Chair: Greg Ettl
Committee Members: David Beck, Beth Gardner, Eric Turnblom, Andrew Gray
Title: Mining Public Data to Enhance Forest Assessment, Monitoring, and Modeling
Abstract: The rapidly-increasing availability of data from remote sensing and field-based tree and forest inventory observations is offering unprecedented opportunities for new information about forests to be gleaned and learned from. Three new interventions for data acquisition, processing, and predictive modeling in Pacific Northwest forests will be presented: (1) probabilistic co-registration of point clouds with forest inventory plots; (2) automated delineation and labeling of forest conditions; and (3) assimilating inventory and tree ring data to bring climatic effects to the Forest Vegetation Simulator (FVS) growth-and-yield model.
These research efforts employ and integrate diverse methods ranging from three-dimensional geometric modeling and visualization, cloud-based processing of satellite imagery and lidar point clouds, hierarchical Bayesian modeling, and the application of computer vision and machine learning algorithms. These activities are pursued to develop, document, and demonstrate reproducible open-source tools and workflows for using publicly-available data to assess forest conditions, monitor how they change over time, and generate stochastic growth-and-yield projections that reflect potential climate impacts and forest management options at operational management scales. To maximize potential transferability and utility, new growth-and-yield approaches involve the targeted development of new features for a model (FVS) employed at every state and federal agency that manages forestland in the Pacific Northwest.