ORIGINAL RESEARCH article
Front. Remote Sens.
Sec. Image Analysis and Classification
Volume 6 - 2025 | doi: 10.3389/frsen.2025.1639835
Predicting Above-Ground Biomass of Clemson Experimental Forest in South Carolina Using Machine Learning and Remote Sensing
Provisionally accepted
- 1Department of Plant and Environmental Sciences, College of Agriculture, Forestry and Life Sciences, Clemson University, Clemson, United States
- 2Clemson University, Clemson, United States
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
This study presents an integrated approach to estimate aboveground biomass using Sentinel-1 and Sentinel-2 remote sensing data combined with field data measurements from 191 plots of Clemson Experimental Forest. This study used several imageries to identify the best model for above ground biomass prediction for carbon market baseline determination using machine learning techniques. Tree diameter measurements were used to estimate above ground biomass per plot and extrapolated to Mg/ha, averaging 173.47 Mg/ha (range: 0-336.34 Mg/ha). Remote sensing data, including Sentinel-1 SAR (VV and VH), and Sentinel-2 optical imagery were acquired for August-September 2021. Tree Canopy Cover (TCC), National Agriculture Imagery Program (NAIP) imagery, and Shuttle Radar Topography Mission Digital Elevation Model (SRTM DEM) datasets were also acquired. Spectral bands from Sentinel-1, Sentinel-2, vegetative indices from Sentinel-2, slope, aspect, elevation from STRM DEM and ancillary bands from TCC and NAIP were used as independent variable for model development. Machine learning models such as Random Forest (RF), Extreme Gradient Boost (XGB), Support Vector Machine (SVM), and Artificial Neural Network (ANN) were evaluated, with ANN showing the best performance (adjusted R 2 = 71.4%, Root Mean Square Error = 52.25 Mg/ha; Mean Absolute Error = 41.7 Mg/ha). Above ground biomass was highest in old growth stands, flat areas, and loamy soils. The study demonstrates the value of remote sensing and machine learning for above ground biomass estimation and carbon market dynamic baseline development. Dynamic baseline through average of the change in above ground biomass from 2015 to 2025 with the developed model from this method could provide an accurate baseline for landowners who want to participate in the carbon market.
Keywords: remote sensing, carbon market, Aboveground biomass (AGB), machine learning, Sentinal 1 and Sentinal 2, Clemson Experimental Forest (CEF)
Received: 02 Jun 2025; Accepted: 23 Jul 2025.
Copyright: © 2025 Sharma, Timilsina, Clay, Khanal and Khanal. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence: Sanjeev Sharma, Department of Plant and Environmental Sciences, College of Agriculture, Forestry and Life Sciences, Clemson University, Clemson, United States
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.