Proactive Forest Management Decision Support through Algorithm Developed using Remote-Estimated Biomass with GEDI LiDAR and Evapotranspiration with ECOSTRESS Correlation Analysis

Abstract Submission: Forest biomass, crucial to carbon cycle analysis is linked to evapotranspiration (ET), a key water-use model-factor. LiDAR technology provides precise measurements of vegetation height, a foundation for biomass estimation and newer satellites, ECOSTRESS supports forest ET estimation. The goal of this study is to leverage high-resolution LiDAR data and ultra-high-resolution orthoimagery of forest land cover to estimate above-ground biomass and correlate it with remotely sensed spatial ET to develop an algorithm to remotely estimate Loblolly Pine (LLP) ET for forest management decision support. The study's sub-objectives included: i) using ECOSTRESS data to correlate with in-situ Flux Tower-based ET data to develop spatial ET rasters with broader coverage and ii) develop SWAT hydrologic model to validate remote-estimated Biomass and ET. The data for this study, including NAIP, ECOSTRESS, and GEDI datasets (in GeoTIFF format), were sourced from the USDA NRCS Data Gateway, NASA's AppEEARS, and ORNL DAAC, respectively. The research focused on five Loblolly Pine (LLP) forested sites across Virginia, North Carolina, South Carolina, and Georgia, where eddy flux towers are installed. Cloud-free ECOSTRESS ET data (70m resolution) from 1999 to 2023 were downloaded to align with GEDI biomass and tree height data. NAIP data from the same period were classified using a Fuzzy ARTMap ANN image segmentation algorithm to determine canopy cover. An existing allometric equation was used to estimate biomass based on LLP plant height and canopy area. These biomass estimates were then compared with GEDI-derived biomass data for validation. The ECOSTRESS L3 band ET values were correlated with GEDI-based biomass estimates to develop the ET-Biomass algorithm. QSWAT models were developed to simulate biomass and ET values for 2019 and 2023 to validate with remote estimated study results. Our approach for remotely estimating biomass, ET, and related parameters could be a valuable tool for global forest management and decision-making.

Learning Objectives/Expected Outcome (Optional) : - GeoAI supported geospatial model automation
- Advanced Image Processing approach for GEDI/ECOSTRESS data analyses
- Spatial Correlation Algorithm development
- A software to remotely estimate Forest (lob lobby Pine) ET with new-age satellites