Spatiotemporal Analysis of Natural Grassland NPP in Xinjiang in the Past 20 Years

CHEN Chunbo¹², LI Gangyong², PENG Jian²

¹State Key Laboratory of Desert and Oasis Ecology, Xinjiang Institute of Ecology and Geography, Chinese Academy of Sciences, Urumqi 830011, Xinjiang, China
²Joint Laboratory for Remote Sensing Monitoring of Grassland Ecosystem in Arid Area, Xinjiang Grassland Technical Promotion Station, Urumqi 830049, Xinjiang, China

Abstract

Using MOD17A3HGF NPP data along with temperature and precipitation records, this study employs trend analysis and partial correlation analysis at multiple scales—including the entire Xinjiang region, northern and southern Xinjiang, various prefectures, and different grassland types—to investigate the spatiotemporal dynamics of natural grassland net primary productivity (NPP) and its response to climate change from 2000 to 2018. The results demonstrate that natural grassland NPP exhibited a fluctuating increasing trend over the study period, with a multi-year average of 0.103 kg C·m⁻². Spatially, NPP gradually decreased from the western Junggar Mountains, Ili River Valley, Tianshan Mountains, and Altai Mountains toward the Junggar and Tarim Basins. The mean NPP in northern Xinjiang (0.149 kg C·m⁻²) significantly exceeded that in southern Xinjiang (0.055 kg C·m⁻²), though both regions showed increasing trends. Grassland NPP across all prefectures generally increased, but with distinct interannual variations. Among the eleven grassland types analyzed (excluding alpine desert), NPP showed increasing trends, though with notable differences between types. The warming-wetting trend in Xinjiang since 2000 has generally benefited grassland vegetation growth, but enhanced interannual precipitation variability has intensified fluctuations in grassland NPP. These findings provide fundamental data for assessing grassland carbon budgets and promote sustainable grassland management in Xinjiang.

Keywords: asymmetric warming-wetting; net primary productivity; spatiotemporal dynamics; space-earth observation; natural grassland in Xinjiang

1. Introduction

1.1 Study Area Overview

The Xinjiang Uygur Autonomous Region (hereafter "Xinjiang") is located in northwestern China (34°22′–49°33′N, 73°32′–96°21′E), covering an area of 1.66×10⁶ km², approximately one-sixth of China's total land area. Characterized by a typical arid climate with strong continental influences, Xinjiang experiences low precipitation (annual average ~150 mm), high evaporation, and mean annual temperatures of 9–12°C. The region's topography follows a "three mountains surrounding two basins" pattern, with the Altai Mountains, Junggar Basin, Tianshan Mountains, Tarim Basin, and Kunlun Mountains arranged from north to south. Elevation ranges from -155 m (Ayding Lake in Turpan) to 8611 m (K2 peak). Based on geographic and climatic characteristics, Xinjiang is divided into northern and southern regions. Northern Xinjiang includes Urumqi, Wujiaqu, Shihezi, Karamay, Beitun, Shuanghe, Changji Hui Autonomous Prefecture, Bortala Mongol Autonomous Prefecture, and Ili Kazakh Autonomous Prefecture (including Altay and Tacheng). Southern Xinjiang comprises Aksu, Kashgar, Hotan, Kizilsu Kirghiz Autonomous Prefecture, Bayingolin Mongol Autonomous Prefecture, Turpan, Hami, Aral, Tumshuq, Tiemenguan, and Kunyu.

Xinjiang's unique natural environment, far from oceans and surrounded by mountains with enclosed basins, creates distinct spatial patterns of water and heat resources that support diverse natural grassland ecosystems. The region's grasslands cover 5,725.88×10⁴ hm², ranking among China's five major pastoral areas. According to the Xinjiang grassland classification system, there are 11 major grassland types: alpine desert, alpine steppe, alpine meadow, mountain meadow, temperate meadow steppe, temperate steppe, temperate desert steppe, temperate steppe desert, temperate desert, lowland meadow, and swamp.

1.2 Data Sources and Preprocessing

The primary datasets included MOD17A3HGF NPP data, climate reanalysis data, and auxiliary information. The MOD17A3HGF Version 6 product (annual scale, 500 m spatial resolution) was obtained from NASA's LP DAAC. Grassland type classification data were provided by the Xinjiang Grassland Technical Promotion Station. Climate data (temperature and precipitation) were acquired from the China Meteorological Forcing Dataset, available through the National Tibetan Plateau Data Center. This dataset, based on GEWEX-SRB and Princeton GLDAS with fusion of China Meteorological Administration observations, offers superior accuracy compared to other reanalysis products.

