NASA Logo in the header
Data Science
Operational

Mapping vegetation height in Alaska with an Earth Observation Foundation Model (EOFM) and commercial spaceborne imagery

We demonstrated the ability of frontier Earth Observing Foundation Models to detect subtle vegetation structure changes in Alaska.

Organization
Launch Date
--
Class
Data/Image Computer Models
Website
--
Key Staffs
  • Program Lead
  • Software Developer
  • Software Developer
  • Principal Investigator
  • Project Scientist

Although the far northern latitudes are changing faster than anywhere on earth, the geographic vastness, sparse vegetation, and subtle but meaningful vegetation structure changes make this area notoriously difficult to measure and model. Many earth-observing (EO) computer vision vegetation height models only predict to the nearest 1 meter, masking the subtle (~0.1 m) but important vegetation height changes typical in the boreal-tundra gradient. We experimented with a frontier open source EO Foundation model trained on 493 million high resolution commercial satellite images to see if fine-tuning could address this gap in resolving fine-scaled vegetation structures. We mapped, validated, and quantified upper canopy surface vegetation heights and estimated the area of height classes across 40% of Alaska’s boreal-tundra ecotone, capturing the proportions of various low and tall shrub classes to provide a basis for assessing the variation in the routes and rates of Arctic and sub-Arctic vegetation change.