Assessment of the GPM IMERG and CHIRPS precipitation estimations for the steppe part of the Crimea

Abstract

This paper compares the spatial distribution datasets on monthly precipitation totals derived from the Famine Early Warning System Network FEWS NET service (CHIRPS 2.0 product) and the International Mission of the Global Precipitation Measurement GPM (IMERG v06 product) with ground-based observations of a stationary weather stations located in the steppe region of the Crimean Peninsula in order to assess the representativeness of the precipitation spatial distribution and the applicability of the datasets for water balance calculations and agricultural crop dynamics modeling. A close convergence was observed between the estimated monthly precipitation totals and the precipitation gauge data during the study period (January 2017 -July 2020), with mean correlation coefficients of 0.75 and 0.73 for the GPM IMERG and CHIRPS, respectively. Both products generally overestimated the precipitation values compared to the measured data, with GPM IMERG (final run) exhibiting the greatest overestimations (1.3-2.1 times the weather station values). Our results demonstrate the requirement of GPM-derived precipitation estimations (particularly those from the GPM_3IMERDL v06 daily accumulated late run dataset) to be additionally verified and calibrated based on data from regional weather stations or the CHIRPS 2.0 product (if available).