Monitoring biweekly dynamics of pan-tropical industrial plantations over 6 years using 100 m PROBA-V data
Anthropogenic land conversion profoundly impacts the Earth’s surface, with varying effects across regions. In the tropics, industrial plantations particularly affect natural forests. Monitoring land use and land cover change (LULCC) due to agricultural expansion is crucial for achieving sustainable...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2025-06-01
|
| Series: | Frontiers in Remote Sensing |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/frsen.2025.1575100/full |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Anthropogenic land conversion profoundly impacts the Earth’s surface, with varying effects across regions. In the tropics, industrial plantations particularly affect natural forests. Monitoring land use and land cover change (LULCC) due to agricultural expansion is crucial for achieving sustainable imports into the European Union under the Regulation on Deforestation-free Products (EUDR). Earth observation satellite missions, providing free global imagery with high revisit frequency, are instrumental in monitoring tropical ecosystems and their transformation. However, accurately mapping the correct dates of tree cutting or planting on a global scale remains a challenge. This study addresses this gap by developing a near real-time sensor-agnostic method for monitoring deforestation and plantation rotation. It is developed using 100 m PROBA-V full Collection 2 archive with a 5-day revisit, spanning 2014 to 2020. A novel index enabled distinguishing vegetation from land cleared for plantations. The variability of atmospheric perturbations and both intra- and inter-annual variability of the vegetation spectral signatures were mitigated using spatial standardization. Statistical thresholds identified pixels that deviated from the normal distribution of forest spectral values, capturing LULCC. It results in pan-tropical annual maps series 2015–2020 illustrating the typical dynamics of perennial plantations, from land preparation to mature plantations, including the dates of cutting and planting. Validation using 899 randomly selected samples through confidence-based stratified sampling yielded a global accuracy of 82% ± 2% for new plantation detection. 62% of the detections 1 Bos et al. were accurate to the exact year, which represents a significant 19% improvement over previous studies. Our initial estimates of industrial plantation dynamics suggest that new oil palm plantations cover approximately 3,064 km2 annually, of which 79% is rotation within existing plantations and 21% expansion into new areas. Annual plantations of other perennial plantations cover about 13,875 km2, of which 81% is from rotation and 19% from expansion. This work demonstrates the effectiveness of optical 100 m spatial resolution for near real-time pan-tropical mapping of perennial industrial plantations in cloudy regions. |
|---|---|
| ISSN: | 2673-6187 |