Speaker
Description
Here we report on a methodology to predict the duration and intensity of space weather events using re-analysis of long-term space weather data to develop a computationally inexpensive algorithm fed only by operationally-available, real-time data streams. As a use case, we apply this methodology to relativistic electron events based on 30+ years of NOAA GOES >2 MeV electron fluxes but the methodology is more broadly applicable and extendable.
In the US, NOAA releases alerts when the >2 MeV electron fluxes from either of two GOES satellites exceeds a fixed flux threshold of 1,000 1/(cm2-s-sr). Continuing alerts are issued on subsequent days if the alert level is exceeded again. Several techniques have been developed to predict a daily time series of electron fluxes up to a day or two into the future but electron enhancements typically occur as discrete “events” who’s onset is difficult to predict. Electron events last anywhere from a single day to over a week with longer-lasting events having more adverse effects. Our methodology uses a decision tree to predict likely event duration starting on day 1 allowing an “All Clear” forecast. For events lasting two or more days we use Bayesian inference to refine and update the prediction of event duration.
It is also of interest to spacecraft operators to forecast maximum event intensity and fluence level (time-integrated flux). We define an open-ended scale for these quantities (similar to the NOAA space weather scales) based on the probability that flux (or fluence) will exceed higher threshold levels.
We discuss the performance metrics for these predictions; initial experiences with testing our algorithms in the CCMC real-time environment; and lessons-learned from working with NASA and NOAA to transition the algorithms toward operational readiness. We also consider how this, or similar methodology can be applied to a variety of space weather phenomena that also happen as discrete “events” superposed on a more predictable and slowly-evolving baseline level
