Vegetation dynamics and their effects on surface water-energy balance over the Three-North Region of China

Vegetation dynamics and their effects on surface water-energy balance over the Three-North Region of China

Agricultural and Forest Meteorology 275 (2019) 79–90 Contents lists available at ScienceDirect Agricultural and Forest Meteorology journal homepage:...

4MB Sizes 0 Downloads 3 Views

Agricultural and Forest Meteorology 275 (2019) 79–90

Contents lists available at ScienceDirect

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

Vegetation dynamics and their effects on surface water-energy balance over the Three-North Region of China

T



Cuiling Denga, Baoqing Zhanga, , Linyin Chengc, Leiqiu Hud, Fahu Chena,b a

Key Laboratory of Western China's Environmental Systems (Ministry of Education), College of Earth and Environmental Sciences, Lanzhou University, Lanzhou, Gansu, 730000, China b Key Laboratory of Alpine Ecology (LAE), CAS Center for Excellence in Tibetan Plateau Earth Sciences, Institute of Tibetan Plateau Research, Chinese Academy of Sciences, Beijing 10010, China c Department of Geosciences, University of Arkansas, Fayetteville, Arkansas, 72701, USA d Department of Atmospheric Science, University of Alabama in Huntsville, Huntsville, Alabama, 35899, USA

A R T I C LE I N FO

A B S T R A C T

Keywords: Vegetation restoration Evapotranspiration Water-energy balance Water use efficiency Water-carbon trade-off

The Three-North Shelterbelt Project (TNSP) is one of the largest ecological restoration projects of the world. Although its important role in regulating the ecosystem of northern China has been acknowledged, how this project affects the surface water-energy balance deserves further evaluation. This study characterizes spatiotemporal variation of climate variables and vegetation coverage/density over the Three-North Region (TNR) of China using multiple datasets since the implementation of the TNSP. Of particular importance is that effects of vegetation restored during the TNSP on surface water-energy budget are examined for the study domain. Our results show that accompanied by a significant enhancement of vegetation coverage and density in the TNR, its annual air temperature and precipitation have increased 1.46 °C and 89.1 mm, respectively from 1982 to 2015. We find such rise in the air temperature and precipitation plays a positive role in the TNR’s vegetation restoration in a sense that the vegetation dynamics show positive correlation with its regional air temperature and precipitation, while more sensitive to precipitation. We also find carbon sequestration of the TNR increases at the cost of greater water consumption through evapotranspiration since the TNSP commences. In some arid regions, revegetation accelerates the water deficit due to an excessive rate of restoring vegetation; yet, no substantial water imbalance occurs as a result of enhanced precipitation and water use efficiency during this period. Although the solar radiation increases with decreases in surface albedo over the last few decades, our results do not reveal an appreciably increasing trend in the land surface temperature of the TNR. This is because the improved vegetation can assimilate CO2 (mitigating greenhouse gas emissions) and transpire water by photosynthesis, thereby increasing latent heat flux and reducing the warming effect. This study highlights a mixed consequence of the TNSP by inducing both positive and negative effects on the surface water-energy balance over the TNR.

1. Introduction Vegetation is essential to terrestrial ecosystem (Peng et al., 2016). It affects not only carbon equilibrium (Piao et al., 2001), climate (Pei et al., 2018; Wilmers, 1990), and surface water-energy budget (Kala et al., 2013; Peng et al., 2016; Jiao et al., 2017), but also the basis of food, fiber and wood production, and human welfare (Liang et al., 2015). To improve ecosystems of northern China, the Chinese government launched the Three-North Shelterbelt project (TNSP) in the ThreeNorth Region (TNR) (Li et al., 2012), one of six-key forestry programs in China. This project has a 73-year construction period from 1978 to



2050 (Liu et al., 2014), covering North, North-east and most of Northwest China. It is one of the largest ecological restoration projects in the world by involving 632 counties in 13 provinces, which account for more than 40% of the land area of China (Yan et al., 2013; Peng et al., 2016). The TNR is one of the most water-stressed regions in the world (Yan et al., 2013). Ecosystem of the TNR is particularly vulnerable due to the human intervention (Liu et al., 2003; Piao et al., 2010). Over the past few decades, population explosion and social development have led to significant ecological changes in this region (Jia et al., 2015), including the construction of shelter forest systems wields a large area (Shen

Corresponding author. E-mail address: [email protected] (B. Zhang).

https://doi.org/10.1016/j.agrformet.2019.05.012 Received 25 September 2018; Received in revised form 24 April 2019; Accepted 14 May 2019 Available online 24 May 2019 0168-1923/ © 2019 Elsevier B.V. All rights reserved.

Agricultural and Forest Meteorology 275 (2019) 79–90

C. Deng, et al.

to mitigate the greenhouse effects by assimilating CO2, increasing latent heat flux and decreasing sensible heat flux through transpiration (Zeng et al., 2017) in addition to enhancing regional carbon sequestration. A comparison between cooling effects associated with vegetation transpiration and carbon assimilation and warming effects induced by vegetation albedo on a global scale shows that the cooling effect trumps (Betts et al., 1997; Bounoua et al., 2000; Zeng et al., 2017). The level of albedo warming generally increases with latitude, opposite to that of ET cooling (Li et al., 2015). In semi-arid and arid regions, vegetation transpiration and CO2 assimilation are hindered by limited availability of water, but whether the cooling effects can overwhelm the warming effects is unclear, especially in areas as the TNR of China where are undergoing a notable variation in vegetating. The relationship between climate elements and vegetation at regional and terrestrial scales has been a focus in many other studies (Fang et al., 2016; Guo et al., 2014; Sang and Su, 2009). To move forward, the objectives of this work are twofold: (1) to investigate effects of vegetation restoration on the surface water balance since the implementation of TNSP, and (2) to examine impacts of vegetation dynamics on the surface energy budget and the trade-off between carbon and water in the TNR.

