1. Introduction
As one of the principal physical processes in the land–atmosphere interaction system (Zhang et al. 2003), land surface energy exchange is restricted by the climate system, but it also imposes strong feedbacks on the climate system (Claussen et al. 2001). Energy exchange between land and the atmosphere is mainly determined by sensible heat flux (H), latent heat flux (LE), and net radiation (NR) (Falge et al. 2005). Investigating the spatiotemporal changes in individual energy flux components and their allocation ratio in energy fluxes is essential for an accurate understanding of regional climate and ecological process changes. Besides climate change, the influences of human activities on energy exchanges between land and the atmosphere cannot be ignored (Mahmood et al. 2010; Oki et al. 2013). Climate change exerts influences mainly through changes in air temperature and precipitation, which drive the allocation of energy components, including H and LE, and thus change the regional energy balance (Wilson et al. 2002; Gu et al. 2006). Human activities influence hydrothermal processes by CO2 emissions or by changing the underlying surface through various land use patterns, thus altering the water and heat allocation (Pielke et al. 2002).
Studies on land surface energy balance at the regional scale are of great significance for weather forecasts, hydrological models, irrigation schedules, water management, and climate change investigations (Choi et al. 2009). However, a number of studies have been conducted at the ecosystem (site) scale based on eddy covariance (EC) tower measurements (von Randow et al. 2004; Ibrom et al. 2007), which could represent heat exchange in a certain range in homogeneous underlying surface conditions but provide an inadequate spatial representation over large areas with more heterogeneous environments, such as eastern China. Therefore, it is necessary to conduct such research at the regional scale.
In recent years, numerical simulations based on remote sensing have been developed as an effective method for evaluating large-scale energy exchange, such as the two-source energy balance model (TSEB; Kustas and Norman 1999); mapping evapotranspiration at high resolution with internalized calibration (METRIC; Allen et al. 2007); the Trapezoid Interpolation Model (TIM; Jiang and Islam 2001); the Surface Energy Balance Algorithm for Land (SEBAL; Bastiaanssen et al. 1998); the Surface Energy Balance System (SEBS; Su 2002; the Simplified Surface Energy Balance Index (S-SEBI; Roerink et al. 2000); and the Surface Energy Balance with Topography Algorithm (SEBTA; Gao et al. 2011). Such models can simulate large-scale land–atmosphere energy exchange, but the simulated results have been inconsistent and are incompatible between the various models under different vegetation growth conditions. In addition, the interaction of material circulation and energy flow in the soil–plant–atmosphere continuum (SPAC) is extremely complex. Furthermore, the biophysical effects arising from changes in land cover exacerbate this complexity (Foley et al. 2003; Kabat et al. 2004). Therefore, the models based on remote sensing have some difficulty in accurately expressing the direct interaction process between land and the atmosphere.
Most previous studies have prioritized typical land-cover changes, such as deforestation (Anderson et al. 2011), grassland grazing (Zhao et al. 2010), and farmland reclamation (e.g., Boucher et al. 2004; Lobell et al. 2006, 2009), rather than combining the land surface water and heat exchanges with the land-cover/land use change at the regional scale. In reality, land-cover conversion does not display a single shift, and the ways in which various land-cover type conversions affect land surface water and heat exchange can vary considerably (Lu and Kueppers 2012). Therefore, the combined effects of interactive land-cover type conversions at the regional scale must be determined. Moreover, a number of previous studies were based on hypothetical land-cover scenarios that were generally in ideal conditions (Pongratz et al. 2006; Ibrom et al. 2007; Beltrán-Przekurat et al. 2012). As a result, such study simulations produced uncertainties when applied to actual land-cover data, although they advanced our knowledge of water and heat exchange processes at the regional scale.
Eastern China stretches across multiple moisture and heat zones and has obvious zonal distributions of water and heat balance. The complex terrain conditions cause the spatial combination of water and heat to be even more complex. In addition, continued population growth and intensified human development activities have greatly changed China’s land-cover/land use patterns, resulting in water and heat balance changes. Thus, studies on water and heat balances in eastern China and their spatial and temporal distribution have important theoretical and practical value for understanding the spatial distribution pattern of hydro-meteorological variables and forecasting the response of the land surface energy budget to future climate change.
Based on remote sensing, a new generation of integrated land surface models that couple the exchange of matter and energy between the land surface and atmosphere are capable of reducing the uncertainties and providing more accurate model simulations. With reliable land-cover data sources, this study, through the rational design of simulation experiments, separates the relative contributions of climate change and land-cover change on land surface energy exchange in eastern China over the past nearly three decades, based on the parameter optimization of an integrated land surface model called the Ecosystem–Atmosphere Simulation Scheme (EASS) model (Chen et al. 2007). The results provide a scientific basis for the accurate assessment of land surface energy exchange in eastern China and the impacts of this exchange on future climate change in the context of global change.
The rest of this paper is structured as follows. A brief overview of the study area, the EASS model, validation criteria and model evaluation, experimental design, and datasets are provided in section 2. Section 3 presents the results and discussion, which includes an evaluation of the model based on several typical ecosystems and evaluation of the simulated regional results, spatiotemporal variation analysis of the land surface energy components under different conditions, an analysis of the contribution of climate change and land-cover change to land surface energy exchange, and some prospects for future work. Finally, we provide a short summary and draw conclusions from our simulation results in section 4.
2. Study area and methodology
a. Study area
We performed this study in eastern China (Fig. 1) between 22° and 51°N latitude and 104° to 135°E longitude because of the availability of land cover and climate data over the past nearly 30 years (1981–2005). The zonal climate is characterized by multiple zones, including subtropical, temperate, and subfrigid zones. In the past nearly three decades, the mean annual temperature has varied from −3° to 22°C across the research area (Fig. 3d). The weather was coldest in January (average temperature of 6°C) and warmest in July (average temperature of 20°C). Snowfall occurred mainly in the north. The precipitation was mainly concentrated in the spring and summer (approximately 65% of total) over the entire area. Spatial distribution of annual mean rainfall in the past nearly 30 years varies greatly, approximately from 101.9 mm yr−1 in the north to 2269.2 mm yr−1 in the south. The northern area has experienced extensive reclamation and farmland expansion from a large area of wasteland over the past nearly 30 years. The central and western regions have suffered from dramatic changes in grassland due to overgrazing from the early 1980s to 1990s and restoration after the later 1990s. The southern region has a relatively large area of forest cover with a typical tropical/subtropical monsoon climate and has undergone deforestation, afforestation, and reforestation. However, the urban land in the eastern region has increased expansively because of the extensive effort put into developing a particularly large-scale urban area in recent years.
