**1. Introduction**

The current level of global climate change has been unprecedented in the past decades or even nearly a thousand years. Almost the entire world is experiencing a warming process, which is mainly characterized by rising temperatures, rising sea levels, retreating glaciers, and so on [1,2]. Global warming has exacerbated the global water cycle over the past century, causing a significant increase in the number of extreme weather events, such as storms, heat waves, floods, and droughts [3–9]. Droughts are one of the most threatening natural disasters in the world. They are caused by the

below-average level of precipitation over a long period of time and are generally characterized by their high frequency, long duration, and wide range [10,11]. In the context of global warming, the frequency and intensity of droughts have increased significantly [12,13], which has seriously a ffected agricultural production, water resources, and ecosystems, and have led to economic losses, famines, epidemics, and desertification [14–21]. Because of the complexity of drought variability, it is challenging to objectively quantify the intensity, duration, and spatial extent of droughts [22–24]. Thus, numerous studies have attempted to improve drought detection and monitoring; a few objective indices have been developed on the basis of readily available climate data, such as the Palmer Drought Severity Index (PDSI) [25], Standard Precipitation Index (SPI) [26], and Standard Precipitation Evapotranspiration Index (SPEI) [27]. Among these indices, the PDSI, which is based on the supply and demand in the water balance, is one of the most widely used drought indices in the world. However, the PDSI has several deficiencies, including the strong influence of the calibration period, the limitation in spatial comparability, and the subjectivity in relating drought conditions to the index values [28]. The SPI can e ffectively represent the multiscale characteristics of droughts; however only the precipitation variability is considered in the SPI calculation, and the role of temperature is ignored. The e ffect of temperature is evident in initiating droughts, although droughts are primarily caused by a below-average level of precipitation [29]. Therefore, the SPEI was developed by Vicente-Serrano et al. [27], which not only considered the effects of temperature on drought severity but also considered the multiscale characteristics that were incorporated in the SPI. Since it was proposed in 2010, the SPEI has been widely used to monitor and assess the dry-wet conditions around the world [30–33].

Vegetation, linking the atmosphere, hydrosphere, and biosphere [34], is an important component in the terrestrial ecological system and has an obvious relationship with climate change through the physiological responses of plants, such as plant photosynthesis, respiration, and evapotranspiration. The dynamic changes of vegetation play a predictive role in regional climate change [35,36]. Climate change can also a ffect the spatial-temporal pattern of vegetation. Drought is one of the most frequent natural disasters and the response of vegetation to drought is a considerable scientific problem [37]. In general, an increased frequency of extreme drought was associated with decreased vegetation growth [19,38,39]. For example, Symeonakis et al. (2004) pointed out that drought was the main factor resulting in vegetation and soil degradation in sub-Saharan Africa [39]. Ahmadi et al. (2019) indicated that drought could a ffect the e fficiency of water use in the ecosystem, subsequently disturbing the composition and functionality of terrestrial ecosystems [19]. In Northern China, drought-induced water stress caused a reduction in the terrestrial gross primary production [40]. Studies on the Qinghai-Tibet Plateau (QTP) and the Loess Plateau have revealed that there is a remarkable correlation between vegetation cover reduction and climate change [41,42]. Nevertheless, the magnitude of the response of vegetation to dry-wet conditions remains uncertain due to the complexity of the dry-wet transition and intrinsic drought sensitivity among vegetation types [21,43]. The intensity, duration, and timing of drought partly determine the e ffect of drought on vegetation productivity, where moderate drought with higher temperatures increases the net primary production (NPP), while severe drought causes a delayed response of NPP to precipitation [22,44]. As a satisfactory indicator of vegetation activities, the Normalized Di fference Vegetation Index (NDVI), which was used in this study, has been widely and successfully used to detect the vegetation variations [36–44].

The Yarlung Zangbo River (YZR) basin located in the southeast QTP, is the most important river to understand the water cycle in the QTP because it is not only the largest river system in QTP with the largest mean annual flow (56% flows from the QTP), but also an important moisture transportation channel from the Indian Ocean to the inner region of QTP. Owing to grea<sup>t</sup> spatial heterogeneities of climatic conditions and enormous biological diversity, the YZR basin has always been the crucial area of global diversity and ecological protection [45]. A considerable amount of research on the impact of climate change on vegetation variations in the YZR basin has been conducted, revealing that vegetation and precipitation are positively correlated in the whole basin and that the vegetation cover change is restrained by the dry-wet regime, terrain, and other factors simultaneously [46–48]. Song et al. (2011) and Li et al. (2015) have demonstrated that the warming rate of the YZR basin was significantly higher than that of the global average, and the duration and magnitude of the drought have gradually aggravated [49,50]. The water resources problem would become more severe under the impact of precipitation and temperature due to the significant warming of the YZR basin in the future [51]. The rising temperature drives a basin-wide vegetation cover improvement, however it seems that decreasing precipitation does not inverse the overall vegetation greening trend [51]. Similar results occurred in Nepal, bounded by the Tibetan highland and the Himalaya, indicating that the correlation between NDVI and temperature was significantly positive, while NDVI exhibited a negative relationship with precipitation [52]. To sum up, the dry-wet transition is of grea<sup>t</sup> importance for the vegetation dynamics, whereas the response mechanism of vegetation to dynamic variations is still unclear. In addition, the special terrain, vulnerable ecological environment, and sensitivity to climate change make the YZR basin one of the hotspot regions for the studies of water-ecology-environment sustainable development under global warming. Therefore, it is crucial to investigate the transition characteristics of the dry-wet regime in the YZR basin and quantify its e ffect on the vegetation dynamic variations, which could provide a scientific reference for the sustainable development of the environment and ecosystem in the Qinghai-Tibet Plateau. The primary objectives of this study are: (1) to detect the changes of the dry-wet regime in the YZR basin; (2) to quantify the spatial-temporal variations of vegetation from long-term satellite-based NDVI data; and (3) to investigate the dynamic responses of vegetation to possible drivers of the dry-wet transition in the YZR basin.

#### **2. Materials and Methods**

## *2.1. Study Area*

The Yarlung Zangbo River originates from the Gyama Langdzom Glacier is one of the highest rivers around the world. It is mainly composed of five tributaries, namely the Duoxiong Zangbo River, the Nianchu River, the Lhasa River, the Nyang River, and the Parlung Tsangpo River. The YZR basin, with a latitude of 28◦00 'N–31◦16' N and longitude of 82◦00 'E–97◦07' E, is located in the southeast QTP (Figure 1), with an area of about 24 km<sup>2</sup> and average altitude of more than 4000 m (ranging from 132 m to 7258 m). The climate of the YZR basin is characterized as cold plateau mountain climate with intense solar radiation and low air temperatures. The amount of precipitation gradually increases from northwest to southeast in the basin, which is mainly a ffected by the warm and humid airflow from the Bay of Bengal and the Indian Ocean. The mean annual precipitation in the basin is 300–500 mm, and the rate increases with elevation by 10–30 mm/100 m [52]. The total rainfall from June to September accounts for 60%–90% of the mean annual precipitation in the whole basin, indicating the precipitation has an uneven distribution within a year. Due to the complex topographical features and high altitude, the vegetation cover within the area exhibits distinct vertical zonality and varies from mountain forest, mountain broad-leaved forest, mountain coniferous forest, and subalpine shrub meadow to alpine meadow along with rising elevation [53].

**Figure 1.** Location of the Yarlung Zangbo River basin and distribution of the hydro-meteorological stations.
