Estimating Forest Aboveground Biomass Density Using Remote Sensing and Machine Learning : A RSME Approach
An accurate estimation of aboveground biomass (AGB) density is essential for effective forest management, carbon stock monitoring, and informed land management decisions. This study employs remote sensing datasets and collaborative efforts with ArcGIS to model AGB density across the Terengganu regio...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | en |
| Published: |
John Wiley & Sons Ltd.
2025
|
| Subjects: | |
| Online Access: | http://ir.unimas.my/id/eprint/48892/3/Estimating%20Forest.pdf http://ir.unimas.my/id/eprint/48892/ https://onlinelibrary.wiley.com/doi/10.1002/ldr.70087 https://doi.org/10.1002/ldr.70087 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | An accurate estimation of aboveground biomass (AGB) density is essential for effective forest management, carbon stock monitoring, and informed land management decisions. This study employs remote sensing datasets and collaborative efforts with ArcGIS to model AGB density across the Terengganu region. Integrated with the random forest algorithm in the Google Earth Engine for AGB density modeling at a spatial resolution of 1 km, the methodology incorporates GEDI Level 4, Sentinel-1 radar, Sentinel-2 optical imagery, and elevation/slope maps. The validation results indicated a root mean square error (RMSE) of 51.35 t per hectare and an average training error of 31.82 t per hectare, demonstrating the model's accuracy and reliability. The model's strong predictive performance (R2 = 0.77) implies that the independent variables accounted for 77% of the variability in the AGB. With a standard deviation of 64.52 t per hectare, the average AGB in the Terengganu area was 90.94 t per hectare, with AGB values ranging widely from 16.89 to 206.99 t per hectare across the region. These findings highlight the potential of integrating multiple remote sensing data sources for comprehensive AGB density mapping, thereby enhancing the monitoring of forest carbon stocks and fostering informed management approaches. This study underscores the importance of open-access data and cloud-based technologies, thereby supporting the availability of tools to implement comparable projects. This research illustrates the significance of combining different datasets and machine learning techniques for the remote assessment of forest biomass, thereby facilitating the improved modeling of ecosystem characteristics and sustainability initiatives. By emphasizing the need for advanced technologies and collaborative strategies, this study enhances forest biomass assessments and supports informed environmental management practices. |
|---|