FAO's LCCS explained: the land cover nomenclature behind SDG reporting

If you've spent any time reconciling a national land cover inventory against UN Statistics Division metadata, you've run into LCCS. It shows up in the technical notes for SDG 15.3.1 (land degradation), 11.3.1 (land consumption rate), and 6.6.1 (water-related ecosystems), usually cited without much explanation, as if everyone already knows what it is.

What LCCS is

The Land Cover Classification System was developed by FAO, with UNEP, starting in the late 1990s. It's a set of diagnostic criteria, a classifier that builds class definitions from combinations of attributes like vegetation structure, life form, leaf type, spatial arrangement, and artificiality, rather than a fixed list of land cover types you pick from. Two countries can run the same classifier and land on different legend labels, but because the underlying criteria are shared, the classes can be crosswalked back to a common standard.

That distinction matters more than it sounds. A fixed class list breaks the moment your terrain doesn't match the categories someone else designed. A diagnostic system lets a dry-forest biome and a boreal one both get classified rigorously, using the same logic, without forcing either into categories built for the other.

FAO built LCCS this way so that national land cover products, each with its own legacy nomenclature, its own history of aerial survey conventions, its own ministry-specific categories, could still be translated into something comparable at the international level. That translation step is the part statistics offices actually have to do.

Why it matters for SDG reporting

SDG indicator metadata doesn't ask countries to adopt LCCS wholesale. It asks for land cover change figures that are methodologically transparent and comparable across reporting cycles and, ideally, across countries. LCCS is the reference nomenclature the UNCCD, FAO and UN Statistics Division point to when they need a common basis for that comparability. If your national legend can be crosswalked to LCCS classes, your figures carry weight in a UN custodian agency review. If it can't, you're explaining a bespoke taxonomy every cycle, and that explanation itself becomes a line of scrutiny.

There's a second reason it matters, less about terminology and more about process. SDG 15.3.1 in particular wants land cover change over time, measured consistently. A legend that gets redefined between assessment cycles, even slightly, even for good reasons, breaks the time series. The IPCC land-use categories (forest land, cropland, grassland, wetlands, settlements, other land) sit alongside LCCS in a lot of national methodologies, and the two aren't identical, so part of the analyst's job is maintaining a defensible crosswalk between whichever classification your source imagery or survey used and whichever one the SDG indicator metadata expects back.

In practice, the hard part is keeping the same classification rules, the same thresholds, and the same crosswalk logic applied the same way, year after year, so that a change figure for 2024 and a change figure for 2019 are measuring the same thing. A lot of national statistics have been assembled from whatever classification study happened to be commissioned that year, built by a different contractor with a different legend, which is exactly the inconsistency that reviewers flag.

That's the practical argument for treating your land cover change layer as a standing product rather than a one-off commission: pull an annually refreshed land-cover change baseline built on a consistent multispectral methodology and class definitions that don't shift contractor to contractor, and the crosswalk work you did for one reporting cycle still holds for the next.

None of this requires a new classification exercise before every submission deadline. It requires a baseline that doesn't move under you.

If your office is still rebuilding the classification from scratch each cycle, it's worth looking at what a standing, annually-refreshed layer would save you.