology and chemistry, influences the production, oxidation and transport of CH4 (Bridgham et al., 2013; Whalen, 2005). Climate can also be a significant factor. However, in our study its effect is likely to be small. Considering air temperature, although mean MaySeptember temperatures at the marsh were ∼4 °C higher than at the peatland, which was close to the long-term difference, the seasonal course of air temperature and period of non-freezing temperatures were similar at the two sites (Fig. 1a and b). In addition, the influences of air temperature on CH4 flux were indirect, as shown in our path analysis (Figs. 3 and 4), and CH4 flux response to air temperature (Fig. S5) showed highly divergent patterns at the two sites, even though the range of temperatures was similar. Hence, we conclude that the differences here are largely attributable to ecosystem characteristics. In comparing these two principal wetland types of northeast China, likely the greatest single factor influencing CH4 flux is water table depth, on average ∼16 cm below the surface at the peatland and on average ∼53 cm above the surface at the marsh. Water table depth controls the level of anoxia, which promotes methanogenesis, and the depth of the oxic zone where diffusing CH4 can be oxidized before escaping to the atmosphere (Frenzel and Karofeld, 2000; Kettunen et al., 1999). It also controls the pathway by which CH4 produced in the soil enters the atmosphere, directly by permitting gas bubble transport in saturated environments and indirectly by influencing plant species composition and the dominance of aerenchymateous plants that transport CH4 directly from the soil to atmosphere (Lai, 2009; Le Mer and Roger, 2001). The large difference in total seasonal CH4 emissions found here (16 ∼ 51 times higher for the marsh compared to the peatland) is consistent with previous observations using eddy covariance. The very small seasonal emission for the peatland (0.38–1.27 g C-CH4 m−2) is comparable to those for other peatland studies, for example, Rinne et al. (2007) measured 8.55 g C-CH4 m−2 during the snow free period in a boreal fen in southern Finland, Euskirchen et al. (2014) measured 3.74 g C-CH4 m−2 from April to September at a collapsed scar bog in Alaska, Long et al. (2010) measured 2.4 g C-CH4 m−2 during May-September in a wooded northern peatland in Canada, Nadeau

Fig. 4. Path diagrams for the marsh. Environmental variables include: net ecosystem CO2 exchange (NEE), average soil temperature at 10 cm depths (Ts10), photosynthetic active radiation (PAR) and water level (WL).Other details as in Fig. 3.

indicating photoinibition at high light reducing plant transport of CH4. Yet, the indirect effects (through NEE) offset this direct effects resulting in relatively weak total effects from PAR (−0.05 in 2012 and 0.03 in 2013). To further clarify the effects of the dominant factors influencing CH4 fluxes, we examined CH4 fluxes in relation to changes in the environmental and biophysical factors selected in the path analysis. The highest proportion of variability in CH4 flux was attributed to an exponential relationship between CH4 flux and Tsaverage at the peatland and Ts10 at the marsh (Figs. 5 a and 6 a). For the peatland ecosystem, the change of TD was highly correlated with the change in Tsaverage before Tsaverage reached its maximum around mid-August (Fig. 1a and e). In order to remove this bias, we conducted a linear regression between Tsaverage and TD before Tsaverage peaked and calculated the residuals of TD as the difference between measured and modelled values (TDre). The residuals of CH4 (CH4re) were also calculated as the difference between measured 6

Agricultural and Forest Meteorology xxx (xxxx) xxx–xxx

L. Sun et al.

Fig. 5. The relationships between CH4 flux and temperature for the permafrost peatland. (a) daily CH4 fluxes and average soil temperature (Tsaverage = mean of 10, 15, 20, 40 and 60 cm depths), (b) the residuals of the daily CH4 flux vs Tsaverage and the thaw depth vs Tsaverage before average soil temperature reached maximum, and (c) the residuals of daily CH4 flux vs Tsaverage and thaw depth after average soil temperature reached maximum during the two growing seasons. Daily CH4 fluxes were calculated based on data without gap filling and included a minimum of 12 half hours of data each day.

et al. (2013) measured 3.3 g C-CH4 m−2 from June to September at a bog in northern Canada and Brown et al. (2014) measured 1.63–3.10 g C-CH4 m−2 from May to September at an ombrotrophic bog in southern Canada. On the other hand, total seasonal CH4 emissions at Sanjiang marsh (19.7–21.8 g C-CH4 m−2) were in the range of seasonal totals reported from inundated or seasonally inundated wetlands (marshes and fens) by other studies, e.g., 10.7–14.4 g C-CH4 m−2 at an alpine wetland on the Tibetan Plateau (Song et al., 2015), 21.7 g C-CH4 m−2 at a fen dominated by Carex larsiocarpa in northern Minnesota (Chasar et al., 2000), 37.1–49.2 g C-CH4 m−2 from a coastal marsh dominated by macrophytes in northwestern Ohio (Chu et al., 2014) and 30 g C-CH4 m−2 at a fen in north eastern Germany (Koebsch et al., 2015). Diel pattern of fluxes were largely non-existent for these wetlands, except the marsh exhibited a weak diurnal cycle in the peak growing period of 2013 (Fig. S4d). The lack of a diurnal trend in CH4 flux has been noted for other peatlands (Nadeau et al., 2013; Pypker et al., 2013; Rinne et al., 2007). At our peatland site, the absence of a diurnal cycle may, in part, be due to the small magnitude of the fluxes and relatively large uncertainties in half-hourly data obscuring the true signal. In addition, the processes of CH4 production below the water table and consumption above the water table may be responding to

