Project Code: TOLKO-MA1-009
Program: FRIP
Sponsor:Tolko Industries Ltd.
Project Status:Complete
This project utilized advanced artificial intelligence (AI) to analyze past and current satellite imagery to determine the decline through optical and spectral changes to develop an Outbreak Severity Map over the Area of Interest (AOI). The use of geospatial Artificial Intelligence to remotely monitor and dynamically manage the forest is a new technology and has been unproven to date. Rezatec (UK Services) Limited was approached for this project as they have had some experience and success in determining overall stand and tree health using satellite imagery.
This analysis used datasets based on both historic and current updates to canopy health status from 2020 and 2021 maps of affected areas and severity releases based on:
  • Deciduous / Non-deciduous layers
  • Health for deciduous only
  • Baseline status
  • Defoliation (based on numbers of defoliations at a stand level)
  • Provided a status on deciduous death
Timing was dictated by the availability of suitable cloud-free images while keeping the focus on timing windows around the known lifecycle and associated seasonal changes. This combined with forest inventory ensured that only stands specific to a particular deciduous pest (e.g., Western Tent Caterpillar) were identified. The assessment was completed remotely, while using high level data inputs to provide
quick, cost-effective information on forest health that was scalable across the AOI. The outcome of the project demonstrated that by using Satellite imagery and comparing year over differences can be an effective method in illustrating where deciduous health is on the decline. From the ground plots collected, the decline was not always attributed to defoliation and other factors were causing the decline in health. Other hardwood pests and diseases, flooding (low lying bogs) along with stand age (>100 years) seemed to be the main contributors to the decline in deciduous health. Defoliators could have been present in 2021, but little to no signs of their presence could be found in 2022.
  • Summary of data acquisition, raster maps, polygon summary, number of ground plots established, total area affected