Prototype one day lead time forecasts of the magnetosphere with ENLIL and OpenGGCM

Forecasting the state of the magnetosphere is currently limited to lead times of the order of 30 minutes when data from a L1 monitor are used. In order to improve the forecast window we use ENLIL, an established heliosphere model, to first predict the solar wind state immediately upstream of Earth. The predicted solar wind and magnetic field are then used to drive the OpenGGCM magnetosphere-ionosphere-thermosphere model. With this combination the magnetosphere state, for example ionosphere currents and magnetosphere particle fluxes, can be forecast up to two days in advance. In this presentation we investigate the accuracy of such predictions and we discuss the prospects for continuous magnetosphere ensemble forecasts utilizing novel fast hardware such as the Cell based PS3 game consoles with appropriately optimized model software.