Assessment of 1 month forecasts of weak Indian monsoons based on the NCEP Climate Forecast System (CFS)

This study focuses on analyses and validation of 1 month forecasts (OMFs) of weak Indian monsoons based on 10 member ensemble hindcasts (retrospective forecasts) of the NCEP Climate Forecast System (CFS) model for the period 1981–2008. The weak monsoon episodes chosen for the analysis correspond to summer monsoon months which were characterized by significant deficits in the All-India monthly rainfall of − 20% of the climatological normal. Examination of the CFS-OMFs shows poor skill of the model in capturing the observed rainfall and circulation anomalies during weak monsoons. The present analysis suggests that deficiencies in realistically capturing the ocean-atmosphere coupling in the tropical Indian Ocean (IO) introduces biases in simulating sea surface temperature and rainfall anomalies in the equatorial region, which in turn affects the monsoon precipitation forecasts over the sub-continent. In particular, the mean thermocline in the near-equatorial IO is found to be practically flat in the CFS model, so that the near-equatorial anomalies in the model are not strong enough to weaken the summer monsoon circulation and reduce the monsoon precipitation over India. By examining a 100 year free run of the CFS model, it is seen that moderate monsoon-droughts simulated by the model have weak teleconnections with the equatorial IO dynamics. On the other hand, intense monsoon-droughts in the CFS-model are found be remarkably linked with the equatorial IO anomalies. It is suggested that improving the slope of the equatorial IO thermocline and allowing for more realistic IO-monsoon coupling in the CFS-model would be an important step for improving the skill of extended-range monsoon forecasts