Each country's status is an aggregation of its Ramsar wetlands' condition trends — the share of sites whose satellite-derived indicators are significantly improving versus declining over recent decades. Explore the map or the table; select a country to see its sites.
| Country | Sites | Improving | Declining | Stable | Not enough evidence | Net trend |
|---|
For every Ramsar site, each condition (state) indicator with at least 8 years of data is tested for a monotonic trend with the non-parametric Mann-Kendall test (Kendall's τ). The indicators that vote are surface-water extent (abiotic: the JRC water area averaged over the months a satellite actually observed, so the denser Landsat record from 2013 does not read as more water; in Australia also DEA Water Observations), vegetation greenness relative to the surrounding 2-10 km landscape and canopy moisture (biotic) and, in Australia, waterbird counts (biotic). Other condition indicators (the JRC annual maximum and mean water area, water permanence, the DEA WIT cover classes, bare soil) are shown but do not vote. Climate and catchment drivers (rainfall, runoff, drought, temperature, groundwater) are context and never vote — otherwise a wetland drained during a wet decade scores as improving. Fire does not vote either: it is natural in some wetlands and damaging in others. Each τ is direction-adjusted, weighted by its statistical confidence (1 − p) so noisy trends shrink toward zero, and then averaged by concept so that several ways of measuring surface water count as “water” once. An indicator with the same value in every year (for example, no open water ever detected) has no trend to test and does not vote. A site is improving or declining by the sign of that score. Where nothing is significant we separate two different statements: with a record long enough to have detected a strong trend (judged on the longest record in each concept, so one long water record is not outweighed by several short ones), no trend means stable; without one, we report not enough evidence instead of calling it stable. A country's net trend = (improving − declining) ÷ assessed sites (improving + declining + stable) — sites we cannot resolve are excluded from the denominator rather than being counted as if they were stable. Recently-listed sites without geolocation are omitted from the map.