Landscape Optimization for Prescribed Burns in Wildfire Mitigation Planning

Abstract

Wildfires have increased in extent and severity, and are posing a growing threat to people’s well-being and the environment. Prescribed burns (burning on purpose parts of the landscape) are one of the key mitigation strategies available to reduce the potential damage of wildfires. However, where to conduct prescribed burns has long been a problem for domain experts. With the advancement of forest science, weather science, and computational modeling, there produced powerful fire simulators that can help inform how wildfires will start and grow. In this paper, we model the problem of selecting where to perform a set of prescribed burns across a large landscape into a multi-objective optimization problem. We build a surrogate objective function from simulation data and solve the multi-objective optimization problem with genetic algorithms. We name our solution as Spatial Multi-Objective for Prescribed Burn (SMO-PB). We also investigate three variants of the approach that further consider spatial fairness. With a case study of Dogrib, Canada, we show that our formulations can successfully provide solutions capable of real world deployment, and showed how fairness can be reached without diminishing the performance a lot.