China’s deserts greening and response to climate variability and human activities

affected by both East Asian monsoons and global westerlies, and the climate variability that occurs within a year is critically influenced by the interplay of monsoonal circulation between summer and winter [5]. As a potential result of global warming [6], desert ecosystems as one of the fragile ecosys- tems are changing dynamically [3]. To clarify the changes of desert ecosystems can help to understand their ecological balance, and to take effective measures in time. It is important to strengthen ecological construction and take effective countermeasures in time. Due to the sen- sitivity to climate variability and good performance in arid areas [7–9], vegetation can be an ideal indicator to detect changes in desert ecosystems dynamics [10]. The normalized differ- ence vegetation index (NDVI), which provides information on vegetation conditions, can identify low biomass and an increase in vegetation cover than other indexes [11], so it has become the most widely used index in studies on vegetation dynamics [12,13]. In recent decades, several studies have focused on NDVI variations in arid and semiarid regions [14– 16]. Their results showed that the NDVI showed greening trends at the annual scale in several regions, such as the Gurbantunggut Desert, which showed an increasing trend that fluctuated from 1981–2003 [17]. At the same time, the NDVI for Central Asia significantly increased, with a value of 11.35% [18]. In global semiarid areas, the average NDVI rise from 1981 to 2007 by about 0.015 NDVI units [19]. Seasonal differences were seen in the global NDVI, for exam-

ple, the upward trend of NDVI in spring, summer, and autumn was observed in some mid- and high-latitude regions of the Northern Hemisphere, and the downward trend of NDVI in summer and autumn was observed in some arid and semi-arid regions of the Southern Hemi- sphere [20]. Variations in the NDVI that occur within a year can be observed at the monthly scale. For instance, the monthly mean NDVI increased between January and August and decreased in December in Nigeria [15]. However, the monthly mean NDVI in Central Asia is highest in January and December [21]. Variations in the NDVI are influenced by multiple factors, which can be divided into envi- ronmental and anthropogenic factors [22–24]. Among various environmental factors, varia- tions in temperature and precipitation, which are forms of climate variability, are considered to be the two most influential factors [25–27]. In most cases, the effect of temperature is posi- tive, but sometimes it has a negative effect. For example, the NDVI is positively associated with the temperature when energy is limited to vegetation growth in higher latitudes and higher altitudes in the North American continent, especially at the beginning of the season, but usu- ally in the mid-season the relationship is negative for lower latitudes when water is the limit of vegetation growth [28]. Temperature has a positive correlation with NDVI at both ends of the growing seasons but a weak negative correlation in the middle of the growing season in the

central Great Plains [29]. Precipitation positively influences the NDVI, especially in arid and semiarid regions that lack water [30–32]. The impact of climate variability also has seasonal differences, as temperature is the dominant factor in spring that lowers the NDVI in inner Asia, which is an environment appropriate for vegetation growth and causes an increase in the melting of winter snow and ice, thereby increasing moisture in the spring [33,34]. Sufficient rainfall during summer benefits the growth of vegetation in arid areas more [35]. The effect of climate has a time lag that is distributed differently in different regions. For example, there was a 3-month lag between the NDVI and temperature at the biome scale in China [36], and there was a 2-year lag in temperature at the global scale [37]. Human activities such as cropland and forest expansion in China and India have increased their relative areas by approximately 25% and 6.8%, respectively, and have played a critical role in global greening [22]. However, a dra- matic decline in NDVI was observed in the Yangtze River and Pearl River deltas due to rapid urbanization [36]. PLOS ONE China’s deserts greening and response to climate variability and human activities PLOS ONE | https://doi.org/10.1371/journal.pone.0256462 August 30, 2021 2 / 20 20dataset%20for%20china%20(1901-2017) precipitation: https://data.tpdc.ac.cn/en/data/ faae7605-a0f2-4d18-b28f-5cee413766a2/?q=1-km %20monthly%20precipitation%20dataset%20for %20China%20(1901-2017). Future researchers first need to click "Apply for Download" on the right

side of the above two web pages, register their accounts and submit instructions on how to use the data, etc., and then obtain the permission to download the data after approval. Funding: This work was supported by Strategic Priority Research Program of Chinese Academy of Sciences “Pan-third pole environmental change and the construction of green Silk Road” [grant numbers XDA20040000, XDA20090000]. Competing interests: The authors have declared that no competing interests exist.