1. Introduction
Land use/cover change (LUCC) has been widely investigated in the past few decades, and it has been found that more than half of the land surface on Earth has been altered by human activities (Pitman et al. 2009; Baldocchi 2014) such as deforestation, agricultural expansion, and urbanization. Forest losses amplify the diurnal temperature variation and increase the average and maximum air temperature (Alkama and Cescatti 2016), they also increase catchment erosion and nutrient loss, surface runoff, and annual discharges (Woodward et al. 2014). Extensive urbanization creates an urban heat island effect (McCarthy et al. 2010; Lin et al. 2016) and affects precipitation P considerably (Wang et al. 2018). However, the effects of anthropogenic land use on local climate are complicated and differ between regions.
LUCC can directly affect terrestrial evapotranspiration (ET): recent studies have shown that large-scale vegetation restoration has occurred in many regions due to human intervention, resulting in great impact on ET and surface water balance (Murray et al. 2012; Zhang et al. 2018; Shao et al. 2019). Moisture flux from the land surface determines the overlying distribution of atmospheric water vapor, cloud properties, and precipitation (Wang-Erlandsson et al. 2014; Zemp et al. 2014; X. L. He et al. 2020). It is essential to understand how evapotranspiration interacts with the land surface and atmospheric processes, because ET plays a decisive role in mediating hydrological flows and atmospheric feedback (Dore 2005; Gordon et al. 2005; Rockström et al. 2009; Trenberth 2011; Bagley et al. 2012).
The phenomenon of moisture from evapotranspiration entering the atmosphere and returning as precipitation is known as water recycling (Brubaker et al. 1993; Trenberth 1999; Dominguez et al. 2006), and needs to be taken into account when considering water security and land use management (Berger et al. 2014; Keys et al. 2017). In addition, moisture recycling estimates can be used reliably to indicate the sensitivity of climate to land use change and land–atmosphere feedback. The recycling strength has been considered as an indicator which controls soil moisture and local ET dynamics on regional climate (Eltahir and Bras 1994; Burde and Zangvil 2001). The boundary layer and mesoscale circulation perturbation caused by LUCC may lead to a change in surface ET and can increase or decrease P locally (Seneviratne et al. 2010; Guillod et al. 2015; Wang-Erlandsson et al. 2018). Observational and modeling studies show that afforestation can increase convection and P in the tropics (Woodcock 1992; Blyth et al. 1994; Trenberth 1999). Besides, simulations have also indicated that P increases as the vegetation density increases (Mintz 1984; Clark and Arritt 1995) and dynamic vegetation feedback increases the magnitude of the predicted rainfall increase later in the rainy season in West Africa (Wang and Alo 2012). Lee et al. (2015) indicated that the positive effect of vegetation cover through agricultural activities on regional precipitation could lead to a positive feedback between the vegetation and climate in the water-limited Sahel region. The vegetation change, especially the replacement of forests with croplands, can cause local or regional climate changes as significant as the enhancement of atmospheric greenhouse gases by human activities (McPherson 2007). Paul et al. (2016) found that deforestation results in weakening of the Indian summer monsoon rainfall (ISMR) because of the decrease in ET and subsequent decrease in the recycled component of precipitation in Indian. A study by Keys et al. (2012) suggested that livelihoods in some semiarid regions are particularly affected by the change in upwind moisture source regions, and the social dynamics of moisture recycling has also been investigated in Mongolia, Niger, and Bolivia (Keys and Wang-Erlandsson 2018). Furthermore, the change in moisture transport and recycling can trigger disastrous weather, such as drought, locally or in adjacent regions (Herrera-Estrada et al. 2019; Roy et al. 2019). A classic example of the land cover change causing drought occurred in the 1960s–80s in the Sahel. A series of modeling and observational efforts have investigated the influence of land degradation on the decadal drought in Sahel (Charney 1975; Zeng et al. 1999; Wang et al. 2004; Yu et al. 2017).
Arid and semiarid regions account for over 40% of the global land and supports over 2.5 billion people (Reynolds et al. 2007). In these regions, particularly those reliant on rainfed agricultural production, land use change that even modestly reduces evaporation and subsequent precipitation, could significantly affect human well-being (Keys et al. 2016). The agro-pastoral ecotone of northern China (APENC) is a region where a great deal of vegetation restoration has been carried out over the past few decades. To mitigate grassland degradation, desertification, and biodiversity loss, from 1998 until now, the Chinese government has implemented numerous policies and programs to restore ecosystems and improve environmental protection. These include the Grain for Green Program, the Three-North Shelter Forest Program, and the Beijing–Tianjin Sand Control Program, returning croplands to grassland and forest land (Li et al. 2015; Zhou et al. 2013). The impact of LUCC on regional water balance has been a growing area of research (Bagley et al. 2014; Bryan et al. 2015; Paul et al. 2016; Wang et al. 2020). Recent studies have observed reduced streamflow in the Yellow River from 1950 to 2015 due to large-scale vegetation restoration programs over the Loess Plateau (Zhang et al. 2016). Wang et al. (2020) found that LUCC led to a significant increase in summer ET in the agro-pastoral ecotone, northwest China. Although there have been some studies on the pattern of LUCC in the APENC (Liu et al. 2011; Cao et al. 2015; Zhao et al. 2017), and studies have indicated the effect of LUCC on P or other variables within the water cycle, nonetheless, the mechanisms of increasing or decreasing precipitation in the moisture recycling processes are still unclear, especially in the APENC. There is a need to clarify and highlight the importance of anthropogenic vegetation modification on regional water balance and intraregional moisture recycling in the APENC.
In this study, we examine how large-scale vegetation restoration modulates intraregional moisture recycling to change the regional water cycle in the agro-pastoral ecotone of northern China. First, we analyze the variation in spatial and temporal patterns of precipitation and evapotranspiration in the APENC during 1995–2015. Then, we use a dynamic recycling model (DRM) to investigate the changes in precipitation recycling in this region. Finally, we analyze the relationship between water balance and intraregional moisture recycling in the study region.
