Controls on winter ecosystem respiration at mid-and high-latitudes

Introduction Conclusions References


Conclusions References
Tables Figures

Back Close
Full

Abstract
Winter CO 2 fluxes represent an important component of the annual carbon budget in northern ecosystems.Understanding winter respiration processes and their responses to climate change is also central to our ability to assess terrestrial carbon cycle and climate feedbacks in the future.The factors influencing the spatial and temporal pattern of winter respiration (RECO) of northern ecosystems are poorly understood.For this reason, we analyzed eddy covariance flux data sets from 57 ecosystem sites ranging from ∼ 35 • N to ∼ 70 • N. Deciduous forests carry the highest winter RECO ratios (9.7-10.5 g C m −2 d −1 ), when winter is defined as the period during which air temperature remained below 0 • C. By contrast, wetland ecosystems had the lowest winter ). Evergreen needle-leaved forests, grasslands and croplands were characterized by intermediate winter RECO values of 7.4-7.9g C m −2 d −1 , 5.8-6.0 g C m −2 d −1 , and 5.2-5.3g C m −2 d −1 , respectively.Cross site analysis showed that winter air or soil temperature, and the seasonal amplitude of the leaf area index inferred from satellite observation, which is a proxy for the amount of litter available for RECO in the subsequent winter, are the two main factors determining spatial pattern of daily mean winter RECO.Together, these two factors can explain 71% (T air , ∆LAI) or 69% (T soil , ∆LAI) of the spatial variance of winter RECO across the 57 sites.The spatial temperature sensitivity of daily winter RECO was determined empirically by fitting an Arrhenius relationship to the data.The activation energy parameter of this relationship was found to decrease at increasing soil temperature at a rate of 83.1 KJ • C-1 (r = −0.32,p < 0.05), which implies a possible dampening of the increase in winter RECO due to global warming.The interannual variability of winter RECO is better explained by soil temperature than by air temperature, likely due to the insulating effects of snow cover.The increase in winter RECO with a 1 • C warming based calculated from the spatial analysis was almost that double that calculated from the temporal analysis.Thus, models that calculate the effects of warming on RECO based only on spatial analyses could be over-estimating the impact.

Introduction
The processes controlling the winter carbon cycle of northern ecosystems, mainly ecosystem respiration, have received much less attention than processes active during the growing season.The long-time standing view of marginal wintertime biological activity (e.g.Coyne and Kelley, 1971;Steudler et al., 1989) suggests that winter respiration is very small compared to growing season respiration.Recent field studies refuted this view by unravelling significant wintertime respiration rates in Arctic tundra, bog, and mountain ecosystems (e.g.Oechel et al., 1997;Fahnestock et al., 1998;Grogan and Chapin, 1999;Panikov and Dedysh, 2000;Aurela et al., 2002;Monson et al., 2006;Bergeron et al., 2007).These studies suggested that winter ecosystem respiration should not be ignored when attempting to quantify and understand the annual carbon balance of ecosystems (Hobbie et al., 2000;Grogan and Jonasson, 2005;Johansson et al., 2006).However, due to large variations in carbon storage and functioning across northern ecosystems and the limited spatial representativeness of individual studies, winter ecosystem respiration remains incompletely understood.
In general, mid and high-latitude ecosystems in general contain large amounts of soil carbon (Post et al., 1982).Part of this soil organic carbon (SOC) mass could be decomposed more actively than fresh input (e.g.litterfall) in response to future warming.The increased high-latitude warming projected by climate models includes winter warming (Serreze et al., 2000;Giorgi et al., 2001) and has already been observed over the past 30 years (IPCC, 2007).The response of SOC balance to warming differs widely among coupled climate-carbon models (Friedlingstein et al., 2006).This is because the net balance in these models depends on two fluxes of opposite directions, i.e., the litter input that may increase under warming if vegetation net primary productivity increases, and the soil carbon microbial decomposition rate that also responds positively to warming (e.g.Jones et al., 2005).Therefore, it is important to disentangle from available observations how temperature and vegetation productivity separately control winter respiration.Introduction