Land-cover changes for the four study periods: (a) 1989, (b) 1995, (c) 2000, and (d) 2005. Where the belt is cut, the land surface data are not available. (top right) The boundary of China, within which the area in blue is our study domain.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
Land-cover changes for the four study periods: (a) 1989, (b) 1995, (c) 2000, and (d) 2005. Where the belt is cut, the land surface data are not available. (top right) The boundary of China, within which the area in blue is our study domain.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
Land-cover changes for the four study periods: (a) 1989, (b) 1995, (c) 2000, and (d) 2005. Where the belt is cut, the land surface data are not available. (top right) The boundary of China, within which the area in blue is our study domain.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
b. Model description
The EASS model used in this study is a third-generation land model (Chen et al. 2007) that is based on a single-layer vegetation canopy overlying a seven-layer soil, including physically based energy and moisture fluxes transferred from and through the vegetation canopy. The thermal dynamic in EASS is treated distinctively between vegetation and the underlying ground (Dickinson et al. 2002). Moreover, because “big leaf” models (Jarvis 1995; Sellers et al. 1997) have self-acknowledged limitations, EASS stratifies the vegetation canopy as sunlit and shaded leaves (Melesse et al. 2008; de Pury and Farquhar 1997; Chen et al. 1999). EASS has been referred to as a “two leaf” canopy model (Chen et al. 2007).
Pixel-by-pixel simulation can be performed for EASS over a defined domain, so it can be adopted at different scales based on the available input data. In addition, EASS has flexible spatiotemporal resolutions as long as the input data of each pixel are defined.
The major characteristics of EASS are briefly summarized as follows: (i) Exchange of carbon, energy, and water between the atmosphere and land were fully coupled based on explicit links between photosynthesis, evapotranspiration, and stomatal conductance. (ii) Vegetation cover was treated as a single layer, and the model stratified the vegetation canopy as sunlit and shaded leaves to minimize the biases from the big-leaf assumption. A foliage clumping index (Ω) and leaf area index (LAI) were used to characterize the effects of the three-dimensional canopy structure on radiation, water, and carbon fluxes. (iii) The key geoscience physical parameters of the model were parameterized based on remote sensing quantitative retrieval, which increased the accuracy of the regional simulation. (iv) A multilayer scheme for energy exchange and water transfer through the soil layers and/or snowpack (if present) was introduced in EASS. The number of snow and soil layers and the depth of each layer were defined by the user according to physical soil structures, snow depth, and application objectives. (v) Finally, the dynamics of the snowpack and freeze–thaw cycle in the soil profile were also emphasized in EASS. EASS was forced by near-surface meteorological variables at a reference level zref within the atmospheric boundary layer, including incident solar radiation (rad), surface air temperature (tmp), relative humidity (rhm), wind speed (wnd), and precipitation (pre). For detailed descriptions, please consult the paper by Chen et al. (2007).
c. Validation criteria and model evaluation
Considering that the impact of land-cover change on land surface energy exchange was explored in this study, parameter optimizations at various vegetation functional types were required. The study conducted parameter optimizations mainly based on evergreen broadleaf forest, evergreen needleleaf forest, shrubland, grassland, wetland, and cropland. The selected calibration sites were mainly located in Asia, Europe, and the United States, including Qianyanzhou (CN-Qia; 26.733°N, 115.050°W, evergreen needleleaf forest); Puechabon (FR-Pue; 43.741°N, 3.596°W, evergreen broadleaf forest); Neustift/Stubai Valley (AT-Neu; 47.116°N, 11.320°W, grassland); University of California, Irvine, 1989 burn site (CA-NS6; 55.917°N, 98.964°W, shrubland); Kaamanen wetland (FI-Kaa; 69.141°N, 27.295°W, wetland); and Bondville (US-Bo1; 40.006°N, 88.290°W, cropland) (Table 1).
EASS validation results for the H, LE, and NR of several typical ecosystem types.
We evaluated the simulated results for H, LE, and NR against those from several of the sites mentioned above. Generally, the H, LE, and NR simulated by EASS corresponded well with the measured daily H, LE, and NR, although there were several uncertainties, particularly in the estimation of H (Table 1).
d. Experimental design
After the model was calibrated and evaluated and the sensitivity analysis was conducted, a regional-scale analysis was conducted using land cover for four study periods and long time-series climate data, as described in sections 2e(1) and 2e(2).
To minimize comparison bias between model experiment runs, the model was spun up about 50 times (over 50 years) using the very first year (1981) data to reach an equilibrium state [<1% changes for all state variables (e.g., soil layer temperature, moisture, etc.)]. The equilibrated state variables for each experiment were then taken as the initial values for the following three experiments.
Experiment one was conducted under the “real” scenarios for expressing the impacts of combined climate change and land-cover change on the land surface energy exchange. We made some assumptions that the land cover did not change for a certain period when land-cover data were available. The land-cover data for 1989, 1995, 2000, and 2005 were used for the periods of 1981–1990, 1991–1995, 1996–2000, and 2001–2005, respectively.
Experiment two was designed to discern the impact of climate change only on the land surface energy exchange. We assumed that the land-cover condition and LAI remained unchanged in the past nearly 30 years. Therefore, we used the data in 1989 to calculate the land surface energy fluxes for the period of 1981–2005.
Experiment three was designed to explore the impact of land-cover change only on the land surface energy fluxes. We assumed that the climate condition remained unchanged during the past nearly 30 years. The atmospheric forcing data of 2000 were used for forcing the model with the four land-cover datasets and LAI data in the same year for the years of 1989, 1995, 2000, and 2005.
In the three groups of simulation experiments, the land cover, LAI, and climate data were varied with time, while other inputs were kept constant. The bulk soil properties changed very slowly and thus were not expected to change obviously within the time frame of this study. To make comparison between the three experiments’ results, we only analyzed the simulated results in 1989, 1995, 2000, and 2005. The 1989 run was taken as a reference run, the model results from the three experimental runs were compared with that run, and interannual change of each experimental run was analyzed. The spatial patterns over the study region were analyzed to estimate the impacts of climate change and land-cover change.
e. Input data
1) Land-cover data
The land-cover datasets used in this study were produced by the Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences. The database was generated based on Landsat Thematic Mapper/Enhanced Thematic Mapper (TM/ETM) images, which were geometrically corrected and orthorectified utilizing ground-control points and high-resolution Digital Elevation Model (DEM) data. Extensive field surveys were carried out for validating the interpretation of Landsat TM/ETM images and land-cover classifications, and the classification accuracy assessment was conducted employing a variety of methods and indices (Zhuang et al. 1999). Four periods of land-cover data over the past nearly 30 years have already been generated, including the years 1989, 1995, 2000, and 2005; and the missing scenes were supplemented using the data before or after the years (e.g., 1987/1990 data were used for 1989, 1995/1996 data were used for 1995, and 1999/2000 data were used for 2000) (Liu et al. 2005a,b). To be compatible with the regional climate models, land-cover data were upscaled without a loss of information by conserving the area ratio, considering the rationality of the spatial position, and using similar spatial configuration features. The spatial distribution change during the four study periods is shown in Figs. 1 and 2. In addition, land-cover types were further differentiated to support the required International Geosphere–Biosphere Program (IGBP) land-cover classification system used in the EASS model, including evergreen needleleaf, evergreen broadleaf, deciduous needleleaf, deciduous broadleaf, mixed forest (MF), shrub, grassland, permanent wetlands, crop, urban, bare or sparse vegetation cover, and open water. Based on the existing four periods of land-cover data, this study explored the impacts of land-cover change on land surface energy flux exchange over the past nearly 30 years.
