[Insert Figure 7 here]

3.5 Spatial linkage of increased precipitation to enhanced water vapor circulations

Previous studies have identified the main sources of moisture for precipitation in the TP, including regions from the Indian sub-continent to the Southern Hemisphere, the Bay of Bengal and the northwestern part of TP. (Chen et al. 2012; Zheng et al. 2019). It is proven that these circulation changes in the increase or decrease of different moisture sources provide favorable water vapor conditions for rainfall on the TP. Thus, the changes in moisture source contributed to different precipitation changes across the entire TP during different periods. As indicated in the preceding results, the lake water levels increased rapidly during the period 2016-2018 relative to the previous stage. The phenomenon was analyzed with the distribution of water vapor transport on the TP.
The differences can be observed with the four years when the water vapor is transported by abnormal northwesterly, southeast wind and south wind in 2015, 2016, 2017 and 2018, respectively (Fig. 8B, Fig. 8C, Fig. 8D, Fig. 8E). In comparison, the total water vapor increases with the most significant anomaly in 2017-2018. The moisture transported to the northwestern part of the TP has increased significantly in last several years due to the strengthening of the westerlies. In this study, precipitation in the NTP is more obviously influenced by the East Asian monsoon and the Indian summer monsoon. The summer monsoon circulation was enhanced, and the meridional water vapor transport from the south plateau to the Bay of Bengal was significantly enhanced. Similarly, precipitation in the CTP was more influenced by the Asian summer monsoons. In addition, the water vapor flux from the Indian Ocean to the plateau was enhanced. From Fig. 8D, we can also find that there is a phenomenon of internal water vapor circulation and retention, which might cause water vapor subsidence and significant rainfall over the western CTP. In general, the change in water vapor transport during the study period supports the changing characteristics of precipitation and lake level for TP as analyzed above.
In this study, as shown in Fig. 6 and Table 3, we can find the rising rates of lake levels at different parts of TP are different. The water level rising rates for some typical lakes exceed 0.80 m/a in the northern TP such as XijingUlan Lake, Mingjing Lake, Lexiewudan and Ayakkum. There are some glaciers distributed in the northern TP, and there is rich rainfall in the rainy season. Therefore, the rapid increase of water level is affected by both rainfall and glacial meltwater. The rising rate of lake level in central TP is mainly divided into two categories. For example, at Zhari Namco and Dagze Co in the western CTP, the rising rates are 1.34 m/a and 1.43 m/a respectively. The rising rates of lakes in the central and eastern CTP are slower, generally less than 0.40 m/a. The reason for this difference may be because the phenomenon of internal water vapor circulation and retention mainly occur in the western CTP. In NWTP which was mainly affected by the westerlies, the distinction in water vapor transport distribution in different years led to a significant difference in the water level rising rates. As indicated in the preceding texts, the differences in water level rise rate are caused by different water vapor transport patterns and main sources in different regions.