TY - JOUR
T1 - Solar Particle Event Dose Forecasting Using Regression Techniques
AU - Lovelace, Alan
AU - Rashid, Al
AU - Wet, Wouter C. de
AU - Townsend, Lawrence W
AU - Hines, J Wesley
AU - Moussa, Hanna
PY - 2018/8/18
Y1 - 2018/8/18
N2 - Doses from solar particle events can be a serious threat to the wellbeing of crews traveling
through space. Therefore, methods for predicting the time such events will take place, methods for
forecasting the dose buildup over time, and methods for forecasting the potential total dose from such
events are needed to enable crews to take actions to mitigate the effects by entering a shielded area
designed for their protection. This work focuses on forecasting the total dose expected for an event, based
upon doses obtained very early in the event, using the kernel regression method. The model uses tables of
calculated doses for historical solar particle events augmented with hypothetical events similar to the
actual ones for training purposes. Reasonably accurate predictions of the total dose expected for an event
can be made within the first hour after event onset. Predictive accuracies generally increase as the event
progresses in time. The only inputs required are doses and times since
AB - Doses from solar particle events can be a serious threat to the wellbeing of crews traveling
through space. Therefore, methods for predicting the time such events will take place, methods for
forecasting the dose buildup over time, and methods for forecasting the potential total dose from such
events are needed to enable crews to take actions to mitigate the effects by entering a shielded area
designed for their protection. This work focuses on forecasting the total dose expected for an event, based
upon doses obtained very early in the event, using the kernel regression method. The model uses tables of
calculated doses for historical solar particle events augmented with hypothetical events similar to the
actual ones for training purposes. Reasonably accurate predictions of the total dose expected for an event
can be made within the first hour after event onset. Predictive accuracies generally increase as the event
progresses in time. The only inputs required are doses and times since
U2 - 10.1029/2017SW001773
DO - 10.1029/2017SW001773
M3 - Article
SP - 1073
EP - 1085
JO - Space Weather Journal
JF - Space Weather Journal
ER -