Interannual spatial changes in land cover from the base year (1989) for the four study periods.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
Interannual spatial changes in land cover from the base year (1989) for the four study periods.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
Interannual spatial changes in land cover from the base year (1989) for the four study periods.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
2) Meteorological data
The meteorological data, including incoming shortwave radiation, air temperature, relative humidity, precipitation, and wind speed, were obtained from the medium-range forecast (MRF) Global Flux Archive of the National Centers for Environmental Prediction (NCEP), which is distributed by the National Center for Atmospheric Research (NCAR). The hourly meteorological inputs at a 30-km resolution were spatially and temporally interpolated from the 6-h and daily NCEP dataset. Spatial NCEP data were downscaled to site scale first and then were interpolated into 30 km × 30 km grids using inverse distance weighting (IDW). To interpolate the 6-h interval NCEP reanalysis data into hourly data, different methodologies were used for different variables. The 1-h relative humidity, wind speed, and precipitation data were linearly interpolated from the 6-h NCEP data. Considering the incoming shortwave radiation is larger than the actual observations, especially at higher latitudes, we calibrated daily NCEP data first using the established relationship between NCEP data and the observational data following Hicke (2005), and then the hourly solar radiation values were estimated based on a function of the solar zenith angle. The hourly temperature was determined using the 6-h values and the maximum/minimum values in NCEP data, assuming that the maximum temperature occurred at 1400 local solar time (LST) and its diurnal variation followed the cosine curve trend. Daily meteorological stations data and hourly flux tower site data were used to verify the downscaled meteorological data at the grid level. The interannual spatial and temporal changes in the five variables are shown in Fig. 3, and regional annual average temperature and annual accumulative precipitation of the four periods are shown in Fig. 4.
Interannual spatial change of 1995, 2000, and 2005 from the base year 1989 and annual average spatial distribution of meteorological forcing variables for the four study periods across eastern China, including (a) precipitation (mm yr−1), (b) incident solar radiation (W m−2), (c) relative humidity (%), (d) air temperature (°C), and (e) wind (m s−1).
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
Interannual spatial change of 1995, 2000, and 2005 from the base year 1989 and annual average spatial distribution of meteorological forcing variables for the four study periods across eastern China, including (a) precipitation (mm yr−1), (b) incident solar radiation (W m−2), (c) relative humidity (%), (d) air temperature (°C), and (e) wind (m s−1).
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
Interannual spatial change of 1995, 2000, and 2005 from the base year 1989 and annual average spatial distribution of meteorological forcing variables for the four study periods across eastern China, including (a) precipitation (mm yr−1), (b) incident solar radiation (W m−2), (c) relative humidity (%), (d) air temperature (°C), and (e) wind (m s−1).
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
Changes in annual average temperature (°C) and annual precipitation (mm) in eastern China over the past nearly 30 years.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
Changes in annual average temperature (°C) and annual precipitation (mm) in eastern China over the past nearly 30 years.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
Changes in annual average temperature (°C) and annual precipitation (mm) in eastern China over the past nearly 30 years.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
3) LAI data
The two-week LAI data before 2000 were acquired from the Climate and Vegetation research group, Boston University (http://sites.bu.edu/cliveg/files/2014/08/gimms_list.txt) (Ganguly et al. 2008). The dataset was evaluated through direct comparisons to ground data and intercomparisons with similar datasets. This indirect validation indicated satisfactory agreement with existing LAI products, such as the Moderate Resolution Imaging Spectroradiometer (MODIS). Therefore, the data were selected to fill a vacancy for MODIS LAI products. LAI data from 2000 onward were acquired from the Land–Atmosphere Interaction Research Group at Beijing Normal University (http://globalchange.bnu.edu.cn/research/lai/) (Yuan et al. 2011). These LAI datasets were generated by reprocessing the MODIS LAI products. Compared to the LAI reference maps and MODIS LAI data, these datasets displayed higher accuracy. The original 8- or 1-km LAI data were then resampled to 30 km for the model run. The interannual spatial and temporal changes in LAI are shown in Fig. 5.
Interannual spatial change of 1995, 2000, and 2005 from the base year 1989 and annual average spatial distribution of LAI (m2 m−2) across eastern China for the four study periods.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
Interannual spatial change of 1995, 2000, and 2005 from the base year 1989 and annual average spatial distribution of LAI (m2 m−2) across eastern China for the four study periods.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
Interannual spatial change of 1995, 2000, and 2005 from the base year 1989 and annual average spatial distribution of LAI (m2 m−2) across eastern China for the four study periods.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
4) Soil texture data
A 1:1 000 000 scale soil map of China provided by the Institute of Soil Science, Chinese Academy of Sciences was used as a model input for soil texture data. There were 94 000 soil spots and 7 292 soil profile attributes in the integrated database, and they were used to generate different levels of soil property spatial distributions. The soil profile was divided into 0–10-, 10–20-, 20–30-, and 70-cm depth categories. The original 10 km was resampled to 30 km. These data provided the fraction of sand, clay, and silt for each soil layer and each pixel. Following the proportion of the different components of the composition, soil texture types were classified into 11 categories, corresponding to the soil texture requirements in the model.
3. Results and discussion
a. Model evaluation
Table 1 presents the statistical comparisons of the H, LE, and NR values simulated with EC tower measurements. The coefficients of determination R2 for the six stations representing six different typical ecosystems were above 0.75. The values of indices of agreement d were higher than 0.8, except for H in grassland. (Table 1). The statistical comparisons between simulated and observed values (slope and RMSE) suggest the model performed well, except for grassland and cropland. In addition, the measured land surface energy was not closed at several EC sites, especially FR-Pue and US-Bol. The incomplete closure of measured surface energy will bring uncertainties to model verification. The inconsistencies between modeled and observed values for the FR-Pue and US-Bol sites are much higher than for other sites. This suggests the inconsistencies would also be influenced by the energy closure of measured fluxes. Overall, the EASS model performed well and was acceptable across different ecosystems. For additional evaluation of the EASS model, please consult the paper by Chen et al. (2007).
