Abstract: The skills of seasonal forecast of East-Asian precipitation, especially for summer seasons were studied. In the perfect model scenario, the potential predictability of East-Asian precipitation in seasonal time scale using signal-to-noise ratio and information-based indices were commented. The most predictable components and corresponding spatial patterns were also derived by PrCA analysis. The results show that the relationship of forecasts and observations displayed by correlation coefficient is characteristic of higher level in ocean region at low latitudes than in land region at high latitudes. While root mean square error skill shows that forecast diverging degree from observations in ocean region is much more than that in land region. Both agree with the current level for forecast. By estimating potential predictability, it shows that the potential predictability exists a spatial variation and it decays from low to high latitude and from ocean to land region. Further analysis of signal and noise components reveals the important role of marine forcing. The skill of seasonal rainfall is quite limited for lacking of apparent marine forcing in the mainland. By projecting the time series of the leading two modes onto the SSTA field, it shows that the forcing of ocean is the main source of potential predictability of East-Asian precipitation prediction in seasonal time scale. Despite the complexity of ENSO, it is clear that it has great influence on East-Asian monsoon, and the teleconnection between East-Asian precipitation prediction and ENSO may reveal the internal relations of them.

Key words: East-Asian precipitation, seasonal forecast, potential predictability, most predictable component