Conclusions References
Tables Figures

Back Close
Full Many studies suggested that major drivers of temporal and spatial variations of winter carbon respiration are temperature (Clein and Schimel, 1995;Hobbie, 1996;Mikan et al., 2002), substrate quality (Grogan et al., 2001;Grogan and Jonasson, 2005) and snow depth (Monson et al., 2006;Nobrega and Grogan, 2007).For example, Grogan and Jonasson (2005) found that both the amount of substrate available for respiration and soil temperature determine landscape level variation of winter ecosystem respiration of birch forest and heath tundra.Monson et al. (2006) observed that interannual variations of a montane forest winter carbon balance were strongly coupled with soil temperature and variability in snow depth changes.These studies are very valuable for understanding site-specific processes, but their results cannot be readily extrapolated across sites and climate gradients to infer regional sensitivities.
Eddy covariance measurements of CO 2 fluxes collected continuously together with climate variables are available across temperate, boreal, and arctic ecosystems (Baldocchi et al., 2001(Baldocchi et al., , 2008)).These data represent a valuable source of information for the analyses of winter RECO spatial and temporal variability.In this study, we focus on northern hemisphere sites from 36 • N to 70 • N, covering a climate gradient of 24 • C of mean annual temperature.In the first part, we investigate the importance of winter respiration, winter RECO and its contribution to annual RECO for different ecosystem types.The results are based on two different definition of the winter season.In the second part, we analyzed the temperature dependency of daily winter RECO site by site using an Arrhenius type model.We also considered a total of 252 site-years that have been aggregated to quantify the sensitivity of anomalies of winter RECO to temperature on inter-annual time scale.In the final part, which tries to open the path for spatial analysis of winter RECO, we examine the relationships between winter RECO and climate variables and productivity-related variables across sites.Introduction

Conclusions References
Tables Figures

Back Close
Full The eddy covariance data used in this study are extracted from La Thuile FLUXNET synthesis database which contains 965 site years processed according to standardized protocols (Papale et al., 2006) (http://www.fluxdata.org).Air temperature, upper soil temperature (between 2 and 10 cm depth) and precipitation measured both on a halfhourly and daily bases at each site were used.The net ecosystem exchange (NEE, in g C m −2 day −1 ) was measured at a 30 min time step, and flux toward the surface was considered to be negative in sign.The NEE time series can be separated into the gross fluxes GPP (gross primary productivity) and RECO (ecosystem respiration) where, by definition NEE = − GPP + RECO (Reichstein et al., 2005).This flux-partitioning algorithm adopted short-term temperature sensitivity to extrapolate night-time respiration to daytime, and this approach avoids significantly biased estimates of RECO which can be obtained using long-term temperature sensitivity affected by confounding factors such as growth dynamics (Reichstein et al., 2005).This partitioning yields information about the ecosystem-scale processes controlling spatiotemporal variation in the measured NEE.Nighttime NEE on half-hourly time step is used for the derivation of winter RECO temperature sensitivity at each site, and other analyses are based on daily RECO.Ancillary observations of maximum LAI from site measurements were used in this study.
Out of the 255 sites located north of 36 • N, a subset of 57 was identified which accomplished the following requirements: having at least two years of air and upper soil temperature, precipitation and RECO measurements and the missing gap is below 30% of the data for each site year.A total of 57 sites containing 256 site years of data were selected, covering evergreen needleleaf forest (100 site × years), deciduous Introduction

Conclusions References
Tables Figures

LAI dataset
Leaf Area Index data were retrieved for each investigated site from MODIS-Aqua downloaded from the ORNL-DAAC MODIS -Collection-5 LAI ASCII data (MYD15A2) (https://daac.ornl.gov).These LAI data have a spatial resolution of 1 km and a temporal resolution of 8-days.They also include quality control (QC) information about cloud and data processing conditions.Coarser resolution LAI data from a 1 km × 1 km area around each site were also collected during the observation period.Only LAI data without significant cloud contamination described in the LAI user's guide (http: //landweb.nascom.nasa.gov/cgi-bin/QAWWW/) within an area of 1 km × 1 km centered on each site were retained for each 8-day period to obtain the maximum and minimum LAI values for each site year.The seasonal amplitude (∆LAI) is defined as the difference between maximum and minimum of LAI and can be considered as a proxy for recent carbon inputs to soil, i.e. substrate available for sustaining winter respiration.