different drivers that have asynchronous diurnal patterns (Whalen, 2005). Finally, the strong dependence of daily CH4 flux on average soil temperature (10–60 cm) rather than surface temperature may indicate the importance of CH4 production from deep soil layers where diurnal variation of soil temperature was not significant. The weak diel trend observed at the marsh site during peak growing season, but not during other time periods, is a common feature in previous studies. Indeed, several authors have observed a diel pattern during some periods of the growing season and not others (Long et al., 2010; Suyker et al., 1996), while Koebsch et al. (2015) found that the diurnal cycle (i.e., peak efflux) shifted during the course of the growing season. Inconsistency in the diurnal trend during the growing season has been attributed to shifting, asynchronous controls on the production and transport of CH4, such as plant transport increasing with PAR (Whiting and Chanton, 1996), deposition of organic substrates for CH4 production coupled to plant productivity (Hatala et al., 2012), thermal stability of the water column reducing diffusion of CH4 (Koebsch et al., 2015), or contrasting lags between environmental drivers and CH4 flux (Sturtevant et al., 2016). We observed that soil temperature in the marsh lagged behind radiation and air temperature by several hours during the peak growing period, therefore, the average nighttime Ts10 (16.5 °C) was higher than 7

Agricultural and Forest Meteorology xxx (xxxx) xxx–xxx

L. Sun et al.

Fig. 6. Examination of environmental and biophysical controls on CH4 flux at the marsh site. (a) daily CH4 fluxes and soil temperature at 10 cm depth (Ts10) and (b) the residuals of daily CH4 fluxes vs Ts10 and daily NEE vs air temperature during the two growing seasons. Daily CH4 and NEE fluxes were calculated based on data without gap filling and included a minimum of 12 half hours of data each day.

and positive control over wetland CH4 emission (Iwata et al., 2015; Lai et al., 2014; Long et al., 2010; Koebsch et al., 2015; Song et al., 2015; Rinne et al., 2007; Turetsky et al., 2014). The high R2 and similar temperature sensitivity for both years at the marsh suggests that this control does not vary greatly between years (Fig. 6a). At the permafrost peatland the average soil temperature of the soil layers from 10 to 60 cm depths gave the strongest relationship with daily CH4 flux (Fig. 5a), which to the best of our knowledge has not been reported in other studies. We surmise that this average temperature better represents CH4 production instead of any single layer soil temperature because it contains information from below the water table where production is occurring and from above the water table where the greatest daily fluctuations in temperature are occurring. The important secondary variables related to CH4 flux at these sites were thaw depth at the peatland and NEE at the marsh. This difference was not surprising given the soil and vegetation differences between sties. High productivity of the aerenchymous sedge plants which dominate the marsh facilitates plant transport of CH4 from the zone of production directly to the atmosphere (Joabsson et al., 1999) and increases the belowground supply of organic substrate for methanogenesis (Chanton and Whiting, 1996; Hatala et al., 2012; Whiting et al., 1991). Previous studies have suggested that plant productivity is a good predictor of CH4 fluxes either across a wide range of wetlands (Whiting and Chanton, 1993) or within a single species (Chu et al., 2014; Lai et al., 2014; Song et al., 2015; Whiting et al., 1991; Whiting and Chanton, 1992). NEE was not an important correlate of CH4 flux at the permafrost peatland, even though some plants within the peatland have aerenchyma they do not dominate the vegetation community and the water table depth was often below their rooting zone, suggesting that this mode of transport would be negatively impacted. The relationship between CH4 flux and thaw depth at the peatland was puzzling. While thaw depth has been used to explain spatial variations in CH4 flux (Iwata et al., 2015; Kim 2015; Vourlitis et al., 1993), those studies that have examined the temporal patterns of CH4 flux and thaw depth have found a strong positive relationship only in the early to peak growing