Data preprocessing involved: (1) converting MOD17A3HGF data from HDF to GeoTIFF format, (2) mosaicking and projecting images, and (3) clipping to Xinjiang's administrative boundaries using ArcGIS 10.2. Climate data processing included converting precipitation rates to annual totals and calculating mean annual temperatures. All spatial operations employed the Geospatial Data Abstraction Library (GDAL) implemented in Python.

1.3 Research Methods

1.3.1 Statistical Analysis of Grassland NPP

Using the MOD17A3HGF NPP dataset, we calculated annual statistics (mean, maximum, minimum, and standard deviation) at multiple scales: prefecture-level, northern/southern Xinjiang, and entire region. The multi-year average spatial pattern was derived from 2000–2018 NPP data. Grassland-type-specific NPP values were extracted by masking the NPP data with each grassland type classification. All spatial data processing, statistical calculations, and output operations were automated using GDAL in Python.

1.3.2 Interannual Variation Analysis

Interannual variation rates were quantified using linear regression (least squares method) for each pixel:

$$k = \frac{n \times \sum_{i=1}^{n}(i \times NPP_i) - (\sum_{i=1}^{n}i) \times (\sum_{i=1}^{n}NPP_i)}{n \times \sum_{i=1}^{n}i^2 - (\sum_{i=1}^{n}i)^2}$$

where n is the study period length (19 years), i is the year index (1–19), and NPPᵢ is the NPP value in year i. The slope k represents the annual change rate: positive values indicate increasing trends, negative values indicate decreasing trends.

1.3.3 Correlation Analysis with Climate Factors

Partial correlation analysis was conducted at the pixel scale to quantify relationships between NPP and climate variables (temperature, precipitation) while controlling for the other variable:

$$r_{xy} = \frac{\sum_{i=1}^{n}(x_i - \bar{x})(y_i - \bar{y})}{\sqrt{\sum_{i=1}^{n}(x_i - \bar{x})^2 \sum_{i=1}^{n}(y_i - \bar{y})^2}}$$

where rₓᵧ is the correlation coefficient between parameter x (NPP) and parameter y (temperature or precipitation), xᵢ and yᵢ are annual values, and x̄ and ȳ are multi-year means.

2. Results

2.1 Spatiotemporal Patterns of Natural Grassland NPP

2.1.1 Interannual Variation

From 2000 to 2018, Xinjiang's natural grassland NPP showed a fluctuating increasing trend, with a multi-year average of 0.103 kg C·m⁻². The 2018 value (0.122 kg C·m⁻²) represented a 34.1% increase from 2000 (0.091 kg C·m⁻²). The study period exhibited significant interannual variability, with alternating positive and negative growth years. The maximum annual increment occurred in 2010 (0.031 kg C·m⁻²), while the minimum occurred in 2004 (-0.014 kg C·m⁻²). After 2010, interannual fluctuations intensified, with the standard deviation increasing from 0.014 kg C·m⁻² (2000–2009) to 0.022 kg C·m⁻² (2010–2018).

2.1.2 Spatial Distribution Pattern

The multi-year average NPP displayed substantial spatial heterogeneity (Figure 4). High-value areas (>0.50 kg C·m⁻²) were concentrated in the Ili River Valley, Tianshan Mountains, and Altai Mountains, while low-value areas (<0.05 kg C·m⁻²) occupied the central Junggar Basin and Tarim Basin margins. Regions with NPP >0.20 kg C·m⁻² accounted for only 19.5% of total grassland area, primarily in the Ili River Valley, Altai Mountains, and Tianshan ranges. Conversely, 45.6% of grassland area had NPP <0.05 kg C·m⁻², distributed in the central Junggar Basin, northern Tianshan slopes, and Tarim Basin periphery. Approximately 80.5% of grassland area exhibited NPP below 0.20 kg C·m⁻², indicating generally low productivity across Xinjiang.

The spatial pattern of interannual change rates (Figure 5) revealed that 63.1% of grassland area showed increasing trends, particularly in the Ili River Valley where increments exceeded 0.40 kg C·m⁻². Decreasing trends were observed in 36.9% of area, mainly in the central Junggar Basin and eastern Tianshan regions.