et al., 2015; Zheng et al., 2016; Zheng and Zhu, 2017), the transition of grassland and wasteland to farmland in the northeast and oasis regions in the northwest (Zhou et al., 2017), and large-scale revegetation (designated “Grain-For-Green’’) over the Loess plateau by returning sloped farmland (> 15°) and wasteland to forest or grassland (Wang et al., 2016a,b; Feng et al., 2016; Zhang et al., 2016a; Zhang and He, 2016; Zhang et al., 2018). Research reveals that vegetated areas exhibit higher evapotranspiration (ET) than non-vegetated ones, leading to greater water consumption (Peel et al., 2010), particularly over arid and semi-arid regions. Questions of particular interest thus focus on the role of such artificial vegetation disturbance in the functioning of the local ecosystem (Cao, 2008; Gao et al., 2018a,b,c; He et al., 2015; Qu et al., 2014; Jian et al., 2015). Previous studies make exploration from aspects including (I) whether the TNSP achieves initial success in improving vegetation coverage and density (He et al., 2015); (Π) whether the TNSP increases water consumption and consequently intensifies water stress in this region (Li et al., 2012); (Ш) how the TNSP affects the surface energy budget of the TNR (He et al., 2015; Li et al., 2012); and lastly, (IV) how vegetation construction influences the trade-off between carbon and water (Zhang et al., 2016, 2017; Zhang et al., 2018). Nevertheless, the aforementioned studies mainly reply on in-situ observations from particular point or partial areas of TNR. Here we apply multiple datasets covering the entire TNR to deliver a more inclusive examination of vegetation dynamics and their effects on this region’s surface water-energy balance. Vegetation can be affected by multiple environmental factors, such as air temperature (T), precipitation (P), carbon source, etc. (Campos et al., 2013; He et al., 2015; Pei et al., 2018; Yan et al., 2013). Anthropogenic activities lead to continuously elevated CO2 concentration over the last few decades (IPCC-FAR5, 2014); and the distribution of P and T of the TNR is no longer stationary since 1970s reported by intensive studies (Jiapaer et al., 2015; Qian and Lin, 2005; Shi et al., 2007; Xu et al., 2010). These environmental variabilities concern changes in vegetation conditions of the TNR. Restoration of vegetation under climatic variations has been investigated using vegetation indices including net primary production (NPP) and leaf area index (LAI) for some areas of the TNR (Dan et al., 2004; Duan et al., 2011; Peng et al., 2011; Huang et al., 2018). A lack of consensus is on how this project affects the entire domain of the TNR as the fourth phase of the TNSP (of a total of seven phases) was not completed until recent (He et al., 2015). It is commonly recognized that vegetation restoration is at the cost of water consumption through ET. As one key element in the watercarbon cycle, ET is closely linked to vegetation growth and watercarbon trade-off (Feng et al., 2017; Vinukollu et al., 2011; Zhang et al., 2001). The total ET consists of bare soil evaporation (Eb), interception evaporation (Ei) and transpiration (Et). It is a useful indictor for measuring water consumption during vegetation restoration as the regional water balance and trade-off between carbon and water can be inferred by examining each component of ET. Research shows that increased vegetation coverage and density can lead to higher Et (Williams et al., 2012; Kala et al., 2013; Williams and Torn, 2015). Under the influence of rising temperature, water use efficiency (WUE, the ratio of carbon assimilation to water loss) of vegetation is likely to enhance in some regions (Bounoua et al., 2010); and bare soil evaporates (Eb) less effective as a consequence of better shading with improved vegetation condition (Williams et al., 2012; Williams and Torn, 2015). Presumably, these effects that act in opposite direction would compensate for an increase in water consumption. As such, how variations in the vegetation of the TNR have affected its regional water balance and the interactions between carbon sequestration and water consumption remains to be determined. Apart from increasing water consumption, vegetation is considered as one of the major factors bearing upon the surface energy budget (Jeong et al., 2009). On the one hand, heavy restoration of vegetation can result in a warming effect by dampening the land surface albedo (Bounoua et al., 2000; Myhre et al., 2005). On the other hand, it helps

2. Study area Terrain types of the TNR include meadow steppes, typical steppes, desert steppes, the Gobi desert, and typical desert (Li et al., 2012). Four sandy areas and eight deserts of China are located in this region, whereof the sandy areas account for 85% of the total sandy area of the country (Shi et al., 2007). Water resource in the TNR is only about one third of the national average, so it is classified as one of the most waterstressed regions in the world (Yan et al., 2013). Bounded by the Helan Mountain, difference in P between the west and the east of the region is distinguishable (Duan et al., 2011). The annual total P in the western area, characterized by desert and wind erosion landforms (Wang et al., 2016a,b) falls below 200 mm·yr−1, making this region a typical arid area with sparse vegetation in most regions. Most area of the eastern region is covered by grass, shrub and forest (Wang et al., 2016a,b) and receives abundant P, ranging from 200 to 400 mm·yr−1. The TNR has sufficient supply of solar radiation (Duan et al., 2011), in which case water scarcity is regarded as a primary limitation for the plant growth and vegetation restoration. The location, land cover types (2003), and major afforestation belt of TNR are shown in Fig. 1. 3. Data and methodology 3.1. Datasets In this study, we use the China Meteorological Forcing Dataset (CMFD, ITPCAS LSM Forcing Data, http://westdc.westgis.ac.cn) to investigate climatic variation and its effect on vegetation over the TNR. This dataset merges China Meteorological Administration in situ meteorological observations, Tropical Rainfall Measuring Mission (TRMM) satellite P analysis, the Global Energy and Water Exchanges Surface Radiation Budget (GEWEX-SRB), and Princeton forcing data. It has a temporal resolution of three hours, covering the period of 1979 to 2015 at a spatial resolution of 0.1° × 0.1°. The CMFD is widely used for communities of surface water budget, land surface modeling, land data assimilation, and other terrestrial hydrological cycling (Chen et al., 2011). It includes annual mean T, cumulative P, and downward shortwave and longwave radiations produced by the Institute of Tibetan Plateau Research (ITP) and the Chinese Academy of Sciences (CAS). The CMFD P are compared with in-situ P observations from 222 meteorological stations located within or around the TNR, which are obtained from National Meteorological Information Center, China (http://data.cma.cn/). Fig. S1 shows that the CMFD P and observed P are highly consistent with each other, demonstrating that the CMFD P 80

Agricultural and Forest Meteorology 275 (2019) 79–90

C. Deng, et al.

Fig. 1. Location of Three-Northern Shelterbelt Region (TNR), main mountains (where Great Khingan is Great Khingan Mountains, Lesser Khingan is Lesser Khingan Mountains, Qilian is Qilian Mountains and Altun is Altun Mountains), basins and plateaus in TNR and its main land cover type in 2001.

Numerical Terradynamic Simulation Group (http://www.ntsg.umt. edu/project/et). The MODIS ET datasets are estimated using the improved ET algorithm (Mu et al., 2011) based on the Penman-Monteith equation, which has a spatial resolution of 500 m and a temporal coverage of 2000-2014. This dataset is download from https://modis.gsfc. nasa.gov/data/dataprod/mod16.php. The monthly PML ET was provided by the Commonwealth Scientific and Industrial Research Organization Land and Water Division (https://doi.org/10.4225/08/ 5719A5C48DB85; Zhang et al., 2016), which has a spatial resolution of 0.5° and a temporal coverage of 1982-2012. The above three ET products are all based on satellite remote sensing ET algorithms. The fifth ET product, i.e., the MTE ET, was produced by upscaling in situ ET observations across global Fluxnet towers with an adaptive machine learning approach (Jung et al., 2009, 2010). The MTE ET covers a period of 1982–2011 and contains a spatial resolution of 0.5°. We validate the five gridded ET data against in-situ observations at seven tower flux stations located in the vicinity of the TNR (see Fig. S2, Table S1, and Table S2 in the Supplementary Materials for details). The Pearson correlation coefficient (R) and root mean square error (RMSE) are selected as standard to evaluate the performance of different gridded ET products at each tower flux stations. The results show that GLEAM agrees the best with the eddy covariance-based ET measurements at each station, so the GLEAM ET dataset is used for the analysis in this work. In addition, WUE, calculated in a way that divides NPP by ET, is employed to assess the relationship between terrestrial water and carbon cycles herein (Hu et al., 2008). All gridded datasets have been interpolated into 0.1° spatial resolution using the nearest interpolation method to calculate the ensemble mean ET. We also examined the ET from GLEAM with the spatial resolution of 0.25°. We find our result is largely unaffected by a change in the spatial resolution (see Fig. S2 and Table S2), and the performance of the nearest interpolation method is satisfactory. For energy balance analysis, surface temperature, albedo, land cover type and carbon dioxide data are in use. Data including 8-day land surface albedo (MCD43C3) from 2000 to 2015, monthly Ts (MYD11C3) from 2003 to 2015 and yearly land cover type (MCD12Q1) from 2001 to 2013 are obtained from MODIS at a 500-m spatial resolution (https://modis.gsfc.nasa.gov/data/). All these products have been validated and widely applied in others’ research (e.g. Liang et al.,

