Sun-to-magnetosphere modeling: CISM forecast model development using linked empirical methods

The Wang-Sheeley-Arge (WSA) method is used to predict the solar wind speed (and certain other parameters) near the Earth's orbit based upon solar surface measurements. This approach gives a predicted solar wind time series with a lead time of three to four days. Such forecasted solar wind conditions can then be convolved with linear and nonlinear filters in order to provide a predicted set of geomagnetic indices or various particle flux estimates. In order to illustrate the method in a concrete way, we present here a demonstration of an end-to-end empirical forecast of relativistic electrons in the outer Van Allen radiation belt. Past work has shown that radiation belt electron fluxes are highly dependent on the speed of the solar wind striking the magnetosphere. We develop filters that predict electron fluxes using the WSA estimates of solar wind speed at Ll, which allows for 3-4 days lead times. We compare the prediction efficiency (PE) provided by these filters with filters developed to use 3-4 day old values of the solar wind velocity measured at Ll and 3-4 day old values of the measured electron fluxes themselves. It is found that the WSA method provides PEs of the electron flux that are slightly lower than that provided by using old Ll or the autocorrelated electron flux data. (C) 2004 Elsevier Ltd. All rights reserved.