Standardizing full-map evaluation method

Background: Evaluating wildfire forecasting models should ideally mimic operational deployment through full-map inference rather than relying solely on curated test datasets.

Question / Future Work: Future work should continue to establish full-map-based inference as the primary methodology for evaluating data-driven Fire Danger Index models, focusing on both fire detection accuracy and the rate of false alarms.

Why It Matters: Establishing a clear standard for operational evaluation methods is essential for translating research advances into reliable real-world decision support systems.

Evidence: Future work in this direction should focus on full-map-based evaluation as the primary method of model evaluation for both fire detection and false alarm rates.