Related Experiment Videos
Predicting dose-time profiles of solar energetic particle events using Bayesian forecasting methods.
1University of Tennessee, Knoxville, TN 37996, USA. nealjs1@ornl.gov.
IEEE Transactions on Nuclear Science
|May 30, 2002
Summary
Bayesian inference and Markov chain Monte Carlo methods predict energetic solar particle event dose-time profiles using early event data. This approach models dose growth and provides event predictions.
Area of Science:
- Space physics
- Computational statistics
Background:
- Energetic solar particle events pose risks.
- Accurate dose-time profile prediction is crucial for mitigation.
Purpose of the Study:
- To develop a Bayesian inference framework for predicting dose-time profiles of solar particle events.
- To utilize early event measurements for predictive modeling.
Main Methods:
- Application of Bayesian inference techniques.
- Employing Markov chain Monte Carlo (MCMC) sampling.
- Development of hierarchical models for surrogate dose values.
- Assumption of nonlinear, sigmoidal dose growth.
Main Results:
- Demonstrated prediction of dose and dose rate using Bayesian posterior predictive distributions.
- Successfully modeled relationships among similar solar particle events.
- Provided example predictions for specific historical events (Nov 8, 2000; Aug 12, 1989).
Conclusions:
- Bayesian inference coupled with MCMC is effective for predicting solar particle event dose-time profiles.
- Early dose and dose-rate measurements are valuable inputs for these predictive models.
- The sigmoidal dose growth model provides a robust framework for event analysis.