Leveraging a Wildfire Risk Prediction Metric with Spatial Clustering

Fire authorities have started widely using operational fire simulations for effective wildfire management. The aggregation of the simulation outputs on a massive scale creates an opportunity to apply evolving data-driven approach closely estimate risks even without running computationally expensive simulations. In one our previous works, we validated application with probability-based risk metric that gives series probability values starting at start location under given weather condition. indicate how likely it is will fall into different categories. considered each as unique entity. Such provision in could expose scalability issues when used larger geographic area and consequently make hugely intensive compute. this work, investigative effort, investigate whether spatial clustering locations based historical areas can address issue significantly compromising accuracy metric. Our results show spatially all Tasmania three clusters leverage by reducing computational requirements theoretical factor thousands mere compromise approximately 5% two categories high low, thereby validating possibility clustering.