the average daytime value (16.1 °C). As Sturtevant et al. (2016) pointed out, the underlying mechanisms of diurnal CH4 fluxes in wetlands is still poorly understood, but has important implications for developing process models of CH4 flux. These two wetlands showed considerable difference in seasonal pattern of the CH4 flux (Fig. 2). Although it is difficult to isolate the exact causal factors, the early-July increase in daily CH4 fluxes at the peatland corresponded to the start of the period when increases in thaw depth were most rapid during the study period (Fig. 1e). Soil temperatures increased throughout the summer and peaked in mid- to lateAugust corresponding to the late season peak in CH4 flux in 2013. We did not see the same peak in 2012, likely as a result of limited precipitation and a sharp draw-down of the water table in August 2012, which may have promoted increased oxidation of CH4 (Bubier et al., 1993). In contrast, the more distinctive trend in daily CH4 flux at the marsh, peaking about two month earlier (late-June) than at the peatland and closely following the trends in soil/air temperatures (Figs. 2 c, d and 1 b), is indicative of the contrasting CH4 transport mechanisms at this site. Peak temperatures favor transport by diffusion and ebullition (Waddington et al., 2009; Noyce et al., 2014). In addition, this pattern, similar to CH4 fluxes measured by chamber technique at the site, is strongly associated with the growth and maximum LAI of the dominant sedge in the marsh (Sun et al., 2013). Hence, plant associated transport also likely contributed to the strong seasonal trend.

4.2. Influence of environmental and biophysical factors Although key factors and processes that control wetland CH4 emission often differ among wetlands, soil temperature, water level and plant phenology are generally identified as the primary drivers of CH4 exchange (Hendriks et al., 2010; Long et al., 2010; Olson et al., 2013) because they represent, to a large extent, the main controls on CH4 production, oxidation, and transport (Le Mer and Roger, 2001). Our results (Figs. 5 a and 6 a) agree with many others from a range of wetland ecosystems that have found soil temperature as the most direct 8

Agricultural and Forest Meteorology xxx (xxxx) xxx–xxx

L. Sun et al.

spatial variations in wetland CH4 emissions (Sun et al., 2013), 2) differences caused by different estimation techniques and methods, and 3) the high sensitivity of CH4 emission from rice paddies to temperature change (van Groenigen et al., 2013). Undoubtedly more long term ecosystem scale CH4 measurements from natural wetlands and rice paddies are needed to reduce the uncertainty. Although land use pressure on the mountain permafrost peatlands of northeast China is not as great as on the Sanjiang Plain marshes, climate warming is having a significant impact on regional permafrost. Since the 1970’s the permafrost area in northeast China has decreased by more than 35% and the southern border of permafrost has moved northward 50–120 km (Jin et al., 2009). Deepening of the active layer and complete loss of permafrost could dramatically influence the GHG budget in these peatlands. Our results suggest that soil warming and deepening of the active layer will increase CH4 emission (Fig. 5a and c). This view is supported by incubation studies that show 1) CH4 content of the deep soil cores (50–100 cm) from this peatland was significantly higher than those from the upper layer of 30–40 cm depth (Wang et al., 2013) and 2) that permafrost soil have higher decomposability than the active layer soil and will emit more CH4 when warmed (Song et al., 2014). Further, complete loss of permafrost can create thermokarst features; these saturated, carbon-rich environments are noted for high emissions of CH4, many times larger than the intact permafrost peatlands (Christensen et al., 2004; Matveev et al., 2016). Thus, continued warming in the mountainous area of northeast China will likely release more CH4 to the atmosphere from these peatlands and monitoring these changes will be important for tracking China’s GHG budget in the coming century.

period (Heyer et al., 2002; Sturtevant et al., 2012; Tagesson et al., 2012). We found the opposite, the influence of thaw depth on CH4 flux was weak in the early- and mid-growing periods (active layer < 60 cm), but a strong positive influence was found in the late growing period when soils were warm and actively > 60 cm (Fig. 5c). One difference is that the previous studies occurred in saturated or near-saturated environments (a fen and drained lake beds, respectively). In our peatland, as suggested by Iwata et al. (2015), this relationship may indicate the growth of methanogenic population in the vertical extent, which is regulated by soil thaw, thus it is only in the late season that methanogen activity is at its maximum and increases the base CH4 emission rate. Although water table depth was the most important factor leading to the significant differences in CH4 emissions between these two wetlands, variations in water level had little direct influence on the seasonal CH4 flux at either site (Figs. 3 & 4). At marsh site, WL was above surface at all times. Others have shown that water table variations had little or no relationship to CH4 flux in continually flooded ecosystems (Chu et al., 2014; Song et al., 2011; Strachanm et al., 2015; Sturtevant et al., 2016; Sun et al., 2013). Although there may be some influence by depth of standing water on bubble pulses of CH4 (Sturtevant et al., 2016), continuous inundation eliminates the dependence of CH4 flux on the variation in the oxidation pathway when water table is below the surface. The absence of a direct relationship between water table and CH4 flux has also been reported for some northern peatlands and was attributed to 1) small seasonal variation in the water table depth (Christensen et al., 2003; Jackowicz-Korczyński et al., 2010; Sachs et al., 2008) or 2) the contrasting seasonal patterns of temperature and water table (Rinne et al., 2007). The latter effect is most likely the reason for the lack of a correlation between WL and CH4 flux for the peatland in this study. It should be noted that, path analysis indicated a small but significant indirect influence of water level (i.e., indirect path coefficient 0.12–0.18 for the peatland, Fig. 3). Although WL was always below the peatland surface (−36.5 to −3.0 cm), a higher water table could indirectly promote CH4 emission by its positive effect on thaw depth through enhanced soil water content and thermal conductivity in this soil (Brown, 1963).

