Realities, Challenges, and Innovation for Solar Flare Forecasting
Abstract
In light of recent head-to-head evaluations of operational flare forecasting facilities, some encouraging trends have been identified, as have some specific challenges. The use of modern magnetic data, for example, can be helpful in some situations -- but having a Forecaster in the Loop still provides additional skill as compared to fully automated methods. Challenges include the fact that generally speaking all methods score below 0.5 on a 0.0--1.0 scale across numerous standard metrics and all methods pretty much fail to correctly identify and predict upcoming variations in flaring activity (the first flare / last flare challenge). As the studies have demonstrated, many different implementation options have been tried; in the context of human-oriented operational forecasts as they are presently defined, perhaps we've gotten as good as we're going to?
At the same time, numerous efforts have been recently published or are underway to establish more promising approaches and algorithms. One difficulty is that the timescales and performance benchmarks thus far are defined by humans: for example defining the required forecast as being for whether or not the Sun output a minimum flux of certain-wavelength light as detected by a human-built satellite sometime in an upcoming human-convenient time period. But to improve forecasts, we probably need to turn to more physics-based timescales -- yet still keep operational requirements in mind such as data-availability reality, the difficulties of turning super-posed epoch analysis into successful forecasts, and how to properly (and fairly) judge performance. In this talk I will motivate the challenges and highlight some ways forward now that we know where we really stand operationally and where the flare-forecasting needs are the clearest. Leka, K. D., et al 2019 a, ApJSupp in press Leka, K. D., et al 2019 b, ApJ in press Park, S.-H, et al 2019, ApJ submitted Support for the workshops upon which some of this talk is based is acknowledged from the Center for International Collaborative Research (CICR), at the Institute for Space-Earth Environmental Research (ISEE), Nagoya University, Japan.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2019
- Bibcode:
- 2019AGUFMSH34A..02L
- Keywords:
-
- 7519 Flares;
- SOLAR PHYSICS;
- ASTROPHYSICS;
- AND ASTRONOMY;
- 7524 Magnetic fields;
- SOLAR PHYSICS;
- ASTROPHYSICS;
- AND ASTRONOMY;
- 7924 Forecasting;
- SPACE WEATHER;
- 7959 Models;
- SPACE WEATHER