Application of machine-learning methods for estimating growth of boreal forest in the Mackenzie River Valley

Abstract:

Innovation Lab at the Computational and Information Sciences and Technology Office (CISTO, NASA GSFC) is dedicated to provide technical solutions to accelerate the research programs by leveraging Hight-End Computing and ML/AI technologies.

We will present an example of the collaboration between Innovation Lab and domain experts. The application is building a ML model based on sub-meter MAXAR stereo remote sensing data, Landsat stand age and environmental covariates, in order to establish mapped estimates of growth potential for remote northern forests in the Mackenzie River Valley of the Northwest Territories, and to understand local topographic and climatic effects on canopy structure and growth.

The presentation will concentrate on technique aspects of the application, including data preparation, model design and evaluation, comparison among ML algorithms, and deployment strategy.