5. Conclusions This is the first detailed comparison of ecosystem-scale CH4 exchange in a permafrost peatland and a freshwater marsh, representing the two main typical wetland ecosystems of northeast China. For the peatland ecosystem, average soil temperature of different depths acted as the dominant control of the seasonal CH4 variability with the influence of thaw depth significantly intensified during the late growing season, while for the marsh ecosystem, soil temperature at 10 cm depth and NEE were the main controlling factors. Total seasonal CH4 emission from the peatland was an order of magnitude smaller compared with that of the marsh. The large differences in the seasonal total emissions from different wetlands suggest the need for more long-term measurements of CH4 exchange over different wetland types. Moreover, the different responses to changes in environmental and biophysical conditions should be considered when estimating CH4 exchange from different wetland types, with particular emphasis on climate warming and land cover change response to better predict future CH4 feedback from northern wetlands. Our results also provide an important data basis for quantifying and modelling the regional CH4 budget from China’s northern wetlands.

4.3. Implications The differences in CH4 flux magnitudes and sensitivities to environmental variables for these two wetlands have important implications for future regional greenhouse gas (GHG) emissions and their estimation, especially in view of on-going large-scale land surface alteration and climatic change. Data analysis from long-term (1960–2006) climate records indicate that while China’s overall mean annual temperature has increased significantly over the past few decades, the largest warming is found in northeast China (Piao et al., 2010). Current estimates of GHG emission for China indicate that inundated wetlands in northeast China, despite being significant sinks for CO2, have a positive effect on atmospheric warming potential because of the high CH4 emissions (Tian et al., 2011). The exponential response of CH4 emission to temperature at the marsh (Fig. 6a) suggests that as the climate continues to warm these ecosystems will emit more CH4 to the atmosphere. Land use is also strongly affecting natural wetlands in northeast China, especially on the Sanjiang Plain. Since the 1950’s more than 75% of wetlands on the Sanjiang Plain have been converted to farmland (Wang et al., 2011). Much of these conversions were to rice paddies and despite a number of environmental restoration projects conducted from the end of the 1990s to 2010, there was still a substantial reduction of natural wetlands and expansion of rice paddies (Dong et al., 2015). While previous research suggests that marshland conversion to paddy land in northeastern China resulted in a cumulative reduction of CH4 emissions (Jiang et al., 2009; Huang et al., 2010), there is uncertainty in such estimates. Several factors contribute to this uncertainty: 1) the large temporal and

Acknowledgements This work has been jointly supported by the National Key R&D Program of China (2016YFA0602303), National Nature Science Foundation of China (41730643, 41771102) and Key Research Program of Frontier Sciences, CAS (QYZDJ-SSW-DQC013). We gratefully acknowledge two anonymous reviewers and Associate Editor Dr. Tim Griffis for their constructive comments on an earlier version of this paper. We would like to thank Fuxi Shi and the staff at the Sanjiang station for their assistance throughout our field work. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.agrformet.2017.11.009. 9

