Northward propagation of convection over the Indian region in multiphysics multimodel ensemble

Abstract

An assessment of a multiphysics multimodel ensemble (MPMME) strategy is provided to simulate the critical aspects of the Indian summer monsoon and its intraseasonal variability. Using various physics combinations of Climate Forecast System (CFS) and its atmospheric component Global Forecast System (GFS), a 15‐year hindcast for May–October is generated. Three different convective parametrizations, simplified Arakawa–Shubert (sas), revised deep‐convection sas (nsas), and revised sas with modified shallow‐convection (nsas_sc) are coupled with two microphysics schemes Zhao and Carr (zc) and Ferrier. Spatiotemporal characteristics of predicted Indian summer monsoon climatology and 20–70‐day periodic intraseasonal oscillations (ISOs) are evaluated using observations. MPMME members reproduce the overall characteristics of the seasonal mean, but they have significant biases over different regions. Pattern correlations reveal that CFS_nsaszc performs best among MPMME in simulating the observed characteristics of rainfall ISOs and providing significant ISO forecast up to pentad 3 lead. A diagnostic based on the vorticity budget equation during strong convective events (SCEs) associated with ISOs is used to understand better the mechanism of northward‐propagating ISOs and the responsible factors that develop a vorticity tendency to the north of convection maxima. The tilting term in the vorticity equation shows northward propagation and leads precipitation maxima by about a week over the Bay of Bengal. Vertical shear of mean zonal winds and meridional gradients of vertical winds are found to be essential in developing vorticity tendency. SCEs are better represented in CFS than in GFS. Notably, along with CFS_nsaszc, two CFS_sas members capture the occurrence of SCEs reasonably well. However, errors in vertical shear of mean zonal winds are remarkably high after pentad 2 lead in CFS_sas and GFS, explaining their relative weakness in simulating ISOs during June–September. This study demonstrates that the MPMME strategy could utilize individual physical schemes' strengths to provide better subseasonal forecasts.