Historical insect disturbance maps from 1985 onwards for Canadian forests derived using earth observation data

Area of Science:

• Forestry
• Remote Sensing
• Ecology

Background:

• Insect outbreaks significantly impact Canadian boreal forests and carbon monitoring.
• Current monitoring relies on costly aerial surveys, lacking long-term, consistent data for severe canopy loss.

Purpose of the Study:

• To develop and evaluate annual maps of severe boreal forest insect pest disturbances in Canada from 1985 to 2024.
• To provide a standardized, long-term record of severe pest-related defoliation using satellite imagery.

Main Methods:

• Leveraged Landsat satellite imagery with 30m spatial resolution.
• Developed a methodology for mapping annual severe pest disturbances in coniferous forests.
• Evaluated prediction accuracy for moderate and severe pest vs. non-pest classes.

Main Results:

• Achieved an overall prediction accuracy of 90% for aggregated moderate and severe pest classes.
• Created a historical dataset of annual severe disturbances from 1985 to 2024.
• Demonstrated the utility of Landsat data for consistent disturbance mapping.

Conclusions:

• The developed methodology provides a valuable, standardized dataset for monitoring severe insect pest disturbances.
• This historical record offers insights for forest ecology, carbon modeling, and forest management.
• Satellite remote sensing offers a cost-effective alternative to aerial surveys for long-term disturbance monitoring.