Using gradient boosting regression to improve ambient solar wind model predictions

TL;DR

This paper introduces a machine learning model using gradient boosting regression that improves the accuracy and speed of predicting Earth's ambient solar wind conditions, outperforming existing models.

Contribution

The study presents a novel, fast, and well-validated machine learning approach that leverages magnetic models of the solar corona to enhance solar wind predictions.

Findings

Outperforms existing models in most validation metrics

Provides a fast and open-source forecasting tool

Offers insights into the physical processes of solar wind

Abstract

Studying the ambient solar wind, a continuous pressure-driven plasma flow emanating from our Sun, is an important component of space weather research. The ambient solar wind flows in interplanetary space determine how solar storms evolve through the heliosphere before reaching Earth, and especially during solar minimum are themselves a driver of activity in the Earth's magnetic field. Accurately forecasting the ambient solar wind flow is therefore imperative to space weather awareness. Here we present a machine learning approach in which solutions from magnetic models of the solar corona are used to output the solar wind conditions near the Earth. The results are compared to observations and existing models in a comprehensive validation analysis, and the new model outperforms existing models in almost all measures. In addition, this approach offers a new perspective to discuss the role of different input data to ambient solar wind modeling, and what this tells us about the underlying physical processes. The final model discussed here represents an extremely fast, well-validated and open-source approach to the forecasting of ambient solar wind at Earth.