Does “right” simulated extreme rainfall result from the “right” representation of rain microphysics?

Abstract

AbstractUsing the observations from the two‐dimensional video disdrometer and polarimetric radar, a detailed process‐based evaluation of five bulk microphysics schemes in the simulation of an extreme rainfall event over the mountainous coast of South China is performed. Most schemes reproduce one of the heavy rainfall areas, and the National Severe Storms Laboratory (NSSL) scheme successfully simulates both heavy rainfall areas in this event. However, our analysis reveals that even the NSSL simulation still cannot accurately represent the rain microphysics for this event. Observational analysis shows that abundant small‐ and medium‐sized (1–4 mm) raindrops are the main contributors to the extreme rainfall. All the simulations tend to underpredict raindrops for diameter around 3 mm. The Lin, WSM6, and Morrison simulations agree better with the observed drop size distribution (DSD) for diameters between 1 and 2 mm for higher rain rates. The Thompson simulation shows a relatively narrow distribution with overpredicted small‐sized (1–2 mm) raindrops. The NSSL simulation has a broad distribution with more large (>4 mm) raindrops probably related to its efficient rain self‐collection process at the low levels, which is conducive to producing extreme rainfall. Proper rain evaporation rate is important in generating cold pools with favorable strength for the maintenance of a convective system in this event. Similar results are obtained in the simulations of two additional extreme rainfall cases, in which the NSSL simulation also overpredicts large raindrops while the Thompson simulation produces more small raindrops. This study indicates that more efforts are needed to improve the representation of rain self‐collection/breakup, rain evaporation processes, and DSD for extreme rainfall over South China. It also highlights the importance in careful consideration of rain DSD in addition to radar reflectivity and surface precipitation when analyzing simulations of extreme rainfall in order to avoid “wrong” interpretation of “right” results.