Investigation of physical processes in heavy rainfall events over the Western Indian region in an ensemble‐based modeling framework and comprehensive rating metrics
Vijay Vishwakarma,Sandeep Pattnaik
DOI: https://doi.org/10.1002/qj.4869
2024-10-31
Quarterly Journal of the Royal Meteorological Society
Abstract:The study utilizes Weather Research and Forecasting (WRF) for the heavy rainfall event (HRE) over Gujarat (2022) with the aim to elucidate the new scientific insights causing this event using the multiphysics ensemble modeling framework up to a forecast lead time of 96 hr. Based on the comprehensive rating metrics (CRM), the top three (i.e., E6, E2, EISRO) and bottom three (i.e., EUSGS, E8, E7) ensemble models are segregated. In general, the ensemble has superior performance compared to deterministic models. The mixed‐parameterized model (EISRO) produced the most realistic structure, emphasizing the importance of land‐use land cover. It is revealed that, during the initial hours (up to day 2) of the HRE, dynamical processes mainly regulate the convective parameters and as the system matures thermodynamical processes are dominant with the least contribution from transient eddy processes. Further analysis demonstrated that dynamical, thermodynamical, and transient eddy processes show unique patterns regulating the moisture distribution over the region. Daily bar plot of dynamical (a), thermodynamical (b), and eddy processes (c) for the top six ensembles over the inner domain (d02) throughout the 96 forecast hours (0000 UTC 10–14 July 2022). The study utilizes the Weather Research and Forecasting (WRF) model to simulate the heavy rainfall event (HRE) over the Gujarat in 2022. A multiphysics ensemble modeling framework is used to elucidate the new scientific insights causing this event up to a forecast lead time of 96 hr. The ensembles comprise six schemes for planetary boundary layer, cloud microphysics, cumulus, and two land‐use and land surface parameterizations (a total of 44 experiments). Based on the comprehensive rating metrics (CRM), the top three (i.e., E6, E2, EISRO) and bottom three (i.e., EUSGS, E8, E7) ensemble models are segregated. Results indicate that among all the CRM, total precipitation (PRCPTOT) and simple daily intensity index (SDII) contribute majorly toward the ranking of a model. In general, the ensemble has superior performance compared to deterministic models. Dynamical parameters (i.e., vertically integrated moisture flux transport [VIMFT], low‐level convergence, wind, and relative vorticity at 850 hPa) in all the ensembles captured the HRE signature robustly on day 1 and day 2 compared to day 3 and day 4. However, the mixed‐parameterized model (EISRO) produced the most realistic structure, emphasizing the importance of LULC. It is revealed that, during the initial hours (up to day 2) of the HRE, dynamical processes mainly regulate the convective parameters and as the system matures thermodynamical processes are dominant with the least contribution from transient eddy processes. Further analysis demonstrated that dynamical, thermodynamical, and transient eddy processes show unique patterns regulating the moisture distribution over the region. Beside the distinct north–south alignment of localized convection, the dynamical processes are noted to have the strongest correlation with the VIMFT followed by thermodynamical and eddy processes over the impacted region.
meteorology & atmospheric sciences