Arrhenius equation to describe the temperature dependency of RECO
The temperature dependency of daily winter RECO at each site was analyzed using the Arrhenius type equation (Lloyd and Taylor, 1994): where, ) is a reference respiration rate at the reference temperature (T ref , K) related both to the amount of substrate available for decomposers, and its quality (Lloyd and Taylor, 1994).E 0 (KJ mol −1 ) is the activation Introduction

Conclusions References
Tables Figures

Back Close
Full energy parameter and represents the ecosystem respiration sensitivity to temperature, and Re the universal gas constant.Across sites, this model was reformulated by adding the dependency of reference respiration on the site characteristics.The model accuracy was then assessed by a cross-validation technique: one site at a time was excluded using the remaining subset as the training set and the excluded one as the validation set, and the model was fitted against the training set and then applied to calculate the modeled value for the validation set.

Winter season definition
Two winter season definitions were tested to estimate the effect of an arbitrary choice to define the winter season in analyzing winter RECO drivers.The first one (D1) defines winter as the period during which the 10-day smoothed air temperature remained below 0 • C for at least five consecutive days.It should be noted that the calculation of winter length according to D1 is based on mean air temperature of all available years for each site.The second definition (D2) defines winter as the 90-day period from 1 December to 28 February.to 29.0 (median = 6.1, 25th percentile = 4.0, and 75th percentile = 8.7) for D1 and D2, respectively.

Winter RECO and its ratio to annual RECO among ecosystem types
Table 2 provides the statistics of winter RECO rates and cumulative winter RECO for different ecosystem types.Deciduous broadleaf forests have the highest winter RECO (rates: 0.97-1.05g C m −2 d −1 ; cumulative: 86.9-96.2g C m −2 ), whereas wetlands have the lowest (rates: 0.21-0.23 g C m  and Schimel, 1995) and wetland has the lowest soil temperature (−3.6 • C).Showing generic values for winter RECO among different ecosystem types is as important as improving the estimates of primary production often derived from remote-sensing indices (e.g.Bunn et al., 2007;Goetz et al., 2005Goetz et al., , 2007) ) when trying to address the evolution of the total carbon balance, as pinpointed by Piao et al. (2008) focusing on autumn warming.
The p-ratio varies among ecosystem types (Table 2).Using definition D1, the highest p-ratio values (%) are found in the Arctic wetlands (10.3 ± 3.2, N = 13), intermediate values in evergreen needleleaf forest (8.3 ± 4.2, N = 100), deciduous broadleaf forest (9.0 ± 2.9, N = 65) and mixed forest (8.3 ± 2.0, N = 19), and the lowest values in grassland (6.8 ± 6.1, N = 34) and cropland (4.7 ± 2.3, N = 24).In contrast, when using definition D2 with a much shorter winter duration in the high latitudes, wetland (5.3 ± 1.7, N = 13) have the lower p-ratio.Anther type of p -ratio is also presented in Table 2, which is defined as the ratio of winter RECO rates to annual mean RECO.Compared to p-ratio, p -ratio (%) is much less varied among different ecosystem types, and it Introduction