2. Data and methods
a. Study area and dividing regions
The APENC (33.5°–48.6°N and 101°–126.5°E, Fig. 1a) is located on the southeast edge of the Mongolian Plateau and northern part of the Loess Plateau, including 11 provinces and autonomous regions and metropolitan area (Heilongjiang, Jilin, Liaoning, Hebei, Beijing, Inner Mongolia, Shanxi, Shaanxi, Ningxia, Gansu, and Qinghai), with a total area of 7.25 × 105 km2. The mean annual air temperature is about 2°–8°C. About 60%–70% of annual precipitation occurs during summer (June–August), with mean annual precipitation of 250–450 mm. The main land cover types include grassland, farmland, forest and shrub land, and desert, and active conversion between grassland and farmland occurs frequently (Wang et al. 2020). The soil types are mainly sandy soil and loess, brown calcic soil, and chestnut soil with a sporadic distribution (Wei et al. 2018). This is one of the world’s largest ecotones, highly sensitive to changes in climate conditions and surface physical properties (Wang et al. 2020).
(a) Location of the agro-pastoral ecotone of northern China. Background information for the agro-pastoral ecotone of northern China: land use and land cover types in (b) 1995 and (c) 2015, (d) land use and land cover type conversion between 1995 and 2015 (C is croplands, F is forests, G is grasslands, U is unused lands, B is built-up lands, W is water body), and (e) pattern of different ecological conservation programs (BSCP represents the Beijing–Tianjin Sand Control Program, GFGP represents the Grain for Green Program, TNSF represents the Three-North Shelter Forest System Project, NFCP represents the Natural Forest Conservation Program).
Citation: Journal of Hydrometeorology 22, 6; 10.1175/JHM-D-20-0125.1
Land use and cover data with 1 km grids were divided into six categories (croplands, forests, grasslands, water body, build-up lands, and unused lands). Land use and land cover types in 1995 and 2015 are shown in Figs. 1b and 1c. Grasslands are the most widespread land use type in APENC. Figure 1d shows the land use and land cover type conversion between 1995 and 2015. The different ecological conservation programs (Beijing–Tianjin Sand Control Program, Grain for Green Program, Three-North Shelter Forest System Project) are shown in Fig. 1e.
The amount of regional recycled rainfall is related to spatial characteristics and scales of the study region (Trenberth 1999, Dominguez et al. 2006; Bisselink and Dolman 2008). Dirmeyer and Brubaker (2007) found a linear relationship in logarithm coordinates between recycled rainfall and the area of the study region. A similar result was found in China by Hua et al. (2017). Thus, we divided the APENC into four distinct subregions for comparative analysis according to the topographical features and land use classes in the APENC. Figure 1a shows the four dominant subregions, and Table 1 shows the main characteristics of them.
The characteristics of the four subregions in the APENC.
Region 1 is mainly located in the northeast of the Qinghai–Tibet Plateau, and is the transition zone from the Qinghai–Tibet Plateau to the Loess Plateau. The high altitude and complex terrain are the main features of this region. The average elevation of the area is 2601 m. The landforms are mainly mountains and plateaus. Alpine meadow dominates this region (Lin et al. 2011), and Qinghai spruce is the main forest vegetation type (Gao et al. 2018); croplands include spring wheat, highland barley, and rape seed (Zhang et al. 2008). Region 2 is mainly located in the middle reaches of the Yellow River, which is a typical area of the Loess Plateau. This area serves as a microcosm of the Loess Plateau where the policy of “Grain to Green” was implemented to rehabilitate severe soil erosion, frequent drought, and sparse vegetation cover (Zhang et al. 2013). Region 3 covers part of the central Inner Mongolia Plateau, and the main vegetation type is grassland, dominated by semidesert steppe (51.6%). It is subject to the arid and semiarid continental monsoon climate in the middle temperate zone; with strong winds and frequent sandstorms, and desertification (Wang et al. 2007; Yang et al. 2015). Region 4 is located in the northeast plain of China, with flat terrain and the average elevation of 472 m; and the dominant vegetation includes semiarid shrubs, grassland, and temperate steppe (Mao et al. 2012).
b. Datasets
The ERA-Interim reanalysis was the main dataset used in this study. ERA-Interim is a global atmospheric reanalysis dataset from 1979 to 2019 produced by the European Centre for Medium-Range Weather Forecasts (ECMWF) (Dee et al. 2011). Variables include precipitation, evaporation, land surface pressure, wind, and specific humidity on a 0.25° × 0.25° global grid. The residual of ERA-Interim in the global hydrological budget is the smallest among reanalysis datasets (Trenberth et al. 2011). ERA-Interim shows the highest fidelity according to reproducing precipitation and its interannual variability in the East Asia (Lin et al. 2014).
The China Meteorological Forcing Dataset (CMFD) produced by the Institute of Tibetan Plateau Research, the Chinese Academy of Sciences was also used in this study from 1979 to 2015 at 0.1° spatial resolution (Yang et al. 2010; Chen et al. 2011; J. He et al. 2020; Yang and He 2019). The CMFD has been widely used for land surface modeling, hydrological modeling, and terrestrial data assimilation (Chen et al. 2011).
Land use and cover data derived from the Moderate Resolution Imaging Spectroradiometer (MODIS), the MODIS enhanced vegetation index (EVI), and normalized difference vegetation index (NDVI) at 16-day intervals with a resolution of 500 m and the MODIS leaf area index (LAI) and albedo at 8-day intervals with a resolution of 500 m from January 2001 to December 2015, AVHRR LAI and global inventory modeling and mapping studies (GIMMS) NDVI data for 1995–2000 with a resolution of 0.05° and global EVI data were used for calculating ET. A digital elevation model (DEM) at a resolution of 30 m was obtained from the National Geomatics Center of China. The data were converted into a 0.1° grid according to New et al. (2000) and Yang et al. (2004). In addition, we also used temperature, pressure, vapor pressure, relative humidity, and radiation data to drive model for estimating ET over the agro-pastoral ecotone zone of northern China.
c. The PT-JPL model and experimental design
A list of derived variables with their derivation equation. RH is the relative humidity (%); fIPAR is the fraction of photosynthesis active radiation (PAR); fAPAR is the fraction of PAR absorbed by the canopy; Tmax is the maximum air temperature (°C); Topt is the optimum plant growth temperature (°C); Rn is the surface net radiation; Rnsoil is the net radiation of surface soil (W m−2);
To explore the impact of large-scale vegetation restoration on ET, we ran two models. The first simulated actual ET with dynamic vegetation (DV) using the PT-JPL model. In this situation, AVHRR LAI, EVI, global inventory modeling and mapping studies (GIMMS) NDVI, and albedo during 1995–2000, and the MODIS EVI, NDVI, LAI, and albedo during 2001–15 were used for calculating DV-ET. The second simulated ET by the same PT-JPL assuming no dynamic vegetation (no-DV). In this situation, we assume that revegetation projects had not been implemented, the vegetation index (VI) and albedo in 1995 were used for calculating no-DV-ET during the entire period of 1995–2015. Both simulations used same climate data. Therefore, the difference between the two situations represents the net impact of revegetation projects on evaporative moisture.