Currently, there is no relative analysis at the regional scale in the literature about land surface energy budgets in the study area. To evaluate the accuracy of our simulation, the modeled values at a grid were compared with the tower-observed values at six ChinaFLUX sites located in the study area. These comparisons show that the surface energy exchange components agree with the observations well for most ecosystem types, except for grassland and cropland (Table 2). Most of the statistical R2 and d indices are over 0.75 and 0.8, respectively. Slopes are almost in the accepted range statistically. Aside from bias and RMSE over grassland and farmland ecosystem types, which are overestimated, other ecosystem types are almost less. The considerable biases for grassland and farmland probably resulted from (1) bias in parameterization; (2) gap-filling errors from the EC data; and (3) land surface energy unclosure in the EC measurements (Li et al. 2005).
Comparison between modeled corresponding raster values and observed site annual values in 2005 from ChinaFLUX located in the study domain.
b. Spatiotemporal variation analysis of the land surface energy components under the combined impacts of land-cover change and climate change
From interannual variation, each energy balance variable continued to increase until 2000. However, NR slightly decreased, LE dramatically dropped, and H abruptly increased after 2000 (Fig. 6d). LE was higher than H throughout the study period, mainly because precipitation was relatively abundant in eastern China, and vegetation was dominated by farmland and woodland. For the regional average seasonal variation, H displayed a sinusoidal trend, with the peak value (approximately 48 W m−2) appearing in March or April, the lowest value (approximately 15 W m−2) occurring in June or July, and a second peak (approximately 27 W m−2) occurring in November (Fig. 6a). This trend corresponded to the single-point observations obtained for the farmland ecosystem in China (Li et al. 2007). The trend with double peaks for H is possibly because a large portion of energy partitioned into LE (rainfall season) during April to October and therefore reduced the H in summer and produced double peaks in spring and fall. The seasonal variation of each month in 2005 was much higher than in other years, because the regional average annual precipitation was reduced in 2005. As shown in Fig. 2, the farmland expanded considerably in 2005, leading to a lower roughness length, decreased LAI, and a subsequently increased H. Comparatively, LE had only one peak (approximately 100 W m−2), which occurred in June or July (Fig. 6b). Furthermore, the peak (approximately 120 W m−2) in 2000 was the highest, because forest, as the dominant vegetation type in the study area, had much higher evapotranspiration than the farmland and other vegetation types. However, the seasonal variations patterns for H, LE, and NR were generally consistent in the four periods.
(a),(b),(c) Seasonal and (d) interannual variation of simulated H, LE, and NR and variation in eastern China for the four study periods under the influence of the combined impacts of land-cover change and climate change.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
(a),(b),(c) Seasonal and (d) interannual variation of simulated H, LE, and NR and variation in eastern China for the four study periods under the influence of the combined impacts of land-cover change and climate change.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
(a),(b),(c) Seasonal and (d) interannual variation of simulated H, LE, and NR and variation in eastern China for the four study periods under the influence of the combined impacts of land-cover change and climate change.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
The spatial pattern changes of annual H, LE, and NR for 1995, 2000, and 2005 were analyzed using the annual average spatial distribution of the baseline year (1989). Figure 7 presents the changes in spatial patterns over the past nearly 30 years. Except for the increased spread of NR to the southern regions in 2005, the interannual variation in the spatial pattern of NR over the past nearly 30 years was relatively small (Fig. 7a). The spatial variation, corresponding with seasonal changes, was in a good agreement with the spatial distribution of air temperature and incident solar radiation, which were the dominant factors affecting the interannual variability of NR. The largest changes in H between 1989 and 1995 were observed in the Henan and Anhui provinces; the maximum of 25 W m−2 resulted from the increased incident solar radiation in the regions (Fig. 3b) and the increased NR (Fig. 7a). In 2000, an obvious increase in the regional H mainly occurred in Chongqing city and the Guizhou and Sichuan provinces (12 W m−2). Although the increase in 2005 was not as large as that in 2000, the increase was extended throughout almost the entire study area. Overall, the distributions of increases in H over the past nearly 30 years changed from being relatively concentrated to spreading throughout almost the entire study area (Fig. 7b). The spatial and temporal changes of LE were largely different from those of H. A clear increase in LE occurred in the middle of the study area. The pattern was mainly controlled by the interannual spatial variation of incident solar radiation and LAI. The value of NR increased as more incident solar radiation was received; similarly, the value of the stomatal conductance was higher with further increases in the LAI, thus increasing the amount of transpiration. Although the precipitation over the past nearly 30 years decreased, the change of simulated soil moisture was not as obvious as in the previous studies (Wang et al. 2009; Bing et al. 2012). The changes of LE for farmland are likely to be affected by the large area of irrigation in eastern China.
Interannual spatiotemporal changes in (a) NR (W m−2), (b) H (W m−2), and (c) LE (W m−2) for the years 1995, 2000, and 2005 from the baseline year of 1989 and the annual average of the four study periods across eastern China under the influence of the combined impacts of land-cover change and climate change.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
Interannual spatiotemporal changes in (a) NR (W m−2), (b) H (W m−2), and (c) LE (W m−2) for the years 1995, 2000, and 2005 from the baseline year of 1989 and the annual average of the four study periods across eastern China under the influence of the combined impacts of land-cover change and climate change.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
Interannual spatiotemporal changes in (a) NR (W m−2), (b) H (W m−2), and (c) LE (W m−2) for the years 1995, 2000, and 2005 from the baseline year of 1989 and the annual average of the four study periods across eastern China under the influence of the combined impacts of land-cover change and climate change.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
Under the combined effects of climate change and land-cover changes, there was no discernible relationship between individual energy components and land cover from the spatial distribution pattern, implying that the land-cover change was not a control factor of the interannual variability for H, LE, and NR. In other words, the spatial distribution of land surface energy fluxes did not respond to land-cover spatial distribution patterns across the study area because of the dominant effects of climate change.
c. Contribution of land-cover change to land surface energy partitioning
The regional average seasonal variations of NR, LE, and H in the four periods were nearly consistent with the simulated results under the combined effects of land-cover change and climate change (Figs. 6 and 8). The monthly H in 1995 was almost coincident with that in the year 1989. In 2000, the monthly H increased slightly, but the monthly H in 2005 was much higher than that during the other three periods (Fig. 8a). One very likely reason is that the cropland area increased, which was transferred from wetlands and woodlands in this year (Fig. 2). On the contrary, the LE in 1995 and 1989 almost reached their maximum values, and LE was the lowest in 2005. The largest difference of LE between the four periods occurred in the peak period (Fig. 8b). Monthly NR fluctuated but did not show large differences between the four periods (Fig. 8c).
