Global Annual Land Use and Land Cover Maps at 10m Resolution, 2017-2023

Impact Observatory, Microsoft, and Esri produced a global map of land use and land cover (LULC) derived from ESA Sentinel-2 imagery at 10-meter resolution for the years 2017 to 2023. Each annual map is a composite of predictions for 9 classes, generated by applying a deep learning model trained on billions of human-labeled pixels to Sentinel-2 annual scene collections. The maps have an assessed average accuracy of over 75% and are aligned to the ESA UTM tiling grid.

Use Cases

• Monitor global land cover change over time based on the annual time series from 2017-2023.
• Analyze transitions between land classes such as 'Bare' and vegetative classes ('Trees', 'Crops', 'Flooded Vegetation', 'Rangeland').
• Train or validate machine learning models for land classification using the 9-class, 10-meter resolution maps.
• Conduct regional environmental assessments by accessing data for a specific Area of Interest (AOI).

Strengths

• Global coverage at a consistent 10-meter spatial resolution.
• Annual time series spanning 7 years (2017-2023).
• Assessed average accuracy of over 75% for the classification.
• Model trained on billions of human-labeled pixels curated by the National Geographic Society.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and total dataset size are unknown, which may limit suitability assessment.

Provenance

Source

Impact Observatory, Microsoft, and Esri.

Collection Method

Generated by applying a deep learning model to ESA Sentinel-2 annual scene collections from the Microsoft Planetary Computer.

Time Range

2017-2023

Freshness

Last update date is unknown; freshness unverified.

Geography

Global