A Digital Twin for Wildfire risk adaptation planning: DT-WILDFIRE

Wildfires pose a growing threat to populated areas of the Mediterranean basin. Rural abandonment has increased fuel loads, creating appropriate conditions for large wildfires. The hot and dry conditions caused by climate change have exacerbated the risk, extent, and severity of wildfires. The rising number of homes in the wildland-urban interface (WUI) implies increasing impacts on lives and property from wildfires. The need for mitigation and adaptation measures against wildfire risk is thus becoming more urgent. The Barcelona Metropolitan Area, a large metropolis with an extended WUI (with more than 20000 inhabitants), is particularly vulnerable. A part of its population and infrastructure is located near the border of the Collserola Natural Park (8000 hectares with 6 million visitors yearly), an extended and concurred forested area, and could be potentially threatened by large forest fires, becoming at the same time a threat for the whole metropolitan area. 

This study presents a Digital Twin (DT) framework for the Barcelona Metropolitan Area, designed to assess the risk of extreme wildfires, and how it is impacted by heatwaves and droughts under different future emission scenarios. The DT-WILDFIRE leverages high-resolution climate model projections, satellite data, local observations, and advanced machine learning (ML) techniques to provide a granular understanding of future climate risks and their cascading impacts on wildfires. 

To quantify the fire risk, we calculate the Fire Weather Index (FWI), a widely recognized metric used to assess the potential for wildfire occurrence and spread, based on prevailing meteorological conditions. We calculate FWI over Catalonia at a resolution of 1.5 km during the historical period, using the EMO1 database. Validation against ERA5Land-derived FWI shows good agreement. This high-resolution FWI will then be used to downscale future FWI projections from climate models, thereby providing greater spatial detail in analyses of future climate change impacts on wildfires in the region. 

Further assessment of wildfire risk is provided by the wildfire susceptibility prediction model, based on the machine learning algorithm XGBoost.  The model is implemented over Catalonia and trained using diverse variables, including population density, electrical power infrastructure, terrain elevation, Normalized Difference Vegetation Index, land cover classifications, FWI, and historical burned area data. The model generates daily wildfire susceptibility maps at regional scale. Model evaluation based in the quadratic weighted Kappa metric indicates moderate to good predictive skill over most of the domain, except in high-elevation areas. Further detailed investigation in these regions is ongoing.  

Future climate risk related to wildfire drivers, such as droughts and heatwaves, is also assessed. To achieve the required resolution, we apply deep learning downscaling methodologies to produce future climate projections at very high resolution (0.8km). 

Finally, the DT aims at quantifying physical damage to residential and commercial real estate, including damage from smoke and business interruption. Ultimately, DT-Wildfire aims at helping authorities and society design participatory risk reduction measures, including nature-based solutions, according to the different climate scenarios.