Conclusions References
Tables Figures

Back Close
Full ranges from 21.3 to 34.5 using D1 and from 21.2 to 31.2 using D2.Similar to p-ratio, the high productive cropland carries the lowest p -ratio values (D1: 21.3 ± 9.9; D2: 21.2 ± 9.6), which might be related to agricultural management practices that removed crop residuals for winter respiration.With winter definition D1, p-ratio increase with latitude (r = 0.20, p = 0.05, data not shown) since higher-latitude sites often have longer winter length (r = 0.51, p < 0.01, data not shown).This implies that winter RECO in colder regions carries a higher relative contribution to total annual RECO than at warmer sites using D1 and thus further stresses the importance of winter RECO in the carbon balance of arctic and boreal ecosystems, as already pointed out by other studies (e.g.Oechel et al., 1997;Fahnestock et al., 1998;Bergeron et al., 2007) It should be noted that the use of the open-path gas analyzers for eddy covariance estimates of small fluxes relaying on the WPL correction (Webb et al., 1998) can introduce the errors (e.g.Kondo and Tsukamoto, 2007), and CO 2 releases can be systematic underestimated during off-season periods especially in cold-climate ecosystems due to instrumental surface heating effects that are associated with open-path gas analyzers (e.g.Licor 7200) (Burba et al., 2008).For example, Burba et al. (2008) 2a).This indicates that substrate availability and quality exerts a significant control on the spatial variation of R ref across sites, and thus supports the conclusions of Grogan and Jonasson (2005) who found that R ref was significantly reduced after removing plant and litter in a birch and heath tundra.
The activation energy E 0 increases with decreasing soil temperature across the sites (excluding the Arctic wetland site US-Atq) (Fig. 2b).The value of the correlation coefficient is very small, but significant.It has been assumed that microorganisms at temperatures around −10 • C in Siberia are in a state of anabiosis (Vorobyova et al., 1997), and thus low E 0 (23.4KJ mol −1 ) might be expected due to microbial metabolic activity constrained by very cold winter soil temperature (around −11 • C in US-Atq).R ref positively correlate also with soil temperature (r = 0.27, p < 0.01, data not shown).
Since R ref can be interpreted as a proxy for substrate quality (Lloyd and Taylor, 1994;Mikan et al., 2002), it can be argued that the colder sites may contain a higher fraction of recalcitrant soil carbon, which decreases R ref .
Therefore, the inferred increase in E 0 at colder soil temperatures might be related to the presence of more recalcitrant soil organic compounds, with higher activation energy than labile compounds, according to Introduction

Conclusions References
Tables Figures

Back Close
Full kinetic theory (Bosatta and Agren, 1999;Davidson and Janssens, 2006).If US-Atq is not considered, a decreasing rate of 83.1 KJ mol −1 K −1 for E 0 has been estimated.The effects of such a decreased sensitivity of winter RECO to soil temperature at warmer sites could be tested in terrestrial ecosystem models.Extrapolating the relationship in Fig. 2 from space to time would imply that the future warming trends reduce the activation energy of winter soil C decomposition, hence dampening the potential increase of RECO with temperature.

Temperature sensitivity of winter RECO on inter-annual time scale
Owning to the short length of RECO and temperature records, temporal correlations between winter RECO and predictor climate variables are not applicable to study the inter annual sensitivity of RECO to climate in details at each site.Instead, we calculate annual winter RECO and winter temperature anomalies at each site, by removing the long-term mean RECO value from yearly values, and perform a least squares regression between all site-years anomalies of winter RECO and winter temperature to quantify the response of winter RECO to interannual variations in temperature.Results from this analysis show that winter RECO anomalies positively correlate with winter (D1 and D2) soil temperature anomalies (Fig. 3), which explained more (D1: R 2 = 0.18, p < 0.01, Fig. 3a; D2: R 2 = 0.14, p < 0.01, Fig. 3b) than air temperature (D1: R 2 = 0.09, p < 0.01; D2: R 2 = 0.05, p < 0.01, data not shown).The soil temperature was a better driver of winter RECO likely because of the influence of snow cover which acts as a thermal insulator controlling soil microbial activity (Zhang et al., 2005).This is consistent with the results of a six-year record of eddy covariance measurements at the Niwot Ridge Ameriflux site in the Rocky Mountains, Monson et al. (2006) showed that interannual variability of net carbon exchange is not controlled by air temperature anomalies, but rather by soil temperature anomalies which were controlled by snow Introduction

