68.8 tonnes of carbon per hectare in 2018 declined to 65.0 tC/ha in 2024 before a slight recovery to 66.6 tC/ha in 2025 in Oudomxay province, Laos. The dataset, created by ketsana phommavong and last updated in April 2026, analyzes changes in aboveground carbon stock and its correlation with the Normalized Difference Vegetation Index (NDVI) using Google Earth Engine and machine learning models.
Use Cases
- Monitoring forest carbon stock changes over time based on the described temporal analysis from 2018 to 2025.
- Evaluating the predictive relationship between NDVI and carbon stock based on the reported regression analysis.
- Comparing the performance of ensemble tree-based machine learning models (XGBoost, Random Forest, Gradient Boosting) for carbon estimation.
- Assessing the spatial variability of carbon stocks based on the reported standard deviation changes.
- Cross-annual validation of carbon stock prediction models based on the described stability analysis.
Strengths
- Provides specific carbon stock values (e.g., 68.8 tC/ha in 2018) and temporal trends over a 7-year period.
- Includes detailed model performance metrics, such as XGBoost achieving a test R² of 0.614.
- Identifies key predictive features, with NDVI noted as the principal predictor.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
- Data is specific to Oudomxay province, Laos, limiting generalizability.
Provenance
- Source
- figshare
- Collection Method
- Analysis using Google Earth Engine and machine learning models.
- Time Range
- 2018 to 2025
- Freshness
- Last updated 2026-04-24 05:01:41; freshness should be verified.
- Geography
- Oudomxay province, Laos