Evaluating CMIP model simulations of the Indian Ocean Dipole - the role of monsoon air-sea feedback on its skewness

The Indian Ocean Dipole (IOD) is the dominant mode of interannual variability in the tropical Indian Ocean, characterized by surface cooling (warming) of the eastern Indian Ocean in its positive (negative) phase. Observed IOD events exhibit distinct amplitude asymmetry in relation to negative nonlinear dynamic heating (NDH) and asymmetric thermal damping. NDH is associated with the asymmetric nonlinear temperature advection and sea surface temperature (SST) response to thermocline changes. The thermal damping is caused by the cloud-radiation-SST feedback, which develops minor additional damping after cold SST anomaly leading to a cloud-free condition. In CMIP5, nearly all models simulate less-skewed IOD compared to the observed. On the other hand, for CMIP6, 6 out of 20 models show realistic skewness.

Further examination reveals that the improvement of IOD skewness in CMIP6 over CMIP5 can be attributed to the former's improved model mean states, particularly the western tropical Indian Ocean (WTIO) SST. Analyses of "less skewed models" indicate that when background SST in WTIO exceeds the SST for maximum precipitation sensitivity, the convective response to SST becomes asymmetric, leading to less zonal wind response and weaker zonal nonlinear advection during positive IOD. The warm SST bias in turn is associated with the weaker Indian summer monsoon. Additionally, ocean stratification also influences the IOD skewness: stronger stratification can lead to larger nonlinear SST response to thermocline changes; this is due to larger vertical temperature gradient resulting in larger nonlinear vertical advection. Our findings imply that substantial efforts should be made in reducing the Indian summer monsoon bias as well as advancing the upper-ocean stratification in the eastern Indian Ocean, for properly representing IOD in Earth System Models.