Page Summary
-
This dataset, provided by the Forest Data Partnership, estimates the per-pixel probability of rubber tree presence using a machine learning model at 10-meter resolution for selected countries.
-
The dataset is intended to support the Forest Data Partnership's goal of reducing forest loss from commodity production through improved monitoring and tracking.
-
Availability spans from 2020 to 2023, and it is important to note that this dataset has not yet been peer-reviewed.
-
Separate terms of use apply for commercial and non-commercial users of Earth Engine for this dataset.
This dataset is part of a Publisher Catalog, and not managed by Google Earth Engine. Contact forestdatapartnership@googlegroups.com for bugs or view more datasets from the Forest Data Partnership Catalog. Learn more about Publisher datasets.
- Catalog Owner
- Forest Data Partnership
- Dataset Availability
- 2020-01-01T00:00:00Z–2023-12-31T23:59:59Z
- Dataset Producer
- Produced by Google for the Forest Data Partnership
- Tags
Description
**Note: This dataset is not yet peer-reviewed.
Please see this GitHub README for more information.**
This image collection provides estimated per-pixel probability that the underlying area is occupied by the commodity. The probability estimates are provided at 10 meter resolution, and have been generated by a machine learning model. For details, see the technical documentation on the Forest Data Partnership repo on Github.
The primary purpose of this image collection is to support the mission of the Forest Data Partnership which aims to halt and reverse forest loss from commodity production by collaboratively improving global monitoring, supply chain tracking, and restoration.
This dataset currently covers the following countries: Thailand, Indonesia, Viet Nam, Côte d'Ivoire, China, Malaysia, Brazil, Mexico, and Sri Lanka.
This community data product is meant to evolve over time, as more data becomes available from the community and the model used to produce the maps continuously improves. If you would like to provide general feedback or additional datasets to improve these layers, please reach out through this form.
Limitations: Model output is limited to selected countries as calendar year composites for 2020 and 2023. Not all regions of the output are well represented by training data. Accuracy is reported in aggregate, and will vary geographically and with user chosen thresholds. Sensor artifacts based on data availability, cross-track nonuniformity, or cloudiness may be visually apparent in output probabilities and result in classification errors at some thresholds.
Note that this dataset has separate terms of use for commercial users of Earth Engine. Please see "Terms of Use" tab for details.
The difference between the 2025a and 2025b versions is that the 2025b version is based on AlphaEarth Foundations Satellite Embeddings, which enables expanded geographic and temporal coverage over 2025a. Additionally, 2025b includes additional input data and an evolved input data processing pipeline. Note that the 2025a model may perform better in some contexts.
Bands
Bands
Pixel size: 10 meters (all bands)
| Name | Min | Max | Pixel Size | Description |
|---|---|---|---|---|
probability |
0 | 1 | 10 meters | Probability that the pixel includes rubber trees for the given year. |
Terms of Use
Terms of Use
For non-commercial users of Earth Engine, use of the dataset is subject to CC-BY 4.0 NC license and requires the following attribution: "Produced by Google for the Forest Data Partnership".
Commercial use of the dataset is subject to the Forest Data Partnership Datasets Commercial Terms of Use.
Contains modified Copernicus Sentinel data [2015-present]. See the Sentinel Data Legal Notice.
Citations
Forest Data Partnership. 2025. Community models 2025a. Online
Explore with Earth Engine
Code Editor (JavaScript)
Map.setCenter(106.48584, 11.17099, 11); var collection = ee.ImageCollection( 'projects/forestdatapartnership/assets/rubber/model_2025a'); var r2020 = collection.filterDate('2020-01-01', '2020-12-31').mosaic(); Map.addLayer( r2020.selfMask(), {min: 0.5, max: 1, palette: 'white,blue'}, 'rubber 2020'); var r2023 = collection.filterDate('2023-01-01', '2023-12-31').mosaic(); Map.addLayer( r2023.selfMask(), {min: 0.5, max: 1, palette: 'white,green'}, 'rubber 2023');