Figures 2a and 2b show the spatial distribution of mean LAI and NDVI during 1995–2015 in APENC. The Sen’s linear variation trends in the LAI and NDVI in the APENC are presented in Figs. 2c–f. Figures 2c and 2d present the spatial patterns of LAI and NDVI variation trends in APENC during 1995–2015. Results show that approximately 52.64% and 49.32% of the APENC displayed a significant increase in LAI (slope ≥ 0.01 yr−2, p < 0.05) and NDVI (slope ≥ 0.005 yr−2, p < 0.05), respectively, during 1995–2015. The Sen’s linear trends suggest that the annual mean LAI and NDVI increased at a rate of 0.013 yr−2 (p < 0.05) and 0.005 yr−2 (p < 0.05), respectively, during 1995–2015.
Spatial distribution of mean (a) LAI and (b) NDVI and the Sen’s slope trend of (c) LAI (yr−2) and (d) NDVI (yr−2) and the variation of spatial averaged (e) LAI and (f) NDVI in APENC during 1995–2015.
Citation: Journal of Hydrometeorology 22, 6; 10.1175/JHM-D-20-0125.1
The simulated ET was evaluated with the observed ET values from 12 flux towers (Shao et al. 2019). Figure S2 in the online supplemental material shows the scatter diagram for the observed ET and simulated ET. The coefficients of correlation r and NSE for monthly ET are 0.82 (p < 0.05) and 0.55, according to the scatter diagram. This suggests that PT-JPL model performed well in simulating ET at the regional scale.
d. WRF Model
The Weather Research and Forecasting (WRF) Model (version 4.2) was used in this study to simulate the precipitation under no-DV situation. A single-nested grid system is adopted, and the domain is centered at 40°N and 110°E with dimensions of 320 × 320 horizontal grid points at a spatial resolution of 10 km. The 6-h, 0.25° × 0.25°ERA-Interim data provided initial and lateral boundary conditions for the WRF Model. The simulation time began at 1200 UTC 1 September 1994 and ended at 1200 UTC 31 December 2015; the time from 1 September 1994 to 31 December 1994 was considered as spinup time. The schemes of configuration of the model referred to Wang et al. (2020). Referring to no-DV-ET experiment, in the WRF simulation, we replaced the vegetation data in the model using 1995 vegetation data including land use data, LAI, vegetation index and so on. In this no-DV simulation experiment, we think the simulated precipitation was not affected by DV. Therefore, under the same natural forcing, the difference between simulated P and observed P can be taken as response caused by local vegetation project.
e. Precipitation recycling ratio
f. Mann–Kendall test and Sen’s slope estimator
The Mann–Kendall (MK) trend test proposed by Mann (1945) and Kendall (1948) is a nonparametric method. The advantage of the MK test is that it does not require data to follow any specific distribution. The MK test uses the Zs statistic to indicate if a time series has an increasing or decreasing trend and the trend’s significance. Positive Zs value indicates an increasing trend and negative value means a decreasing trend. The trend is significant at confidence levels of 95% and 99% when the absolute value of the Zs statistic exceeds 1.96 and 2.58 (|Zs| ≥ 1.96 and |Zs| ≥ 2.58), respectively (Gocic and Trajkovic 2013). Sen (1968) developed the nonparametric procedure for estimating the slope of the trend in meteorological time series by using a linear model. The MK test and Sen’s slope were widely used to assess the significance of trends in hydrometeorological data (Yue et al. 2002; Hamed 2008; Atta-ur-Rahman and Dawood 2017). In this study, the MK test and Sen’s slope estimator were used to detect the trends of meteorological variables.
3. Results
a. Changes in precipitation in the APENC
Figure 3 shows the spatial distribution of average P and the Sen’s linear trend of annual P (Ptrend) during 1985–1995 and 1995–2015 in APENC. From Figs. 3c and 3d, we found that there was a decreased trend from 1985 to 1995 in the east of APENC but an obviously increased trend during 1995–2015. During 1985–1995, the APENC experienced a reduction in P in region 4 (Fig. 3c), but in 1995–2015, 83% of the APENC showed an increased tendency in P, involving the subregions 2 and 4. This upward trend is statistically significant (Ptrend ≥ 3 mm yr−2, p < 0.05).
Spatial distribution of average P (mm yr−1) during (a) 1985–1995 and (b) 1995–2015 and the Sen’s linear trend of annual P (mm yr−2) during (c) 1985–1995 and (d) 1995–2015 in APENC.
Citation: Journal of Hydrometeorology 22, 6; 10.1175/JHM-D-20-0125.1
Figure S3 shows the regional average time series for precipitation from 1995 to 2015 in the four subregions of the APENC. There are continuous undulating increasing trends in the APENC during 1995–2015, the positive P trends were significant at the 95% significance level according to the statistical tests (Zs = 2.63, 2.39, 0.39, and 1.96, see Table 3) in all subregions but subregion 3. The annual trend for P in the four subregions exhibits an upward slope of 4.63, 5.38, 1.83, and 5.85 mm yr−2 from 1995 to 2015 in the regions 1 to 4, respectively.
Results of the statistical tests for P, ET (including DV-ET and no-DV-ET), PRR, and recycled rainfall over the period 1995–2015. The Mann–Kendall test is denoted by Zs. The * indicates a statistically significant trend at the 95% significance level (Zs > 1.96), and ** indicates a statistically significant trend at the 99% significance level (Zs > 2.58).
b. Changes in ET in the APENC
The spatial pattern of annual mean ET with no dynamic vegetation (no-DV-ET) and dynamic vegetation (DV-ET) scenarios are shown in Figs. 4a and 4b. From Figs. 4a and 4b, we found that ET increased gradually from the northwest to the southeast in the APENC. Figures 4c and 4d show the Sen’s linear trend of annual no-DV-ET (no-DV-ETtrend) and DV-ET (DV-ETtrend) during 1995–2015. Due to the implementation of large-scale vegetation restoration, most regions of the APENC showed a rising ET during 1995–2015. Comparing no-DV-ETtrend and DV-ETtrend (Figs. 4c,d), we found a significantly increased DV-ET (DV-ETtrend >1 mm yr−2, p < 0.05) over 50.36% of the APENC. The region with a significantly decreased (DV-ETtrend < −1 mm yr−2, p < 0.05) ET only constituted 34.65% of the entire APENC.
