Recalculate the global estimate
Global Mangrove Watch version 3.0, described by Bunting and colleagues in 2022, estimates mangrove extent at 152,604 km² in 1996 and 147,359 km² in 2020. Subtraction gives a net reduction of 5,245 km². Dividing by the 1996 estimate gives 5,245 ÷ 152,604 × 100 = 3.44%, consistent with the paper’s rounded 3.4%.
Dividing the net loss by 24 gives approximately 219 km² per year, but this is only an interval average. It does not show that each year lost that amount. Nor is net loss the same as total clearing: gains in some locations can partly offset losses elsewhere.
Why change pixels deserve more scrutiny than the extent map
The product uses JAXA L-band radar mosaics for 11 observation epochs from 1996 to 2020, updating a baseline with a map-to-image change-detection approach. The reported overall extent accuracy is 87.4%, while estimated accuracies for the gain and loss classes are 58.1% and 60.6%. A broadly credible map of where mangroves occur can therefore still be much less reliable at identifying the smaller set of pixels that changed.
The study identifies misregistration and fragmented areas around aquaculture ponds as sources of error. A shifted boundary can create adjacent apparent loss and gain without a real ecological change. Readers should inspect both dates and the patch geometry rather than treating every colored change pixel as a confirmed clearing.
The uncertainty changes the conclusion
The paper reports a net-change uncertainty interval from −13,587 to +1,444 km² around the central estimate of −5,245 km². That interval includes zero. The defensible statement is that the mapped central estimate is a decline, with substantial uncertainty—not that the exact global loss is known to the nearest square kilometre.
For a local restoration claim, the global result is a starting point. Confirm the project boundary, the same canopy definition at both dates and field or high-resolution evidence. A new patch of mapped canopy does not by itself establish habitat quality, long-term persistence or a verified carbon credit.
Bunting et al. (2022), Global Mangrove Extent Change 1996–2020: Global Mangrove Watch Version 3.0, Remote Sensing 14, 3657. Published estimates, methods and accuracy statistics are taken from the study abstract. Percentage and interval-average calculations are reproduced above. This is a reading of the published result, not a new pixel-level classification. Source checked 30 October 2026; analysis updated on that date.
A coastline map is a classified observation
Optical and radar imagery can map mangrove cover, but tides, clouds, shadows, mixed pixels and plantations affect classification. A change map should disclose sensor, dates, spatial resolution, training data and accuracy assessment. A global layer is useful for regional screening, not a parcel boundary.
Match each claim to evidence
| Claim | Useful records | Key limitation |
|---|---|---|
| Extent changed | Comparable classified imagery and field checks | Method changes can mimic land-cover change |
| Coast migrated | Repeat shoreline or vegetation-edge positions | Tidal stage and positional uncertainty matter |
| Carbon stock is high | Biomass plots, sediment cores, depth and bulk density | Stock is not an annual sequestration rate |
| Storm risk fell | Water levels, waves, vegetation structure and damage data | Protection varies with event and landscape geometry |
Migration needs room and sediment
Sea-level rise does not produce one universal mangrove response. Landward migration depends on elevation gradients, barriers and land use; persistence also depends on sediment elevation keeping pace with relative sea-level change. Local subsidence can be as important as ocean-level change.
Do not convert every benefit into one number
Habitat, fisheries, carbon and wave attenuation use different baselines and units. Avoid adding modeled values without checking overlap, uncertainty and geographic transfer.
Evidence task
Choose one mapped mangrove change. Record the two dates, sensor resolution, classification definition, tidal context, accuracy metric and one alternative explanation. Then state whether the record supports change in extent, condition or both.
Sources and limits
- Global Mangrove Watch: mapped extent and change products.
- NOAA Ocean Service: Mangrove forests.
- Donato et al. (2011): field-based mangrove carbon synthesis.
Sources accessed 10 October 2026. This page is not a navigation chart, property survey or project-credit verification.