Statistical analysis on the current capability to predict the Ap Geomagnetic Index

Abstract

In this work, forecasting results of the daily Ap geomagnetic index by different Space Weather Prediction Centers (SWPCs) covering the time period from October 2014 up to July 2020, are considered. A three-day forecast of this index obtained from the space weather prediction centers, is analyzed. Standard forecast verification measures, descriptive statistics with correlation coefficients, and error analysis between forecasts and observations have been performed to evaluate the quality or the skill of the predictions. In particular, an error analysis using fit performance metrics, such as the mean average error (MAE) and the root mean squared error (RMSE) among others, as well as the threshold performance metrics, such as the Hansen-Kuipers, Gilbert and Heidke skill scores, the probability of detection (POD), the probability of false detection (POFD) and the area under the curve in receiving the operating characteristic (ROC) plot among others, are calculated. For the Ap geomagnetic index predictions during Day-0, the Pearson's correlation coefficient is ranging between the values of 0.57 and 0.79, while during the Day-2 it is decreased to the range from 0.37 to 0.44. In conclusion, the majority of the used SWPCs perform quite accurately on the conditions of Space Weather in active as well as in quiet periods showing a reliable effort for predicting geomagnetic storms.