Multi-Modal Foundation · allenai
OlmoEarth v1 FT — Mangrove (Base)
An Earth observation model fine-tuned to map mangrove extent for coastal conservation and habitat monitoring.
What OlmoEarth v1 FT — Mangrove (Base) is
OlmoEarth v1 FT — Mangrove (Base) is an Earth observation model created by allenai. Fine-tuned from the foundational OlmoEarth-v1-Base architecture, it specialises in mapping the spatial extent of mangrove ecosystems across coastal environments.
Mangroves are vital coastal ecosystems that serve critical roles in shoreline protection, biodiversity preservation, and blue-carbon accounting—the measurement of carbon sequestered in marine and coastal vegetation. By providing dedicated mapping capabilities, the model assists researchers and environmental scientists in coastal conservation and habitat surveillance.
How it works
The model builds upon OlmoEarth-v1-Base through task-specific fine-tuning. It was trained on a selected subset of the Global Mangrove Watch v4 Reference Samples to classify and delineate mangrove cover from satellite and Earth observation imagery.
Key facts
- Base model
- OlmoEarth-v1-Base
- Category
- Earth-Observation Fine-Tune
- Training data
- Global Mangrove Watch v4 Reference Samples (subset)
- Release date
- October 2025
What you can use it for
- Mangrove Extent Mapping
- Delineating boundaries and spatial coverage of mangrove forests across coastal zones.
- Blue-Carbon Accounting
- Supporting the estimation and tracking of carbon storage in coastal wetland ecosystems.
- Coastal Conservation and Monitoring
- Tracking habitat changes and informing spatial protection strategies for vulnerable coastal environments.
Run it here
We list this model but cannot run it yet: no pipeline_tag — needs a human.
Everything on this page still applies. Running it needs a loader for this repository, not a bigger machine.
Strengths and limits
Good at
- Tailored directly for coastal ecosystem analysis through fine-tuning from OlmoEarth-v1-Base.
- Trained using reference data curated by the Global Mangrove Watch initiative.
Watch out for
- Trained only on a subset of the Global Mangrove Watch v4 Reference Samples rather than the full reference dataset.
- Specialised for mangrove extent mapping, offering no evaluated capabilities for other coastal or terrestrial wetland vegetation types.
Inputs and outputs
You provide
Sequence or text
A protein or nucleotide sequence, or free text — whichever this model was trained on.
You get back
Mean-pooled representation, one vector per input.
Licence
Check the licence before commercial use · See model card
The See model card licence carries terms worth reading before you depend on it.
Questions
Can I use OlmoEarth v1 FT — Mangrove (Base) commercially?
Check the licence before commercial use. The See model card licence carries terms worth reading before you depend on it.
What is the underlying foundation model?
The model is fine-tuned from OlmoEarth-v1-Base.
What dataset was used during fine-tuning?
It was trained on a subset of the Global Mangrove Watch v4 Reference Samples.
What research domains benefit from this model?
The model is aimed at earth science, climate research, and ecological conservation, specifically for blue-carbon accounting, habitat tracking, and coastal ecosystem monitoring.
Written from OlmoEarth v1 FT — Mangrove (Base)'s own documentation (model card). Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.