2.2 Regional Variations in Grassland NPP

2.2.1 Northern vs. Southern Xinjiang

Both northern and southern Xinjiang exhibited increasing NPP trends, but with notable differences. Northern Xinjiang's multi-year average NPP (0.149 kg C·m⁻²) was substantially higher than southern Xinjiang's (0.055 kg C·m⁻²), though northern Xinjiang showed greater interannual variability. The Ili Kazakh Autonomous Prefecture achieved the highest NPP (0.365 kg C·m⁻²), followed by Bortala Mongol Autonomous Prefecture (0.182 kg C·m⁻²) and Tacheng (0.166 kg C·m⁻²). In southern Xinjiang, NPP ranged from 0.03–0.09 kg C·m⁻² across prefectures, with Turpan showing the lowest values.

2.2.2 Prefecture-Level Variations

All prefectures displayed increasing NPP trends, but with distinct interannual patterns (Figure 7). Northern prefectures (except Karamay) consistently outperformed southern prefectures. NPP values ranged from 0.037 kg C·m⁻² (Hotan) to 0.30 kg C·m⁻² (Ili), with most southern prefectures clustering below 0.09 kg C·m⁻².

2.3 Grassland Type-Specific NPP Trends

Analysis of 11 grassland types revealed increasing trends for all categories except alpine desert (Figure 8). Mountain meadow showed the highest multi-year average NPP (0.385 kg C·m⁻²), followed by swamp (0.322 kg C·m⁻²) and temperate meadow steppe (0.223 kg C·m⁻²). Alpine desert exhibited minimal change (0.002 kg C·m⁻²). Temperate desert steppe and temperate desert showed moderate values (0.041–0.042 kg C·m⁻²), while temperate steppe desert displayed slight increases.

2.4 Response of Grassland NPP to Climate Factors

Spatial correlation analysis (Figure 9) revealed complex relationships between NPP and climate variables. NPP-temperature correlations were mixed: 41.74% of area showed positive correlation (primarily on northern Tianshan slopes in Changji and Urumqi), while 58.26% showed negative correlation, especially in southern Xinjiang. The mean correlation coefficient was -0.076. In contrast, NPP-precipitation correlations were predominantly positive (79.59% of area), with 20.41% showing high positive correlation concentrated in the Ili River Valley, northern Tianshan slopes, and western Junggar Mountains. The mean precipitation correlation coefficient was 0.076.

3. Discussion

Under the combined influence of climate change and human activities, Xinjiang's natural grassland NPP showed an overall increasing trend after 2000, but with significant spatiotemporal heterogeneity. The warming-wetting trend in Xinjiang, particularly pronounced since 2000, has generally benefited grassland vegetation growth. However, enhanced interannual precipitation variability has intensified NPP fluctuations, consistent with previous findings on climate extremes in the region.

The spatial pattern of NPP closely mirrors precipitation distribution, decreasing from mountainous areas toward basins. This relationship aligns with studies showing that Xinjiang's grassland productivity is strongly controlled by water-heat dynamics. The higher NPP in northern Xinjiang reflects greater precipitation from Atlantic air masses and orographic effects, while southern Xinjiang's lower productivity results from continental air masses and intensified aridity. The positive NPP-precipitation correlation across most regions underscores precipitation's critical role in grassland growth, particularly in this arid environment.

Since 2000, implementation of the "Returning Grazing Land to Grassland" program and ecological compensation policies has reduced human disturbance, alleviated grassland degradation, and promoted vegetation recovery. These measures, including grazing bans, rotational grazing, and reseeding, have contributed to the observed NPP increases. The consistency between remote sensing observations and policy implementation effects suggests that ecological restoration efforts have been effective in reversing historical degradation trends.

However, the study reveals that climate warming affects northern and southern Xinjiang differently. Northern Xinjiang's temperature rise extends the growing season, while southern Xinjiang's warming increases evapotranspiration demand, potentially exacerbating water stress despite precipitation increases. This asymmetric response highlights the complexity of climate-vegetation interactions across the region.

4. Conclusion

This study utilized MOD17A3HGF NPP data and auxiliary datasets to analyze natural grassland NPP dynamics across multiple spatial scales and its response to climate factors in Xinjiang from 2000 to 2018. The main conclusions are:

1) Natural grassland NPP in Xinjiang showed a fluctuating increasing trend, with a multi-year average of 0.103 kg C·m⁻². Spatially, NPP gradually decreased from the Ili River Valley, Tianshan Mountains, Altai Mountains, and Kunlun Mountains toward the Junggar and Tarim Basins.

