Impact of reduced ENSO variability and amplitude on ISMR prediction in the long‐lead forecasts of monsoon mission CFS

Abstract

Despite the reduction in interannual variability and amplitude of El Nino Southern Oscillation (ENSO) after 2000, its influence on Indian summer monsoon rainfall (ISMR) is stronger than in the previous decades. Meanwhile, our analysis with high-resolution Monsoon Mission Climate Forecast system (MMCFS, 38 km horizontal resolution) indicates a sharp decrease in ISMR skill and ENSO-ISMR teleconnection since 2000 at long lead (3 months, Feb IC) hindcasts. At the same time, ISMR skill is intensified for a short lead (1 month, Apr IC). The persistence skill of ENSO decreased after 2000, but the decrease is more severe for January-March initial months. So the ENSO skill will be lower for models initialized during these months compared with other months. Also, the Feb IC hindcast has a similar ENSO teleconnection pattern during both periods and hence it failed to capture the modified influence of ENSO after 2000, such as the absence of dipole anomalies in the Indian Ocean and stronger subsidence over India. Thus, the model hindcasts that depend heavily on ENSO teleconnection for ISMR simulation have reduced skill in recent years. The role of other predictors along with ENSO for Apr IC hindcasts is hence evident in the current period. It is shown that after 2000, tropical Atlantic SST anomalies contributed significantly to ISMR interannual variability in observations. In these, Atlantic Meriodnal Mode (AMM) influences ISMR through off-equatorial circulation in absence of ENSO during the second period. Atlantic zonal mode has a weaker influence in absence of ENSO during both periods. Atlantic SST influence on ISMR is captured by April IC hindcasts during period 2, while Feb IC has the opposite relationship. Thus, the presence of another predictor in the form of tropical Atlantic SST along with strong ENSO-ISMR teleconnection makes the short lead hindcast of ISMR more skilful in the recent period for MMCFS.

Monsoon Mission Coupled Forecast System version 2.0: model description and Indian monsoon simulations

“…The convective parameterization scheme used in the atmospheric part of MMCFSv1 and MMCFSv2 is based on the Arakawa-Schubert scheme, with orographic gravity wave, drag, and momentum mixing. Pillai et al (2022) showed that the prediction skill for El Niño-Southern Oscillation (ENSO) was lower for MMCFSv1 initialized with February (3-month lead time) initial conditions compared with when it was initialized with April (1-month lead time) initial conditions. They showed that models which depend on ENSO teleconnection for ISMR interannual variability (MMCFSv1 in their case) have better ISMR prediction skills with April initial conditions.…”

“…Recent studies (Ramu et al, 2016;George et al, 2016;Pillai et al, 2022) have shown that the seasonal prediction skill of monsoon in MMCFSv1 is significantly impacted by the El Niño-Southern Oscillation (ENSO)-monsoon relationship. MMCFSv1 also has some limitations in representing the relationship between Indian Ocean SST and monsoon.…”

Monsoon Mission Coupled Forecast System Version 2.0: Model Description and Indian Monsoon Simulations

“…The convective parameterization scheme used in the atmospheric part of MMCFSv1 and MMCFSv2 is based on the Arakawa-Schubert scheme, with orographic gravity wave, drag, and momentum mixing. Pillai et al (2022) showed that the prediction skill for El Niño-Southern Oscillation (ENSO) was lower for MMCFSv1 initialized with February (3 months lead time) initial conditions compared to when it was initialized with April (1 month lead time) initial conditions. They showed that models which depend on ENSO teleconnection for ISMR interannual variability (MMCFSv1 in their case) have better ISMR prediction skills with April initial conditions.…”

“…Recent studies (Ramu et al, 2016;George et al, 2016;Pillai et al, 2016Pillai et al, , 2022 have shown that the seasonal prediction skill of monsoon in MMCFSv1 is significantly impacted by the ENSO-monsoon relationship. MMCFSv1 also has some limitations in representing the relationship between Indian Ocean SST and monsoon.…”