Abstract

This research examines the neighborhood-level relationship between urban tree canopy coverage and air conditioning (A/C) usage in San Pedro Sula, Honduras, a rapidly growing city with high heat exposure. A computer vision model was retrained on local aerial imagery to segment tree canopies, which were combined with building footprints from OpenStreetMap (OSM) to calculate a neighborhood-level tree-to-building ratio. These data were integrated with census information on A/C usage and socioeconomic indicators (e.g., households with no unsatisfied basic needs, roof type). Results show that while overall tree coverage in the city is about 45%, canopy distribution is highly uneven. A/C usage shows a similar disparity, with 26.1% of households reporting at least one unit. Spatial regression analyses (e.g., Spatial Durbin Model, Leroux binomial model) reveal a strong inverse relationship between the tree-to-building ratio and A/C usage: doubling or tripling the ratio is associated with a 21–45% reduction in A/C usage. Also, the tree-to-building ratio proved a more reliable indicator of cooling benefits than overall tree coverage, which tends to align with wealthier areas that still display high A/C usage. These findings highlight the need for targeted greening in heat-exposed neighborhoods, supporting nature-based solutions for climate resilience in low-income tropical cities.