Evaluating the Skill of Satellite and Reanalysis Precipitation Products in Monitoring Different Levels of Drought Severity at the Global Scale

Abstract

The profound effects of severe drought on ecosystems, agricultural output, and economic stability underscore the importance of precise monitoring across various drought intensities. In this study, we computed the Standardised Precipitation Index (SPI) at different temporal scales and different levels of severity by using monthly precipitation data recorded by 9036 in situ gauge stations across the globe from 2001 to 2018. The skill of six satellite precipitation products (IMERG-Early, IMERG-Late, IMERG-Final, ERA5, GSMaP, and MERRA2) for drought monitoring was evaluated on a global scale based on the probability of detection (POD), false alarm ratio (FAR), and critical success index (CSI). Our findings revealed the following: (1) IMERG-Final demonstrated a greater ability for global drought monitoring among the six products across all timescales (average POD = 0.77, FAR = 0.25, CSI = 0.62). (2) The performance of each product in drought monitoring varied across different regions and timescales. The performance of the IMERG-Final was best in the European Union region and decreased as the timescale increased. The performance of the GSMaP was best in the western Northern Hemisphere land area and improved as the timescale increased. (3) Each product's performance in monitoring drought varied not only across different severity levels, with generally more accurate monitoring for mild drought conditions than for moderate and severe ones, but also revealed that the discrepancy in monitoring various drought intensities expanded as the timescale lengthened. This suggests an increasing divergence in drought monitoring capabilities of different products over longer periods.