Abdus Salam International Center for Theoretical Physics, Italy

An ensemble of AMIP‐type experiments with prescribed interannual varying sea surface temperature (SST) and different initial conditions is used to study the relationship between Indian summer monsoon extreme conditions and the El Nino Southern Oscillation (ENSO). Based on the selection of extreme monsoon rainfall years ‘In Phase’ or ‘Out of Phase’ with respect to the observations, this study identifies specific SST and atmospheric circulation patterns responsible for the remote forcing on the monsoon. A clear common characteristic of externally forced extreme monsoon years is identified with an ENSO pattern having summer SST anomalies of the same sign in the tropical Pacific Ocean but also in the Indian and Atlantic tropical sectors. This finding assumes that the SST pattern in summer is enough to modulate the Walker circulation and consequently to suppress or enhance convection over South Asia, even if it does not evolve into an ENSO event. The analysis of the ‘Out of Phase’ cases (i.e. when the model reproduces a weak monsoon instead of a strong one, or the reverse) reveals how the model wrongly responds to the SST forcing, ignoring other processes like ocean–atmosphere coupling. Once ENSO is linearly removed the main source of remote forcing for strong (weak) monsoon characteristics over India is the tropical Atlantic with negative (positive) anomalies, and with weak anomalies of the same sign located in the south Indian Ocean. The results of the role of the forcing from the tropical Atlantic are also confirmed by a set of atmospheric model experiments where interannually varying SST is prescribed only in the Atlantic Ocean, while the rest of the SST is climatological.

A multi-model set of atmospheric simulations forced by historical sea surface temperature (SST) or SSTs plus Greenhouse gases and aerosol forcing agents for the period of 1950–1999 is studied to identify and understand which components of the Asian–Australian monsoon (A–AM) variability are forced and reproducible. The analysis focuses on the summertime monsoon circulations, comparing model results against the observations. The priority of different components of the A–AM circulations in terms of reproducibility is evaluated. Among the subsystems of the wide A–AM, the South Asian monsoon and the Australian monsoon circulations are better reproduced than the others, indicating they are forced and well modeled. The primary driving mechanism comes from the tropical Pacific. The western North Pacific monsoon circulation is also forced and well modeled except with a slightly lower reproducibility due to its delayed response to the eastern tropical Pacific forcing. The simultaneous driving comes from the western Pacific surrounding the maritime continent region. The Indian monsoon circulation has a moderate reproducibility, partly due to its weakened connection to June–July–August SSTs in the equatorial eastern Pacific in recent decades. Among the A–AM subsystems, the East Asian summer monsoon has the lowest reproducibility and is poorly modeled. This is mainly due to the failure of specifying historical SST in capturing the zonal land-sea thermal contrast change across the East Asia. The prescribed tropical Indian Ocean SST changes partly reproduce the meridional wind change over East Asia in several models. For all the A–AM subsystem circulation indices, generally the MME is always the best except for the Indian monsoon and East Asian monsoon circulation indices.