Sensitivity of cloud microphysics on the simulation of heavy rainfall in WRF- a case study for the 7–10 August 2019 event over Kerala, India

Abstract

This work investigates the mechanism of the extremely heavy rainfall event during 7–10 August 2019 in Kerala. As the skill of WRF model for simulating the heavy rainfall strongly depends on the choice of cloud microphysics (CMP), sensitivity analysis is performed using cloud resolving WRF simulations with different CMP schemes to assess their performance. Total five sensitivity experiments are conducted by varying CMP schemes namely Thompson, Goddard, WSM6, Kessler and Purdue-Lin. Model results along with the fifth generation European Reanalysis (ERA5) data indicate that the heavy rainfall occurred in association with the monsoon depression, intensification of off-shore trough, and low-level jet (LLJ) phenomena and associated moisture transport and convergence. Our results clearly suggest that the heavy rainfall occurred during 7–10 August 2019 was less intensive compared to the August 2018 event. One of the prime reasons for meagre intensity is the presence of the upper tropospheric cyclonic flow over southeast Arabian Sea off Kerala, which reduced the intensity of low-level westerly flow, moisture transport and convection in 2019 event relative to the 2018 event. Comparison of simulated rainfall against station and gridded rainfall data indicate that the spatial and temporal variations of rainfall could be reproduced except that the model slightly over-estimated rainfall at a few stations in the central parts of Kerala. Out of the five CMP schemes, Thompson scheme simulates large amounts of cloud, rain, and snow during the period of peak rainfall with relatively low amounts of ice. Moreover, Thompson indicates a high warming in the 4–6 km layer due to large snow production while Lin and Goddard configurations exhibit altitude shift in this pattern i.e., more heating in the upper troposphere (9–15 km layer) due to producing more solid hydrometeor distributions and associated freezing processes. Among the CMP schemes, Thompson followed by Goddard better simulated the spatial and temporal distribution of rainfall as they produced the strongest convergence, CAPE and vertical motions compared to WSM6, Lin and Kessler schemes. Moreover, the simulated hydrometeor mixing ratios in the 2019 heavy rainfall event are less by 60–80% as compared to that of 2018 due to differences in the synoptic settings.