Principal component analysis in the modeling of HILDCAAs during the Solar Minimum of Cycle 23/24

Abstract

In this article, we propose a new approach to model the high-intensity, long-duration, continuous AE (Auroral Electrojet) activity (HILDCAA) by relaxing one of the criteria originally designed, based on the interplanetary features during the unusual Solar Minimum of Cycle 23/24 (SMC23/24). This relaxation does not intend to suppress or modify the original HILDCAAs’ conception, but propose a new view of the same phenomena by enlarging the sample of events, which in turn may improve space weather monitoring and prediction programs. To assess and classify the Alfvénity associated with HILDCAAs, the values of 4h-Windowed Pearson Cross-Correlation (4WPCC) between the IMF components and the solar wind speed components observed in situ at the Lagrangian point L1 (1 AU) were evaluated. The principal component analysis (PCA) was performed on the dataset and, from the first three principal components, which represent ~65% of the accumulative percent variance, we applied principal component regression (PCR) in each of the following parameters: the AE index, the Interplanetary Magnetic Field (IMF) components, the plasma density, the solar wind speed, the temperature, the IMF magnitude, and the SYM-H geomagnetic index. Furthermore, we applied Multiple Linear Regression (MLR) to establish a linear model to express the AE index in terms of the PCR-based model parameters. The AE MLR-based model demonstrated to hold a prognosis potential for HILDCAAs. Despite that, this model is only suitable for the SMC23/24. In this sense, this model might be implemented a real-time analysis for short-term HILDCAA prognosis in the near future.