Original title: The spatial anatomy of urban wildfire vulnerability: a spatially validated GeoAI framework reveals the roles of building density and vegetation moisture in structure loss during the 2025 Palisades Fire
Parastoo Farajpoor, Mohammadreza Narimani · August 2026
When wildfires hit cities, predicting which homes burn is hard. Researchers studied the January 2025 Palisades Fire to learn why some houses survived while others did not. They combined real damage reports with satellite images and maps of buildings, plants, and terrain. Then, they trained computer models to find patterns across neighborhoods. The biggest danger was nearby housing density. When a home had more buildings within 100 m, its odds of being destroyed jumped sharply. Plants also mattered, but in opposing ways. Moist vegetation 100-300 m away shielded homes, while greener plants 30-100 m away actually raised destruction odds once moisture was factored in. Overall, hazards depended heavily on neighborhood conditions 100-300 m around each house, showing how local plant moisture and density drive fire risk.
Imagine you are deciding which houses in a neighborhood to protect from a coming storm. You might think that houses with more trees around them are at higher risk. But the real key is whether those trees are soaked with water or bone-dry, and how close the houses are to each other. Here, the researchers found that the number of nearby buildings was the strongest predictor of fire damage, and that plant moisture, not just greenness, determined protection. So focusing only on vegetation could mislead.
Homeowners and urban planners often assume that more greenery reduces wildfire risk, but this study of the 2025 Palisades Fire shows that vegetation moisture matters far more than greenness alone. Dense building clusters near homes drastically increase destruction odds, suggesting that thinning buildings and maintaining plant moisture could save thousands of structures. The findings directly inform where to focus fire mitigation efforts and how to design fire-resilient neighborhoods.
The abstract reports that spatial block validation drops performance from 0.92 to 0.75, indicating the model overfits to specific neighborhoods. The analysis is based on a single wildfire event (2025 Palisades Fire), so results may not generalize directly to other fires with different conditions. The post-fire track mapped burn severity and vegetation recovery but was not validated for other purposes. The study identifies predictive factors but does not test the effectiveness of the suggested mitigation strategies.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
The spatial anatomy of urban wildfire vulnerability: a spatially validated GeoAI framework reveals the roles of building density and vegetation moisture in structure loss during the 2025 Palisades Fire — Parastoo Farajpoor, Mohammadreza Narimani
Abstract licence: CC0
Explanation generated by Synthesis from the abstract — it is not the original paper. This is a preprint, so it hasn't necessarily been peer reviewed.