Abstract
Solar flares are the most energetic manifestations of solar activity and can significantly affect Earth’s magnetosphere, ionosphere, and technological systems. Therefore, anticipating these events remains a fundamental challenge in heliophysics. Soft X-ray observations from the GOES satellites have long been used to monitor and characterize solar flares. Among the various forecasting approaches, the Flare Anticipation Index (FAI), originally proposed by Hudson (2025), has established itself as a promising diagnostic tool for thermal activity preceding flares. The FAI is based on the detection of a Hot Onset Precursor Event (HOPE), characterized by a gradual increase in plasma temperature, and the emission measure before the impulsive phase of a flare. In this work, we performed a statistical validation of the FAI using a representative dataset of approximately 8,000 days between 1980 and 2025. Plasma temperatures and emission measurements were derived from GOES/XRS observations, and FAI-based alerts were generated using predefined thresholds correlated with GOES solar flares within a 30-minute time interval from flare start to peak. A total of 48,344 flares of different classes were analyzed, yielding varying detection rates depending on the flare class. The parameter sets were chosen to minimize the number of false positives and increase the detection rate of large flares (M and X). The results suggest that the FAI is particularly sensitive to medium and large solar flares and has potential for near real-time prediction, while also highlighting the need to optimize the thresholds to improve its predictive performance for each class.