performs well in the TNR (The mean R of all stations is 0.90). Although Moderate Resolution Imaging Spectroradiometer (MODIS) products have higher spatiotemporal resolution than Advanced Very High Resolution Radiometers (AVHRR) products, AVHRR products cover a period of 1982–2011, much longer than the MODIS products that only trace back to year 2000. Therefore, we decide to use two vegetation indicators including LAI and NPP, respectively obtained from MODIS and AVHRR sensors. These indicators are applied to examine the spatiotemporal variation of vegetation coverage and density. We use yearly NPP data of MODIS (MOD17A3) and 8-day LAI data (MOD15A2H) between 2000 and 2015 (https://lpdaac.usgs.gov/) at the 1000-m and 500-m spatial resolutions, respectively. The AVHRR’s LAI data from 1982 to 2011 is derived from the Global Inventory Modeling and Mapping Studies (GIMMS) LAI3 g product. This is one of the longest LAI time series running from July 1981 to December 2011 with 15-day intervals and a 1/12 ° spatial resolution (Zhu et al., 2013; Pfeifer et al., 2014). The data has been validated against 29 field observation sites and other satellite-based LAI data (Cook and Pau, 2013). The AVHRR’s NPP data from 1982 to 2011, taken from Liang et al. (2015) has a spatial resolution of 8 km × 8 km and time intervals of 15 days. Monthly ET was obtained from multiple global ET products, including the Global Land Evaporation – the Amsterdam Method ET product (GLEAM Verison 3.1; Miralles et al., 2011; Martens et al., 2017), the AVHRR ET product (AVHRR; Zhang et al., 2010), the MOD16A2 ET product (MODIS, Mu et al., 2011), the Penman-MonteithLeuning ET product (PML; Zhang et al., 2016), and the Model Tree Ensembles ET product (MTE; Jung et al., 2010). The GLEAM ET estimate the components including actual evaporation, Et, Eb, Ei, openwater evaporation, and sublimation. GLEAM also computes surface and root-zone soil moisture, potential evaporation and evaporative stress conditions. It is a suitable tool for studying the water-energy effects and water-carbon interactions during vegetation restoration as all components of ET are available (Martens et al., 2017; Miralles et al., 2011). The GLEAM ET has a spatial resolution of 0.25° and a temporal coverage of 1982–2015 (Martens et al., 2017; Miralles, De Jeu, et al., 2011; Miralles, Holmes, et al., 2011), and was acquired from https://www. gleam.eu. The monthly AVHRR ET has a spatial resolution of 8 km and covers a period of 1982-2013. This product was obtained from the 81

Agricultural and Forest Meteorology 275 (2019) 79–90

C. Deng, et al.

investigate change rate of climatic elements and vegetation indexes at the pixel scale.

Table 1 Summary of detailed information for various datasets used in this study. Datasets

Time range

Temporal resolution

Spatial resolution

yearly

0.1°

2000–2015 1982–2010

China Meteorological Forcing Dataset MOD17A3 ——

yearly yearly

1 km 8 km

2001–2015 1982–2011 2000–2015 2003–2015

MOD15A2H AVHRR MCD43C3 MYD11C3

8-day 15-day 8-day monthly

500 m 1/12° 500 m 500 m

2001–2013 1979–2015 1982–2015 1982–2015 1982–2013 1982–2011 1982–2012 2000–2014

MCD12Q1 NOAA GLEAM GLEAM AVHRR MTE PML MODIS

yearly yearly daily daily monthly monthly monthly 8-day

500 m —— 0.25° 0.25° 1/12° 0. 5° 0. 5° 500 m

Temperature (T) Precipitation (P)

1982–2015

Vegetation net primary production (NPP) Leaf area index (LAI) Albedo Land surface temperature (Ts) Land cover CO2 density Evapotranspiration (ET)

Data name

Θ=

i=n

i=n

i=n

i=n n∑i = 1 i 2

i=n (∑i = 1

i)2

n∑i = 1 iXi − ∑i = 1 i∑i = 1 Xi −

(1)

where Θ is the change rate, i is the order of year from 1 to n, n is the number of years, and Xi represents the value of different variables when the number of the year is i. When Θ > 0, it indicates the variable increases over time, otherwise it decreases. The larger the absolute value, the more significant the change is. To a large extent, NPP can be affected by both T and P. We thus employed the partial correlation to examine the relationship between climatic elements and the vegetation status. Partial correlation measures the strength and direction of a linear relationship between two variables whilst controlling for the effect of one or more other variables. In this way, it helps reveal the individual role of T and P in vegetation dynamics. The Pearson’s partial coefficient (R) is computed to measure the strength of partial correlation between climatic variables and NPP

Rij, h =

2015; Peng et al., 2016; Wang et al., 2016a,b). The annual mean carbon dioxide data are derived from the National Oceanic and Atmospheric Administration (NOAA) Earth System Research Laboratory (ESRL) (https://www.esrl.noaa.gov/gmd/ccgg/trends/gl_data.html). All these quantitative data are interpolated into a spatial resolution of 0.1°. Monthly series of LAI and Ts are generated with the maximum value composite method (Holben, 1986) which can reduce noise caused by the instability of satellites, cloud contamination and uncorrected atmospheric effects. The data are then aggregated to inform an annual frequency. Details on the datasets herein are shown in Table 1.

Rij (1 −

2 R ih )(1

− R2jh)

(2)

where Rij, h is the Pearson’s partial correlation coefficient between i and j with the control of h, Rij , Rih and Rjh are the Pearson’s correlation coefficient between these two elements. When Rij, h > 0, it implies a positive relationship between climatic elements and NPP, otherwise it is negative. We also examine the significance of this correlation by performing Hypothesis Testing.

t=

r n−q−2 1 − r2

(3)

where t is the t-test value; n refers to sample size, and herein n is 29; q is the degree of freedom; and r is the partial correlation coefficient. When t < 0.05, the partial correlation is statistically significant (at a 95% statistical significance level).

3.2. Statistical analysis methods We employ a simple regression model to determine the interannual trends of climatic factors including T and P, and vegetation parameters. The least-squares method, based on a linear regression, is applied to

Fig. 2. Interannual changes of mean T (a), interannual changes of annual cumulative P (b), spatial distribution of mean annual T (c), and annual cumulative P trends (d) in the Three-North Region between 1982 and 2015. 82

Agricultural and Forest Meteorology 275 (2019) 79–90

C. Deng, et al.

4. Results and discussion

anthropogenic activities, land cover type in the TNR has changed drastically from 2001 to 2013, during which the area of forest and farmland is enlarged by respective 26.2% and 28.6%, and grassland area declines roughly 4 000 km2 (0.1%) (Table 2). In areas close to the Tianshan Mountains, the margins of the Junggar Basin, most regions of the Loess Plateau, the North-east Region, and the Three-North Shelterbelt of Xinjiang, the NPP shows an appreciable growth, more than 3 gC·m−2·yr-1. We find the largest increase exceeding 5 gC·m−2·yr−1 at the boundary of the North-east Regions and the southern loess plateau (Fig. 3e). All these regions are strongly influenced by the TNSP and “Grain for Green” projects. In addition to the project’s effect, the regional climate warming accelerates snow melt in the Mountains, yielding more water for vegetation restoration. On the contrary, the NPP tends to decline in the Horqin Grassland, the Qaidam Basin, and areas near the Altun Mountains where no large vegetation restoration program takes place; and the largest reduction surpasses 1 gC·m−2·yr-1. Additional results in Figs. 3 and S3 reveal consistency in the changing pattern between LAI and NPP.

