Forecasting Geomagnetic activity (Dst Index) using the ensemble kalman filter

Abstract

ABSTRACT A novel Ensemble Kalman Filter (EnKF) method is adopted to make reliable forecasts of geomagnetic activity (Dst index) for real-time applications. In this method, an educated estimate (forecast) of Disturbance storm time (Dst) is made based on the ring current dynamics like injection and decay rates. This estimated Dst is further updated with the assimilation of real-time Dst values or ∆H values from a single ground based magnetometer observations. The forecasted Dst values by this EnKF method are validated with the true Dst during the severely disturbed period that consists of 2 super (Dst< -250 nT), 8 intense (Dst< -100 nT), and 13 moderate geomagnetic storms (Dst< -50 nT). It is found that the EnKF method implemented in this work offers the best accurate forecast of Dst with a root mean square error (RMSE) and regression coefficient (R) of 4.3 nT and 0.99, respectively.