Next Day Wildfire Spread: A Machine Learning Dataset to Predict Wildfire Spreading From Remote-Sensing Data

Predicting wildfire spread is critical for land management and disaster preparedness. To this end, we present `Next Day Wildfire Spread,' a curated, large-scale, multivariate data set of historical wildfires aggregating nearly decade remote-sensing across the United States. In contrast to existing fire sets based on Earth observation satellites, our combines 2D with multiple explanatory variables (e.g., topography, vegetation, weather, drought index, population density) aligned over regions, providing feature-rich machine learning. demonstrate usefulness set, implement neural network that takes advantage spatial information predict spread. We compare performance other learning models: logistic regression random forest. This can be used as benchmark developing propagation models remote sensing lead time one day.