4.1. Changes in climate and vegetation 4.1.1. Regional climate characteristics Fig. 2 shows annual mean T and cumulative P in the TNR, both of which increase significantly from 1982 to 2015. The average T has increased 1.46 °C over the past 34 years, equivalent to 0.43 °C /10 yr (R = 0.454, P < 0.01). The maximum T having a magnitude of 5.56 °C occurs in 2006 and the minimum value of 3.94 °C happens in 1996 (Fig. 2a). Similar to T, annual cumulative P increases over time, roughly 2.62 mm·yr−1 (R = 0.381, P < 0.01); and the increment has a total of 89.1 mm ending in 2015 (Fig. 2b). Due to the vast area and complicated geological conditions, the TNR exhibits an apparent heterogeneity in its regional climate (Fig. 2c and d). Most regions of the TNR feature a rising T during the last few decades. Over 89.2% of the whole area, particularly in the Turpan Depression, the Qaidam Basin, and the eastern Junggar Basin, the amplification of T is significant, at a rate of 1.2 °C/10 yr on average. The region that marks with a decreased T only constitutes 10.8% of the entire domain, wherein places deserving highlight are the Kunlun Mountains and the land adjacent to the Qaidam Basin and the Tarim Basin. Their T is characterized by a decreasing rate of approximate 1.2 °C /10 yr, and the decline at the rest areas is negligible (Fig. 2c). Regions that experience a reduction in P account for 8% of the total area (Fig. 2d), involving the southern Altai Mountains and Great Khingan, parts of the Loess Plateau and the Northern Plain. 92% of the total area shows a tendency of increased P since 1982. This upward trend is statistically significant over the mountainous region. For instance, the annual enhancement over the Kunlun Mountains, the Altun Mountains and the Tianshan Mountains is more than 12 mm. Comparing Fig. 2c with Fig. 2d, excepting the south of the Loess Plateau and the southern North-east Region, we find the increase of T and P bears a highly spatial consistency, and such change is likely to yield positive effects for vegetation growth.

4.2. Relationships between climatic elements and vegetation growth To examine each individual effect of T and P on vegetation growth (i.e., T and P), we perform the partial correlation analysis in Fig. 4a. Our result indicates that about 64% of areas have a positive correlation between NPP and T (Fig. S4a), among which 17% shows significantly positive (P < 0.05) correlation, mainly involving the southern Lesser Khingan Mountains, the North-east Plain, the Loess Plateau, southern Xinjiang Province and regions close to mountains. Over these regions, their partial correlation coefficient is approximately 0.6 (P < 0.05). Explanation for such positive correlation is that at high latitude or altitude, the role of relatively low T is to inhibit vegetation growth (Fig. S5), whereas increased T contributes to enhancing photosynthesis and lengthening the growing season (Xu et al., 2013), thus boosting NPP. Other regions in Xinjiang province display a negative correlation (Fig. S4a), in particular the west and north-east of the Tarim Basin and Turpan Depression where a significantly negative correlation (P < 0.05) exists between NPP and T (Fig. 4a). Depending upon their special continental desert climate type and unique location, the aridity of the Tarim Basin and the Turpan Depression is exceptional to the extent that annual cumulative P barely reaches 17.4 mm in some extreme cases (Chen et al., 2006). Since the potential evapotranspiration in these areas can easily mount to 3000 mm·yr−1, water shortage takes a great responsibility for limiting vegetation growth. Regarding the partial correlation between NPP and P, we find regions that carry a positive partial correlation are more than 84% of the whole area (Fig. S3b). Significant positive relationship accounts for 28%, mainly distributed in the main afforestation area, regions close to mountains and eastern Inner Mongolia, and in eastern Inner Mongolia (Fig. 4b). 16% of the whole study area has a negative correlation, scattered over the Greater Khingan Mountains, Qaidam Basin, margin of Tarim Basin and Junggar Basin (Fig. S4b). Statistically significant negative correlation is not common. Areas where NPP and T are significantly partial correlated account for 17% of the total; and areas where NPP and P are significantly partial correlated take up 28%. Hence, the result suggests that P plays a more important role in vegetation dynamics than T over the TNR.

4.1.2. Changes in vegetation coverage and density Mean annual NPP in the TNR ranges from 0 to 766 gC·m−2 with an average of 145 gC·m−2 during 1982–2010 (Fig. 3a). Tightly connected to the spatial distribution of P, our results show a gradual reduction in the amount of NPP from the east to the west. Briefly, the NPP of the Tianshan Mountains, the Kunlun Mountains, the Altai Mountains, the Loess Plateau, the North-east Region as well as the main afforestation areas has a magnitude larger than 300 gC·m−2. Moving to the west, the NPP in the Hexi Corridor and in most areas of Xinjiang province is below 100 gC·m−2. This result supports our finding that water stress is a key factor for reduced vegetation growth and carbon assimilation in these regions. We also examined the mean annual LAI in Fig. 3b, the distribution of which is analogous to that of NPP. The annual averaged LAI is 0.67, ranging from 0.11 to 2.8. Parts of the Loess Plateau and the north-east region have the highest LAI, far exceeding 1.4. By contrast, the LAI is less than 0.2 in the Hexi Corridor and in Xinjiang. Upward trend is noticed in both NPP and LAI from 1982 to 2015. This trend is more pronounced during the last 15 years (P < 0.05) (Fig. 3c and d). Analysis using both AVHRR and MODIS renders consistent results (Fig. S2 in the Supplementary Materials). We modified NPP and LAI data from MODIS based on polynomial curve fitting method, which is employed by using the best fitted polynomial generated by fitting data from a common period which both MODIS and AVHRR all have data on the MODIS data during 2011–2015 and 2012–2015, respectively. Both NPP and LAI witnessed a significant increase by 1.77 gC·m−2·yr-1 (P < 0.01) and 1.60·yr-1 (P < 0.01), respectively (Fig. 3c and d). The Chinese government endorses the TNSP in 1978 and the “Grain for Green” project in northern China in 1999. The latter brings significant changes in land cover type and improvement in vegetation coverage and density. As a result of intense

4.3. Surface water-energy variation after vegetation restoration 4.3.1. Effects of vegetation restoration on regional water budget Values of ET, NPP, and WUE vary with land cover types. The order for annual mean amount of ET, NPP, and WUE values for different land cover types, from largest to smallest, is farmland, forest, and grassland (Fig. 5). This trend implies that farmland has greater water consumption than forest or grassland. The average values of ET, NPP, and WUE 83

Agricultural and Forest Meteorology 275 (2019) 79–90

C. Deng, et al.

Fig. 3. Spatial distribution of mean NPP between 1982 and 2010 from AVHRR (a), mean LAI between 1982 and 2011 from AVHRR (b), interannual changes of NPP (c), and LAI (d). Spatial distribution of mean annual NPP trends between 1982 and 2010 (e) and spatial distribution of mean annual LAI trends between 1982 and 2011 (f) in the Three-North Region from AVHRR. Table 2 Annual areas of major land cover types in the TNR from 2001 to 2013 based on the MODIS land cover dataset. Land cover type (km2)

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013

Forest Grassland Farmland

47.5 151.6 38.7

52.6 150.4 45.9

51.1 152.2 44.1

60.2 177.1 51.9

51.1 152.2 44.1

51.1 152.2 44.1

51.1 152.2 44.1

51.1 152.2 44.1

51.1 152.2 44.1

51.1 152.2 44.1

60.0 151.2 49.8

60.0 151.2 49.8

60.0 151.2 49.8

Fig. 4. Spatial distribution of the partial correlation coefficients between annual NPP and annual mean T (a), and between annual NPP and annual cumulative P (b) over the period of 1982–2010 (the region in white indicates the partial correlation is not significant at a 95% statistical significance level). 84

Agricultural and Forest Meteorology 275 (2019) 79–90

C. Deng, et al.

Fig. 5. The distribution of yearly ET (a), NPP (b), and WUE (c) among three land cover types (Farmland, Forest and Grassland).

