Multi-week prediction of South-East Asia rainfall variability during boreal summer in CFSv2

Abstract

In this paper we analyze the multi-week prediction bias and skill from the National Centers for Environment Prediction (NCEP) Climate Forecast System version 2 (CFSv2) based on its hindcasts for 1999–2012. The analyses focus on the prediction of the rainfall variability over South-East Asia during boreal warm seasons and the dependence of the prediction on the activity of intrasesaonal leading modes. It is shown that the prediction skill measured by anomaly correlation is comparable between the total anomalies and intraseasonal anomalies during the first 2 weeks. After week 2, the prediction skill drops substantially and the skill for total anomalies is largely from the prediction for the interannual variability. Moreover, the forecast skill tends to be higher when the amplitude of the Madden–Julian Oscillation and the Boreal Summer Intraseasonal Oscillation (BSISO) is larger, especially for the BSISO. It is noted that the prediction skill over South-East Asia depends on the phase of the BSISO. One deficiency in the CFSv2 is that the northward propagation of the forecast BSISO is generally slower than the observed.