Estimating canopy fuels across Europe with satellite data and allometric equations
Keywords: remote sensing, geospatial data, forest structure analysis, vegetation monitoring, fire modelling, wildfire risk assessment
Abstract. This study presents results regarding the estimation of two critical variables for modelling fire behaviour and fire danger: the canopy base height (CBH) and the canopy bulk density (CBD). Both variables have been mapped as raster datasets at a 100-meter spatial resolution across Europe, harmonizing data for all EU countries. Therefore, these canopy fuels are subsequently used for further processing regarding the identification of fire danger assessment, being a key input for forest fire prevention actions. A more in-depth analysis of these findings has been submitted to a journal and is currently in a revision phase. We present here a summary of the results and ideas for future developments. The overall study consists of estimating CBH and CBD using Earth observation products combined with artificial intelligence and species-specific allometric equations, applied geo-spatially using a tree species map of Europe encompassing the 16 most important tree species. Validation was carried out by comparing the results with higher-accuracy sampling methods, combining LiDAR data and field measurements in different European latitudes, typically applied on a smaller scale and with greater detail. Results show, as expected, a higher level of uncertainty than local methods, but they are still applicable to the European scale for which they were implemented. The accuracies reported in our study, when considering aggregated data on the 7 areas in Portugal were the following: R = 0.75, RMSE = 0.890 m, and MAPE = 54% for the mean CBH, and R = 0.93, RMSE = 0.020 kg m−3, and MAPE = 57% for the mean CBD.