Spatial distribution of average annual (a) no-DV-ET (mm yr−1) and (b) DV-ET and the Sen’s linear trend of annual (c) no-DV-ET (mm yr−2) and (d) DV-ET during 1995–2015. (Values exceeding the 95% confidence level are marked by the diagonal line).
Citation: Journal of Hydrometeorology 22, 6; 10.1175/JHM-D-20-0125.1
The regional average time series for ET under the DV (DV-ET) and no-DV (no-DV-ET) are shown in Fig. 5. From 1995 to 2015, there was a significantly increasing trend (blue line) of DV-ET at the 95% significance level in the APENC (Zs = 3.60, 5.10, 3.35 and 3.89 in all four subregions). ET increased at a rate of 1.57, 3.58, 1.53, and 1.84 mm yr−2 in the four subregions, respectively. However, no-DV-ET exhibited an increasing trend (red line) only in regions 1 and 2 with rates of 0.16 and 1.61 mm yr−2, respectively. In regions 3 and 4, there was no obvious trend in ET. To further examine the difference in ET caused by the vegetation restoration, we performed the correlation analysis between the ET and LAI as shown in Fig. 6. The results indicate that there is a significantly positive correlation (r = 0.78, p < 0.05) between the DV-ET and LAI in contrast to with no-DV-ET (r = 0.42). Additionally, the paired-samples t test (Table 4) shows that there was a statistically significant (p < 0.05) difference between the DV-ET and no-DV-ET. Comparing the scenarios of the DV and no-DV, the results show that the vegetation restoration projects distinctly enhanced evaporative moisture in the APENC.
Annually averaged ET variation in the APENC: the blue and red lines represent DV and no-DV, respectively, for (a)–(d) subregions 1–4. The R is Pearson’s correlation coefficient, and ** indicates a statistically significance at the 99% significance level.
Citation: Journal of Hydrometeorology 22, 6; 10.1175/JHM-D-20-0125.1
The relationship between LAI and no-DV-ET, DV-ET, no-DV-ETc, and DV-ETc. The R is Pearson’s correlation coefficient, and ** indicates a statistically significance at the 99% significance level.
Citation: Journal of Hydrometeorology 22, 6; 10.1175/JHM-D-20-0125.1
Statistical paired samples t test for ET and PRR under DV and no-DV situations in the four subregions during 1995–2015.
The spatial distribution of average canopy transpiration Ec (mm yr−1) and soil evaporation Es (mm yr−1) and the Sen’s linear trends of annual Ec and Es (mm yr−2) with no-DV and DV during 1995–2015 are shown in Fig. 7 and Fig. S4, respectively. An obviously increased trend in the DV-Ec and decreased trend in the DV-Es were found in Fig. 7d and Fig. S4d by comparing the scenarios of the DV and no-DV. Annually averaged DV-Ec in APENC ranged from 0 to 364.56 mm yr−1 with an average of 142.63 mm yr−1 during 1995–2015 (Fig. 7b). Comparing the scenarios of DV and no-DV, over 73.58% of the whole APENC experienced a significant increase in DV-Ec (DV-Ectrend ≥ 1 mm yr−2, p < 0.05), while only 9.53% of the APENC showed a significantly decreased DV-Ec (DV-Ectrend ≤ −1 mm yr−2, p < 0.05) (Fig. 7d). However, as shown in Fig. S4, 88.26% of the APENC displayed a significantly decreased trend in DV-Es (DV-Es trend ≤ −1 mm yr−2, p < 0.05).
Spatial distribution of average canopy transpiration Ec (mm yr−1) with (a) no-DV and (b) DV and the Sen’s linear trend of annual Ec (mm yr−2) with (c) no-DV and (d) DV during 1995–2015. (Values exceeding the 95% confidence level are marked by the diagonal line).
Citation: Journal of Hydrometeorology 22, 6; 10.1175/JHM-D-20-0125.1
Figure 8 shows the spatial distribution of the percentages of annual mean Ec, Es, and Ei under the situations of the DV and no-DV. Figure 9 shows temporal variation of the percentages of Ec, Es, and Ei under the situations of the DV and no-DV during 1995–2015 in the APENC. Under the no-DV scenario, Ec, Es, and Ei accounted for 35.11%, 55.21% and 9.68%, respectively, of the total ET. However, under the DV situation, Ec, Es, and Ei accounted for 45.98%, 42.38%, and 11.70%, respectively, of the total ET. The Ec display a significant upward trend with 0.93% yr−2 (p < 0.05), while Es exhibited a downward trend with −1.11% yr−2 (p < 0.05) for the period of 1995–2015. This clearly shows that the large-scale vegetation restoration projects increased the Ec in the APENC since 1995.
Percentage of annual mean ET partitioning components-canopy transpiration Ec, soil evaporation Es, and interception evaporation Ei under DV and no-DV situations for (a) no-DV-Ec, (c) no-DV-Es, and (e) no-DV-Ei, and (b) DV-Ec, (d) DV-Es, and (f) DV-Ei.
Citation: Journal of Hydrometeorology 22, 6; 10.1175/JHM-D-20-0125.1
Percentage of annual ET partitioning components—canopy transpiration Ec, soil evaporation Es, and interception evaporation Ei—under no-DV and DV situations.
Citation: Journal of Hydrometeorology 22, 6; 10.1175/JHM-D-20-0125.1
c. PRR variation in the APENC
The mean annual PRR in the four subregions during 1995–2015 was calculated using Eq. (10) and is shown in Fig. 10. The annual mean DV-PRR are 10.15%, 9.30%, 11.01%, and 12.76% for all subregions 1–4, respectively, in the APENC during 1995–2015. As shown in Fig. 7 and Table 3, during 1995–2015, there were undulating increasing trend (Zs = 4.16, 4.74, 4.80, and 3.96 in all subregions 1–4, respectively, at the 99% significance levels) in the DV-PRR during 1995–2015 comparing with the no-DV-PRR in the APENC. The recycled rainfall is shown in Fig. 10 (bar chart) for the four subregions. It is clear that there is an increasing trend (Zs = 3.04, 3.11, 1.78, and 2.32 in all subregions, respectively) for recycled rainfall, significant at the 99% significance level, and the trends of recycled rainfall are 0.62, 0.73, 0.52, and 0.93 mm yr−2, respectively, for subregions 1–4. As shown in Table 4, the paired-samples t test indicates a statistically significant (p < 0.05) difference between the DV-PRR and no-DV-PRR.
