Classification and mapping of European fuels using a hierarchical, multipurpose fuel classification system

Abstract

Abstract. Accurate and spatially explicit information on forest fuels becomes essential to designing an integrated fire risk management strategy, as fuel characteristics are critical for fire danger estimation, fire propagation, and emissions modelling, among other aspects. This paper proposes a new European fuel classification system that can be used for different spatial scales and purposes (propagation, behaviour, and emissions). The proposed classification system is hierarchical and encompasses a total of 85 fuel types, grouped into six main fuel categories (forest, shrubland, grassland, cropland, wet and peat/semi-peat land, and urban), plus a nonfuel category. For the forest cover, fuel types include two vertical strata, overstorey and understorey, to account for both surface and canopy fuels. In addition, this paper presents the methods to map fuel types at the European scale, including the first-level of the classification system. Land cover, biogeographic datasets, and bioclimatic modelling were used. The final map, which is publicly available (https://doi.org/10.21950/YABYCN; Aragoneses et al., 2022a), included 20 fuel categories at 1 km spatial resolution. A first assessment of this map was performed using field information obtained from LUCAS (Land Use and Coverage Area frame Survey), complemented with high-resolution data. This validation exercise provided an overall accuracy of 88 % for the main fuel types and 81 % for all mapped fuel types. To facilitate the use of this fuel dataset in fire behaviour modelling, an assignment of fuel parameters to each fuel type was performed by developing a crosswalk to the standard fuel models defined by Scott and Burgan (2005; FBFMs – Fire Behavior Fuel Models), considering European climate diversity. This work has been developed within the framework of the FirEUrisk project, which aims to create a European-integrated strategy for fire danger assessment, reduction, and adaptation.

A method to produce a flexible and customized fuel models dataset

“…There are limitations to produce a wide European common fuel model system since fuel models are site specific and should be applied in the regions for which they were developed to produce more realistic fuel mapping [47] . More recently, a hierarchical fuel classification system consisting of 20 fuel types (both surface and canopy) was developed at the European level, based on land cover, biogeographic datasets and bioclimate modelling [48] . However, one limitations of this fuel map is its spatial resolution of 1-km, which may undermine its accuracy and the quality of fire spread and behavior simulations.…”

Towards an Integrated Approach to Wildfire Risk Assessment: When, Where, What and How May the Landscapes Burn

“…To consider the great variety of European fuel conditions, first a hierarchical fuel classification system was developed, including surface and canopy fuels, which could be used for both target scales, ET and PS [343]. This FirEUrisk classification system includes 80 fuel categories, structured into the following seven main fuel types: (a) Forests (areas with canopy cover ≥ 15% and canopy height ≥ 2 m); (b) Shrublands (areas with shrubs, scrub, garrigue, and maquis); (c) Grasslands (areas with herbaceous non-cultivated vegetation); (d) Cropland: (areas with cultivated vegetation); (e) Wet and peat/semi-peat land (which in turn comprises areas with a permanent mixture of vegetation and water, including marshes; moorland/heathland; peatlands and peat bog, and moss and lichens; (f) Urban (areas with ≥15% of built-up structures and/or buildings); (g) Nonfuel: (areas with permanent water bodies, open sea, snow, ice, bare soil, sparse vegetation (<10% of terrain cover).…”

“…Urban fuels were split into continuous and discontinuous fabric. This classification system was used to generate all fuel type datasets for the FirEUrisk project, including the whole ET [343] (Figure 7 left) and PS (Figure 7b right). The former were obtained from existing land cover and biophysical models, while the latter were generated from high resolution satellite image classification [344], with resolutions ranging from 100 to 20 m. Both ET and PS fuel type maps were associated to existing fuel models: a preliminary assignment for the ET area was done using the Fire Behavior Fuel Model (FBFM) standard models [88], where for the PS the fuel models were estimated from field data or biophysical regional estimations.…”

“…Outputs of Rothermel equations were: Rate of Spread (m/s), Reaction intensity (kW/m 2 ), Fireline Intensity (kW/m) (see Figure 9) and Flame Length (m). These variables were computed for the ET area using the two reference weather scenarios, with the DFMC modelled for those conditions and the fuel type dataset described in Section 4.3.3, using the adaptation to the Scott and Burgan fuel models [88], done by Aragoneses et al [343]. Elevation data was downloaded and rescaled from NASA Shuttle Radar Topography Mission data (https://lpdaac.usgs.gov/products/srtmgl1v003/, accessed on 30 March 2023).…”

“…The resulting dataset is the first long term Europespecific land use scenario set that is aligned with a historic reconstruction, improving the ability of the DGVMs to forecast future fire regimes [155] at a fine spatial resolution. While the final time series is based on a 1-km 2 resolution, the results were converted to a 9-km 2 resolution to match the resolution that is used natively in the DGVMs and harmonized the land cover class legend to the fuel map generated within the project [343]. Data are available at DataverseNL (https://dataverse.nl/dataset.xhtml?persistentId=doi:10.34894/ALZTYS, accessed on 20 April 2023).…”