Simulation of the African ITCZ during austral summer seasons and ENSO phases : application of an RCM derived from stretched grid ESM

Abstract

INTRODUCTION : Climate predictability across timescales in a changing climate presents a unique opportunity and challenges for state-of-the-art climate models. The use of regional climate models (RCMs) forced with interactively coupled Earth System Models (ESMs) for the sub-seasonal, seasonal, and decadal predictions is an actively growing research area. METHODS : The study explores a stretched-grid RCM constrained with an ESM which integrates a climate change signature. Spectral relaxation paradigm is applied to limit the climate drift within the range of the multi-model sea-surface temperature (SST) and sea-ice concentration (SIC) variability. The model retroactive ensemble simulations for November initialization are evaluated on the seasonal migration of the ITCZ during El-Niño and La-Niña phases, exploring both the spatial and zonal positions. The model is also evaluated on the ITCZ process’ characteristics that include the Hadley cell (HC), stream function and the subtropical jet stream (STJ) using quantitative methods. RESULTS : The RCM and the driving ESM demonstrate skillful performance in identifying the seasonal trajectory of both the spatial and zonal migration of the ITCZ during El-Niño and La-Niña. Moreover, the RCM also demonstrates a good skill in determining both the descending edge of the HC and the STJ with the highest mean percentage error of 16.3 and 7.5% for the HC and STJ latitudes, respectively. CONCLUSIONS : The November initialization of the RCM skillfully simulates the seasonal migration of the ITCZ (and related characteristics) aligned to the observations and reanalysis datasets. Notwithstanding, the RCM manifests a tendency of more dynamic error growth relative to its driving ESM as the lead time increases. Furthermore, the RCM is also out of phase with a southerly shift of the stream function compared to the 500 hPa reanalysis stream function. The modeling framework offers process oriented and teleconnection studies. It also provides great potential for climate applications with suitable bias corrections techniques, albeit the source and mechanism of its dynamic error growth deserve further investigation.