availability (Wa, defined as Wa = P − ET ) increases by 2.02 mm·yr−1 in the TNR, though less than the temporal increases of P (Fig. 7a). In contrast, regions becoming drier in terms of P show a noticeable reduction in Wa, a matter of the increased ET in these regions. Fig. 6c shows that ET in most regions of the TNR features a substantial increase, but Wa in these regions does not correspond to a declined amount in the recent past. This is mainly a reflection of the increased P and WUE, which alleviate the intensified water stress during the vegetation restoration. We find vegetation restoration can accelerate the regional water consumption given the result that the rising rate of ET over regions where the increase of NPP is greater than 3 g.C·m−2·yr−1 is 0.15 mm·yr−1 annually, more than ten times of the regions (0.01 mm·yr−1) where the increased NPP is less than 3 g.C·m−2·yr−1 (Fig. 7c). The correlation between NPP and Wa is largely negative (Fig. 7d), especially for Class 1 NPP changes (see caption for Fig. 7 for details of classes) which show a decreasing rate of 0.52 mm·g.C−1 m-2. The Wa has a negative value in most regions, indicating that the water resource in these regions is inadequate for vegetation restoration. The correlation between NPP and Wa for Class 3 dictates a negative relationship; nonetheless, the Wa is positive in most regions. For Class 2, there is no obvious correlation between NPP and Wa. A suggestion derived from our results is that to develop a sustainable vegetation

for farmland may be related to the variety in irrigation practices and land management methods. Our results are in a good agreement with Peng et al. (2015). Annual accumulated ET in the TNR increases at a rate of 0.35 mm·yr−1. Its variation has a range between the minimum of 211.51 mm in 2000 and the maximum of 258.74 mm in 2012 (Fig. 6a). Comparing Fig. 3 and Fig. 6c, we find the ET increases in regions where NPP and LAI also increase, such as the Tianshan Mountains, the Kunlun Mountains, the Qilian Mountains, the Qaidam Basin, and the Loess Plateau. Vegetation restoration in these regions of TNR causes the increasing of water consumption by ET, thus raising water stress. The WUE in the TNR enhances from 1982 to 2015. Using AVHRR and MODIS data, we estimate its time-varying change which is 0.034 and 0.055 gC·m−2 mm-1 per decade, respectively (Fig. 6b). The spatial distribution of these changes is not uniform (Fig. 6d) in a sense that area showing increasing trend fills more than 99% of the whole region. Regions with a relatively high NPP observe a remarkable growth of WUE. Figs. 6b and 6d show that the ET in areas where WUE has dramatically increased is not necessary to coincide with an increase. This finding indicates that the improved WUE is capable to partially balance against the growth of water stress caused by vegetation restoration. In accordance with the augmented P, annual surface water

Fig. 6. Interannual changes of ET (a) and spatial distributions of ET trends (c) from 1982 to 2015; and interannual changes of WUE (b) between 1982 and 2015 and spatial distributions of WUE trends (d) in the Three-North Region between 1982 and 2006. 85

Agricultural and Forest Meteorology 275 (2019) 79–90

C. Deng, et al.

Fig. 7. Interannual changes in annual surface water availability (Wa) (a), spatial distribution of Wa (b) from 1982 to 2015, interannual changes of ET in different NPP change classes (trendNPP < 3: where NPP change rate is below 3 g.C·m−2·yr−1, trendNPP≥3: where NPP change rate is over or equal to 3 g.C·m−2·yr−1) (c), and relationship between Wa and NPP in different classes (Class 1: where T < 0 °C and P < 200 mm; Class 2: where T ranges from 0 to 5 °C and P ranges from 200 to 400 mm; Class 3: when T ranges from 5 to 10 °C and P ranges from 400 to 600 mm and Class 4: when T > 10 °C and P > 600 mm, but there are no sites meeting this criteria) (d) in 1997 (P is the lowest) over the Three-North Region.

Inner Mongolia increases at a dramatic rate, approximately over 5.3 × 103 W·m−2·yr-1 (Fig. 8b). T in these regions concur with a significant enhancement (Fig. 2c). On the contrary, radiation decreases in the Altun Mountains, north-western Qilian Mountains and the Northwest Plain. The concentration of CO2 is mounting continuously by 1.8 ppm·yr-1 (R = 0.99, P < 0.05) (Fig. 8c), which can intensify the regional greenhouse effect. Affected by changes in vegetation and human activities, the land

restoration project in regions where P is below 200 mm excessive water consumption and soil water deficit should be avoided.

4.3.2. Vegetation effects on regional energy budget Our analysis reveals that the mean annual radiation profiles an increasing trend of (R = 0.268, P < 0.05) 1.3×103 W·m−2·yr-1 (Fig. 8a). The radiation of eastern Xinjiang, the western Qaidam Basin and the Junggar Basin, southern Great Khingan, the Loess Plateau, and southern

Fig. 8. Interannual changes in Net radiation (a), interannual changes of CO2 concentration (b), and spatial distribution of Net radiation trends (c) in the Three-North Region between 1982 and 2015. 86

Agricultural and Forest Meteorology 275 (2019) 79–90

C. Deng, et al.

Fig. 9. Interannual changes in albedo (a) and spatial distributions of albedo trends (b) in the Three-North Region from 2000 to 2015.

surface albedo shows a reduction of 0.00007·yr−1 (R = 0.005, P < 0.05) (Fig. 9a). Fig. 9b gives the spatial pattern of albedo changes. The albedo reduces more than 0.002·yr−1 over the regions including the Qilian Mountains, the Tianshan Mountains, Inner Mongolia, the middle of the Loess Plateau, and the northern Junggar Basin. At the Tarim Basin and the Qaidam Basin, the albedo increases slightly about 0.001·yr−1 (Fig. 9b). With the increase in radiation and CO2 concentration and a decrease in land surface albedo, the TNR is expected to experience a warming trend of Ts. Fig. 10 shows variations of Eb, Et, and Ei, components of ET over the TNR. The result reveals a sharp decline in the ratio of Eb/ET and Ei/ET, in contrast to the growing Et/ET ratio (Fig. 10d). The Yinshan Mountains, the northern Loess Plateau, the Junggar Basin, the Altun Mountains, the Qilian Mountains, the border of the Qaidam Basin and the main afforestation areas of the TNSP show a significant decrease in Eb/ ET. This is particularly noticeable over the main afforestation areas and the Altun Mountains (Fig. 10a). The distribution of transpiration changes is opposite to that of the bare soil evaporation and in regions where NPP and LAI increase, the Et/ET has a growing tendency with time illustrated by Fig. 10b and Fig. 3. In contrast, Et/ET decreases

considerably over places where a great increase of WUE is realized. One reason is, as improvement of vegetation continues, increased WUE begins to offset the increase in water consumption. Except for the eastern Inner Mongolian Plateau, the areas of the Qilian Mountains, the Loess Plateau and the Tianshan Mountains where Ei/ET increases are largely consistent with those showing an increase in LAI. Regions where Ei/ET decreases mainly encompass the north-eastern regions, the middle of the Inner Mongolian Plateau, the border of the Tarim Basin and the Tarim Basin itself, over which a reduction in NPP and LAI is observed. An effect of vegetation restoration is that the enhanced latent heat flux can partially compensate for the increased radiation absorbed by the land surface, therefore attenuating the increase in sensible heat flux. It should be noted that the spatial mean Ts in the TRN does not show apparent increasing trend (Fig. 11a). Fig. 11b reveals that 59.4% of the entire area experience a slight increase in Ts (0.08 °C·yr−1), covering the middle and west of the study area, and middle of the Inner Mongolian Plateau, the south-eastern Junggar Basin, the east of the Tarim Basin and parts of the Tianshan Mountains. Over the Loess Plateau, the Tianshan Mountains, the Kunlun Mountains, the boundary of the Junggar Basin, the north-eastern Tarim Basin and the main

Fig. 10. Spatial distribution of Eb/ET (a), Et/ET (b), and Ei/ET (c) trends in the Three-North Region and interannual changes of Eb/ET, Et/ET, Ei/ET from 1982 to 2015 (d). 87

Agricultural and Forest Meteorology 275 (2019) 79–90

C. Deng, et al.

Fig. 11. Interannual changes of Ts (a) and spatial distribution of Ts trends (b) in the Three-North Region between 2003 and 2015.