Agricultural and Forest Meteorology xxx (xxxx) xxx–xxx

L. Sun et al.

References

Jiang, C.S., Wang, W.S., Hao, Q.J., Song, C.C., 2009. Effect of land-use change on CH4 and N2O emissions form freshwater marsh in Northeast China. Atmos. Environ. 43, 3305–3309. Jin, H.J., Wang, S.L., Lu, L.Z., Yu, S.P., 2009. Features of permafrost degradation in Hinggan Mountains, Northeastern China. Sci. Geogr. Sin. 29 (2), 223–228. Joabsson, A., Christensen, T.R., Wallén, B., 1999. Vascular plant controls on methane emissions from northern peat forming wetlands. Trends Ecol. Evol. 14, 385–388. Kettunen, A., Kaitala, V., Lehtinen, A., Lohila, A., Alm, J., Silvola, J., Martikainen, P.J., 1999. Methane production and oxidation potentials in relation to water table fluctuations in two boreal mires. Soil Biol. Biochem. 31, 1741–1749. Kim, Y., 2015. Effect of thaw depth on fluxes of CO2 and CH4 in manipulated Arctic coastal tundra of Barrow, Alaska. Sci. Total Environ. 505, 385–389. Kljun, N., Calanca, P., Rotach, M.W., Schmid, H.P., 2004. A simple parameterisation for flux footprint predictions. Bound-Lay. Meteorol. 112 (3), 503–523. Koebsch, F., Jurasinski, G., Koch, M., Hofmann, J., Glatzel, S., 2015. Controls for multiscale temporal variation in ecosystem methane exchange during the growing season of a permanently inundated fen. Agric. For. Meteorol. 204, 94–105. Kormann, R., Meixner, F.X., 2001. An analytical footprint model for nonneutral stratification. Bound-Lay. Meteorol. 99, 207–224. Lafleur, P.M., 2009. Connecting atmosphere and wetland: trace gas exchanges. Geogr. Compass 3, 560–585. Lai, D.Y.F., Roulet, N.T., Moore, T.R., 2014. The spatial and temporal relationships between CO2 and CH4 exchange in a temperate ombrotrophic bog. Atmos. Environ. 89, 249–259. Lai, D.Y.F., 2009. Methane dynamics in northern peatlands: a review. Pedosphere 19 (4), 409–421. Le Mer, J., Roger, P., 2001. Production, oxidation, emission and consumption of methane by soils: A review. Eur. J. Soil Biol. 37 (1), 25–50. Limpens, J., Berendse, F., Blodau, C., Canadell, J.G., Freeman, C., Holden, J., Roulet, N., Rydin, H., Schaepman-Strub, G., 2008. Peatlands and the carbon cycle: from local processes to global implications-a synthesis. Biogeosciences 5, 1475–1491. Liu, X.T., 2005. The Wetlands in Northeast China. Chinese Science Press, Beijing, China. Long, K.D., Flanagan, L.B., Cai, T., 2010. Diurnal and seasonal variation in methane emissions in a northern Canadian peatland measured by eddy covariance. Global Change Biol. 16, 2420–2435. Ma, K., You, L.Z., Liu, J.G., Zhang, M.X., 2012. A Hybird Wetland Map for China: A Synergistic Approach Using Census and Spatially Explicit Datasets. PLoS One 7 (10), e47814. http://dx.doi.org/10.1371/journal.pone.0047814. Ma, X.H., Yin, S.C., Wen, B.L., et al., 2013. Carbon Reserves and Emissions of Peatlands in China. China Forestry Publishing House, Beijing, China. Matveev, A., Laurion, I., Deshpande, B.N., Bhiry, N., Vincent, W.F., 2016. High methane emissions from thermokarst lakes in subarctic peatlands. Limnol. Oceanogr. 