Skip to main content ->
Ai2

OlmoEarth's impact: Global Mangrove Watch

Mangroves aren’t just trees by the sea—they’re living infrastructure that shields coastal communities from storms and erosion. Their knotted roots knock down incoming waves and reduce storm-surge height. They’re also carbon-storing powerhouses: on average, a single hectare holds ~1,000 tons of carbon in wood and soils, far more than the Amazon’s ~400 tons per hectare.

Mangrove forests are at risk from coastal development, sea-level rise, and extreme weather. Conservationists, governments, and practitioners need up-to-date, accurate maps to see where mangroves are thriving—and where they’re disappearing.

Mangroves growing back (in blue) after restoration projects in Guinea-Bissau; GMW data from 1990-2024. Courtesy Global Mangrove Watch.

The Global Mangrove Watch (GMW) – a collaboration between Aberystwyth University, soloEO, Wetlands International, and The Nature Conservancy – maps the world’s mangroves from year to year. This herculean effort has traditionally needed 12–18 months of expert checks across millions of locations per refresh. By the time a map ships, parts of the coastline have often already changed.

GMW has piloted the OlmoEarth Platform to refresh maps on a regular cadence with higher accuracy and less labeling burden. By combining model-output generation with built-in review, the team can produce more accurate maps faster with less manual effort.

“An annual map that arrives with a few years’ delay doesn’t really help you as a protected area manager,” says Lammert Hilarides, Senior Technical Officer at Wetlands International. “For deforestation alerts, timing is even more critical. You need to know as early as possible if mangroves are being cleared or degraded, so you can act quickly.”

More up-to-date extent layers – maps showing the location of mangroves – directly support restoration planning, biomass and carbon accounting, and local engagement. They also provide a stronger starting point for historical loss analysis and near-real-time monitoring.

To cut time-to-truth (the lag between when loss is detected and when experts can confirm it), GMW is training change-detection models for monthly alerts using examples of verified mangrove loss.

An animation of the mangrove loss alerts (in orange/red) in Rhakine, Myanmar; GMW data from 2022-2025. Courtesy Global Mangrove Watch.

Accurate, near-real-time loss events enable natural resource managers and governments to respond more quickly to threats – from anomalous logging to storm damage – and plan restoration where it will have the biggest impact.

“Until now, updating our global mangrove maps was a months-long, manual process that limited us to covering only half the world’s mangroves,” says Chris Sheridan, Program Manager, Coasts and Detlas at Wetlands International. “With our beta of OlmoEarth, we’re seeing accuracy levels around 97% and could cut processing time in half, allowing us to reach global scale for the first time. That kind of speed and reliability completely changes what’s possible for conservation teams on the ground.”

35 years of mangrove cover change on the coast of Surinam; GMW data from 1990-2024. Courtesy Global Mangrove Watch.

Because OlmoEarth integrates data acquisition, modeling, and review, it can replace a patchwork of manual tools with a coherent, repeatable workflow. The result is practical: better maps, delivered faster, with less friction—putting timely information in the hands of the people safeguarding mangroves worldwide.