A Data-Driven Fire Spread Simulator: Validation in Vall-llobrega's Fire

Abstract

While full-physics fire models continue to be unsuitable for wildfire emergency situations, the so-called operational fire spread simulators are incapable of providing accurate estimations of the macroscopic fire behavior while quickly reacting to a change of governing spread mechanisms. A promising approach to overcome these limitations are data-driven simulators, which assimilate observed data with the aim of improving their forecast with affordable computation times. Although preliminary results obtained by several data-driven simulators are promising, this scheme needs intensive validation. Detailed studies of the particular aspects related to data assimilation are essential to gain insight about the applicability of this approach to operational wildfire simulation. This paper presents the validation of the simulator presented in Rios et al. (2014b, 2016, 2018) with a large scenario of real complexity with intricate terrain. The study case corresponds to a wildfire of significant repercussions occurred in Catalonia in March 2014. We employed as reference data the event reconstruction performed by the Catalan Fire Service and validated with operational observations. Detailed information about fuel and meteorology was collected by the fire brigades and allowed reconstructing the fire development with Farsite, a widely employed simulator. Subsequently, our simulator was tested without a detailed description of the fuel and wind parameters, i.e., imitating its intended deployment conditions. It proved capable of automatically estimating them and correctly simulating the fire spread. Additionally, the effect of the assimilation window on the forecast accuracy was analyzed. These results showed that the simulator is able to correctly handle complex terrain and wind situations to successfully deliver a short-term fire-front forecast in those real and complex scenarios.

Thermal Infrared Video Stabilization for Aerial Monitoring of Active Wildfires

“…This information has subsequently been used for additional fire behaviour analysis and evaluation of suppression activities [10]- [12]. Furthermore, there is a growing trend to incorporate observed fire perimeter evolution into operational fire spread simulators in order to improve forecasts using data assimilation techniques [13]- [17].…”

Data-driven surrogate model with latent data assimilation: Application to wildfire forecasting

“…System parameters are often obtained using calibrations of local fire scenarios [14,15]. Furthermore, running physics-based simulations for large-scale wildfires can be computationally expensive, even when coupled with discrete event modelling (e.g., Discrete EVent System [44], Cellular Automata [56]) and parallelization computing, leading to difficulties for real-time fire progression monitoring [64]. Much effort has been devoted to improving the prediction accuracy and efficiency via data-driven surrogate models, including the use of machine learning (ML) [49] and data assimilation (DA) techniques [52].…”

Evaluating the Performance of Fire Rate of Spread Models in Northern-European Calluna vulgaris Heathlands

“…In light of this alternative, more broadly applicable alternative models would be valuable. Fully physical-based approaches are often costly and require powerful computational hardware [47]. These methods, however, are becoming more viable as computational resources become more advanced and can provide intricate analyses of the effects of characteristics, such as fuel heterogeneity on fire behaviour [48].…”