61, S150–S164. Mauder, M., Foken, T., 2004. Documentation and Instruction Manual of the Eddy Covariance Software Package TK2. Arbeitsergebn. University of Bayreuth, Dept. of Micrometeorology ISSN 1614-8916. Meng, H.N., Song, C.C., Miao, Y.Q., Mao, R., Wang, X.W., 2014. Response of CH4 emissions to moss removal and N addition in boreal peatland of northeast China. Biogeosciences 11, 4809–4816. Miao, Y.Y., Song, C.C., Wang, X.W., Sun, X.X., Meng, H.N., Sun, L., 2012. Greenhouse gas emissions from different wetlands during the snow-covered season in Northeast China. Atmos. Environ. 62, 328–335. Mitra, S., Wassmann, R., Vlek, P.L.G., 2005. An appraisal of global wetland area and its organic carbon stock. Curr. Sci. 88 (1), 25–35. Mitsch, W.J., Gosselink, J.G., 2007. Wetlands, 4th edn. Wiley, Hoboken. Mitsch, W.J., Bernal, B., Nahlik, A.M., Mander, Ü., Zhang, L., Anderson, C.J., Jørgensen, S.E., Brix, H., 2013. Wetlands, carbon, and climate change. Landsc. Ecol. 28 (4), 583–597. Nadeau, D.F., Rousseau, A.N., Coursolle, C., Margolis, H.A., Parlange, M.B., 2013. Summer methane fluxes from a boreal bog in northern Quebec, Canada, using eddy covariance measurements. Atmos. Environ. 81, 464–474. Niu, Z.G., Gong, P., Cheng, X., et al., 2009. Geographical characteristics of China’s wetlands derived from remotely sensed data. Sci. China Ser. D-Earth Sci. 52 (6), 723–738. Noyce, G.L., Varner, R.K., Bubier, J.L., Frolking, S., 2014. Effect of Carex rostrate on seasonal and interannual variability in peatland methane emissions. J. Geophys. Res. Biogeosci. 119, 24–34. Olefeldt, D., Turetsky, M.R., Crill, P.M., McGuire, A.D., 2013. Environmental and physical controls on northern terrestrial methane emissions across permafrost zones. Global Change Biol. 19, 589–603. Olson, D.M., Griffis, T.J., Noorments, A., Kolka, R., Chen, J., 2013. Interannual, seasonal, and retrospective analysis of the methane and carbon dioxide budgets of a temperate peatland. J. Geophys. Res. Biogeosci. 118, 226–238. Piao, S., Ciais, P., Huang, Y., et al., 2010. The impacts of climate change on water resources and agriculture in China. Nature 467 (7311), 43–51. Pypker, T.G., Moore, P.A., Waddington, J.M., Hribljan, J.A., Chimner, R.C., 2013. Shifting environmental controls on CH4 fluxes in a sub-boreal peatland. Biogeosciences 10, 7971–7981. Rinne, J., Riutta, T., Pihlatie, M., Aurela, M., Haapanala, S., Tuovinen, J.-P., Tuittila, E.S., Vesala, T., 2007. Annual cycle of methane emission from a boreal fen measured by the eddy covariance technique. Tellus B 59, 449–457. Rogerson, P.A., 2001. Statistical Methods for Geography. Sage, Cromwell Press Ltd, London, pp. 124–153. Roulet, N.T., 2000. Peatlands, carbon storage, greenhouse gases, and the Kyoto Protocol: prospects and significance for Canada. Wetlands 20, 605–615. Sachs, T., Wille, C., Boike, J., Kutzbach, L., 2008. Environmental controls on ecosystem-

