How ecosystem extent is calculated

Every ecosystem extent in the Atlas is an area measurement, and area is one of the easier things in geospatial work to get quietly wrong. This note explains how the figures are produced and what they can and cannot support.

Why area is harder than it looks

A raster pixel does not represent a constant area on the ground. In a standard geographic projection, a pixel near the equator covers far more surface than a pixel of the same size near the poles. Counting pixels and multiplying by a nominal size therefore overstates high-latitude ecosystems and understates tropical ones.

Projections and equal-area grids

To avoid that, extents are computed on an equal-area projection, where every pixel represents the same real-world area regardless of latitude. The difference is not a rounding detail — for global summaries it moves headline coverage figures by several percentage points.

Chunking large geometries

Very large or highly detailed boundaries are split into chunks before analysis. Beyond a certain size, a geometry has to be resampled to be processed in one pass, and resampling silently drops narrow features — the exact coastal and riparian ecosystems that tend to matter most.

Marine, coastal and inland boundaries

Country figures are reported against two boundaries. The first follows the Exclusive Economic Zone and includes marine and coastal ecosystems; the second covers land only. Both are published because they answer different questions, and mixing them is the most common way to misread a national profile.

Reading the numbers with care

Two figures accompany every extent: how much of the area is covered by any source dataset, and how much of that coverage has been classified. A low extent in a poorly covered region usually reflects a gap in the source maps rather than an absence of the ecosystem itself.