Machine Learning (ML) of Forest Stand Metrics to Quantify Carbon Storage

Description

This research project focuses on quantifying basic forest stand metrics through the application of ML to remotely sensed data. The project will leverage global data to develop understanding of forest growth and successional conditions at a local level. Numerous environmental variables and forest inventory data must be incorporated to train ML algorithms on high performance computing systems (HPCs) to achieve resolutions that lead to understanding of carbon stores at a local level (e.g., a single DOD installation). Knowing that understanding dominant forest habitat type and forest volume (as calculated from tree height, diameter, and density) will yield significant understanding to forest carbon storage, the purpose of this work is to demonstrate that basic forest inventory metrics (e.g., tree diameter and density) may be effectively quantified from ML. The Government is not expecting the periods of performances to overlap. Objectives: The objectives of the project for the initial year are as follows:1. Develop technical team and identify initial study area(s) of interest.2. Develop and test a proof of concept outlining novel methods to quantify basic forest stand metrics.3. Compile a repository of forest inventory data from national and international partners. 4. Validate accuracy of resulting, prototype forest stand metrics. The objectives of the project for Optional Year 1 are as follows:1. Expand the study area(s) and refine the prototype novel methods (developed during initial year) to quantify basic forest stand metrics.2. If required, expand the repository of forest inventory data from national and international partners to cover the second year’s study area.3. Validate accuracy of resulting, large area forest stand metrics by prioritized areas of interest. 4. Generate peer-reviewed journal article with ERDC researchers to describe the application of novel methodologies to quantify basic forest stand metrics developed during initial year of the project. The objectives of the project for Optional Year 2 are as follows:1. Conduct a final accuracy assessment and if required, refine the established methods to increase basic forest stand metric accuracy.2. Generate a peer-reviewed journal article(s) in conjunction with ERDC researchers integrating all study conclusions.3. Develop and present public seminars based on study findings. Successful applicants should have expert knowledge of: 1) forestry, natural resources, and carbon storage; 2) field data collection capabilities; 3) compiling national and global forest inventory databases; 4) experience developing novel approaches to machine learning of forest characteristics. Areas of expertise that may be required in combination to perform this study include:1) Capacity to collect and/or compile forest inventory data at up to global scales.2) Advanced computing capabilities for ML applications to characterize forest metrics.3) Development of novel ML approaches to improve forest inventory, forest characterization, and/or forest carbon storage research with local and global applications.

Eligibility

Grants.gov lists this opportunity under eligibility category code 25. These codes correspond to applicant types (state/local government, tribal organization, nonprofit, educational institution, individual, small business, etc.) defined in Grants.gov's own eligibility reference. See the current Grants.gov eligibility categories or check the official listing below for this opportunity's exact eligibility statement.

Agency Contact

Key Dates