Comparison of different gridded precipitation data sets over a sub-region of High Mountain Asia.

Precipitation is an important component of the water cycle of High Asia and here particularlycrucial for the glacier mass balance and the seasonal dynamics of snow cover. To successfullymodel (i) past and future response of glaciers to climate forcing and (ii) the development ofsnow cover, reliable precipitation data are needed. However, this is especially difficult in highmountain areas where observational data, particularly at higher and complex terrain, arescarce. Beyond station data, other, spatial explicit data sources for precipitation are (i)simulated rainfall from models and (ii) retrievals from weather satellites. While precipitationpatterns are highly heterogeneous in both, time and space they are difficult to simulatecorrectly. With satellite data on the other hand, the assignment of rates of liquid andespecially solid precipitation is yet uncertain. To assess the differences among theseveral precipitation data sets, we compare selected accessible gridded data setsin High Mountain Asia. We present a comparison of five precipitation products(ERA-Interim, ERA5, HAR, JRA-55 and MERRA-2) and one satellite-based precipitationproduct (IMERG) for the period of April 1st to October 31st, 2014. The study area(80-87˚ E and 28-31˚ N) includes parts of India, a part of the Himalayas and partsof the Tibetan Plateau. The main statistical metrics used for comparison are thecorrelation coefficients r of different data sets and the corresponding coefficients ofdetermination r2 for different methods of temporal and spatial data aggregation. Wefurther compare the data sets to existing observational data available to the public. [ABSTRACT FROM AUTHOR]