TNSP is acceptable in terms that it brings more advantages than disadvantages to the ecosystem of the study area. This work also seeks further understanding on the trade-off between carbon sequestration and water-energy consumption resulting from vegetation restoration in the TNSP.

afforestation areas, especially the Lesser Khingan Mountains and the Great Khingan Mountains, Ts shows a tendency of decrease at a rate of 0.16 °C ·yr−1. With the impact of TNSP and “Grain for Green” project, ET and NPP in the north-east, the Loess Plateau, and the main afforestation areas increase significantly while Ts decreases. This evidence is supportive of our finding that vegetation restoration in these regions can lower the temperature of land surface by increasing latent heat flux and mitigate the greenhouse effect by assimilating more CO2.

Acknowledgements This work is jointly supported by the National Natural Science Foundation of China (41790421, 41877150, and 51609111), the Strategic Priority Research Program of Chinese Academy of Sciences (XDA20100102), the Natural Science Foundation of Qinghai Province in China (2018-ZJ-936Q), and the Academic Department Consultation Program of Chinese Academy of Sciences (DX-2017-B05).

5. Conclusions In this study, we examine the spatiotemporal variation of various climate variables and vegetation coverage/density over the TNR of China using multiple datasets. Our purpose is to deliver a better understanding of vegetation dynamics and its effects on the surface waterenergy balance and the trade-off between carbon-water during one of the grand projects for vegetation restoration in history. We find both vegetation density and coverage in the TNR are significantly improved, affirmed by the enhanced NPP and LAI after the start of the TNSP. Annual mean air temperature (T) and accumulated precipitation (P) increase by 1.46 °C and 89.1 mm from 1982 to 2015, respectively. Such long-term changes in air temperature and precipitation can be tightly related to the land-atmosphere coupling. For instance, through the impact on the partitioning of the incoming energy in the latent and sensible heat fluxes, land soil moisture and vegetation coverage have several impacts on climate processes, involving air temperature and precipitation. The main point of this study is to evaluate how the surface water-energy balance may be altered as a result of the vegetation restoration, rather than to determine which factor is the driver of such change. To a large degree, vegetation dynamics positively correlate with T and P, though more sensitive to P. The improved vegetation coverage and density result in more carbon assimilation at the cost of increasing water consumption on the order of 0.35 mm·yr−1. However, the surface water availability in this region does not commensurate with a decrease (Fig. S6 and Table S3 also show that the observed annual streamflow over most river basins located in the TNR have no substantial reduction), as the increased P and water use efficiency operate in the opposite direction. Vegetation restoration accelerates the water deficit, especially in arid regions undergoing a high vegetation recovery rate. These results call attention on developing sustainable vegetation restoration in the arid regions. The absorbed solar radiation that heats the land surface has been increasing over the past few decades. This is an effect of the intensified solar radiation and decreased land surface albedo associated with the vegetation restoration. Land surface temperature does not show a significant increasing trend in regions where vegetation has been recovered significantly. This is due to the cooling effects, a result of the increased latent heat flux and decreased sensible heat flux through vegetation transpiration, and the reduced greenhouse effects with more carbon sequestration. Collectively, our results demonstrate that the initial performance of

Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.agrformet.2019.05. 012. References Betts, R., Cox, P.M., Lee, S.E., Woodward, F.I., 1997. Contrasting physiological and structural vegetation feedbacks in climate change simulations. Nature 387 (6635), 796–799. https://doi.org/10.1038/42924. Bounoua, L., Collatz, G.J., Los, S.O., Sellers, P.J., Dazlich, D.A., Tucker, C.J., Randall, D.A., 2000. Sensitivity of climate to changes in NDVI. J. Clim. 13 (13), 2277–2292. https://doi.org/10.1175/15200442(2000)013<2277:SOCTCI>2.0.CO;2. Bounoua, L., Hall, F.G., Sellers, P.J., Kumar, A., Collatz, G.J., Tucker, C.J., Imhoff, M.L., 2010. Quantifying the negative feedback of vegetation to greenhouse warming: a modeling approach. Geophys. Res. Lett. 37 (23), 817–824. https://doi.org/10.1029/ 2010GL045338. Campos, P., Guillermo, E., Moran, M.S., et al., 2013. Ecosystem resilience despite largescale altered hydroclimatic conditions. Nature 494 (7437), 349–352. https://doi.org/ 10.1038/nature11836. Cao, S.X., 2008. Why large-scale afforestation efforts in China have failed to solve the desertification problem. Environ. Sci. Technol. 42 (6), 1826–1831. https://doi.org/ 10.1021/es0870597. Chen, Y.N., Takeuchi, K., Xu, C.C., Chen, Y.P., Xu, Z.X., 2006. Regional climate change and its effects on river runoff in the Tarim Basin. China. Hydrol. Process. 20 (10), 2207–2216. https://doi.org/10.1002/hyp.6200. Chen, Y.Y., Yang, K.K., He, J., Qin, J., Shi, J.C., Du, J.Y., He, Q., 2011. Improving land surface temperature modeling for dry land of China. J. Geophys. Res. Atmos. 116 (D20), D20104. https://doi.org/10.1029/2011JD01592. Cook, B.I., Pau, S., 2013. A global assessment of long-term greening and browning trends in pasture lands using the GIMMS LAI3g dataset. Remote Sens. (Basel) 5 (5), 2492–2512. https://doi.org/10.3390/rs5052492. Dan, L., Ji, J.J., Ma, Z.G., 2004. The variation of net primary production and leaf area index over Xinjiang Autonomous Region and its response to climate change. Acta Ecol. Sin. 27 (9), 3582–3592 in Chinese with English abstract. Duan, H., Yan, C.Z., Tsunekawa, A., Song, X., Li, S., Xie, J.L., 2011. Assessing vegetation dynamics in the Three-North Shelter Forest region of China using AVHRR NDVI data. Environ. Earth Sci. 64 (4), 1011–1020. https://doi.org/10.1007/s12665-011-0919-x. Fang, O.Y., Wang, Y., Shao, X.M., 2016. The effect of climate on the net primary productivity (NPP) of Pinus koraiensis, in the Changbai Mountains over the past 50 years. Trees 30 (1), 281–294. https://doi.org/10.1007/s00468-015-1300-6. Feng, H.H., Zou, B., Luo, J.H., 2017. Coverage-dependent amplifiers of vegetation change on global water cycle dynamics. J. Hydrol. (Amst) 550, 220–229. https://doi.org/10.

