Tracking precipitation events in time and space in gridded observational data

Abstract

Abstract Novel precipitation event data sets are created by tracking 3‐hourly observed rainfall in both time and space in the TRMM 3B42 and ERA‐Interim data sets. Relative to TRMM, ERA‐Interim data undersimulate the total number of events (factor of ∼0.5), and oversimulate the frequency of events lasting >5 days (factor of 1.6). Longer‐lasting events tend to have larger spatial footprints and higher intensity precipitation at any point in their lifetime, and thus contribute significantly more to total precipitation than shorter events. Precipitation changes in selected tropical and subtropical regions are attributed to different event characteristics: some to changes in event rainfall intensity, others to changes in the number of events. In the 40°S–40°N spatial average, the number of events lasting 1–5 days significantly increased from 1998 to 2014 in both the TRMM 3B42 and ERA‐Interim data. The event data sets, analysis scripts, and selected processed data are freely available online.