Reduction of the GPM IMERG Final Run Underestimation in the Eastern Himalaya

Abstract The Yarlung Zsangbo Grand Canyon (YGC) in the eastern Himalaya is one of the deepest canyons in the world. Satellite precipitation products should be assessed and calibrated before their applications in this remote mountain area. A new rain gauge network was installed in the YGC in November 2018, since then the network observation data were utilized to calibrate the Integrated Multi-satellitE Retrievals for Global Precipitation Measurement V06 Final Run (IMERG-F) product. The evaluation demonstrated that the IMERG-F data reasonably captured the observed seasonal and diurnal variations in the precipitation but with much weaker seasonal and diurnal variations compared with the gauge data. The IMERG-F overestimated/underestimated the hourly light/heavy precipitation frequency, leading to a significant underestimation of daily and monthly rainfall amounts. The rainfall produced by the two layers of cloud in the mountainous region cannot be captured by the IMERG-F algorithm, which causes the underestimation of total rainfall. To address this issue, we applied a cumulative distribution function (CDF) calibration, which successfully reduced the mean bias of hourly and monthly rainfall for IMERG-F from −0.11 mm h −1 and −95.0 mm month −1 to 0.03 mm h −1 and −5.2 mm month −1 . The mean biases of the daily light, moderate, and heavy rainfall decreased from −0.93, −1.02, and 4.71 to 0.13, −0.13, and 3.24 mm day −1 , respectively. The CDF method can effectively correct the underestimation bias in IMERG-F. This study has implications for the application of satellite rainfall products to global mountain areas. Significance Statement The Yarlung Zsangbo Grand Canyon (YGC) is one of the deepest canyons in the world. Precipitation in the YGC often brings natural disasters to local communities, which affect their livelihood. Remotely sensed precipitation products can be valuable for this region if they are adequately calibrated and assessed using rain gauges. A new rain gauge network was installed in the YGC in November 2018, and more than 3 years of observations were utilized to evaluate the Integrated Multi-satellitE Retrievals for Global Precipitation Measurement (GPM) (IMERG) precipitation product. The evaluation results demonstrated that the IMERG significantly underestimated rainfall at the daily and monthly time scales. Some possible mechanisms for this underestimation were investigated to help scientists improve the satellite precipitation products for this region.