88

Agricultural and Forest Meteorology 275 (2019) 79–90

C. Deng, et al.

Mu, Q.Z., Zhao, M.S., Running, S.W., 2011. Improvements to a MODIS global terrestrial evapotranspiration algorithm. Remote Sens. Environ. 115 (8), 1781–1800. https:// doi.org/10.1016/j.rse.2011.02.019. Peel, M.C., McMahon, T.A., Finlayson, B.L., 2010. Vegetation impact on mean annual evapotranspiration at a global catchment scale. Water Resour. Res. 46 (9), 2095–2170. https://doi.org/10.1029/2009WR008233. Pei, F.S., Wu, C.J., Liu, X.P., et al., 2018. Monitoring the vegetation activity in China using vegetation health indices. Agric. For. Meteorol. 248, 215–227. https://doi.org/ 10.1016/j.agrformet.2017.10.001. Peng, D.L., Wu, C.Y., Zhang, B., et al., 2016. The influences of drought and land-cover conversion on inter-annual variation of NPP in the Three-North Shelterbelt Program Zone of China based on MODIS data. PLoS One 11 (6), e0158173. https://doi.org/10. 1371/journal.pone.0158173. Peng, H.Y., Thevs, N., Rozi, A., Zerbe, S., Abdusalih, N., 2015. Water allocation and water consumption of irrigated agriculture and natural vegetation in the Aksu-Tarim river basin, Xinjiang. China. J. Arid Environ. 112, 87–97. https://doi.org/10.1016/j. jaridenv.2014.05.028. Peng, S.S., Chen, A.P., Xu, L., et al., 2011. Recent change of vegetation growth trend in China. Environ. Res. Lett. 6, 044027–044040. https://doi.org/10.1088/1748-9326/ 6/4/044027. 2011. Pfeifer, M., Lefebvre, V., Gonsamo, A., et al., 2014. Validating and linking the GIMMS leaf area index (LAI3g) with environmental controls in Tropical Africa. Remote Sens. (Basel) 6 (3), 1973–1990. https://doi.org/10.3390/rs6031973. Piao, S.L., Ciais, P., Huang, Y., et al., 2010. The impacts of climate change on water resources and agriculture in China. Nature 467 (7311), 43–51. https://doi.org/10. 1038/nature09364. Piao, S.L., Fang, J.Y., Guo, Q., 2001. Terrestrial net primary production and its spatiotemporal patterns in China during 1982–1999. Acta Scicentiarum Naturalum Universitis Pekinesis 37 (4), 563–569. https://doi.org/10.13209/j.0479-8023.2001. 102. (in Chinese with English Abstract). Qian, W.H., Lin, X., 2005. Regional trends in recent precipitation indices in China. Meteorol. Atmos. Phys. 90, 193–207. https://doi.org/10.1007/s00703-004-0101-z. Qu, J.J., Cao, S.X., Li, G.S., Niu, Q.H., Feng, Q., 2014. Conservation of natural and cultural heritage in Dunhuang. China. Gondwana Res. 26, 1216–1221. https://doi.org/ 10.1016/j.gr.2013.08.017. Sang, W.G., Su, H.X., 2009. Interannual NPP variation and trend of Picea schrenkiana forests under changing climate conditions in the Tianshan Mountains, Xinjiang. China. Ecol. Res. 24 (2), 441–452. https://doi.org/10.1007/s11284-008-0521-1. Shen, Q., Gao, G.Y., Fu, B.J., Lu, Y.H., 2015. Responses of shelterbelt stand transpiration to drought and groundwater variations in an arid inland river basin of Northwest China. J. Hydrol. (Amst) 531, 738–748. https://doi.org/10.1016/j.jhydrol.2015.10. 053. Shi, Y.F., Shen, Y.P., Kang, E.S., 2007. Recent and future climate change in Northwest China. Clim. Change 80 (3–4), 379–393. https://doi.org/10.1007/s10584-0069121-7. Vinukollu, R.K., Wood, E.F., Ferguson, C.R., Fisher, J.B., 2011. Global estimates of evapotranspiration for climate studies using multi-sensor remote sensing data: evaluation of three process-based approaches. Remote Sens. Environ. 115 (3), 801–823. https://doi.org/10.1016/j.rse.2010.11.006. Wang, H., Liu, G.H., Li, Z.S., Ye, X., Wang, M., Gong, L., 2016a. Impacts of climate change on net primary productivity in arid and semiarid regions of China. Chinese Geogr. Sci. 26 (1), 35–47. https://doi.org/10.1007/s11769-015-0762-1. Wang, S., Fu, B.J., Piao, S.L., Lu, Y.H., Ciais, P., Feng, X.M., Wang, Y.F., 2016b. Reduced sediment transport in the Yellow River due to anthropogenic changes. Nat. Geosci. 9, 38–42. https://doi.org/10.1038/NCLIMATE3092. Williams, C.A., Reichstein, M., Buchmann, N., 2012. Climate and vegetation controls on the surface water balance: synthesis of evapotranspiration measured across a global network of flux towers. Water Resour. Res. 48 (6), W06523. https://doi.org/10. 1029/2011wr011586. Williams, I.N., Torn, M.S., 2015. Vegetation controls on surface heat flux partitioning, and land-atmosphere coupling. Geophys. Res. Lett. 42, 9416–9424. https://doi.org/10. 1002/2015GL066305. Wilmers, F., 1990. Effects of vegetation on urban climate and buildings. Energ. Buildings 15 (3), 507–514. https://doi.org/10.1016/0378-7788(90)90028-H. Xu, L., Myneni, R.B., Chapin III, F.S., et al., 2013. Temperature and vegetation seasonality diminishment over northern lands. Nat. Clim. Change 3, 581–586. https://doi.org/ 10.1038/nclimate1836. Xu, L.G., Zhou, H.F., Liang, C., Du, L., Li, H., 2010. Spatial and temporal variability of annual and seasonal precipitation over the desert region of China during 1951–2005. Hydrol. Process. 24 (20), 2947–2959. https://doi.org/10.1002/hyp.7708. Yan, S.H., Dong, S.C., Li, Z.H., Liu, J.J., Wang, J.N., 2013. Carrying capacity of water resources for three-north shelterbelt construction in China. J. Resour. Ecol. 4 (1), 50–55. https://doi.org/10.5814/j.issn.1674-764x.2013.01.007. Zhang, L., Dawes, W.R., Walker, G.R., 2001. Response of mean annual evapotranspiration to vegetation changes at catchment scale. Water Resour. Res. 37 (3), 701–708. https://doi.org/10.1029/2000WR900325. Zhang, B.Q., He, C.S., Burnham, M., Zhang, L.H., 2016. Evaluating the coupling effects of climate aridity and vegetation restoration on soil erosion over the Loess Plateau in China. Sci. Total Environ. 539, 436–449. https://doi.org/10.1016/j.scitotenv.2015. 08.132. Zhang, B.Q., He, C.S., 2016. A modified water demand estimation method for drought identification over arid and semiarid regions. Agric. For. Meteorol. 230–231, 58–66. https://doi.org/10.1016/j.agrformet.2015.11.015. Zhang, B.Q., Wang, Z.K., Chen, G., 2017. A sensitivity study of applying a two-source potential evapotranspiration model in the standardized precipitation evapotranspiration index for drought monitoring. Land Degrad. Dev. 28, 783–793. https://

