Dynamical Seasonal Prediction of Indian Summer Monsoon using AGCM: Weighted Ensemble Mean Approach

Abstract

This study examines the fidelity of portable unified model's Atmospheric General Circulation Model in ensemble seasonal prediction of Indian Summer Monsoons of 1999-2004, driven by May SST anomaly persistent boundary conditions. Simple Ensemble-mean (EM) is inappropriate due to the presence of large deviation among the ensemble members in simulation of Indian Summer Monsoon Rainfall (ISMR). Thus 'Weighted Ensemble Mean' (WEM) method is used in the present study. In WEM method, weights are determined for all ensemble members at each model grid point using daily precipitation anomaly to distinguish the most reliable members and outliers among the ensemble members. Ensemble-mean then obtained by weighted combinations of all ensemble members is referred as 'Weighted Ensemble Mean'. The WEM prediction of ISMR better matches with observations than EM in majority of the monsoons. Further, WEM estimated using monthly and seasonal mean weights are assessed with respect to WEM from daily mean weights. WEM with monthly/seasonal weighting is found to be similar to EM in most of the monsoons and hence daily weighting is more suited approach than monthly/seasonal weighting.