Annually averaged regional precipitation recycling ratio (blue line) and recycled rainfall (bar chart) during 1995–2012 in the APENC [(a)–(d) for subregions 1, 2, 3, and 4, respectively]. The blue line (broken line) represents DV-PRR (the trend of DV-PRR) and the black line (broken line) represents no-DV-PRR (the trend of no-DV-PRR). The orange bar (red broken line) represents DV recycled rainfall (the trend of DV recycled rainfall). The superscript ** indicates a statistically significant trend at the 99% significance level.
Citation: Journal of Hydrometeorology 22, 6; 10.1175/JHM-D-20-0125.1
d. Water balance in the APENC
The atmospheric moisture circulation patterns over the APENC with respect to seasonal precipitation are shown in Fig. S5. The vectors represent the climatological seasonal vertically integrated moisture flux. The circulation in spring and winter is characterized by westerlies, and in summer and fall by the Indian Ocean monsoon and westerlies. However, the Indian Ocean monsoon only contributes to the summer precipitation of regions 1 and 2. From Fig. S5, we can see that water vapor is transported from the west to the east in the APENC, however, the water vapor is mostly not converted into precipitation during the transportation process; the evaporated water is transported from regions 1 and 2 into 3 or 4.
Figure 11 illustrates the regional moisture cycle. All components were standardized by dividing by the annual averaged P then multiplying by 100 (Guo et al. 2018). The PRR was highest in region 4 (12.76%), followed by regions 3 and 1 (11.01% and 10.15%) and lowest in region 2 (9.30%). The results are consistent with Hua et al. (2016). The difference between annual mean P and ET in the four subregions was not significant. However, the horizontal moisture flux into the four subregions (Fin) did exhibit significant differences. The horizontal moisture flux Fin was 145.13 × 6.61 × 102, 313.04 × 6.29 × 102, 518.90 × 5.55 × 102, and 496.09 × 6.38 × 102 kg s−1 in the four subregions, respectively, indicating that the water available for precipitation gradually increases from region 1 to region 4 in the APENC. However, Fin witnessed a decrease trends by −0.76 × 6.61 × 102, −1.60 × 6.29 × 102, −2.94 × 5.55 × 102, and −3.19 × 6.38 × 102 kg s−1 yr−1 in the subregions 1–4, respectively. We can derive the conversion ratio of Fin to P by calculating Pa/Fin, the conversion ratio for the four subregions were 61.91%, 28.97%, 17.15%, and 17.59%, respectively. In region 1, the moisture entering the region was lowest, but the conversion ratio was the highest, while much more water vapor entered the other subregions, but the conversion ratio was much lower.
Schematic representations of the mean annual water cycle over the four subregions [(a)–(d) for regions 1, 2, 3, and 4, respectively]. The “T” is the trends of variables. The superscript * indicates a statistically significant trend at the 95% significance level, and ** indicates a statistically significant trend at the 99% significance level.
Citation: Journal of Hydrometeorology 22, 6; 10.1175/JHM-D-20-0125.1
Table 5 further presents the relationships of the variables listed in Fig. 11 and vegetation index. For all subregions, according to the above analysis, we summarized the characteristics of the moisture cycle in APENC during 1995–2015 in Fig. S6. From the opinion of view of moisture flux transport, although Fin display a decreased trend, however, the opposite trend for Pa has a positive effect on precipitation in APENC. On the other hand, vegetation restoration projects have positive effects on ET. The enhanced ET magnified P through recycling processes. The PRR in the APENC was reinforced during the period of 1995–2015, which suggests that the precipitation in the APENC tends to gain a more and more proportion of moisture from local ET through the land–atmosphere interaction processes. Figure 12 shows the variation of the difference of recycled rainfall between DV and no-DV situations during 1995–2012 in the APENC. There was a significantly increasing trend (blue line) of the difference of recycled rainfall at the 99% significance level in the APENC. This analysis further indicates that revegetation projects accelerates the local recycling P and contributed more precipitable water over the APENC.
Summary of statistical correlation of the variables in APENC during 1995–2015. The * indicates statistical significance at the 95% confidence level, and ** indicates statistical significance at the 99% confidence level.
The variation of the difference of recycled rainfall between DV and no-DV situations during 1995–2015 in the APENC. The superscript ** indicates a statistically significant trend at the 99% significance level.
Citation: Journal of Hydrometeorology 22, 6; 10.1175/JHM-D-20-0125.1
4. Discussion
Overgrazing and reclamation can cause serious land desertification and soil erosion. To address these problems, China has implemented many large-scale vegetation restoration programs in semiarid and arid areas since 1998. Implementation of such vegetation restoration programs has improved the ecosystems of the APENC (Wang et al. 2020). In this study, we investigated how vegetation restoration modulates the intraregional moisture cycle to change regional water balance. Specifically, we examined the changes in precipitation, ET, and moisture recycling in the APENC. We found that, over the period of 1995–2015, there were varying but increasing trends in the annual P in the APENC, a finding similar to previous studies (Xue et al. 2019). Annual mean regional ET exhibited significant increasing trends from 1995 to 2015, indicating that the revegetation has enhanced regional ET, confirming the results of previous study (Wang et al. 2020). While there was an increase in ET during 1995–2015, the increasing moisture from ET have a positive feedback on precipitation by moisture recycling, and the recycled rainfall showed a remarkable increase.
Our results indicate PRR showed a significant increasing trend over the entire region during 1995–2015. A similar result of an increasing trend in PRR over the arid and semiarid region of China during 1979–2010 was also reported by Hua et al. (2016). In addition, Li et al. (2018) also pointed out that, in Asian arid regions, the PRR increased during 1981–2000. However, a decreasing trend was found in the west of the Tibet Plateau during 1979–2008 (Guo and Wang 2014). It worth noting that the PRR is determined by the size of the study region and moisture recycling process are closely related to the scale of region (Eltahir and Bras 1994; Dominguez et al. 2006). In addition, the uncertainty of analysis data has effects on PRR in different regions (Guo et al. 2018). Furthermore, land use/cover change induced by human activities also strongly affects the PRR via changing surface ET process. Our results found revegetation has a positive effect on PRR via enhancing canopy transpiration. Vervoort et al. (2009) found that the PRR in irrigation region is larger than in grasslands in Australia. Irrigation has a positive impact on the PRR by increasing soil moisture and supplying abundant moisture for ET in North America (Harding and Snyder 2014).
