| The Kerala Flood, the South India in 2018 is known as an unprecedented flood disaster resulted in more than 400 casualties, caused by abnormal rainfall in August. To mitigate the flood disaster, rainfall prediction is one of the most important nonstructural measures. We have investigated the possibility on reproducing the heavy rainfall diagnostically (boundary condition of ERA5) and prognostically (boundary condition of GFS and GEFS) via Weather Research and Forecasting (WRF) model. The target area is Chalakudy River subbasin of Periyar River in the central Kerala state.
The abnormal rainfall is supported by large moisture flux from the Arabian See west of Kerala, maintained by an axis of lowlevel jet originated by a cyclonic vortex in Bay of Bengal to the central India. As a result, heavy rainfall was developed along the windward side of the mountain range. However, the simulated rainfall underestimated two raingauge observations (Upper Sholayar Dam and Lower Sholayar Dam). Those raingauges are located between the foothill and top of the mountain range, but they are outside of our simulated heavy rainfall core area. When we tried finer resolution of model grid intervals from 5km to 1.66km, there was some improvement. An ensemble forecasting with the lateral boundary condition of GEFS forecast showed heavy rainfall core in the windward of the mountains as well, but it still has underestimation at the raingauge sites. Radar observation images showed rainbands transported from western ocean to the river basin, and provided heavy rainfall from coastal to mountainous area. Some ensemble members showed more rainfall in the foothill area than other members. The ensemble forecast could show the possibility of rainfall area in the foothill to the mountain range.
|