2) Northern Xinjiang's grassland NPP (0.149 kg C·m⁻²) significantly exceeded that of southern Xinjiang (0.055 kg C·m⁻²), with both regions showing increasing trends. Prefecture-level analysis revealed generally increasing NPP but with distinct interannual variations. Among grassland types, ten of eleven categories showed increasing NPP, while alpine desert remained essentially unchanged.

3) Grassland NPP exhibited both positive and negative correlations with temperature, while responses to precipitation were predominantly positive. Positive temperature correlations were mainly distributed on northern Tianshan slopes, whereas positive precipitation correlations concentrated in the Ili River Valley, northern Tianshan slopes, and western Junggar Mountains.

The "asymmetric warming-wetting" trend in Xinjiang—where the region overall becomes warmer and wetter but local areas experience warming-drying—has important implications for grassland management. While climate change and reduced human interference have benefited grassland vegetation, intensified precipitation variability poses challenges for sustainable grassland utilization. These findings provide a scientific basis for grassland health assessment, carbon budget evaluation, and sustainable management strategies in Xinjiang.

References

[1] Zhang Li, Zhou Guangsheng, Ji Yuhe, et al. Spatiotemporal dynamic simulation of grassland carbon storage in China[J]. Scientia Sinica (Terrae), 2016, 46(10): 1392-1405.

[2] Zhao Peng, Chen Tao, Wang Qian, et al. Quantitative analysis of the impact of climate change and human activities on grassland ecosystem NPP in Xinjiang[J]. Journal of University of Chinese Academy of Sciences, 2020, 37(1): 51-62.

[3] Costanza R, Michael M. What is a healthy ecosystem?[J]. Aquatic Ecology, 1999, 33(1): 105-115.

[4] Wu J G. Urban ecology and sustainability: The state of the science and future directions[J]. Landscape and Urban Planning, 2014, 125: 209-221.

[5] Wu Jianguo, Guo Xiaochuan, Yang Yu, et al. What is sustainability science?[J]. Chinese Journal of Applied Ecology, 2014, 25(1): 1-11.

[6] Chen Zhuoqi, Shao Quanqin, Liu Jiyuan, et al. Analysis of net primary productivity of terrestrial vegetation on the Qinghai Tibetan Plateau, based on MODIS remote sensing data[J]. Scientia Sinica (Terrae), 2012, 42(3): 402-410.

[7] Li Bo, Zhao Bin, Peng Ronghao. Principles of terrestrial ecosystem ecology[M]. Beijing: Higher Education Press, 2005.

[8] Chen Chen, Jing Changqing, Xing Wenyuan, et al. Desert grassland dynamics in the last 20 years and its response to climate change in Xinjiang[J]. Acta Prataculturae Sinica, 2021, 30(3): 1-14.

[9] Cai Zonglei, Miao Zhenghong, Chang Xue, et al. Research on grassland fractional vegetation coverage inversion method based on drone large quadrat data and domestic satellite[J]. Acta Agrestia Sinica, 2019, 27(5): 1431-1440.

[10] Yang Feng, Qian Yurong, Li Jianlong, et al. Degradation characteristics and causes of desert grassland in the northern Tianshan Mountains[J]. Journal of Natural Resources, 2011, 26(8): 1306-1314.

[11] Xu Wenxuan, Lian Zhongmin, Xu Ting, et al. Degradation pattern of desert steppe plant community around a piosphere in Junggar Basin of Xinjiang[J]. Chinese Journal of Ecology, 2016, 35(1): 104-110.

[12] Chen Jiao, Zhang Li. Estimating grassland coverage based on hyperspectral remote sensing in the northern Tianshan Mountains[J]. Pratacultural Science, 2017, 34(1): 30-39.

[13] Sun Shize, Wang Chuanjian, Yin Xiaojun, et al. Estimating aboveground biomass of natural grassland based on multispectral images of unmanned aerial vehicles[J]. Journal of Remote Sensing, 2018, 22(5): 848-856.

[14] Zhou Lilei, Zhu Huazhong, Zhong Huaping, et al. Spatial analysis of soil bulk density in Yili, Xinjiang Uygur Autonomous Region[J]. Acta Prataculturae Sinica, 2016, 25(1): 64-75.