In this study, we analyzed the moisture transport process in the APENC. In general, the moisture in precipitation was derived from both remote inputs and local recycling. Here, we found that there was no obvious increase in the input of moisture from westerlies and the Indian Ocean monsoon during 1995–2015 (Fig. S5 and Fig. 11). A previous study has shown that precipitation over south and east China is mainly attributable to the oceanic moisture sources, while the precipitation over northern China mainly comes from the terrestrial sources (Zhao et al. 2016). However, the moisture entering the four subregions gradually increased from region 1 to region 4, but the conversion ratio Pa/Fin was 61.91%, 28.97%, 17.15%, and 17.59% for the four regions, respectively, which indicates the west is more likely to experience rainfall than the east in the APENC. The local recycling moisture exhibited a significantly increasing trend during 1995–2015 (Figs. 10 and 11), indicating that the increased ET plays an important role in regional hydrological processes.
Charney (1975) and Charney et al. (1977) originally proposed the mechanism for the feedback of LUCC on precipitation variability, and stated that albedo enhancement has a positive feedback to drought. However, LUCC also affects the local ET, providing a second feedback mechanism. Vegetation restoration intensifies the processes in which vegetation transports soil moisture from subsoil to the leaves and loses water through ET. The increased ET decreases the vapor-pressure deficit (VPD), which, in turn, enhances regional moisture recycling processes. Besides, regional warming may also accelerate above processes (Zeng et al. 1999; Yu et al. 2017). In view of this, we analyzed the variation of temperature T with ET, LAI, and P (Fig. 13 and Table 5). Figure 13a shows the spatial distribution of the Sen’s linear trend of temperature in APENC during 1995–2015, however, there was no significant increased trend in the temperature in the region during 1995–2015. The statistical correlations between T and ET, between LAI and ET, and between P and ET in APENC during 1995–2015 were shown in Figs. 13b–d. It can be seen from the Figs. 13b and 13c that there is a significant statistical correlation between ET and LAI, but the correlation of ET with T is not significant. This analysis further indicates that vegetation change mainly accelerate increase in ET and the recycle processes.
(a) Spatial distribution of the Sen’s linear trend of temperature T (°C yr−2) and the statistical correlation between (b) T and ET, (c) LAI and ET, and (d) P and ET in APENC during 1995–2015. (Values exceeding the 90% confidence level are marked by the points).
Citation: Journal of Hydrometeorology 22, 6; 10.1175/JHM-D-20-0125.1
Our results indicate that ecological vegetation restoration increased local ET (Fig. 5), and the increasing moisture had a positive feedback on precipitation based on the significant increase in the PRR in the APENC during 1995–2015 (Fig. 10). Vegetation restoration does not directly increase local P, but does so indirectly by increasing land surface ET and internal moisture recycling (Fig. 12 and Fig. S6). The increased ET is not only a source of water for more precipitation but also affects the thermodynamic structure of the lower atmosphere, which can enhance convection, thus potentially changing the regional water-energy circulation to produce more P (Schär et al. 1999; Alexandre and Gonzalo 2014; Wang-Erlandsson et al. 2014). Due to the indirect influence on the regional atmospheric boundary structure and moisture circulation, a part of advected moisture that otherwise would have been merely blown across the study area is retained and produced total P, thus becoming a source for the additional P induced by land surface–atmosphere interactions, in addition to local ET (Alexandre and Gonzalo 2014). The higher extraction of moisture from the large-scale flow attributable to the effect of ET fluxes can be as relevant as the direct impact of ET fluxes alone as mere suppliers of water (Koster and Suarez 2001; Koster et al. 2004; Wei et al. 2016). In addition, these changes in ET in a region can also significantly alter the P that falls downwind (Tuinenburg et al. 2012; Lo and Famiglietti 2013; Wei et al. 2013). However, the uncertainty about the impacts of vegetation restoration on circulation change also exists. Vegetation restoration can alter surface albedo and roughness, thereby affect the moisture flux processes in the boundary layer (Foley et al. 2003; Bagley et al. 2012; Alkama and Cescatti 2016). Pitman and Hesse (2007) found that vegetation changes surface roughness length that affects the surface frictional drag, wind velocity, and moisture convergence in the Australian monsoon region. While Yu and Notaro (2020) found that in observation, vegetation anomaly in the Australian monsoon region causes a dynamic response in the regional pressure system and monsoonal flow. Yu et al. (2017) indicated that positive anomalies of remotely sensed vegetation greenness favor enhanced evapotranspiration, precipitable water, convective activity and rainfall, in the Sahel during the late and postmonsoon periods. Therefore, further studies are needed to assess the uncertainty of LUCC on circulation mechanism in difference seasons.
5. Conclusions
In this study, we analyzed the impacts of anthropogenic LUCC on intraregional moisture recycling processes in the agro-pastoral ecotone of northern China. Multiple datasets and methods were used to estimate how anthropogenic vegetation restoration affects the intraregional hydrological processes and the land surface–atmosphere interaction over regional scales. The main findings are highlighted below: First, the trends in precipitation showed that regional available water was gradually increasing, benefitting local ecosystems and sustainable development. Second, while ET shows an upward trend during the period of 1995–2015, the increasing moisture from ET had a positive feedback on precipitation via moisture recycling. Finally, our study shows that vegetation restoration strengthens the intraregional moisture recycling processes during the period of 1995–2015 according to the vegetation–rainfall feedbacks dominated by moisture recycling mechanism. Our analysis shows the importance of considering LUCC on par with water recycling to identify anthropogenic influences on water resources. Our findings help better understanding of the impacts of LUCC on local water resources, and in turn support local water resource management and decision making.
Acknowledgments
The project is partially funded by the National Natural Science Foundation of China (Grants 42030501, 41530752, and 41877150) and Scherer Endowment Fund of Department of Geography, Western Michigan University. We also acknowledge kind support by the National Field Station for Grassland Ecosystem in Ordos, Inner Mongolia, China and Supercomputing Center of Lanzhou University.
REFERENCES
Alexandre, R. E., and M. M. Gonzalo, 2014: Moisture recycling and the maximum of precipitation in spring in the Iberian Peninsula. Climate Dyn., 42, 3207–3231, https://doi.org/10.1007/s00382-013-1971-x.
Alkama, R., and A. Cescatti, 2016: Biophysical climate impacts of recent changes in global forest cover. Science, 351, 600–604, https://doi.org/10.1126/science.aac8083.
Atta-ur-Rahman, and M. Dawood, 2017: Spatio-statistical analysis of temperature fluctuation using Mann-Kendall and Sen’s slope approach. Climate Dyn., 48, 783–797, https://doi.org/10.1007/s00382-016-3110-y.
