Precipitation variability in High Mountain Asia from multiple datasets and implication for water balance analysis in large lake basins

For the period 1979-2011, eight gridded monthly precipitation datasets, including GPCP, CMAP-1/2, TRMM, PREC/L, APHRODITE, NCEP-2 and ERA-Interim, are inter-compared with each other and station observations over High Mountain Asia (HMA). The precipitation variability from the first six gauge-based or merged analysis datasets agree better with each other than with the two reanalysis data. The long-term trend analysis of GPCP, CMAP-1, PREC/L and APHRODITE precipitation datasets consistently reveals moderate increases in the inner and northeastern Tibetan Plateau (TP) and northwest Xinjiang, and obvious decreases in the southeast TP. However, in the Himalayas and Karakorum, there are large discrepancies among different datasets, where GPCP and APHRODITE precipitation datasets show significant decreases along the Himalayas while other datasets show strong spatial heterogeneity or slight variations. The larger uncertainties in the rugged area may be largely attributed to scarce station observations, as well as the stronger snow-induced scattering by microwave measurement. To assess which precipitation datasets tend to be more suitable for hydrologic analysis in HMA, we further investigate the accuracy of precipitation estimates at basin scale by comparing with gauge-based observations, and examine the coherences of annual lake water budgets and precipitation variability over four large closed lake catchments. The results indicate that two reanalysis precipitation datasets show evidently weaker correlations with station observations; the other six datasets perform better in indicating inter-annual variations of lake water budgets. It suggests that these merged analysis precipitation datasets, especially for GPCP, CMAP-1/2 and PREC/L, have the potential in examining regional water balances of the inner basins in HMA.