Local diagnostics of aurora presence based on intelligent analysis of geomagnetic data

Abstract

Despite the existing variety of approaches to monitoring space weather and geophysical parameters in the auroral oval region, the issue of effective prediction and diagnostics of auroras as a special state of the upper ionosphere at high latitudes remains virtually unresolved. In this paper, we explore the possibility of local diagnostics of auroras through mining of geomagnetic data from ground-based sources. We assess the significance of indicative variables and their statistical relationship. So, for example, the application of Bayesian inference to the data from the Lovozero geophysical station for 2012–2020 has shown that the dependence of a posteriori probability of observing auroras in the optical range on the state of geomagnetic parameters is logarithmic, and the degree of its significance is inversely proportional to the discrepancy between empirical data and approximating function. The accuracy of the approach to diagnostics of aurora presence based on the random forest method is at least 86 % when using several local predictors and ~80 % when using several global geomagnetic activity indices characterizing the geomagnetic field disturbance in the auroral zone. In conclusion, we discuss promising ways to improve the quality metrics of diagnostic models and their scope.

Keywords

• auroras
• geomagnetic variations
• geomagnetic data
• ascaplots
• machine learning
• data mining
• bayesian inference
• random forest