Figure 3

3.2 Temporal variations of precipitation, AET, and TWSA in the EIB

The monthly series from the average of five GRACE TWSA products and the GPM precipitation data were used to simulate the monthly AET, utilizing the hydrologic budget method in the EIB and its 16 closed basins. The difference between precipitation and AET (described as (P – ET)) was also analyzed. There is a direct physical relationship between the TWS and (P – ET); according to the water balance within a closed basin, the change in TWS is equivalent to the accumulated (P – ET). Figure 4 presents the temporal variations in precipitation, AET, (P – ET), and TWSA in the EIB. There was little difference in the mean precipitation of the EIB between months (Figure 4(b)). In terms of the intra-annual variation of precipitation at the basin scale, there were three major findings. First, there was little difference in the mean precipitation for each month in the CSB and BLB, particularly the CSB, in which the coefficient of variation was 0.24 (Table 1). This was also the major cause of the marginal difference in the monthly precipitation of the EIB, as this precipitation accounted for a large proportion of the EIB. Second, the monthly precipitation from June to October was lower November to May of the following year in the HRB, IIRB, ASB, ISB, and TuRB. Precipitation from June to October accounted for less than 20% of the annual precipitation, in which the former two basins accounted for less than 10%. Third, under the impacts of the East Asian summer monsoon, the monthly precipitation in nine basins (i.e., GHCB, IMPB, JB, MPLB, MPIRB, QB, QPB, TaRB, and TB) in East Asia varied from month to month. The coefficient of variation exceeded 0.90. Precipitation in these basins was mainly concentrated from May to October, and this precipitation accounted for more than 80% of the annual precipitation, while it was >90% for the QB and QPB.
The mean annual precipitation and AET of the EIB for 2002–2020 was 353.4 and 358.7 mm, respectively. This indicates that water consumed by the AET was greater than the precipitation that occurred in the EIB. The number of years where P < AET and P > AET was similar, although the mean (P – ET) of the former (21.4 mm) was greater than the latter (10.1 mm). During November to March of the following year, precipitation in the EIB was able to meet the AET demand (i.e., P > AET); however, monthly precipitation was less than the AET from April to October. In other words, precipitation in the EIB was insufficient to meet the demand from the AET during this period.
The annual and monthly TWSA of the EIB was generally decreasing over the 2002–2020 period, particularly for 2005–2015 (Figures 8(c) and 8(d)). Overall, the results were consistent among the five TWSA products for the annual and monthly series. The accumulated (P – ET) was highly correlated with the TWSA series; the R2 between the monthly series of accumulated (P – ET) and the mean of the five TWSA products was 0.97, and reached up to 1.00 between the two annual series. According to the standardized monthly series, the TWSA was lagging by approximately one month compared with the (P – ET).