Event-based Verification of IMERG Precipitation Estimates Over Complex Terrain in the Southern Appalachian Mountains

Abstract Partitioning precipitation by event type provides a robust analysis and physically-based error diagnostic study when comparing in-situ observations with gridded remotely sensed precipitation estimates. We outline a novel event-based method for comparing in-situ and satellite-based observations. In this sense we are able to provide a robust, independent verification of satellite-based precipitation estimates. Specifically, more than a decade of precipitation observations from the Duke Great Smoky Mountain Rain Gauge Network (GSMRGN) are compared with half-hourly and daily satellite-based estimates from the Integrated Multi-satellitE Retrievals for the Global Precipitation Measurement Mission (IMERG). The ability of IMERG to detect and estimate precipitation in a complex terrain region is analyzed and the errors are documented with respect to the season and direction of the events. IMERG demonstrated a positive bias relative to the GSMRGN throughout the study domain with variations in magnitude based on the direction and season. The bias was reduced for events propagating from the northwest consistent with under-observing variability from the orographic precipitation enhancement and rain shadow effects. Accordingly, the GSMRGN demonstrated spatial nonlinearities across gauges’ northwest precipitation profiles that do not meet the assumptions of the Kalman filter used in IMERG’s data assimilation.