Twenty‐four hour predictions of the solar wind speed peaks by the probability distribution function model

Abstract

Abrupt transitions from slow to fast solar wind represent a concern for the space weather forecasting community. They may cause geomagnetic storms that can eventually affect systems in orbit and on the ground. Therefore, the Probability Distribution Function (PDF) model was improved to predict enhancements in the solar wind speed. New probability distribution functions allow for the prediction of the peak amplitude and the time to the peak while providing an interval of uncertainty on the prediction. It was found that 60% of the positive predictions were correct, while 91% of the negative predictions were correct, and 20% to 33% of the peaks in the speed were found by the model. This represents a considerable improvement upon the first version of the PDF model. A direct comparison with the Wang-Sheeley-Arge (WSA) model shows that the PDF model is quite similar, except that it leads to fewer false positive predictions and misses fewer events, especially when the peak reaches very high speeds.