Can IMERG Data Capture the Scaling of Precipitation Extremes With Temperature at Different Time Scales?

Abstract

This study evaluates the validity of the Integrated Multisatellite Retrievals for Global Precipitation Measurement (IMERG) in scaling extreme precipitation with temperature (termed scaling factor, SF). To this end, we use Hadley-Integrated Surface Database (HadISD) data set at 1-hourly, 6-hourly, and 24-hourly scales between 2000 and 2020 over the contiguous United States (CONUS) as a reference data set. Our findings reveal that IMERG can capture changes in precipitation extremes with the temperature at 6-hourly, and 24-hourly scales, especially over dry regions. The difference between IMERG's SF and HadISD's SF at 6-hourly scale is less than 0.3%/°C in dry areas. However, IMERG underestimates SF in wet regions, especially at 1-hourly scale, up to ∼5%/°C. It is found that IMERG's performance is dependent on temperature, particularly at 1-hourly scale. IMERG overestimates extreme precipitations when temperatures are below 5°–11°C while the opposite is true at higher temperatures, indicating the need for considering temperature in adjusting IMERG.