Improving landscape fire frequency estimates by integrating public land fire data and satellite imagery

This is a Preprint and has not been peer reviewed. This is version 2 of this Preprint. You must log in to post a comment. There are no comments or no comments have been made public for this article. This is a Preprint and has not been peer reviewed. This is version 2 of this Preprint. Add a Comment You must log in to post a comment. Comments There are no comments or no comments have been made public for this article. Background Contemporary fire regimes are changing rapidly, and effective fire management requires knowledge of fire history, often derived from satellite imagery. Satellites, however, are not well suited to detecting low intensity fires, meaning fire history data are often inaccurate. Aims We aimed to improve satellite fire frequency estimates by incorporating data from fire history on public land and environmental co-variation. Methods Using a generalisable workflow, we applied boosted regression trees, generalised linear, and generalised additive models to predict fire frequency in an eastern Australia case study. Key results Relative to unprocessed satellite data, generalised linear and generalised additive models improved correlations with public land fire data by 0.39 and 0.25, respectively. Generalised linear models estimated low fire frequencies well (≤2 fires), whereas generalised additive models predicted higher fire frequencies (≥3 fires) more accurately. Conclusions For mapping land as burnt or unburnt, generalised linear models are most appropriate. For understanding the total number of fires over time, and for most vegetation types, predictions from generalised additive model are most appropriate. Implications Our approach can improve the accuracy of fire frequency estimates from satellite data, to assist fire management and conservation. However, model selection will depend on the application and vegetation type. https://doi.org/10.32942/X24331 Ecology and Evolutionary Biology fire management, fire scar mapping, Landsat, predicitive modelling, satellite fire data, Sentinel, species distribution modelling, remote sensing Published: 2025-04-23 18:14 Last Updated: 2026-01-23 18:55 CC-BY Attribution-NonCommercial 4.0 International Data and Code Availability Statement: Open data/code has been made available for peer-review only as an archived Zenodo repository (Charles and Smith 2025) https://doi.org/10.5281/zenodo.15133643. Language: English