PROSWIN: Probabilistic Solar Wind Speed Forecasting Using Deep Distributional Regression From Solar Images
PROSWIN forecasts hourly solar wind speed four days ahead with calibrated probabilistic uncertainties.
Researchers present PROSWIN, a deep distributional-regression model that forecasts hourly solar wind speed at Earth four days ahead from solar images and magnetograms. On 14 years of data, forecast uncertainties deviate by less than 1% on average. Timeline CRPS is 41.0 km/s, high-speed-stream peak CRPS is 45.3 km/s, and a new prediction score is 42.3 km/s. The authors report it is the only compared model accurate for both timeline and peak values.
- Forecasts hourly solar wind speed at Earth four days ahead
- Timeline CRPS is 41.0 km/s and HSS peak CRPS is 45.3 km/s
- Average uncertainty calibration error is under 1 percent
- The 171 Å channel complements usual 193 Å and 211 Å inputs
- AI models
- PROSWIN
From the original article
1 min read at arxiv.orgAccurately predicting fast solar wind conditions is challenging, as uncertainties are large and unquantified by traditional single-value prediction models. In particular, the risks of high-speed solar wind streams (HSSs), which can cause damage to technological infrastructure, cannot be reliably assessed without probabilistic forecasts.
By Daniel Collin · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.