Seasonal variation of the simulated (a) H, (b) LE, and (c) NR in eastern China under the influence of land-cover change for the four study periods.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
Seasonal variation of the simulated (a) H, (b) LE, and (c) NR in eastern China under the influence of land-cover change for the four study periods.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
Seasonal variation of the simulated (a) H, (b) LE, and (c) NR in eastern China under the influence of land-cover change for the four study periods.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
The impacts of land-cover change on the spatial distribution of the land surface energy fluxes are shown in Fig. 9. The differences of annual NR values is small (<=5 W m−2) between the four periods (Fig. 9a). The interannual variations of the spatial distributions of H and LE were very discrete and with a mosaic style. There is no an apparent regularity throughout the entire study area (Figs. 9b,c). These mosaic-style spatial distribution characteristics of interannual variations may have occurred because the LAI did not change synchronously with the land-cover change. In other words, these patterns were caused by the uncertainty in the interannual LAI. When compared with the distribution of the interannual variation of the LAI (Fig. 5), we can further affirm that the H and LE distribution changes were mainly controlled by the LAI.
As in Fig. 7, but with the impact of land-cover change only.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
As in Fig. 7, but with the impact of land-cover change only.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
As in Fig. 7, but with the impact of land-cover change only.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
Conversions of different land-cover types are generally associated with the pronounced changes in albedo and consequent changes in land surface energy partitioning. The impact of the EASS model-simulated primary land-cover type conversions from 1989 to 2005 in the study area on H, LE, and NR are summarized in Table 3. These four land-cover change categories were the largest land-cover conversions, accounting for 1.21%, 0.69%, 1.08%, and 0.90% of the entire study area, respectively. In Table 3, we can see that for most conversions from one land cover to another, the energy was kept near balance, except for the conversion from shrub to grassland. One possible reason for this is due to bias in parameterization in our model. This can also be explained with Table 1. The RMSE of grassland in Table 1 is very large, leading to energy imbalance when shrub was converted to grassland. For all of these four conversion scenarios, the changes in H were larger than the changes in LE. The conversion of cropland (CRO) to evergreen needle forest (ENF) led to the greatest decrease in H (5.11 W m−2, 19.39%) within the land-cover change scenarios. Similarly, this change also triggered about a 7.44% (2.66 W m−2) increase in LE. The conversion with the least impact on H was the conversion from deciduous broadleaf forests (DBF) to cropland, with only about 4.83% (2.39 W m−2) increase in H and 3.92% (2.12 W m−2) decrease in LE. NR increased by 1.47% (2.02 W m−2) in that scenario. The simulated abnormal changes can be attributed to the uncertainty in the LAI and model parameterization due to the land surface energy unclosure of site-measured flux data and error of the model itself. By comparing the contribution of these conversion types for the entire study area, we observed that the conversion from shrub to grassland has the greatest effects on H, resulting in approximately 3.65 × 108 kW reduction. The greatest influence on LE (1.54 × 108 kW) and NR (1.54 × 108 kW) was derived from the conversion from cropland to evergreen coniferous forest. The conversion from deciduous broadleaf forest to cropland exerts minimal influence on H (1.03 × 108 kW), whereas minimal impact on LE (0.65 × 108 kW) and NR (0.09 × 108 kW) were noted for the conversion from cropland to grassland (GRA) and from shrub (SHR) to grassland, respectively.
Several primary land-cover conversions and their impacts on the annual average H, LE, and NR. The changes (%) were calculated using the second land-cover type in the year 2005 minus the first land-cover type in the year 1989 (downward arrow means a decrease and an upward arrow means an increase).
There are a few published studies assessing the impacts on land surface energy exchange from the conversion of land-cover types (Twine et al. 2004; Sterling 2005; Mao and Cherkauer 2009; Sterling et al. 2013; Beltrán-Przekurat et al. 2012), and these results presented large differences. Under different local climates and different land surface conditions, the change in land surface energy budget resulting from different conversion categories varies greatly not only in magnitudes but also in directions of change. Moreover, in croplands, there is a great difference between different types of crops (Beltrán-Przekurat et al. 2012). Our modeling experiments focus on the impacts on the land surface energy exchanges of main land-cover conversions (Table 3). In general, forests produce more evapotranspiration than other land-cover types because of their larger canopy cover. Within our land-cover transfer tests, conversions from cropland to evergreen coniferous forest and from deciduous broadleaf forest to cropland show a similar variation trend in LE. LE increased from shrub to grassland because the conversion process occurred mainly in the midwest of the study area, where the grassland experienced a restoration over the past nearly 30 years; therefore, high expansion of grassland led to an increase in LE. These results are consistent with the findings of Li et al. (2009).
d. Contributions of climate change to land surface energy partitioning
As shown in Fig. 10, the impact of climate change on the distribution of the land surface energy fluxes was considerable. The monthly H was almost coincident for the four periods (Fig. 10a). There were very large fluctuations in LE for particular months within the four periods, particularly in the peak periods of June or July (Fig. 10b). Similarly, NR displayed a small fluctuation in corresponding months between the four periods, but compared with LE, these minor changes highlighted the high level of consistency in NR (Fig. 10c). Comparing with the seasonal variations under the combined effects of climate change and land-cover changes, we found that changes in energy fluxes influenced by climate change were similar to those influenced by combined climate and land-cover changes, except for H.
As in Fig. 8, but with the impact of climate change.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
As in Fig. 8, but with the impact of climate change.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
As in Fig. 8, but with the impact of climate change.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
We conducted further analysis to quantify the impacts of climate change on the land surface energy budget at the regional scale. From the spatial distribution of interannual variations (Fig. 11), we can see that the impact of meteorological forcing variables on land surface energy exchange was rather discernible. The change in the spatial distribution pattern of NR displayed a pronounced increase in the central region (Fig. 11a). The interannual variation of H was consistent with the simulated result under combined effects (Fig. 11b). However, the degree of coincidence of LE was not as clear as for H. In particular, the spatial distributions of LE in 2005 were particularly varied between these results and the simulation under combined effects (Fig. 11c). These results also corresponded to the seasonal variation. To clearly demonstrate the spatial control of climate change on the land surface energy budget, we compared the modeled results with the meteorological forcing data and found that NR was mainly controlled by incoming shortwave radiation. Annual variation in NR at the center of the study area was larger than in other regions, but the mean annual NR in the north central region was obviously higher than in the other regions. Similar spatial patterns could be found between modeled LE (Fig. 11c) and incoming shortwave radiation (Fig. 3b). However, except for the effect of incoming shortwave radiation, the spatial patterns of H were similar to those of air temperature. H was relatively lower in the regions with higher multiple-year average of air temperature (Figs. 11b, 3d).
As in Fig. 7, but with the impact of climate change only.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
As in Fig. 7, but with the impact of climate change only.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
As in Fig. 7, but with the impact of climate change only.
