Last decade has seen a tremendous improvement in the forecasting skill of numerical weather prediction (NWP) models. This is attributed to increased sophistication in NWP models, which resolve complex physical processes, advanced data assimilation, increased grid resolution and satellite observations. However, prediction of heavy rains is still a challenge since the models exhibit large error in amounts as well as spatial and temporal distribution. Two state-of-art NWP models have been investigated over the Indian monsoon region to assess their ability in predicting the heavy rainfall events. The unified model operational at National Center for Medium Range Weather Forecasting (NCUM) and the unified model operational at the Australian Bureau of Meteorology (Australian Community Climate and Earth-System Simulator — Global (ACCESS-G)) are used in this study. The recent (JJAS 2015) Indian monsoon season witnessed 6 depressions and 2 cyclonic storms which resulted in heavy rains and flooding. The CRA method of verification allows the decomposition of forecast errors in terms of error in the rainfall volume, pattern and location. The case by case study using CRA technique shows that contribution to the rainfall errors come from pattern and displacement is large while contribution due to error in predicted rainfall volume is least.

Date Published: 3 May 2016
PDF: 10 pages
Proc. SPIE 9882, Remote Sensing and Modeling of the Atmosphere, Oceans, and Interactions VI, 98821O (3 May 2016); doi: 10.1117/12.2223646

Published in SPIE Proceedings Vol. 9882:
Remote Sensing and Modeling of the Atmosphere, Oceans, and Interactions VI
Tiruvalam N. Krishnamurti; Madhavan Nair Rajeevan, Editor(s)