Download (PDF 2.36 MB): https://research.fs.usda.gov/download/treesearch/67993.pdf
Abstract
Multispectral data collected from satellite-based sensors serve as a powerful tool for quantifying forest aboveground biomass. In this paper, we evaluated the performance of models that used satellite-image-derived variables and the models that used a combination of satellite-image-derived variables along with stand-level attributes to quantify forest aboveground biomass of natural mixed-hardwood forests. Imagery from Landsat 8 and Sentinel 2 satellite missions were used to derive the remote-sensing variables. We also identified the important predictors for these models using the recursive feature elimination method. The models that included variables derived from satellite images and stand-level attributes performed better than the models that only used variables derived from satellite images.
Parent Publication
Citation
Jha, Sakar; Yang, Sheng-I; Hale, D. Stuart; Shoch, David; Johnson, Trisha; Hodges, Donald G.; Hoyt, Kevin P. 2024. Quantifying Forest Aboveground Biomass Using Remote Sensing and Ground Measurements for Mixed Hardwood Forests in East Tennessee. In: Bragg, Don C.; Oswald, Brian P.; Koerth, Nancy E., eds. Proceedings of the 22nd biennial southern silvicultural research conference. Gen. Tech. Rep. SRS-274. Asheville, NC: U.S. Department of Agriculture, Forest Service, Southern Research Station: 51–57. https://doi.org/10.2737/SRS-GTR-274-Pap8.