Citation: Journal of Climate 27, 24; 10.1175/JCLI-D-13-00492.1
Similar to the analysis of the impact of land-cover change on the land surface energy fluxes, for the regions where great land-cover changes occurred, we analyzed the impact of climate change on the land surface energy fluxes. Compared with those in the year 1989, both the air temperature and incoming solar radiation increased, and the precipitation decreased in 2005. As a result, simulated net radiation, sensible heat flux, and latent heat fluxes increased (see analysis 1 in Table 4). In addition, we also studied the changes in land surface fluxes under the impact of climate change overall in the whole study area. Results show that the precipitation decreased, the air temperature increased, and solar radiation increased in 2005 contrasting with those in the year 1989, which consequently lead to increasing in net radiation and sensible heat flux and decreasing in latent heat flux (see analysis 2 in Table 4).
Impact of climate change occurring over the same region as in Table 3 and over the whole region on the annual average H, LE, and NR. The changes (%) were calculated using the second land-cover type in the year 2005 minus the first land-cover type in the year 1989 (downward arrow means a decrease and an upward arrow means an increase).
e. Identifying the respective contributions of climate change and land-cover change to land surface energy partitioning
Similar to the analysis of the impacts of various land-cover type conversions on land surface energy fluxes, the contributions of climate change and land-cover change to the simulated land surface energy fluxes were analyzed by employing land surface flux changes between 1989 and 2005 (Table 5). The entire regional annual mean changes in H, LE, and NR were 1.67, 5.61, and 8.71 W m−2, respectively, under the combined effects of climate change and land-cover change. According to the energy balance principle, in addition to a small portion of the NR assigned to the soil heat flux, the main energy was allocated to LE (64.41%). Given the respective contribution (or contribution rate) of the two driving forces, the contributions of climate change to each component of land surface energy were considerably higher than the contribution of land-cover change, accounting for 83.89% (1.77 W m−2), 99.8% (9.51 W m−2), and 97.83% (8.56 W m−2) for H, LE, and NR, respectively. We also observed that the sum of the individual impact of the two driving forces was not exactly equal to the simulated results of the combined effects, implying that land surface energy balance variables might display additive or subtractive impacts of climate change and land-cover change. In other words, the regional effects of land-cover change on the land surface energy budget can offset or magnify the effects of climate change.
Individual and combined contribution rate from climate change and land-cover change to land surface energy components in eastern China.
However, except for these two main factors, other factors, such as soil, could also affect the land surface energy budget. These effects are not included in our model experiments. Other uncertainties may have resulted from the assumptions of our modeling framework. It should be noted that the model experiments designed to discern the proportional contributions of climate change and land-cover change on the land surface energy exchange are based on the assumption that land-cover change is independent of climate change. In fact, there are interactions and feedbacks between the land-cover change and the climate system, and the energy fluxes might show additive or subtractive impacts of climate change under changing land-cover conditions (Mishra et al. 2010). This is not considered in our modeling experiments.
f. Future work and challenges
Although some preliminary results were obtained in this research, a number of important challenges were identified for future studies by our group: (i) Existing simulated results will be analyzed further so as to examine the cause of model–EC divergences and correct the model further, especially to optimize the simulation results for farmland and grassland ecosystem types. The latter obtained EC observation data in the study area were used to verify the model to ensure the accuracy and reliability of EASS in the regional scales’ studies. (ii) Parameterization schemes integrating remote sensing technologies will be designed to optimize the key parameters (such as albedo, roughness length, and tree height) affecting land surface energy exchange at a larger spatial scale to better improve the simulation accuracy. (iii) A number of observed profile tests will be conducted. The footprint model will be introduced to analyze the flux gradient at different transverse sections, making it possible to capture the spatial variation information of each energy component more accurately and effectively. (iv) Finally, a combination of remote sensing retrieving, the mechanistic models, and observation data assimilation will be applied to extend the discrete multisource stations’ profile data to the landscape or regional scale and quantize the spatial and temporal dynamic distribution of water and heat fluxes at different underlying surfaces.
4. Conclusions
We can draw several conclusions from this study: (i) The EASS model has the capacity to simulate different vegetation types. (ii) The interannual variability of H and LE in eastern China over the past nearly 30 years displayed an upward trend before 2000 but tended to decline after that time. (iii) The seasonal variation of the H presented a sinusoidal curve, with the highest peak in March or April and the lowest point in July or August. The seasonal variation patterns of the LE and NR displayed a single peak in June or July. (iv) Changes in energy flux components resulting from different land-cover type conversions vary considerably; thus, the contribution of energy balance components in the study area also changes considerably. (v) Climate change and its impacts on land surface energy balance have been dominant in eastern China in the past three decades, with climate change accounting for more than 80% of the change in land surface energy balance. In term of the impacts of land-cover changes rather than climate change, seasonal variation, especially for H, between each period has obvious differences. Through this study, we conclude that reasonable designed modeling experiments can discern the comparative contributions of climate change and land-cover change on land surface energy exchanges. This will help us to better understand climate change at regional and global scales and provide a reference for accurately predicting future climate change.
Acknowledgments
This work was supported by the National Basic Research Program of China (973 Program) (No. 2010CB950902), the Research Plan of LREIS, CAS (Grant O88RA900PA), the Key Project for the Strategic Science Plan in IGSNRR, CAS (Grant 2012ZD010), the research Grant (41271116) funded by the National Science Foundation of China, a Research Plan of LREIS (O88RA900KA), CAS, the “One Hundred Talents” program funded by the Chinese Academy of Sciences, the Strategic Priority Research Program “Climate Change: Carbon Budget and Related Issues” of the Chinese Academy of Sciences (Grant XDA05040403), and the National High Technology Research and Development Program of China (Grant 2013AA122002). We acknowledge the agencies that supported the operations at the flux towers used here, which are parts of FLUXNET and ChinaFLUX.
REFERENCES
Allen, R. G., M. Tasumi, and R. Trezza, 2007: Satellite-based energy balance for mapping evapotranspiration with internalized calibration (METRIC)—Model. J. Irrig. Drain. Eng., 133, 380–394, doi:10.1061/(ASCE)0733-9437(2007)133:4(380)).
Anderson, R. G., and Coauthors, 2011: Biophysical considerations in forestry for climate protection. Front. Ecol. Environ., 9, 174–182, doi:10.1890/090179.
Baldocchi, D., and Coauthors, 2001: FLUXNET: A new tool to study the temporal and spatial variability of ecosystem-scale carbon dioxide, water vapor, and energy flux densities. Bull. Amer. Meteor. Soc., 82, 2415–2434, doi:10.1175/1520-0477(2001)082<2415:FANTTS>2.3.CO;2.