[15] Xun Qilei, Dong Yiqiang, An Shazhou, et al. Monitoring of grassland herbage accumulation by remote sensing using MOD 09GA data in Xinjiang[J]. Acta Prataculturae Sinica, 2018, 27(4): 10-26.

[16] Yan Junjie, Liu Haijun, Zhao Yu, et al. Spatiotemporal dynamics of grassland net primary productivity in Ili River Valley from 2000 to 2015[J]. Research of Soil and Water Conservation, 2018, 25(5): 390-396.

[17] Zhang Renping, Guo Jing, Zang Yunling. Spatial distribution pattern of NPP of Xinjiang grassland and its response to climatic changes[J]. Acta Ecologica Sinica, 2020, 40(15): 5318-5326.

[18] Du Xuan, Zheng Jianghua, Mu Chen, et al. Correlation analysis of the spatial temporal variation of grassland net primary productivity and climate factors in Xinjiang in the past 15 years[J]. Ecological Science, 2018, 38(13): 4789-4799.

[19] Yan Junjie, Liu Haijun, Cui Dong, et al. Spatiotemporal dynamics of grassland degradation in Yili Valley of Xinjiang over the last 15 years[J]. Pratacultural Science, 2018, 35(296): 508-520.

[20] Ren Xuan, Zheng Jianghua, Mu Chen, et al. Evaluating reliability of grassland net primary productivity estimates using different meteorological interpolation methods[J]. Pratacultural Science, 2017, 34(3): 439-448.

[21] Duan H, Zhao H, Jiang Y, et al. Analysis of seasonal grassland change and its drivers during 1982—2006 in Xinjiang[J]. Rangeland Ecology & Management, 2017, 70(4): 422-429.

[22] Zhang W, Yang X, Manlike A, et al. Comparative study of remote sensing estimation methods for grassland fractional vegetation coverage: A grassland case study performed in Ili Prefecture, Xinjiang, China[J]. International Journal of Remote Sensing, 2019, 40(5): 2243-2258.

[23] Shi Y, Gao J, Li X, et al. Improved estimation of aboveground biomass of disturbed grassland through including bare ground and grazing intensity[J]. Remote Sensing, 2021, 13, 2105.

[24] Guo Yang, Jia Zhibin, Zhang Qi, et al. Study on the spatiotemporal dynamics of forage livestock balance in Hulunbuir grassland of Inner Mongolia based on remote sensing data[J]. Chinese Journal of Grassland, 2021, 43(4): 30-37.

[25] Yang Shuxia, Zhang Wenjuan, Feng Qisheng, et al. Monitoring of grassland herbage accumulation by remote sensing using MODIS daily surface reflectance data in the Qingnan Region[J]. Acta Prataculturae Sinica, 2016, 25(8): 14-26.

[26] Han Qifei, Luo Geping, Li Chaofan, et al. Modeling the grazing effect of grassland on the carbon source/sink in Xinjiang[J]. Acta Ecologica Sinica, 2017, 37(13): 4392-4399.

[27] Yao Junqiang, Chen Jing, Tuoliwubieke Dilinuer, et al. Trend of climate and hydrology change in Xinjiang and its problems thinking[J]. Journal of Glaciology and Geocryology, 2020, 42(3): 1-14.

[28] Yang Jingya, Li Xinguo, Yan Kai, et al. Grassland vegetation dynamics and the relationship between the temperature and precipitation in Hejing County, Xinjiang, based on NDVI[J]. Ecological Science, 2018, 37(6): 38-44.

[29] Zhang R, Liang T, Guo J, et al. Grassland dynamics in response to climate change and human activities in Xinjiang from 2000 to 2014[J]. Scientific Reports, 2018, 8(1): 1-11.

[30] Zhang R, Guo J, Yin G. Response of net primary productivity to grassland phenological changes in Xinjiang, China[J]. Peerj, 2021, 9(1): e10650.

[31] Liu Jiqi, Li Zechun, Qin Xiaoguang. The evolution of the natural environment, climate change and the impact of human activities in Xinjiang[M]. Beijing: China Water Conservancy and Hydropower Press, 2014.

[32] Yang Lipu. Xinjiang water resources and utilization[M]. Urumqi: Xinjiang People's Publishing House, 1981.

