Projecting Fire Intensity: data and code

This repository contains code, data, and output from the paper: "Wildfires will intensify in the wildland-urban interface under near-term warming" by Calum X. Cunningham, John T. Abatzoglou, Todd M. Ellis, Grant J. Williamson, and David M.J.S. Bowman.The repository contains the following folders:1. Code"00_Setup.R" links MODIS hotspots with explanatory variables."01_Analysis.R" conducts the main analysis. Code can be run starting at Section 2 by loading pre-processed data (see below).2. Data"Data_for_model" contains .rds files containing the preprocessed sample of 8.9 million MODIS hotspots. These data can be loaded at Section 2 of the main analysis."Spatial Variables" contains spatial layers of explanatory variables, as described in Table 1 of the paper.3. ModelsThe best-performing generalised additive model for each biome andModel evaluation metrics for each biome.4. OuputContains a geotiff with 3 layers: (i) projected fire radiative power under on FWI93 (2000-2023),(ii) the difference in projected FRP with 1.5°C warming above the preindustrial baseline, and (iii) the difference in projected FRP with 2°C warming above the preindustrial baseline.