Tree Condition and Analysis Program – Detecting Forest Disturbance at the Tree Level across the Contiguous United States with High Resolution Imagery

Abstract

Abstract Effective management of forest insects and diseases requires detection of abnormal mortality, particularly among a single species, sufficiently early to enable effective management. Remote detection of individual trees crowns requires a spatial resolution not available from satellites such as Landsat or Sentinel-2. In the United States, there are currently few operational systems capable of effectively and affordably detecting and mapping tree mortality over broad landscapes using high-resolution imagery. Here, we introduce the Tree Condition and Analysis Program (TreeCAP), an open-source system that uses freely available imagery from the National Agriculture Imagery Program (NAIP) to create maps of tree condition (healthy or damaged). We demonstrate the potential applications of TreeCAP in four study sites: (1) beetle-killed pines in California, (2) emerald ash borer progression in Wisconsin, (3) hemlock wooly adelgid mortality in Pennsylvania, and (4) drought damage in Texas. We achieved an average overall accuracy of 87% across all study sites.