Bagley, J. E., A. R. Desai, P. A. Dirmeyer, and J. A. Foley, 2012: Effects of land cover change on moisture availability and potential crop yield in the world’s breadbaskets. Environ. Res. Lett., 7, 014009, https://doi.org/10.1088/1748-9326/7/1/014009.
Bagley, J. E., A. R. Desai, K. J. Harding, P. K. Snyder, and J. A. Foley, 2014: Drought and deforestation: Has land cover change influenced recent precipitation extremes in the Amazon? J. Climate, 27, 345–361, https://doi.org/10.1175/JCLI-D-12-00369.1.
Baldocchi, D., 2014: Managing land and climate. Nat. Climate Change, 4, 330–331, https://doi.org/10.1038/nclimate2221.
Berger, M., R. van der Ent, S. Eisner, V. Bach, and M. Finkbeiner, 2014: Water Accounting and Vulnerability Evaluation (WAVE): Considering atmospheric evaporation recycling and the risk of freshwater depletion in water footprinting. Environ. Sci. Technol., 48, 4521–4528, https://doi.org/10.1021/es404994t.
Bisselink, B., and A. J. Dolman, 2008: Precipitation recycling: Moisture sources over Europe using ERA-40 data. J. Hydrometeor., 9, 1073–1083, https://doi.org/10.1175/2008JHM962.1.
Blyth, E. M., A. J. Dolman, and J. Noilhan, 1994: The effect of forest on mesoscale rainfall: An example from HAPEX–MOBILHY. J. Appl. Meteor., 33, 445–454, https://doi.org/10.1175/1520-0450(1994)033<0445:TEOFOM>2.0.CO;2.
Brubaker, K. L., D. Entekhabi, and P. S. Eagleson, 1993: Estimation of continental precipitation recycling. J. Climate, 6, 1077–1089, https://doi.org/10.1175/1520-0442(1993)006<1077:EOCPR>2.0.CO;2.
Bryan, A. M., A. L. Steiner, and D. J. Posselt, 2015: Regional modeling of surface-atmosphere interactions and their impact on Great Lakes hydroclimate. J. Geophys. Res. Atmos., 120, 1044–1064, https://doi.org/10.1002/2014JD022316.
Burde, G. I., and A. Zangvil, 2001: The estimation of regional precipitation recycling. Part I: Review of recycling models. J. Climate, 14, 2497–2508, https://doi.org/10.1175/1520-0442(2001)014<2497:TEORPR>2.0.CO;2.
Cao, Q., D. Y. Yu, M. Georgescu, Z. Han, and J. G. Wu, 2015: Impacts of land use and land cover change on regional climate: A case study in the agro-pastoral transitional zone of China. Environ. Res. Lett., 10, 124025, https://doi.org/10.1088/1748-9326/10/12/124025.
Charney, J. G., 1975: Dynamics of deserts and drought in the Sahel. Quart. J. Roy. Meteor. Soc., 101, 193–202, https://doi.org/10.1002/qj.49710142802.
Charney, J. G., W. J. Quirk, S. H. Chow, and J. Kornfield, 1977: A comparative study of the effects of albedo change on drought in semi-arid regions. J. Atmos. Sci., 34, 1366–1385, https://doi.org/10.1175/1520-0469(1977)034<1366:ACSOTE>2.0.CO;2.
Chen, Y. Y., K. Yang, J. He, J. Qin, J. C. Shi, J. Y. Du, and Q. He, 2011: Improving land surface temperature modeling for dry land of China. J. Geophys. Res., 116, D20104, https://doi.org/10.1029/2011JD015921.
Clark, C. A., and R. W. Arritt, 1995: Numerical simulations of the effect of soil moisture and vegetation cover on the development of deep convection. J. Appl. Meteor., 34, 2029–2045, https://doi.org/10.1175/1520-0450(1995)034<2029:NSOTEO>2.0.CO;2.
Dee, D. P., and Coauthors, 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553–597, https://doi.org/10.1002/qj.828.
Dirmeyer, P. A., and K. L. Brubaker, 2007: Characterization of the global hydrologic cycle from a back-trajectory analysis of atmospheric water vapor. J. Hydrometeor., 8, 20–37, https://doi.org/10.1175/JHM557.1.
Dominguez, F., P. Kumar, X. Z. Liang, and M. F. Ting, 2006: Impact of atmospheric moisture storage on precipitation recycling. J. Climate, 19, 1513–1530, https://doi.org/10.1175/JCLI3691.1.
Dore, M. H. I., 2005: Climate change and changes in global precipitation patterns: What do we know? Environ. Int., 31, 1167–1181, https://doi.org/10.1016/j.envint.2005.03.004.
Eltahir, E. A. B., and L. B. Bras, 1994: Precipitation recycling in the Amazon basin. Quart. J. Roy. Meteor. Soc., 120, 861–880, https://doi.org/10.1002/qj.49712051806.
Feng, F., and Coauthors, 2015: Validity of five satellite-based latent heat flux algorithms for semi-arid ecosystems. Remote Sens., 7, 16 733–16 755, https://doi.org/10.3390/rs71215853.
Fisher, J. B., K. P. Tu, and D. D. Baldocchi, 2008: Global estimates of the land–atmosphere water flux based on monthly AVHRR and ISLSCP-II data, validated at 16 FLUXNET sites. Remote Sens. Environ., 112, 901–919, https://doi.org/10.1016/j.rse.2007.06.025.
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, https://doi.org/10.1890/1540-9295(2003)001[0038:GSHTEC]2.0.CO;2.
Gao, L. L., X. H. Gou, Y. Deng, Z. Q. Wang, F. Gu, and F. Wang, 2018: Increased growth of Qinghai spruce in northwestern China during the recent warming hiatus. Agric. For. Meteor., 260-261, 9–16, https://doi.org/10.1016/j.agrformet.2018.05.025.
Gocic, M., and S. Trajkovic, 2013: Analysis of changes in meteorological variables using Mann-Kendall and Sen’s slope estimator statistical tests in Serbia. Global Planet. Change, 100, 172–182, https://doi.org/10.1016/j.gloplacha.2012.10.014.
Gordon, L. J., W. Steffen, B. F. Jönsson, C. Folke, M. Falkenmark, and A. S. Johannessen, 2005: Human modification of global water vapor flows from the land surface. Proc. Natl. Acad. Sci., 102, 7612–7617, https://doi.org/10.1073/pnas.0500208102.
