Predicting Cosmic Ray Intensity Variations in Solar Cycle 24 Using Solar Wind and IMF Parameters

Abstract

Cosmic ray intensity at Earth fluctuates with the changing solar wind and interplanetary magnetic field. Solar Cycle 24, known for its unusually weak activity, offers a clear view of these variations. This study uses Oulu neutron monitor data together with solar wind and IMF parameters from NASA’s OMNI database to build daily prediction models. Multiple linear regression, Ridge regression, and Random Forest regression are tested using rolling-window validation so that each section of the cycle is evaluated independently. IMF magnitude and solar wind speed emerge as the strongest predictors, while Random Forest captures nonlinear behavior—especially during Forbush decreases—far better than the linear approaches. The findings demonstrate that data-driven models can reliably support short-term cosmic-ray forecasting when real-time measurements are available.