Conclusions References
Tables Figures

Back Close
Full can be expected since the snow characteristics at site level can not be truly reflected by a remote sensing product at a spatial resolution of 25 × 25 km 2 .In addition, we found no significant relationship between winter RECO and precipitation anomalies (winter definition D1: r = 0.00, p = 0.49; D2: r = 0.00, p = 0.71) and no correlation with ∆LAI anomalies (D1: r = 0.00, p = 0.44; D2: r = 0.00, p = 0.818).
The winter RECO sensitivity to soil temperature defined by the linear regression slope q1 (g C m −2 d −1 • C −1 ) calculated from the spatial relationship across sites (D1: y = 0.14x + 0.71, R 2 = 0.55, p < 0.01, q1 = 0.14; D2: y = 0.12x + 0.72, R 2 = 0.49, p < 0.01, q1 = 0.12 data not shown) is higher than the RECO temporal (inter-annual) sensitivity to soil temperature (D1: y = 0.09x − 6E − 18, R 2 = 0.18, p < 0.01, q1 = 0.09; D2: y = 0.07x + 1E − 17, R 2 = 0.14, p < 0.01, q1 = 0.07, Fig. 3a and b).This result might suggest that a 1 • C warming increases winter ecosystem CO 2 loss by 0.12-0.14g C m 2 d −1 based on the spatial relationship (range corresponding to D1 or D2 winter definitions) against only by 0.07-0.09g C m 2 d −1 based on the temporal relationship.The difference between these two winter RECO temperature sensitivities is due to the fact that inter-annual variation in winter RECO temperature sensitivity is mainly driven by direct climate effects, but the spatial variability accounts both for transient climate effects on respiration and long-term climate-induced biological variation in space (e.g.aboveground biomass, soil organic carbon and litter) that support the respiration (Hibbard et al., 2005).For this reason the application of the two sensitivities in models for the future projections of winter RECO may be more complex than thought.Temperature sensitivity as observed in time may be appropriate for simulating short term responses and space sensitivity for representing the long term (equilibrium) responses.Thus, simply projecting winter RECO in response to future warming based on spatial relationships incorporated in models can overestimate CO 2 losses by ecosystems.

Environmental and biotic controls on winter RECO across sites
Under the winter definition D1, winter RECO is found to increase exponentially with increasing air temperature (R 2 = 0.66, p < 0.01; Fig. 4a) and soil temperature (R 2 = 0.65, Introduction

Conclusions References
Tables Figures

Back Close
Full p < 0.01, Fig. 4b) across all sites.The corresponding activation energy values (E 0 ) are of 85.6 KJ mol −1 and 130.5 KJ mol −1 respectively.The comparable E 0 values (85.3 KJ mol −1 for air temperature; 129.5.0 KJ mol −1 for soil temperature) can also be obtained by excluding the two coldest sites (US-Ivo and US-Atq) from the exponential analysis.Besides temperature, ∆LAI, used as a proxy for decomposable carbon in litter and soil pools, also significantly explains spatial variation in winter RECO (y = 0.13x + 0.10, R 2 = 0.29, p < 0.01) (Fig. 4c).Given ∆LAI co-varies with temperature, the significant correlation between winter RECO and ∆LAI can also be inferred from partial regression analysis when controlling for temperature (r = 0.34, p < 0.05).Conversely, we found no significant correlation between winter RECO and total soil carbon concentration at each site (data not shown).This suggests that between-site differences of decomposable carbon, (∆LAI used as proxy) may explain some of the spatial heterogeneity in winter RECO, which was also found by other studies (e.g.Sommerfeld et al., 1996;Mast et al., 1998;Elberling, 2007;Migliavacca et al., 2010).A unit increase of ∆LAI (m −2 m −2 ) explains a winter RECO increase of 0.13 g C m −2 d −1 .
Thus, we enhanced the statistical model of Eq. ( 1) in order to incorporate an additional linear dependency of the reference respiration rate R ref on ∆LAI (Sect.3.3).Therefore, under the winter definition D1, winter RECO in g C m −2 d −1 can be expressed by: where T ai r and T soi l are air and soil temperature, respectively both expressed in K.The fit parameters A1, A2, E 0 and A 1, A 2, E 0 can be calculated by least-square fitting (Bjorck, 1996) to the available data.The coefficients of determination for Eqs. ( 2 (2) and (3) performed using cross-validation approach (Fig. 5) shows that 67% and 62% of winter RECO variance can be explained by Eqs. ( 2) and (3).Given that temperature is the dominant controlling factor of winter RECO across sites and co-varies with some of the other potential drivers, we regressed the fit residuals of Eqs. ( 2) and (3) against winter duration and winter precipitation to determine if these two alternative drivers could explain additional variance.There was no significant correlation between the fit residuals and either winter precipitation (Eq.2: r = 0.00, p = 0.756; Eq. 3: r = 0.01, p = 0.889, data not shown) or winter length (Eq.2: r = 0.00, p = 0.864; Eq. 3: r = 0.06, p = 0.179, data not shown).Both Eqs. ( 2) and ( 3) describe empirically the winter RECO spatial variability and thus have predictive power to extrapolate winter RECO on the continental scale.