1016/j.jhydrol.2017.04.056. Feng, X.M., Fu, B.J., Piao, S.L., et al., 2016. Revegetation in China’s Loess Plateau is approaching sustainable water resource limits. Nat. Clim. Change 6, 1019–1022. https://doi.org/10.1038/nclimate3092. Gao, X.D., Li, H.C., Zhao, X.N., Ma, W., Wu, P.T., 2018a. Identifying a suitable revegetation technique for soil restoration on water limited and degraded land: considering both deep soil moisture deficit and soil organic carbon sequestration. Geoderma 319, 61–69. https://doi.org/10.1016/j.geoderma.2018.01.003. Gao, X.D., Zhao, X.N., Li, H.C., Guo, L., Lu, T., Wu, P.T., 2018b. Exotic shrub species (Caragana korshinskii) is more resistant to extreme natural drought than native species (Artemisia gmelinii) in a semiarid revegetated ecosystem. Agric. For. Meteorol. 263, 207–216. https://doi.org/10.1016/j.agrformet.2018.08.029. Gao, X.D., Liu, Z.P., Zhao, X.N., Ling, Q., Huo, G.P., Wu, P.T., 2018c. Extreme natural drought enhances interspecific facilitation in semiarid agroforestry systems. Agric. Ecosyst. Environ. 265, 444–453. https://doi.org/10.1016/j.agee.2018.07.001. Guo, L.H., Wu, S.H., Zhao, D.S., Yin, Y.H., Leng, G.Y., Zhang, Q.Y., 2014. NDVI-based vegetation change in Inner Mongolia from 1982 to 2006 and its relationship to climate at the biome scale. Adv. Stat. Climatol. Meteorol. Oceanogr. 2014 (4), 79–92. https://doi.org/10.1155/2014/692068. He, B., Chen, A.F., Wang, H.L., Wang, Q.F., 2015. Dynamic response of satellite-derived vegetation growth to climate change in the three north shelter forest region in China. Remote Sens. (Basel) 7 (8), 9998–10016. https://doi.org/10.3390/rs70809998. Holben, B.N., 1986. Characteristics of maximum-value composite images from temporal AVHRR data. Int. J. Remote Sens. 7 (11), 1417–1434. https://doi.org/10.1080/ 01431168608948945. Hu, Z.M., Yu, G.R., Fu, Y.L., Sun, X.M., Li, Y.N., Shi, P.L., Wang, Y.F., Zheng, Z.M., 2008. Effects of vegetation control on ecosystem water use efficiency within and among four grassland ecosystems in China. Glob. Change Biol. Bioenergy 14 (7), 1609–1619. https://doi.org/10.1111/j.1365-2486.2008.01582.x. Huang, W., Feng, S., Liu, C., Chen, J., Chen, J.H., Chen, F.H., 2018. Changes of climate regimes during the last millennium and the twenty-first century simulated by the Community Earth System Model. Quaternary Sci. Rev. 180, 42–56. https://doi.org/ 10.1016/j.quascirev.2017.11.019. Jeong, S.J., Ho, C.H., Jeong, J.H., 2009. Increase in vegetation greenness and decrease in springtime warming over east Asia. Geophys. Res. Lett. 36 (2), L02710. https://doi. org/10.1029/2008gl036583. Jia, K., Liang, L., Liu, J.Y., Li, Q.Z., Wei, X.Q., Yuan, W.P., Yao, Y.J., 2015. Forest cover changes in the three-north shelter forest region of China during 1990 to 2005. Int. J. Agric. Environ. Inf. Syst. 26 (2), 206–215. https://doi.org/10.3808/jei.201400268. Jian, S.Q., Zhao, C.Y., Fang, S.M., Yu, K., 2015. Effects of different vegetation restoration on soil water storage and water balance in the Chinese Loess Plateau. Agric. For. Meteorol. 206, 85–96. https://doi.org/10.1016/j.agrformet.2015.03.009. Jiao, Y., Lei, H.M., Yang, D.W., Huang, M.Y., Liu, D.F., Yuan, X., 2017. Impact of vegetation dynamics on hydrological processes in a semi-arid basin by using a land surface-hydrology coupled model. J. Hydrol. (Amst) 551, 116–131. https://doi.org/10. 1016/j.jhydrol.2017.05.060. Jiapaer, G., Liang, S.L., Yi, Q.X., Liu, J.P., 2015. Vegetation dynamics and responses to recent climate change in Xinjiang using leaf area index as an indicator. Ecol. Indic. 58, 64–76. https://doi.org/10.1016/j.ecolind.2015.05.036. Jung, M., Reichstein, M., Bondeau, A., 2009. Towards global empirical upscaling of FLUXNET eddy covariance observations: validation of a model tree ensemble approach using a biosphere model. Biogeosciences 6, 2001–2013. https://doi.org/10. 5194/bg-6-2001-2009. Jung, M., Reichstein, M., Ciais, P., Seneviratne, S.I., Sheffield, J., Goulden, M.L.…, Zhang, K., 2010. Recent decline in the global land evapotranspiration trend due to limited moisture supply. Nature 467 (7318), 951–954. https://doi.org/10.1038/ nature09396. Kala, J., Decker, M., Exbrayat, J.F., Pitman, A.J., Carouge, C., Evans, J.P., Abramowitz, G., 2013. Influence of leaf area index prescriptions on simulations of heat, moisture, and carbon fluxes. J. Hydrometeorol. 15 (1), 489–503. https://doi.org/10.1175/ JHM-D-13-063.1. Li, M.M., Liu, A.T., Zou, C.J., Xu, W.D., Shimizu, H., Wang, K.Y., 2012. An overview of the “Three-North” Shelterbelt project in China. For. Stud. China 14 (1), 70–79. https:// doi.org/10.1007/s11632-012-0108-3. Li, Y., Zhao, M.S., Motesharrei, S., Mu, Q.Z., Kalnay, E., Li, S.C., 2015. Local cooling and warming effects of forests based on satellite observations. Nat. Commun. 6, 6603. https://doi.org/10.1038/ncomms7603. Liang, W., Yang, Y.T., Fan, D.M., et al., 2015. Analysis of spatial and temporal patterns of net primary production and their climate controls in China from 1982 to 2010. Agric. For. Meteorol. 204, 22–36. https://doi.org/10.1016/j.agrformet.2015.01.015. Liu, J.Y., Liu, M.L., Zhuang, D.F., Zhang, Z.X., Deng, X.Z., 2003. Study on spatial pattern of land-use change in China during 1995–2000. Sci. China Ser. D Earth Sci. 46 (4), 373–384. https://doi.org/10.1360/03yd9033. Liu, W.H., Zhu, J.J., Jia, Q.Q., Zheng, X., Li, J.S., Lou, X.D., Hu, L.L., 2014. Carbon sequestration effects of Shrublands in Three-North Shelterbelt Forest Region. China. Chinese Geogr. Sci. 24 (4), 444–453. https://doi.org/10.1007/s11769-014-0698-x. Martens, B., Miralles, D.G., Lievens, H., et al., 2017. GLEAM v3: satellite-based land evaporation and root-zone soil moisture. Geosci. Model Dev. 10 (5), 1903–1925. https://doi.org/10.5194/gmd-2016-162. Miralles, D.G., Holmes, T.R.H., de Jeu, R.A.M., Gash, J.H., Meesters, A.G.C.A., Dolman, A.J., 2011. Global land-surface evaporation estimated from satellite-based observations. Hydrol. Earth Syst. Sci. Discuss. 15 (2), 453–469. https://doi.org/10.5194/ hessd-7-8479-2010. Myhre, G., Kvalevag, M.M., Schaaf, C.B., 2005. Radiative forcing due to anthropogenic vegetation change based on MODIS surface albedo data. Geophys. Res. Lett. 32 (21), L21410. https://doi.org/10.1029/2005GL024004.

89

Agricultural and Forest Meteorology 275 (2019) 79–90

C. Deng, et al.

465–480. https://doi.org/10.1007/s00704-015-1646-0. Zhou, D.Y., Wang, X.J., Shi, M.J., 2017. Human driving forces of oasis expansion in northwestern china during the last decade-a case study of the Heihe River Basin. Land Degrad. Dev. 28, 412–420. https://doi.org/10.1002/ldr.2563. Zhu, Z., Bi, J., Pan, Y.Z., et al., 2013. Global data sets of vegetation leaf area index (LAI) 3g and Fraction of Photosynthetically Active Radiation (FPAR) 3g derived from Global Inventory Modeling and Mapping Studies (GIMMS) Normalized Difference Vegetation Index (NDVI3g) for the period 1981 to 2011. Remote Sens. (Basel) 5, 927–948. https://doi.org/10.3390/rs5020927. Zeng, Z.Z., Piao, S.L., Li, L.Z.X., et al., 2017. Climate mitigation from vegetation biophysical feedbacks during the past three decades. Nat. Clim. Change 7 (6), 432–439. https://doi.org/10.1038/NCLIMATE3299.

doi.org/10.1002/ldr.2548. Zhang, K., Kimball, J.S., Nemani, R.R., Running, S.W., 2010. A continuous satellite-derived global record of land surface evapotranspiration from 1983 to 2006. Water Resour. Res. 46, W02502. https://doi.org/10.1029/2009WR008800. Zhang, S.L., Yang, D.W., Yang, Y.T., Piao, S.L., Yang, H.B., Lei, H.M., Fu, B.J., 2018. Excessive afforestation and soil drying on China’s Loess Plateau. J. Geophys. Res. Biogeo. 123, 923–935. https://doi.org/10.1002/2017JG004038. Zheng, X., Zhu, J.J., Xing, Z.F., 2016. Assessment of the effects of shelterbelts on crop yields at the regional scale in Northeast China. J. Agric. Food Syst. Community Dev. 143, 49–60. https://doi.org/10.1016/j.agsy.2015.12.008. Zheng, X., Zhu, J.J., 2017. A new climatic classification of afforestation in Three-North regions of China with multi-source remote sensing data. Theor. Appl. Climatol. 127,

90