Automated High‐Frequency Geomagnetic Disturbance Classifier: A Machine Learning Approach to Identifying Noise While Retaining High‐Frequency Components of the Geomagnetic Field

Abstract

We present an automated method to identify high‐frequency geomagnetic disturbances in ground magnetometer data and classify the events by the source of the perturbations. We developed an algorithm to search for and identify changes in the surface magnetic field, dB/dt, with user‐specified amplitude and timescale. We used this algorithm to identify transient‐large‐amplitude (TLA) dB/dt events that have timescale less than 60 s and amplitude >6 nT/s. Because these magnetic variations have similar amplitude and time characteristics to instrumental or man‐made noise, the algorithm identified a large number of noise‐type signatures as well as geophysical signatures. We manually classified these events by their sources (noise‐type or geophysical) and statistically characterized each type of event; the insights gained were used to more specifically define a TLA geophysical event and greatly reduce the number of noise‐type dB/dt identified. Next, we implemented a support vector machine classification algorithm to classify the remaining events in order to further reduce the number of noise‐type dB/dt in the final data set. We examine the performance of our complete dB/dt search algorithm in widely used magnetometer databases and the effect of a common data processing technique on the results. The automated algorithm is a new technique to identify geomagnetic disturbances and instrumental or man‐made noise, enabling systematic identification and analysis of space weather related dB/dt events and automated detection of magnetometer noise intervals in magnetic field databases.

Magnetosphere–Ionosphere Drivers of Transient‐Large‐Amplitude Geomagnetic Disturbances: Statistical Analysis and Event Study

“…The next steps of the algorithm incorporate a filtering process that has requirements derived from the statistical analysis of geophysical and noise‐type events described in McCuen et al. (2023). The first condition is that at least one d B /d t interval identified from the first step in each hour window of data lasts 10 s or more.…”

“…This condition is implemented because many noise-type events in magnetometer data consist of more than 5% concentration of large, second-timescale dB/dt (hundreds, sometimes thousands of dB/dt within an hour period), so this ratio filter excludes instances that are highly likely to be a result of noise interference rather than geophysical source. The 5% ratio threshold is another requirement derived from the analysis of McCuen et al (2023).…”

Catalog of Geomagnetic Storms with Dst Index ≤ −50 nT and Their Solar and Interplanetary Origin (1996–2019)

“…On the other hand, the Starlink satellite failure in February 2022 happened during the course of a weak GS [16] with the exact reasons still being debatable, e.g., see [17] and the references therein. Novel techniques, such as machine learning algorithms for the classification of GSs, are also being tested [18].…”