Predictability of CFSv2 in the tropical Indo-Pacific region, at daily and subseasonal time scales

The predictability of a coupled climate model is evaluated at daily and intraseasonal time scales in the tropical Indo-Pacific region during boreal summer and winter. This study has assessed the daily retrospective forecasts of the Climate Forecast System version 2 from the National Centers of Environmental Prediction for the period 1982–2010. The growth of errors in the forecasts of daily precipitation, monsoon intraseasonal oscillation (MISO) and the Madden–Julian oscillation (MJO) is studied. The seasonal cycle of the daily climatology of precipitation is reasonably well predicted except for the underestimation during the peak of summer. The anomalies follow the typical pattern of error growth in nonlinear systems and show no difference between summer and winter. The initial errors in all the cases are found to be in the nonlinear phase of the error growth. The doubling time of small errors is estimated by applying Lorenz error formula. For summer and winter, the doubling time of the forecast errors is in the range of 4–7 and 5–14 days while the doubling time of the predictability errors is 6–8 and 8–14 days, respectively. The doubling time in MISO during the summer and MJO during the winter is in the range of 12–14 days, indicating higher predictability and providing optimism for long-range prediction. There is no significant difference in the growth of forecasts errors originating from different phases of MISO and MJO, although the prediction of the active phase seems to be slightly better.