Canopy top height and indicative high carbon stock maps for Indonesia, Malaysia, and Phili

10-meter resolution canopy top height and high carbon stock land cover maps for Indonesia, Malaysia, and the Philippines. The maps were generated by a deep convolutional neural network trained on Sentinel-2 optical satellite images and calibrated with GEDI and airborne lidar data, using imagery from September 2020 to March 2021. The dataset was created by researchers from ETH Zurich and is hosted on Zenodo.

Use Cases

• Estimating aboveground biomass and carbon density based on canopy height maps.
• Classifying land cover for conservation planning using the 9-category high carbon stock (HCS) scheme.
• Monitoring forest density changes over time using the young regenerating forest to high density forest categories.
• Mapping specific agricultural land uses like oil palm and coconut plantations from satellite imagery.

Strengths

• Provides a high spatial resolution of 10 meters ground sampling distance.
• Covers three major Southeast Asian countries (Indonesia, Malaysia, Philippines).
• Derived from Sentinel-2 satellite imagery and calibrated with GEDI and airborne lidar data.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• The temporal coverage is limited to a six-month period from September 2020 to March 2021.
• Row count and file size are unknown, which may limit suitability assessment.

Provenance

Source

ETH Zurich research team (Nico Lang et al.), hosted on Zenodo.

Collection Method

A deep convolutional neural network trained on Sentinel-2 images with reference data from GEDI lidar and airborne lidar calibration.

Time Range

September 1, 2020 to March 1, 2021

Geography

Indonesia, Malaysia, Philippines