1 INTRODUCTION
Inland basins, also referred to as endorheic basins, are defined as regions where runoff in the basin has no direct hydraulic connection with the ocean; this means that inland basin runoff is landlocked from the ocean. This runoff eventually enters inland lakes or is taken up by evapotranspiration. These regions are one of the most sensitive to climate change and human activities (Huang, Xu, Guan, Wang, & Guo, 2016; Wang et al., 2018). The Eurasia inland basin (EIB) is the largest inland basin in the world; its accounts for more than one-third of the global inland basin area and spans 20 countries. As the climate in EIBs is characterized by extremely low precipitation and high evaporation, the hydrology and ecosystem of the EIB are sensitive to changes in precipitation, actual evapotranspiration (AET), and water storage. Therefore, changes in and attribution analyses of terrestrial (or total) water storage (TWS) and AET in the EIB are hugely important to water resource management, ecosystem health and sustainable agricultural irrigation in Eurasia.
The Gravity Recovery and Climate Experiment (GRACE) is able to accurately estimate monthly TWS changes in basins that are larger than approximately 200 000 and 100 000 km2 at low and high latitudes, respectively (Rodell et al., 2018). Wang et al. (2018) found that TWS was decreasing in global endorheic basins, using the GRACE TWS product; this TWS decline in the EIB contributed to 70% of the global decline. Based on three GRACE Mascon products, Rodell et al. (2018) analyzed TWS trends in 34 regions from 2002 to 2016; they categorized the drivers of this change as natural inter-annual variability, unsustainable groundwater consumption, climate change, and combinations of these factors.
In terms of basin scale research, there are inconsistent results on the main causes of water storage change, with opposing conclusions for some basins. For example, some studies have suggested that the primary cause for the water level decline in the Aral Sea was increased water consumption from enhanced human activities, particularly irrigation and damming (Yang et al., 2020). However, irrigation diversions increased from 1992 to 2005 and decreased from 2005 to 2016 (Wang et al., 2020). Thus, it was concluded that although irrigation diversion plays a dominant role in this water storage decline, its influence has been gradually weakening (Jia, Lia, Li, & Huang, 2020). Another study reported that the mountain lakes in the source region of the Aral Sea was expanding, which was mainly caused by increased glacial melt induced by elevating air temperature (Zheng et al., 2019). Several studies have shown that increasing evaporation rates play a dominant role in the water level decline in the Caspian Sea (Chen et al., 2017). Arpe, Molavi-Arabshahi and Leroy (2020) found that the decrease in net precipitation over the sea and the mean annual inflow contributed 57% and 43% to this water level decline, respectively. Other studies attributed the water level decline to increased evapotranspiration or two major earthquakes in 2000 (Elguindi & Giorgi, 2006; Ozyavas, Khan, & Casey, 2010).
In addition to the water level declines in the Aral and Caspian Seas, a decline in TWS was found to occur in other basins in the EIB, including the Iran (Joodaki, Wahr, & Swenson, 2014; Khaki et al., 2018; Moghim, 2020), Tarim (Yang, Xia, Zhan, Qiao, & Wang, 2017; Xu et al., 2019) and Turpan basins (Xu et al., 2019). The main cause for this decline in TWS in the Tarim basin is decreased precipitation (Yang et al., 2017; Xu et al., 2019). However, it was found that glacial retreat and increased water resource consumption from human activities also has an influence on TWS decline. The main factors attributed to the TWS decline in Iran differed in previous studies. Joodaki et al. (2014) found that this TWS decline was highest in the Middle East, where human activities where human activities was largely responsible for this decline. Khaki et al. (2018) found that the TWS of Iran was continuing to decrease, despite removing the influence of human activities. As such, this decline in TWS is likely to be influenced by a combination of human activities and climate change.
Increasing trends were also detected for water levels or TWS in several basins within the EIB, including water levels at Balkhash Lake (Duan et al., 2020) and Issyk-Kul Lake (Alifujiang, Abuduwaili, Ma, Samat, & Groll, 2017) and the TWS in the Qaidam (Bibi, Wang, Li, Zhang, & Chen, 2019; Meng, Su, Li, & Tong, 2019) and Qiangtang Plateau (Meng et al., 2019; Liu, Yao, & Wang, 2019) basins. Despite these research findings, there are relatively few studies that have examined the attributes of the underlying factors driving these trends.
There have also been contradictory conclusions about the TWS trends within the same region. For example, Wang et al. (2020) detected a decreasing TWS for 2002–2016 using one total water storage anomalies (TWSA) product in the Gansu Hexi Corridor basin. Cao, Nan, Cheng and Zhang (2018) found that the basin TWS significantly increased during 2002–2013 using another TWSA product. Thus, these opposing conclusions may be a result of researchers utilizing different datasets.
In summary, there continue to be inconsistent conclusions regarding changes in TWS and the main drivers of these changes in the EIB and its sub-basins. This demonstrates the need for further studies on these TWS changes and the main impact factors for these regions. The hydrologic budget (i.e., hydrologic gains and losses) is an effective method to conduct analyses at the basin scale (Liu, Yao, Huang, Wu, & Liu, 2014; Liu et al., 2016; Reager et al, 2016; Liu, Wang, & Yao, 2018). Unlike exorheic basins, which include AET and runoff in hydrologic losses, inland basins only include AET. In other words, the hydrologic budget in inland basins may only be expressed by precipitation, AET, and TWS. As such, this method is more suitable for the analysis of TWS changes and its main attribution in inland basins.
This study used multi-source data on the EIB and its closed basins to achieve three key objectives: 1) simulate the monthly AET series of each inland basin using the hydrologic budget method; 2) detect the spatiotemporal characteristics of annual and monthly precipitation, TWS, and AET using a non-parametric test method in each basin; and 3) identify the main attributions of AET and TWSA changes using the water balance principle in each closed basin.