Sub-seasonal to Seasonal Scale Skill Assessment of NMME Models in Forecasting Precipitation in California

In this presentation we report on the assessment of sub-seasonal (2 weeks to 1 month) to seasonal (>1 month) lead precipitation forecasts from the North American Multimodel Ensemble (NMME) models and the ensemble mean, over California. Skillful forecast of precipitation in this region is much needed by the Department of Water Resources, and other regional and local scale water agencies. A significant fraction of the precipitation in California tends to come from a few big precipitation events (e.g. atmospheric rivers) making the interannual precipitation in this region highly variable and precipitation forecasting a challenge. The importance of large storms to California's precipitation regime makes it a good candidate for evaluating sub-seasonal to seasonal scale forecasts from a multimodel ensemble forecast systems such as the NMME . We examine the skill of the individual NMME models and the ensemble mean precipitation forecast over the hindcast period (1982-2010) and also their performances during specific anomalous climate events such as strong ENSO events that are known to have some influence on the precipitation in this region. Finally we look at the fidelity of the NMME models in forecasting the Madden Julian Oscillations (MJO) that is shown to have influence on the intensity of the precipitation events in this region.