Using Deep Learning for Spatial and Temporal Analysis of Wildfire Start and Progression

Wildfires pose a significant threat to ecosystems, public safety, and property, and understanding the spatial and temporal patterns of their occurrence is of paramount importance for mitigating their devastating impacts. In this study, we address the critical task of wildfire start and progress prediction by leveraging cutting-edge deep learning techniques. We obtain data from remote-sensing and ground based input sources to analyze patterns for next 24 hours using past 24 hours of data. Real time wildfire prediction requires a multi dimensional model, that is able to learn and retain information from our data set. Our approach employs a novel deep learning architecture, which captures both spatial and temporal dependencies in wildfire data, grants us results that provide substantial breakthroughs in using ML to predict wildfire start and progress.