Wildfires pose a major environmental and societal challenge, due to their link with anthropogenic activities and changing climatic conditions. This study aimed to enhance our understanding of the drivers of wildfire occurrence across continental Chile by developing robust predictive models incorporating climatic, land cover, and anthropogenic variables. We leveraged geospatial data on historical fire events, infrastructure, fuels and weather, coupled with historical fire records through Random Forest binary models to ascertain the key drivers of ignition across four distinct ecological zones: North, Central Chile, South, and the Andes. Our analysis explored potential differences between arson and unintended fires within these regions. Model validation, assessed using the Area Under the Curve (AUC), revealed significant regional variations in predictive performance. The southern and northern zones exhibited higher predictive capacity, potentially due to less complex landscapes and fewer ignition sources compared to the densely populated and infrastructure- prone central zone, which showed the lowest AUC. The Andes region displayed intermediate performance. Our results indicated that anthropogenic factors, particularly the distance to power lines, roads, and the wildland-urban interface (WUI), were consistently among the most important predictors of wildfire ignition across the majority of the studied regions. This highlights the significant impact of human accessibility and infrastructure on fire incidence in Chile. In contrast, fuel-related and climatic variables, such as Dry Fuel Moisture Content (DFMC) and its anomaly, showed generally lower importance, although their influence increased notably in the southern zone. Partial dependence plots further elucidated the distinct ways in which these key variables influenced ignition probability across different regions and between arson and unintended fires. The findings emphasize the necessity of adopting region-specific approaches in wildfire modeling and prevention strategies, acknowledging the different interactions between natural and anthropogenic factors across Chile. This research provides a fundamental understanding for future advanced modeling and targeted risk management efforts. Future research should aim to incorporate more detailed socioeconomic data to further refine predictive models and inform effective risk mitigation strategies.

Este estudio desarrolla un marco reproducible para cartografiar la Interfaz Urbano-Forestal (WUI) a escala local usando datos globales abiertos, aplicado en Chile. La metodología integra parámetros espacializados sobre la presencia de edificaciones y la altura, cobertura y continuidad de la masa de combustible para clasificar el territorio en WUI-Interface y WUI-Intermix. La solidez se evalúa mediante un análisis de sensibilidad con 12 combinaciones de los parámetros, del cual se deriva un índice de intensidad y acuerdo de clasificación. Los resultados muestran que el 6,5% de las viviendas en Chile (890.884 edificaciones) se ubican en la WUI. Destacan los sectores de La Araucanía, Los Lagos y Bío-Bío por la exposición asociada a continuidad de combustibles, mientras que en la Región Metropolitana predomina un patrón vinculado mayormente a la expansión urbana.

Fires in Chile are one of the main natural threats to society and the balance of ecosystems. As is common, the cause behind most of them is human action, and their occurrence is linked to numerous climatic, environmental, or accessibility factors. The aim of this study was to predict the daily probability of ignition for all continental Chile, with a spatial resolution of 100 meter, distinguishing between intentional and unintentional fires to study potential disparities in the role of fire drivers. To achieve this, a Random Forest machine learning model was trained and tested using CONAF's ignition data from 2009 to 2019.A total of 100 model realizations were calibrated by combining randomly stratified samples of fire ignition with spatial variables related to accessibility, anthropogenic presence, infrastructure, and dead fuel moisture content. For each realization, we trained and evaluated a binary classification Random Forest model, aggregating their outcomes and predictions to account for uncertainty. Models were also evaluated in terms of prediction ability and residuals independence. The results show differences between intentional and unintentional fires in terms of accuracy (0.86 and 0.83, respectively), but also regarding the role of the drivers. Notable differences in variable importance were also observed, with distance to power lines being the most important variable for intentional fires, while the Wildland-Urban Interface (WUI) played a larger role for unintentional fires. While the importance of WUI had been identified in previous studies, the significance of distance to power lines had not been widely considered, despite its potential impact on the accessibility of remote areas with high fuel loads. Interestingly, dead fuel moisture (DFMC) and fuel types were less important in both models, with DFMC showing surprisingly low relevance, contrary to expectations. The ignition probability maps generated displayed similar small-scale spatial patterns, with high ignition probabilities concentrated in central Chile, where most studies have been conducted. The southern and northern regions showed either negligible or low ignition probabilities, mainly due to a lack of fuel. At a local scale, intentional fire models were clearly associated with power lines and road networks, while unintentional fires were more influenced by proximity to buildings. Areas farther from human activity centers showed higher probabilities for unintentional fires, likely linked to recreational activities.