Speaker
Description
Geomagnetic activity indices, such as the well-known Kp index and the recently developed Hpo index, are essential for monitoring and forecasting space weather. These indices provide a global assessment of geomagnetic disturbances caused by solar activity based on data collected from a network of worldwide distributed geomagnetic observatories. Accurate forecasting of these indices is crucial because geomagnetic storms can significantly impact technological infrastructure, including satellite systems, power grids, and communication networks, as well as human activities in Earth’s atmosphere and near-Earth space.
We present a novel method for forecasting geomagnetic indices using observatory-level data [1]. By accounting for the regional variability of geomagnetic storm impacts across different observatory locations, our approach improves prediction accuracy and provides forecasts tailored to local conditions. The model employs a supervised neural network—specifically, a multi-layer perceptron (MLP) classifier—to predict H30, H60, and K values at each observatory based on solar wind parameters. These localized predictions are then combined using standard procedures to derive the global Hpo and Kp indices.
We evaluate the performance of our model in forecasting the Kp index alongside the high-resolution Hpo indices, Hp60 and Hp30, which offer hourly and half-hourly representations of planetary geomagnetic disturbances. These indices serve a similar purpose to the traditional 3-hourly Kp index but have the added advantage of being continuous and unbounded, lacking the upper limit of 9 imposed on Kp. Our results show good agreement between predicted and observed values, with the model effectively capturing key trends and the overall behaviour of geomagnetic activity—even under conditions of sparse solar wind input. These findings underscore the model’s reliability and potential for enhancing geomagnetic forecasting capabilities. Additionally, the adaptable nature of the framework enables its application to other globally derived geomagnetic indices, such as Dst and AE.
[1] Kervalishvili, G., Michaelis, I., Korte, M., Rauberg, J., Matzka, J (2025). A novel model for forecasting geomagnetic indices using machine learning. Geophysical Research Letters, 52, e2025GL114848. https://doi.org/10.1029/2025GL114848.
| Do you plan to attend in-person or online? | In-person |
|---|
