Forest Fire Propagation Prediction Based on Overlapping DDDAS Forecasts

Abstract

Forest fire devastate every year thousand of hectares of forest around the world. Fire behavior prediction is a useful tool to aid coordination and management of human and mitigation resources when fighting against these kind of hazards. Any fire spread forecast system requires to be fitted with different kind of data with a high degree of uncertainty, such as for example, me- teorological data and vegetation map among others. The dynamics of this kind of phenomena requires to develop a forecast system with the ability to adapt to changing conditions. In this work two different fire spread forecast systems based on the Dynamic Data Driven Application paradigm are applied and an alternative approach based on the combination of both predictions is presented. This new method uses the computational power provided by high performance computing systems to deliver the predictions under strict real time constraints.