Bastiaanssen, W. G. M., H. Pelgrum, J. Wang, Y. Ma, J. Moreno, G. J. Roerink, and T. A. van der Wal, 1998: A remote sensing surface energy balance algorithm for land (SEBAL). Part 2: Validation. J. Hydrol., 212–213, 198–212, doi:10.1016/S0022-1694(98)00254-6.
Beltrán-Przekurat, A., R. A. Pielke Sr., J. L. Eastman, and M. B. Coughenour, 2012: Modeling the effects of land-use/land-cover changes on the near-surface atmosphere in southern South America. Int. J. Climatol., 32, 1206–1225, doi:10.1002/joc.2346.
Bing, L. F., H. B. Su, Q. Q. Shao, and J. Y. Liu, 2012: Changing characteristic of land surface evapotranspiration and soil moisture in China during the past 30 years. J. Geo-Inf. Sci.,14, 1–13, doi:10.3724/SP.J.1047.2012.00001.
Boucher, O., G. Myhre, and A. Myhre, 2004: Direct human influence of irrigation on atmospheric water vapour and climate. Climate Dyn., 22, 597–603, doi:10.1007/s00382-004-0402-4.
Chen, B., J. Chen, and W. Ju, 2007: Remote sensing based ecosystem–atmosphere simulation Scheme (EASS)—Model formulation and test with multiple-year data. Ecol. Modell., 209, 277–300, doi:10.1016/j.ecolmodel.2007.06.032.
Chen, J. M., J. Liu, J. Cihlar, and M. L. Goulden, 1999: Daily canopy photosynthesis model through temporal and spatial scaling for remote sensing applications. Ecol. Modell., 124, 99–119, doi:10.1016/S0304-3800(99)00156-8.
Choi, M., W. P. Kustas, M. C. Anderson, R. G. Allen, F. Li, and J. H. Kjaersgaard, 2009: An intercomparison of three remote sensing-based surface energy balance algorithms over a corn and soybean production region (Iowa, U.S.) during SMACEX. Agric. For. Meteor., 149, 2082–2097, doi:10.1016/j.agrformet.2009.07.002.
Claussen, M., V. Brovkin, and A. Ganopolski, 2001: Biogeophysical versus biogeochemical feedbacks of large-scale land cover change. Geophys. Res. Lett., 28, 1011–1014, doi:10.1029/2000GL012471.
de Pury, D., and G. Farquhar, 1997: Simple scaling of photosynthesis from leaves to canopies without the errors of big-leaf models. Plant Cell Environ., 20, 537–557, doi:10.1111/j.1365-3040.1997.00094.x.
Dickinson, R. E., and Coauthors, 2002: Nitrogen controls on climate model evapotranspiration. J. Climate, 15, 278–295, doi:10.1175/1520-0442(2002)015<0278:NCOCME>2.0.CO;2.
Falge, E., and Coauthors, 2005: Comparison of surface energy exchange models with eddy flux data in forest and grassland ecosystems of Germany. Ecol. Modell., 188, 174–216, doi:10.1016/j.ecolmodel.2005.01.057.
Foley, J. A., M. H. Costa, C. Delire, N. Ramankutty, and P. Snyder, 2003: Green surprise? How terrestrial ecosystems could affect earth’s climate. Front. Ecol. Environ., 1, 38–44, doi:10.2307/3867963.
Ganguly, S., A. Samanta, M. A. Schull, N. V. Shabanov, C. Milesi, R. R. Nemani, Y. Knyazikhin, and R. B. Myneni, 2008: Generating vegetation leaf area index Earth system data record from multiple sensors. Part 2: Implementation, analysis and validation. Remote Sens. Environ., 112, 4318–4332, doi:10.1016/j.rse.2008.07.013.
Gao, Z., C. Liu, W. Gao, and N.-B. Chang, 2011: A coupled remote sensing and the Surface Energy Balance with Topography Algorithm (SEBTA) to estimate actual evapotranspiration over heterogeneous terrain. Hydrol. Earth Syst. Sci., 15, 119–139, doi:10.5194/hess-15-119-2011.
Gu, L., and Coauthors, 2006: Direct and indirect effects of atmospheric conditions and soil moisture on surface energy partitioning revealed by a prolonged drought at a temperate forest site. J. Geophys. Res., 111, D16102, doi:10.1029/2006JD007161.
Hicke, J. A., 2005: NCEP and GISS solar radiation data sets available for ecosystem modeling: Description, differences, and impacts on net primary production. Global Biogeochem. Cycles,19, GB2006, doi:10.1029/2004GB002391.
Ibrom, A., and Coauthors, 2007: Effects of land-use change on matter and energy exchange between ecosystems in the rain forest margin and the atmosphere. Stability of Tropical Rainforest Margins: Linking Ecological, Economic and Social Constraints of Land Use and Conservation, T. Tscharntke, et al., Eds., Springer-Verlag, 463–492.
Jarvis, P. G., 1995: Scaling processes and problems. Plant Cell Environ., 18, 1079–1089, doi:10.1111/j.1365-3040.1995.tb00620.x.
Jiang, L., and S. Islam, 2001: Estimation of surface evaporation map over Southern Great Plains using remote sensing data. Water Resour. Res., 37, 329–340, doi:10.1029/2000WR900255.
Kabat, P., and Coauthors, 2004: Vegetation, Water, Humans and the Climate: A New Perspective on an Interactive System. Springer, 566 pp.
Kustas, W. P., and J. M. Norman, 1999: Evaluation of soil and vegetation heat flux predictions using a simple two-source model with radiometric temperatures for partial canopy cover. Agric. For. Meteor., 94, 13–29, doi:10.1016/S0168-1923(99)00005-2.
Li, Y. J., Z. Xu, Y. Wang, L. Zhou, and G. Zhou, 2007: Latent and sensible heat fluxes and energy balance in a maize agroecosystem. Chin. J. Plant Ecol.,31, 1132–1144.
Li, Z., W.-Z. Liu, X.-C. Zhang, and F.-L. Zheng, 2009: Impacts of land use change and climate variability on hydrology in an agricultural catchment on the Loess Plateau of China. J. Hydrol., 377, 35–42, doi:10.1016/j.jhydrol.2009.08.007.
Li, Z. Q., G. R. Yu, X. F. Wen, L. M. Zhang, C. Y. Ren, and Y. L. Fu, 2005: Energy balance closure at ChinaFLUX sites. Sci. China, 48D,54–72.
Liu, J., M. Liu, H. Q. Tian, D. Zhuang, Z. Zhang, W. Zhang, X. Tang, and X. Deng, 2005a: Spatial and temporal patterns of China’s cropland during 1990–2000: An analysis based on Landsat TM data. Remote Sens. Environ., 98, 442–456, doi:10.1016/j.rse.2005.08.012.