[33] Xu Peng. Xinjiang grassland resources and its utilization[M]. Urumqi: Animal Husbandry Department of Xinjiang Uygur Autonomous Region/Xinjiang Science and Technology Publishing House, 1993.

[34] Luo Lin. Introduction to Xinjiang grassland resources(I)[J]. Xinjiang Xumuye, 1990(5): 32-36.

[35] Running S, Zhao M. Mod17a3hgf Modis/terra net primary production gap filled yearly L4 global 500m sin grid v006[J]. Nasa Eosdis Land Processes Daac, 2019.

[36] Yang K, He J, Tang W J, et al. On downward shortwave and longwave radiations over high altitude regions: Observation and modeling in the Tibetan Plateau[J]. Agricultural and Forest Meteorology, 2010, 150(1): 38-46.

[37] He J, Yang K, Tang W, et al. The first high resolution meteorological forcing dataset for land process studies over China[J]. Scientific Data, 2020, 7(1): 1-11.

[38] Wang Q, Zhai P, Qin D. New perspectives on warming wetting trend in Xinjiang, China[J]. Advances in Climate Change Research, 2020, 11(3): 252-260.

[39] Huang Xiaotao, Luo Geping. Temporal characteristics of evapotranspiration and water use efficiency in grasslands of Xinjiang[J]. Chinese Journal of Plant Ecology, 2017, 41(5): 506-518.

[40] Guo Yanyun, Liu Yan, Li Qiuyue, et al. Validation and adaptability evaluation of grassland ecosystem model CENTURY in the Tianshan Mountain area[J]. Acta Agrestia Sinica, 2020, 28(1): 252-258.

[41] Jiang Y, Guo J, Peng Q, et al. The effects of climate factors and human activities on net primary productivity in Xinjiang[J]. International Journal of Biometeorology, 2020, 64(1): 765-777.

[42] An L, Hao Y, Yeh T J, et al. Annual to multidecadal climate modes linking precipitation of the northern and southern slopes of the Tianshan Mts[J]. Theoretical and Applied Climatology, 2020, 140(1): 453-465.

[43] Li C, Wang R, Ning H, et al. Characteristics of meteorological drought pattern and risk analysis for maize production in Xinjiang, northwest China[J]. Theoretical and Applied Climatology, 2018, 133(3): 1269-1278.

[44] Yao J, Chen Y, Zhao Y, et al. Response of vegetation NDVI to climatic extremes in the arid region of Central Asia: A case study in Xinjiang, China[J]. Theoretical and Applied Climatology, 2018, 131(3): 1503-1515.

[45] Zhang Haiyan, Fan Jiangwen, Shao Quanqin, et al. Ecosystem dynamics in the Returning Rangeland to Grassland programs, China[J]. Acta Prataculturae Sinica, 2016, 25(4): 1-15.

[46] Zhang Haiyan, Fan Jiangwen, Shao Quanqin. Land use/land cover change in the grassland restoration program areas in China, 2000—2010[J]. Progress in Geography, 2015, 34(7): 840-853.

[47] Li Wei. Study on the initial effectiveness of Xinjiang's implementation of grassland ecological protection subsidy and reward mechanism[J]. Xinjiang Xumuye, 2015, 220(10): 22-27.

[48] Hu Zhentong, Kong Deshuai, Wei Tongyang, et al. Grassland eco-compensation: Equivalent relationship between livestock reduction and compensation[J]. Journal of Natural Resources, 2015, 30(11): 1846-1859.

[49] Jiang Ping, Ding Wenguang, Xiao Jing, et al. Altitudinal difference of vegetation NPP and its response to climate change in Xinjiang[J]. Arid Land Geography, 2021, 44(3): 849-857.

[50] Xie Dawei, Zhang Nuo, Su Ying, et al. Industrial development modes and restraining factors of severe poverty region after resident relocation for poverty alleviation: Cases of three prefectures in south Xinjiang[J]. Arid Land Geography, 2020, 43(5): 1401-1408.

[51] Liu Y, Wang Q, Zhang Z, et al. Grassland dynamics in responses to climate variation and human activities in China from 2000 to 2013[J]. Science of the Total Environment, 2019, 690: 27-39.

[52] Naeem S, Zhang Y, Tian J, et al. Quantifying the impacts of anthropogenic activities and climate variations on vegetation productivity changes in China from 1985 to 2015[J]. Remote Sensing, 2020, 12(7): 1113.