Conclusions
Availability of climate and eddy covariance flux data across different ecosystems opens a new opportunity to quantify winter RECO and its empirical spatial and temporal controls across North Hemisphere ecosystems.Given a winter definition based upon temperature below the freezing point, we found an increase in the ratio of winter to annual respiration towards higher latitude.This shows that the relative importance of winter processes in controlling respiration is higher than that of summer processes.More research is therefore needed to include winter RECO in order to assess annual carbon budgets, especially at high latitudes where biogeochemical models often assume zero decomposition in cold winter.
Our empirical characterization of temperature controls on winter RECO has two major implications.First, winter RECO temperature sensitivity obtained on space and temporal scales should be treated differently, since the RECO sensitivity to warming obtained from spatial gradients will definitely be exaggerated when extrapolated to future warming.Second, soil temperature has a larger predictive power than air temperature in explaining interannual variations of winter RECO, likely through snow depth

Conclusions References
Tables Figures

Back Close
Full thermal isolation effects already indicated by individual studies.However, this should be further explored in future studies with an access to more in-situ snow data.This indicates that winter carbon exchanges in a future warmer northern hemisphere can not simply be deduced from air temperature changes, but need to account for interactions between soil carbon, soil temperature, moisture, and snow cover effects (start and end of snow accumulation, snow density and structure) on soil thermal regime.

BGD Introduction Conclusions References
Tables Figures

Back Close
Full  Full      p-ratio and p -ratio is the ratio of cumulative winter RECO to annual total RECO and the ratio of winter RECO rates to annual mean RECO rates, respectively.
SD is the standard deviation, and (SD) within a column followed by different letters ( a , b and c ) were significantly different (p < 0.05).Introduction

Conclusions References
Tables Figures

Back Close
Full Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper |

Figure 1
Figure 1 shows the histogram of cumulative winter RECO based on total winter season -(a) and (b) -and the histogram of p-ratios -(c) and (d) -according to both definitions.The p-ratio is defined as the ratio of cumulative winter RECO to annual total RECO.These histograms contain data from all site-years.The total winter RECO for both winter definitions ranges from 5.1 to 192.9 g C m −2 (median = 66.9, 25th percentile = 45.7 and 75th percentile = 88.5) and from 0.6 to 240.5 g C m −2 (median = 57.5, 25th percentile = 27.9, and 75th percentile = 95.1),respectively.The p-ratio (%) varies from 1.0 to 37.8 (median = 7.4, 25th percentile = 5.1, and 75th percentile = 7.8) and from 0.2 Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | found that flux measurements made with open-path LI-7500 can underestimate CO 2 release by 33% during the winter season compared to closed-path instrument.sensitivity of winter RECO Since daytime RECO is separated from NEE based on the temperature sensitivity of nighttime NEE in La Thuile dataset(Reichstein et al., 2005), for each site and using winter season definition D1, nighttime NEE on a half-hourly time step was regressed against the corresponding (soil or air) nighttime temperature during the whole winter season based on the Arrhenius function (Eq. 1) thus providing site-specific couples of parameters (R ref , E 0 ).Values of reference respiration rate R ref range from 0.18 to 1.75 (g C m −2 d −1 ) when air temperature is used as the predictor, and from 0.16 to 1.22 (g C m −2 d −1 ) when soil temperature is used.Across all the sites, R ref significantly increases with ∆LAI (Fig.
Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | ) and (3) are of 0.71 and 0.69, while the root mean square errors are of 0.20 and 0.22 g C m −2 d −1 , respectively.A cross Discussion Paper | Discussion Paper | Discussion Paper | validation of the regression models in Eqs.
Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | ∆LAI: the average difference between maximum and minimum of MODIS LAI (m 2 m −2 ) from corresponding available years and the MODIS LAI data is only available after year 2000.WRECO are average winter ecosystem respiration rates (g C m −2 d −1 ) for D1 (air temperature < 0 • C) and D2 (December-February) over available years, Discussion Paper | Discussion Paper | Discussion Paper | , MF, GRA, CRO and WET represent evergreen needle leaf forest, deciduous broadleaf forest, mixed forest, grassland, cropland and wetland respectively.

Table 1 .
Sites characteristic in this study.