The three-river source region is in the interior of the Qinghai-Tibet Plateau. As the birthplace of the Yangtze River, Yellow River and Lancang River, this place provides a total of 60 billion cubic metres of water to downstream areas each year, which is why this area is known as the reservoir of China (Li et al., 2019; Zhang et al., 2017). With the unique topography, landform and climate, the three-river source region features many plateau lakes, glaciers and alpine wetlands, marking it as one of the most important water resource accumulation and regulation areas in China (Shen et al., 2021). However, in the past several decades, due to the impact of climate change and human activities, the snow-capped mountains and glaciers there have retreated, lakes and wetlands have shrunk or even dried up, the area of water and soil loss has expanded, desertification and grassland degradation have become increasingly aggravated, and water conservation capacity has declined sharply (Li and Xiao, 2020). According to statistics, taking the total area experiencing disasters as an example, drought disasters strike the three-river source region most, with a proportion of 38%, which exceeds floods, hail, frost, pests and snow disasters, and the increase in drought disasters in the central and northern parts show a frequent trend, with an increasing rate of approximately 8.35×104hm2/10a (Wu et al., 2011).
Global warming is regarded as an indisputable fact, so monitoring drought events has become a hot spot for scientists worldwide. The drought index can represent drought characteristics and be used to assess drought events effectively. Thus, this index can be used as a key indicator for drought monitoring. The standardized precipitation index (SPI) is regarded as the most representative and widely used drought index (Raziei, 2021). Only precipitation data are required for drought condition assessment. The calculation process is simple, and the time scale is flexible, but the impacts on soil evaporation and temperature on drought are not considered (Kalisa et al., 2020). The SPI is calculated based on observational data from meteorological stations, but observational data of the Qinghai-Tibet Plateau are scarce, which makes it difficult to fully describe the regional drought characteristics. On the other hand, in addition to the drought index calculated based on meteorological elements, the anomaly of soil moisture can also effectively reflect the actual dry-wet condition of the land surface and be used to test the drought index or serve as an evaluation standard for the applicability of various drought indices (Hogg et al., 2013; Halwatura et al., 2017). The soil moisture anomaly percentage index (SMAPI) defined by Wu et al. (2011) can reflect the occurrence, development process and severity of actual droughts and better represent historical drought events. However, due to the significant spatial variability in soil moisture, even though the soil moisture observation at each site is accurate, its representativeness and typicality are difficult to assess (Liu et al., 2019).
Water vapour transport is an essential factor leading to regional droughts and floods(Xu et al., 2020). Many studies have focused on the characteristics of water vapour transport on the Qinghai-Tibet Plateau based on isotope analysis, reanalysis data sets, and numerical simulation of atmospheric models. Xu et al. (2002) proposed that the “large triangle” of water vapour transport on the plateau was the main cause of abnormal drought and floods in China and East Asia. Bothe et al. (2011) discussed the relationship between drought and large-scale circulation on the Qinghai-Tibet Plateau. They believed that extreme drought and wet events on the Qinghai-Tibet Plateau were related to anomalous circulation in the North Atlantic and Europe together with the wave train across Eurasia, adjusting the flow fields in the northern and eastern plateau through anticyclone-cyclone dipole and thereby inhibiting the water vapour transport from the Bay of Bengal to the plateau. Zhang et al. (2017) believed that the water vapour carried by the westerly jet and the Indian monsoon caused precipitation on the Qinghai-Tibet Plateau and analysed the characteristics of the two water vapour transport paths.
The Lagrangian method-based flexible particle dispersion model (FLEXPART) has been applied to water vapour transport research. This model overcomes the deficiencies, including the lack of understanding of the Lagrangian method (such as dynamic trajectory, the assumption of how the surface evaporation is distributed in the entire atmospheric column) and its limitations of only tracking the trajectory of the air masses qualitatively, but the water vapour evaporation and dissipation of air masses cannot be quantitatively determined during movement (Sodemann et al., 2008; Li et al., 2020). Gimeno et al. (2012, 2013) discussed the oceanic and continental source regions of terrestrial precipitation on a global scale. Chen et al. (2012) used the Lagrangian backward trajectory model to simulate the paths and potential source regions of water vapour transport on the Qinghai-Tibet Plateau during 2005–2009, indicating that the water vapour in the low-latitude ocean carried by the Indian monsoon was the main water vapour source of the Qinghai-Tibet Plateau. Sun and Wang (2014) used the Lagrangian backward trajectory model to study the water vapour source and transport characteristics of China’s semiarid grassland in cold and warm seasons. Salah et al. (2018), based on the FLEXPART trajectory model and with the SPEI as the drought assessment method, discussed the abnormal characteristics of water vapour transport in two severe drought cases in the history of Fertile Crescent. Zhao et al. (2018) traced the path of water vapour transport from the Qinghai-Tibet Plateau to the Yangtze River based on the model and analysed the influence of plateau convection on heavy rain in the Yangtze River Basin.
The three-river source region is a typical area affected by the mid-latitude westerlies and the South Asia monsoon. The northerly airflow from the north of the plateau and the southerly airflow from the south form a convergence zone in the source region of the three rivers. This flow field makes the three-river source region an area with active weather systems, such as low humidity and shear lines. These low-value systems provide power for water vapour convergence (Sun and Wang, 2018). On the other hand, the decrease in precipitation caused by abnormal water vapour transport both in the source region itself and outside is an important cause of drought. Drought in the three-river source region will not only adversely affect local production, life, and social development but will also have a serious impact on the security of water resources in the middle and lower reaches of the three major rivers, directly threatening the entire Yangtze and Yellow River basins and even Southeast Asia (Dong et al., 2016).
Understanding the dry-wet evolution process of the three-river source region, especially abnormal water vapour transport corresponding to regional drought, has important practical significance for a deep understanding of the land surface process in the source region. Most of the investigations are based on the dry-wet evolution characteristics of the land surface in the three-river source region, that is, describing the intensity, scope, and starting and ending times of drought events, especially extreme drought events. However, few studies have focused on the physical process of drought in the three-river source region from the perspective of atmospheric water resources.
Considering the uncertainty of the land surface dry-wet evolution in the three-river source region and the complex water vapour sources, this work started from the drought index of the source region and based on the observation data from the Qinghai-Tibetan Plateau permafrost and alpine wetland (PAW) monitoring network located in the three-river source region from 2008 to 2018, the applicability of the SPI in the three-river source region was assessed and evaluated. Then, the time series and drought intensity based on the SPI from 1988 to 2020 were obtained. Subsequently, the Lagrangian backward trajectory model was used to quantitatively evaluate the relationship between the drought event and the water vapour transport under different drought events corresponding to different drought indices. Finally, the anomaly of the water vapour source during the extreme drought event was extracted, and the intensity of the source region evaporation was analysed. The results of the research not only helped to better understand the characteristics of the water vapour transport and evaporation source in the three-river source region but also helped to understand the characteristics and mechanism of the land surface dry-wet evolution process corresponding to the water vapour transport.