Revealing Vertical Distribution of Precipitation in the Glacierized Upper Indus Basin Based on Multiple Datasets

Abstract

AbstractThe quality and vertical distribution of precipitation are of utmost importance in hydrological modeling studies in glacierized catchments, which have been facing known challenges in the quality, distribution, and magnitude of observed as well as gridded precipitation due to complex topography. The purpose of the study is to reveal the vertical distribution of precipitation based on multiple datasets and evaluate the suitability of these datasets for hydrological applications and in the upper Indus basin (UIB). The performance of five gridded precipitation datasets, that is, APHRODITE, CFSR, Princeton Global Meteorological Forcing Dataset for Land Surface Modeling (PGMFD), TRMM, and High Asia Refined analysis (HAR), was evaluated against the observed precipitation (OBS) during 2001–07. A corrected reference precipitation dataset was constructed based on the water and mass balance using inverse modeling methods. TRMM was identified as the best dataset to represent the spatial and temporal distribution of OBS precipitation. The runoff coefficients for OBS, APHRODITE, TRMM, and PGMFD were greater than 1, indicating that these datasets were underestimated and unable to close the water balance in UIB. The mean annual corrected precipitation was estimated as 593 mm yr−1, while the main water-producing elevation zone was located between 3900 and 6600 m. CFSR had the highest correlation and lowest bias with corrected precipitation along the vertical profile. It is concluded that the OBS and most of the gridded precipitation are insufficient to sustain the water and mass balance in UIB. It is recommended to correct the precipitation based on mass balance at high altitudes (especially along the main water-producing zone) in glacierized catchments before its application in hydrological modeling studies.