Baer, D.S., Paul, J.B., Gupta, J.B., O′Keefe, A., 2002. Sensitive absorption measurements in the near-infrared region using off-axis integrated-cavity-output spectroscopy. Appl. Phys. B: Lasers Opt. 75 (2–3), 261–265. Bartlett, K.B., Harriss, R.C., 1993. Review and assessment of methane emissions from wetlands. Chemosphere 26 (1–4), 261–320. Bassow, S.L., Bazzaz, F.A., 1998. How environmental conditions affect canopy leaf-level photosynthesis in four deciduous tree species. Ecology 79, 2660–2675. Bridgham, S.D., Megonigal, J.P., Keller, J.K., Bliss, N.B., Trettin, C., 2006. The carbon balance of North American wetlands. Wetlands 26 (4), 889–916. Bridgham, S.D., Cadillo-Quiroz, H., Keller, J.K., Zhuang, Q., 2013. Methane emissions from wetlands: biogeochemical, microbial, and modeling perspectives from local to global scales. Global Change Biol. 19, 1325–1346. Brown, M.G., Humphreys, E.R., Moore, T.R., Roulet, N.T., Lafleur, P.M., 2014. Evidence for a nonmonotonic relationship between ecosystem-scale peatland methane emissions and water table depth. J. Geophys. Res. Biogeosci. 119, 826–835. http://dx.doi. org/10.1002/2013JG002576. Brown, R.J.E., 1963. Influence of vegetation on permafrost. Permafrost International Conference Proceedings. National Academy of Sciences, National Research Council, Washington D.C., USA, pp. 20–25. Bubier, J., Costello, A., Moore, T.R., Roulet, N.T., Savage, K., 1993. Microtopography and methane flux in boreal peatlands, northern Ontario, Canada. Can. J. Bot. 71, 1056–1063. Chang, X.L., Jin, H.J., He, R.X., Yang, S.Z., Yu, S.P., Lu, L.Z., Guo, D.X., Wang, S.L., Kang, X.C., 2008. Advances in permafrost and cold region environments studies in the Da Hinggan Mountains, Northeast China. J. Glaciol. Geocryol. 30, 176–182. Chanton, J.P., Whiting, G.J., 1996. Methane stable isotopic distributions as indicators of gas transport mechanisms in emergent aquatic plants. Aquat. Bot. 54, 227–236. Chasar, L.S., Chanton, J.P., Glaser, P.H., Siegel, D.I., 2000. Methane concentration and stable isotope distribution as evidence of rhizospheric processes: comparison of a fen and bog in the Glacial Lake Agassiz Peatland Complex. Ann. Bot. 86, 655–663. Christensen, T.R., Ekberg, A., Strőm, L., Mastepanov, M., Nicolai, P., Öquist, M., Svensson, B.H., Nykänen, H., Martikainen, P.J., Oskarsson, H., 2003. Factors controlling large scale variations in methane emissions from wetlands. Geophys. Res. Lett. 30, 1414. http://dx.doi.org/10.1029/2002GL016848. Christensen, T.R., Johansson, T., Akerman, H.J., Mastepanov, M., Malmer, N., Friborg, T., Crill, P., Svensson, B.H., 2004. Thawing sub-arctic permafrost: Effects on vegetation and methane emissions. Geophys. Res. Lett. 31, L04501. http://dx.doi.org/10.1029/ 2003GL018680. Chu, H., Chen, J., Gottgens, J.F., Ouyang, Z., John, R., 2014. Net ecosystem methane and carbon dioxide exchanges in a Lake Erie coastal marsh and a nearby cropland. J. Geophys. Res. Biogeosci. 119 (5), 722–740. Dong, G.T., Bai, J., Yang, S.T., Wu, L., Cai, M.Y., Zhang, Y.C., Luo, Y., Wang, Z.W., 2015. The impact of land use and land cover change on net primary production on China’s Sanjiang Plain. Environ. Earth Sci. 74, 2907–2917. Euskirchen, E.S., Edgar, C.W., Turetsky, M.R., Waldrop, M.P., Harden, J.E., 2014. Differential response of carbon fluxes to climate in three peatland ecosystems that vary in the presence and stability of permafrost. J. Geophys. Res. Biogeosci. 119 (8), 1576–1595. Forster, P., Ramaswamy, V., Artaxo, P., et al., 2007. Changes in atmospheric constituents and in radiative forcing. Climate Change 2007: the Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Frenzel, P., Karofeld, E., 2000. CH4 emission from a hollow-ridge complex in a raised bog: The role of CH4 production and oxidation. Biogeochemistry 51, 91–112. Hatala, J.A., Detto, M., Baldocchi, D.D., 2012. Gross ecosystem photosynthesis causes a diurnal pattern in methane emission from rice. Geophys. Res. Lett. 39, L06409. Hendriks, D.M., van Huissteden, J., Dolman, A.J., 2010. Multi-technique assessment of spatial and temporal variability of methane fluxes in a peat meadow. Agric. For. Meteorol. 150, 757–774. Heyer, J., Berger, U., Kuzin, I.L., Yakovlev, O.N., 2002. Methane emissions from different ecosystem structures of the subarctic tundra in Western Siberia during midsummer and during the thawing period. Tellus B 54, 231–249. Huang, Y., Sun, W.J., Zhang, W., Yu, Y.Q., Su, Y.H., Song, C.C., 2010. Marshland conversion to cropland in northeast China from 1950 to 2000 reduced the greenhouse effect. Global Change Biol. 16, 680–695. IPCC, 2013. Summary for policymakers. 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 Univ. Press, Cambridge, U. K., and New York 731 pp. Ibrom, A., Dellwik, E., Flyvbjerg, H., Jensen, N.O., Pilegaard, K., 2007a. Strong low-pass filtering effects on water vapor flux measurements with closed-path eddy correlation systems. Agric. For. Meteorol. 147, 140–156. Ibrom, A., Dellwik, E., Larse, S.E., Pilegaard, K., 2007b. On the use of the Webb-PearmanLeuning theory for closed-path eddy correlation measurements. Tellus B 59, 937–946. Iwata, H., Harazono, Y., Ueyama, M., Sakabe, A., Nagano, H., Kosugi, Y., Takahashi, K., Kim, Y., 2015. Methane exchange in a poorly-drained black spruce forest over permafrost observed using the eddy covariance technique. Agric. For. Meteorol. 214–215, 157–168. Jackowicz-Korczyński, M., Christensen, T.R., Bäckstrand, K., Crill, P., Friborg, T., Mastepanov, M., Ström, L., 2010. Annual cycle of methane emission from a subarctic peatland. J. Geophys. Res. 115, G02009. http://dx.doi.org/10.1029/2008JG000913.

