Models for Short-Term Forecast of Maximum X-ray Class of Solar Flares Based on Magnetic Energy of Active Regions

The accuracy of the M. Aschwanden’s (2020) model for short-term (24 h) prediction of the maximum X-ray class of solar flares based on the power-law dependence on the energy of the potential magnetic field of active regions is checked and assessed. For this purpose, a sample of 275 flares (253 M-class and 22 X-class) in isolated active regions on the solar disk in 2010−2023 is analyzed. Magnetic field extrapolations are made in the nonlinear force-free and potential approximations using the GX Simulator based on photospheric vector magnetograms from the Helioseismic and Magnetic Imager (HMI) instrument onboard the Solar Dynamics Observatory (SDO). It is found that in 6% of cases, Aschwanden’s model underestimates the predicted maximum flare class with respect to the observed one (maximal underestimation by 4.4 times). The accuracy of this model (the average ratio of the observed to predicted maximum flare class) is 0.31 ± 0.47. Four other statistical models are proposed, two of which, like Aschwanden’s model, are based on the power-law dependence of the maximum flare class on the energy of potential magnetic field, and the other two are based on the power-law dependence on the free magnetic energy of active regions. These models give fewer (or no) underestimations of the maximum flare class, but two to three times lower forecast accuracy, ranging from 0.11 to 0.17. Additionally, based on the obtained statistical sample, estimates of the limiting X-ray class of solar flares are made. The five models give different limits ranging from ~X14 to ~X250. The realism of these values and the possibility of refining the models by expanding the sample of events is briefly discussed.