Forecasting SEP Event Properties by Fitting Intensity Profiles with Functional Forms

This work explores the use of functional forms to forecast the properties of solar energetic particle (SEP) events during the rise phase of the solar protons themselves. The functional forms tested are the lognormal, Weibull and inverse Weibull functions, each of which requires 4 parameters to be determined. The fitting process is applied to GOES >10 MeV proton measurements (both light curves and cumulative fluence) for over 200 events, of which 78 are used as the training sample for a forecasting protocol. It is found necessary to fix several of the functional parameters early in the event when few data points are being fitted, and the longitude dependence of the free parameters over the sample is used to provide values in this case. The procedure and the resulting uncertainty in the forecast is discussed: for well-behaved events, the peak flux can be estimated to within a factor of 5 within 1 hour of event onset.