10

Agricultural and Forest Meteorology xxx (xxxx) xxx–xxx

L. Sun et al.

climate warming. J. Geophys. Res. 116, G02011. http://dx.doi.org/10.1029/ 2010JG001393. Turetsky, M.R., Kotowska, A., Bubier, J., et al., 2014. A synthesis of methane emissions from 71 northern, temperate, and subtropical wetlands. Global Change Biol. 20, 2183–2197. van Groenigen, K.J., van Kessel, C., Hungate, B.A., 2013. Increased greenhouse-gas intensity of rice production under future atmospheric conditions. Nat. Clim. Change 3, 288–291. Vourlitis, G., Oechel, W., Hastings, S.J., Jenkins, M., 1993. The effect of soil moisture and thaw depth on CH4 flux from wet sedge tundra ecosystems on the North Slope of Alaska. Chemosphere 26, 329–337. Waddington, J.M., Harrison, K., Kellner, E., Baird, A.J., 2009. Effect of atmospheric pressure and temperature on entrapped gas content in peat. Hydrol. Processes 23 (20), 2970–2980. Wang, Z.M., Song, K.S., Ma, W.H., Ren, C.Y., Zhang, B., Liu, D.W., 2011. Loss and fragmentation of marshes in the Sanjiang plain, Northeast China, 1954–2005. Wetlands 31, 945–954. Wang, X.W., Song, C.C., Wang, J.Y., Miao, Y.Q., Mao, R., Song, Y.Y., 2013. Carbon release from Sphagnum peat during thawing in a montane area in China. Atmos. Environ. 75, 77–82. Whalen, S.C., 2005. Biogeochemistry of methane exchange between natural wetlands and the atmosphere. Environ. Eng. Sci. 22 (1), 73–94. Whiting, G.J., Chanton, J.P., 1992. Plant-dependent methane emission in a subarctic Canadian fen. Global Biogeochem. Cycles 6 (3), 225–231. http://dx.doi.org/10. 1029/92GB00710. Whiting, G.J., Chanton, J.P., 1993. Primary production control of methane emission from wetlands. Nature 364 (6440), 794–795. Whiting, G.J., Chanton, J.P., 1996. Control of the diurnal pattern of methane emission from emergent aquatic macrophytes by gas transport mechanisms. Aquat. Bot. 54, 237–253. Whiting, G.J., Chanton, J.P., Bartlett, D.S., Happell, J.D., 1991. Relationship between CH4 emission, biomass, and CO2 exchange in a subtropical grassland. J. Geophys. Res. 96 (D7), 13067–13071. http://dx.doi.org/10.1029/91JD01248. Zedler, J.B., Kercher, S., 2005. Wetland Resources: Status, Trends, Ecosystem Services, and Restorability. Annu. Rev. Environ. Resour. 30, 39–74. Zhao, K.Y., 1999. Wetlands in China. Chinese Science Press, Beijing, China.

scale CH4 emission from polygonal tundra in the Lena River Delta, Siberia. J. Geophys. Res. 113, G00A03. http://dx.doi.org/10.1029/2007JG000505. Schemske, D.W., Horvitz, C.C., 1988. Plant-animal interactions and fruit production in a neotropical herb: a path analysis. Ecology 69, 1128–1137. Song, C.C., Xu, X.F., Tian, H.Q., Wang, Y.Q., 2009. Ecosystem–atmosphere exchange of CH4 and N2O and ecosystem respiration in wetlands in the Sanjiang Plain, Northeastern China. Global Change Biol. 15, 692–705. Song, C.C., Sun, L., Huang, Y., Wang, Y.S., Wan, Z.M., 2011. Carbon exchange in a freshwater marsh in the Sanjiang Plain, Northeastern China. Agric. For. Meteorol. 151, 1131–1138. Song, C.C., Wang, X.W., Miao, Y.Q., Wang, J.Y., Mao, R., Song, Y.Y., 2014. Effects of permafrost thaw on carbon emissions under aerobic and anaerobic environments in the Great Hing'an Mountains. China. Sci. Total Environ. 487, 604–610. Song, W., Wang, H., Wang, G., Chen, L., Jin, Z., Zhuang, Q., He, J., 2015. Methane emissions from an alpine wetland on the Tibetan Plateau: Neglected but vital contribution of the nongrowing season. J. Geophys. Res. Biogeosci. 120 (8), 1475–1490. Ström, L., Christensen, T.R., 2007. Below ground carbon turnover and greenhouse gas exchange in a sub-arctic wetland. Soil Biol. Biochem. 39, 1689–1698. Strachanm, I.B., Nugent, K.A., Crombie, S., Bonneville, M., 2015. Carbon dioxide and methane exchange at a cool-temperate freshwater marsh. Environ. Res. Lett. 10, 065006. http://dx.doi.org/10.1088/1748-9326/10/6/065006. Sturtevant, C.S., Oechel, W.C., Zona, D., Kim, Y., Emerson, C.E., 2012. Soil moisture control over autumn season methane flux: Arctic Coastal Plain of Alaska. Biogeosciences 9, 1423–1440. Sturtevant, C., Ruddell, B.L., Knox, S.H., Verfaillie, J., Matthes, J.H., Oikawa, P.Y., Baldocchi, D., 2016. Identifying scale-emergent, nonlinear, asynchronous processes of wetland methane exchange. J. Geophys. Res. Biogeosci. 121 (1), 188–204. Sun, L., Song, C., Miao, Y., Qiao, T., Gong, C., 2013. Temporal and spatial variability of methane emissions in a northern temperate marsh. Atmos. Environ. 81, 356–363. Suyker, A.E., Verma, S.B., Clement, R.J., 1996. Methane flux in a boreal fen: Season-long measurement by eddy correlation. J. Geophys. Res. 101 (D22), 28637–28647. Tagesson, T., Molder, M., Mastepanov, M., et al., 2012. Land-atmosphere exchange of methane from soil thawing to soil freezing in a high-Arctic wet tundra ecosystem. Global Change Biol. 18, 1928–1940. Tian, H.Q., Xu, X.F., Lu, C.Q., et al., 2011. Net exchanges of CO2, CH4, and N2O between China’s terrestrial ecosystems and the atmosphere and their contributions to global

11