Guillod, B. P., B. Orlowsky, D. G. Miralles, A. J. Teuling, and S. I. Seneviratne, 2015: Reconciling spatial and temporal soil moisture effects on afternoon rainfall. Nat. Commun., 6, 6443, https://doi.org/10.1038/ncomms7443.
Guo, L., N. P. Klingaman, M. E. Demory, P. L. Vidale, A. G. Turner, and C. C. Stephan, 2018: The contributions of local and remote atmospheric moisture fluxes to East Asian precipitation and its variability. Climate Dyn., 51, 4139–4156, https://doi.org/10.1007/s00382-017-4064-4.
Guo, Y. P., and C. H. Wang, 2014: Trends in precipitation recycling over the Qinghai-Xizang Plateau in last decades. J. Hydrol., 517, 826–835, https://doi.org/10.1016/j.jhydrol.2014.06.006.
Hamed, K. H., 2008: Trend detection in hydrologic data: The Mann-Kendall trend test under the scaling hypothesis. J. Hydrol., 349, 350–363, https://doi.org/10.1016/j.jhydrol.2007.11.009.
Harding, K. J., and P. K. Snyder, 2014: Modeling the atmospheric response to irrigation in the great plains. Part II: The precipitation of irrigated water and changes in precipitation recycling. J. Hydrometeor., 601, 1687–1703, https://doi.org/10.1175/JHM-D-11-099.1.
He, J., K. Yang, W. J. Tang, H. Lu, J. Qin, Y. Y. Chen, and X. Li, 2020: The first high-resolution meteorological forcing dataset for land process studies over China. Sci. Data, 7, 1–11, https://doi.org/10.1038/s41597-020-0369-y.
He, X. L., and Coauthors, 2020: A Bayesian three-cornered hat (BTCH) method: Improving the terrestrial evapotranspiration estimation. Remote Sens., 12, 878, https://doi.org/10.3390/rs12050878.
Herrera-Estrada, J. E., J. A. Martinez, F. Dominguez, K. L. Findell, E. F. Wood, and J. Sheffield, 2019: Reduced moisture transport linked to drought propagation across North America. Geophys. Res. Lett., 46, 5243–5253, https://doi.org/10.1029/2019GL082475.
Hua, L. J., L. H. Zhong, and Z. J. Ke, 2016: Precipitation recycling and soil-precipitation interaction across the arid and semi-arid regions of China. Int. J. Climatol., 36, 3708–3722, https://doi.org/10.1002/joc.4586.
Hua, L. J., L. H. Zhong, and Z. J. Ke, 2017: Characteristics of the precipitation recycling ratio and its relationship with regional precipitation in China. Theor. Appl. Climatol., 127, 513–531, https://doi.org/10.1007/s00704-015-1645-1.
Kendall, M. G., 1948: Rank Correlation Methods. Charles Griffin, 160 pp.
Keys, P. W., and L. Wang-Erlandsson, 2018: On the social dynamics of moisture recycling. Earth Syst. Dyn., 9, 829–847, https://doi.org/10.5194/esd-9-829-2018.
Keys, P. W., R. J. van der Ent, L. J. Gordon, H. Hoff, R. Nikoli, and H. H. G. Savenije, 2012: Analyzing precipitation sheds to understand the vulnerability of rainfall dependent regions. Biogeosciences, 9, 733–746, https://doi.org/10.5194/bg-9-733-2012.
Keys, P. W., L. Wang-Erlandsson, and L. J. Gordon, 2016: Revealing invisible water: Moisture recycling as an ecosystem service. PLoS One, 11, e0151993, https://doi.org/10.1371/journal.pone.0151993.
Keys, P. W., L. Wang-Erlandsson, L. J. Gordon, V. Galaz, and J. Ebbesson, 2017: Approaching moisture recycling governance. Global Environ. Change, 45, 15–23, https://doi.org/10.1016/j.gloenvcha.2017.04.007.
Koster, R. D., and M. J. Suarez, 2001: Soil moisture memory in climate models. J. Hydrometeor., 2, 558–570, https://doi.org/10.1175/1525-7541(2001)002<0558:SMMICM>2.0.CO;2.
Koster, R. D., and Coauthors, 2004: Regions of strong coupling between soil moisture and precipitation. Science, 305, 1138–1140, https://doi.org/10.1126/science.1100217.
Lee, E., Y. Q. He, M. Zhou, and J. J. Liang, 2015: Potential feedback of recent vegetation changes on summer rainfall in the Sahel. Phys. Geogr., 36, 449–470, https://doi.org/10.1080/02723646.2015.1120139.
Li, R. L., C. H. Wang, and D. Wu, 2018: Changes in precipitation recycling over arid regions in the Northern Hemisphere. Theor. Appl. Climatol., 131, 489–502, https://doi.org/10.1007/s00704-016-1978-4.
Li, S., P. L. An, Z. H. Pan, X. M. Li, and Y. Liu, 2015: Farmers’ initiative on adaptation to climate change in the northern agro-pastoral ecotone. Int. J. Disaster Risk Reduct., 12, 278–284, https://doi.org/10.1016/j.ijdrr.2015.02.002.
Lin, R. P., T. J. Zhou, and Y. Qian, 2014: Evaluation of global monsoon precipitation changes based on five reanalysis datasets. J. Climate, 27, 1271–1289, https://doi.org/10.1175/JCLI-D-13-00215.1.
Lin, S., J. M. Feng, J. Wang, and Y. H. Hu, 2016: Modeling the contribution of long-term urbanization to temperature increase in three extensive urban agglomerations in China. J. Geophys. Res. Atmos., 121, 1683–1697, https://doi.org/10.1002/2015JD024227.
Lin, X., and Coauthors, 2011: Response of ecosystem respiration to warming and grazing during the growing seasons in the alpine meadow on the Tibetan Plateau. Agric. For. Meteor., 151, 792–802, https://doi.org/10.1016/j.agrformet.2011.01.009.
Liu, J. H., J. X. Gao, S. H. Lv, Y. W. Han, and Y. H. Nie, 2011: Shifting farming-pastoral ecotone in China under climate and land use changes. J. Arid Environ., 75, 298–308, https://doi.org/10.1016/j.jaridenv.2010.10.010.
Lo, M. H., and J. S. Famiglietti, 2013: Irrigation in California’s Central Valley strengthens the southwestern US water cycle. Geophys. Res. Lett., 40, 301–306, https://doi.org/10.1002/grl.50108.