Toward Improving S2S Forecasts of the Japan Meteorological Agency's Global Ensemble Prediction System (JMA-GEPS)

The Japan Meteorological Agency operates the Global Ensemble Prediction System (JMA-GEPS) to support medium- and extended-range forecasts, including operational One-month Forecasts. The performance of this sub-seasonal to seasonal (S2S) forecast can be influenced by interactions among atmosphere, ocean, land, ice, aerosol and so on.

In particular, it is important to use better sea surface temperature (SST). In the atmospheric model of JMA-GEPS, SST is prescribed as persisting anomaly from the climatological SST throughout the forecast period, which can potentially make a large error for forecasts over weeks 3 and 4. To tackle this issue with limited computational cost, two-tiered SST method has been developed using operationally precomputed SST by Seasonal Ensemble Prediction System. We will show better performance with this method, comparing to the operational JMA-GEPS.

Furthermore, towards improving S2S forecast skill of JMA-GEPS, other feasibility studies on volcanic aerosol in the stratosphere are conducted. The preliminary results will be also demonstrated in this presentation.