Simulation of land-ocean influences on the Indian monsoon in simulations of present-day climate and at doubled carbon dioxide concentrations

A 150 year-long numerical simulation of present-day climate using the Max-Planck Institute’s coupled ocean-atmosphere model ECHAM4-T42-OPYC3 is analysed with regard to the interannual variability of the strength of the Indian summer monsoon and its relation to land-surface and ocean interactions. Individual years are categorised into three classes of monsoons: normal, strong and weak (greater or less than one standard deviation of precipitation over India). The ensembles of anomalous monsoons are then sub-divided into a composite of cases coinciding with sea surface temperature (SST) anomalies in the Pacific related to the El Niño-Southern Oscillation (ENSO) phenomenon and a second composite of anomalous monsoons occurring when no SST anomalies are found in the Pacific. The coupled model shows variations of the SST in the Pacific which are as large as and occur at a similar frequency as in observations. Thus it provides the basis for a realistic simulation of the interannual monsoon variability in ENSO-related conditions, but it overemphasizes the biennial component of occurrence. As in observations, about a third of all cases of weak monsoons occur in the summer when an El Niño begins to develop in the Pacific, while strong monsoons are often associated with La Niña events. This is an improvement from earlier coupled model simulations. In the model simulation, a modulation of the strength of the monsoon is due to a change of the large-scale land/ocean temperature gradient in the Indian Ocean sector in the mid-troposphere. The two composites show different developments during the annual cycle. In ENSO-related strong monsoon cases the atmosphere over land warms up during the spring in association with generally warmer tropics as a remnant from a warm event in the previous winter. During the summer months the warming in the Indian Ocean region is replaced by a cooling in association with the developing La Niña, while the land remains significantly warmer than normal. Therefore, in the coupled simulation the Indian Ocean shows only very small SST anomalies, while observed SSTs may vary in connection with ENSO events at a time lag of four months. For non-ENSO related monsoons a warming also occurs over land in the summer, but then neither the Indian Ocean nor the tropical west Pacific exhibit any significant anomalies during the spring. Simulated temperature anomalies responsible for these modulations are relatively small. Independent of the origin of the monsoon anomaly, strong monsoons differ from weak monsoons by a significant precipitation pattern over India and a modification of the 850 hPa zonal wind field over the maritime continent and the West Pacific. Similar to observations, the anomalous monsoons in the coupled model are related to precursors in the 200 hPa zonal wind field in the spring, but no evidence could be found for a significant influence from the Eurasian snow pack in the spring on the subsequent Indian summer monsoon. While the model shows many realistic features, the variation of the monsoon occurs against a background of a deficient regional rainfall pattern in India and too small a range of Indian Ocean SST variations in conjunction with ENSO events. It is shown how inter-annual variability is modified in a changed climate as associated with carbon dioxide levels projected for the next century (Roeckner et al., 1998). The relative influences from the Asian land mass and the Pacific Ocean changes under these circumstances.