Fighting Forest Pests With AI: A Hemlock Success Story

Fairfax County, Virginia, faced a significant challenge with the hemlock woolly adelgid, a pest threatening its hemlock trees and evergreen canopy. The county’s Urban Forest Management Division (UFMD), tasked with preserving the urban forest, recognized that mapping hemlocks was critical to combat infestations. Collaborating with the GIS Division, they leveraged geospatial artificial intelligence (GeoAI) using ArcGIS technology to analyze aerial photography. By employing a deep learning model and additional filtering techniques—like analyzing greenness, infrared data, and lidar intensity—UFMD pinpointed evergreen trees requiring treatment, significantly narrowing the target area from 116,521 to 33,110 trees.

The GIS Division’s analysis enabled the Forest Pest Branch to efficiently develop a treatment plan and allocate resources. Utilizing ArcGIS Field Maps, the mobile team located and treated infested trees. This streamlined approach, integrating GIS tools with ground efforts, ensured precise targeting and resource optimization. The collaboration highlighted the effectiveness of automating tree detection, laying the groundwork for future projects aimed at preserving Fairfax County’s tree canopy.

Building on this success, Fairfax County used the evergreen data from the hemlock project to train a new deep learning model to map evergreens across the Mason Neck peninsula. This project covered nearly 9,000 acres and identified over 36,000 evergreen trees. The automated process enabled more efficient planning and monitoring, particularly in areas with dense tree clusters. Fairfax County’s innovative use of GeoAI for forest management demonstrates the transformative potential of technology in addressing environmental challenges.