Liu, J., H. Tian, M. Liu, D. Zhuang, J. M. Melillo, and Z. Zhang, 2005b: China’s changing landscape during the 1990s: Large-scale land transformations estimated with satellite data. Geophys. Res. Lett., 32, L02405, doi:10.1029/2004GL021649.
Lobell, D., G. Bala, and P. Duffy, 2006: Biogeophysical impacts of cropland management changes on climate. Geophys. Res. Lett., 33, L06708, doi:10.1029/2005GL025492.
Lobell, D., G. Bala, A. Mirin, T. Phillips, R. Maxwell, and D. Rotman, 2009: Regional differences in the influence of irrigation on climate. J. Climate, 22, 2248–2255, doi:10.1175/2008JCLI2703.1.
Lu, Y., and L. M. Kueppers, 2012: Surface energy partitioning over four dominant vegetation types across the United States in a coupled regional climate model (Weather Research and Forecasting Model 3—Community Land Model 3.5). J. Geophys. Res., 117, D06111, doi:10.1029/2011JD016991.
Mahmood, R., and Coauthors, 2010: Impacts of land use/land cover change on climate and future research priorities. Bull. Amer. Meteor. Soc., 91, 37–46, doi:10.1175/2009BAMS2769.1.
Mao, D., and K. A. Cherkauer, 2009: Impacts of land-use change on hydrologic responses in the Great Lakes region. J. Hydrol., 374, 71–82, doi:10.1016/j.jhydrol.2009.06.016.
Mauder, M., C. Liebethal, M. Göckede, J.-P. Leps, F. Beyrich, and T. Foken, 2006: Processing and quality control of EC data during LITFASS-2003. Bound.-Layer Meteor., 121, 67–88, doi:10.1007/s10546-006-9094-0.
Melesse, A., W. Abtew, and T. Desalegne, 2008: Flow analysis and characterization of the Blue Nile River basin system. Proc. Workshop on Hydrology and Ecology of the Nile River Basin under Extreme Conditions, Addis Ababa, Ethiopia, National Science Foundation, 113–126.
Mishra, V., K. A. Cherkauer, D. Niyogi, M. Lei, B. C. Pijanowski, D. K. Ray, L. C. Bowling, and G. Yang, 2010: A regional scale assessment of land use/land cover and climatic changes on water and energy cycle in the upper Midwest United States. Int. J. Climatol., 30, 2025–2044, doi:10.1002/joc.2095.
Oki, T., E. M. Blyth, E. H. Berbery, and D. Alcaraz-Segura, 2013: Land use and land cover changes and their impacts on hydroclimate, ecosystems and society. Climate Science for Serving Society: Research, Modeling and Prediction Priorities, G. R. Asrar and J. W. Hurrell, Eds., Springer, 185–203, doi:10.1007/978-94-007-6692-1_7.
Oncley, S. P., and Coauthors, 2007: The energy balance experiment EBEX-2000. Part I: Overview and energy balance. Bound.-Layer Meteor., 123, 1–28, doi:10.1007/s10546-007-9161-1.
Pielke, R. A., Sr., G. Marland, R. A. Betts, T. N. Chase, J. L. Eastman, J. O. Niles, D. Niyogi, and S. Running, 2002: The influence of land-use change and landscape dynamics on the climate system: Relevance to climate-change policy beyond the radiative effect of greenhouse gases. Philos. Trans. Roy. Soc. London, A360, 1705–1719, doi:10.1098/rsta.2002.1027.
Pongratz, J., L. Bounoua, R. S. DeFries, D. C. Morton, L. O. Anderson, W. Mauser, and C. A. Klink, 2006: The impact of land cover change on surface energy and water balance in Mato Grosso, Brazil. Earth Interact., 10, doi:10.1175/EI176.1.
Roerink, G. J., Z. Su, and M. Menenti, 2000: S-SEBI: A simple remote sensing algorithm to estimate the surface energy balance. Phys. Chem. Earth, B25, 147–157, doi:10.1016/S1464-1909(99)00128-8.
Sellers, P. J., and Coauthors, 1997: Modeling the exchanges of energy, water, and carbon between continents and the atmosphere. Science, 275, 502–509, doi:10.1126/science.275.5299.502.
Sterling, S. M., 2005: The impact of anthropogenic global land cover transformation on the land–atmosphere fluxes of the water and carbon cycles. Ph.D. dissertation, Duke University, 424 pp.
Sterling, S. M., A. Ducharne, and J. Polcher, 2013: The impact of global land-cover change on the terrestrial water cycle. Nat. Climate Change, 3, 385–390, doi:10.1038/nclimate1690.
Su, Z., 2002: The Surface Energy Balance System (SEBS) for estimation of turbulent heat fluxes. Hydrol. Earth Syst. Sci., 6, 85–100, doi:10.5194/hess-6-85-2002.
Twine, T. E., C. J. Kucharik, and J. A. Foley, 2004: Effects of land cover change on the energy and water balance of the Mississippi River basin. J. Hydrometeor., 5, 640–655, doi:10.1175/1525-7541(2004)005<0640:EOLCCO>2.0.CO;2.
von Randow, C., and Coauthors, 2004: Comparative measurements and seasonal variations in energy and carbon exchange over forest and pasture in South West Amazonia. Theor. Appl. Climatol., 78, 5–26, doi:10.1007/s00704-004-0041-z.
Wang, X. T., W. D. Guo, Z. Zhong, and X. Y. Cui, 2009: Long term trends of soil moisture and temperature change in East China in relationship with climate background. Adv. Earth Sci., 24, 181–191.
Willmott, C. J., 1981: On the validation of models. Phys. Geogr., 2, 184–194, doi:10.1080/02723646.1981.10642213.
Wilson, K. B., and Coauthors, 2002: Energy partitioning between latent and sensible heat flux during the warm season at FLUXNET sites. Water Resour. Res., 38, 1294, doi:10.1029/2001WR000989.
Yuan, H., Y. Dai, Z. Xiao, D. Ji, and W. Shangguan, 2011: Reprocessing the MODIS Leaf Area Index products for land surface and climate modelling. Remote Sens. Environ., 115, 1171–1187, doi:10.1016/j.rse.2011.01.001.
Zhang, Q., S. Wang, and G. Wei, 2003: A study on physical parameters of local land-surface processes on the Gobi in Northwest China. Chin. J. Geophys., 46, 883–895, doi:10.1002/cjg2.408.
Zhao, Y., and Coauthors, 2010: Modeling grazing effects on coupled water and heat fluxes in Inner Mongolia grassland. Soil Tillage Res., 109, 75–86, doi:10.1016/j.still.2010.04.005.
Zhuang, D., J. Liu, and M. Liu, 1999: Research activities on land-use/land cover in the past ten years in China using space technology. Chin. Geogr. Sci., 9, 330–334, doi:10.1007/s11769-999-0006-3.