A linear Markov model has been developed to predict the short-term climate variability of the East Asian monsoon system, with emphasis on precipitation variability. Precipitation, sea level pressure, zonal and meridional winds at 850 mb, along with sea surface temperature and soil moisture, were chosen to define the state of the East Asian monsoon system, and the multivariate empirical orthogonal functions of these variables were used to construct the statistical Markov model. The forecast skill of the model was evaluated in a cross-validated fashion and a series of sensitivity experiments were conducted to further validate the model. In both hindcast and forecast experiments, the model showed considerable skill in predicting the precipitation anomaly a few months in advance, especially in boreal winter and spring. The prediction in boreal summer was relatively poor, though the model performance was better in an ENSO decaying summer than in an ENSO developing summer. Also, the prediction skill was better over the ocean than the land. The model's forecast ability is attributed to the domination of the East Asian monsoon climate variability by a few distinctive modes in the coupled atmosphere-ocean-land system, to the strong influence of ENSO on these modes, and to the Markov model's capability to capture these modes.

The sensitivity of the East Asian Monsoon (EAM) precipitation prediction skill to the variabilities of the tropical Indian and Pacific sea surface temperature (SST), the North Atlantic Oscillation (NAO), and the heating and snow cover over the Tibetan Plateau is evaluated in the framework of a linear Markov model. It is found that the tropical Indo‐Pacific SST helps to improve the prediction of EAM precipitation over oceanic regions and some localized areas over land, while the Indian Ocean alone does not show significant impact on prediction. The remote effects of NAO improve the prediction in the middle and lower reaches of the Yangtze and Huaihe River valleys in boreal spring, fall and winter, and over oceanic areas in boreal summer and fall. The predictive skill of our model is not sensitive to the inclusion of the sensible heating and snow cover over the Tibetan Plateau, probably because their effects are implicitly present in the original EAM model.