Sekonyer river
← Back to blog

Ecosystem Extent Accounting: What It Takes to Answer One Indicator

Sustainability•26 October 2026•TREEO Indonesia•3 min read

Global Biodiversity Framework Headline Indicator A.2 asks a question a country ought to be able to answer: what proportion of your territory is covered by natural and semi-natural ecosystems, and how is that changing? Indonesia was selected as a pilot country to test how it is compiled under the SEEA Ecosystem Accounting framework. The technical work turned out to be far less about mapping and far more about reconciling maps that already exist.

The Assignment Is Not to Make a Map

This surprises people. Indonesia already has an indicative ecosystem map, produced nationally, integrating elevation, ecoregion and vegetation characteristics, land cover, mangrove and karst layers, peat ecological function, topographic base data and soils.

The task is to build a transparent, reproducible evidence base from what exists — a national Ecosystem Extent Account — rather than to generate another competing product. That distinction shapes everything downstream, because an account is not a map. It is a table of areas by ecosystem type at an opening and a closing period, with the change between them attributable to something.

A map answersAn account answers
What is here?How much of each type is there?
At one momentAt two moments, with change between
Cartographic accuracyAccounting consistency across periods

Why Two Reference Years Is Harder Than One

The agreed periods are 2020 opening and 2024 closing. Producing a defensible change figure between them requires the two snapshots to be built the same way — same classification, same minimum mapping unit, same source hierarchy.

If the 2024 layer benefits from better imagery or a revised classification scheme, some of the apparent change is methodological rather than ecological. This is the same problem forest carbon projects face when they re-stratify between monitoring events, and it has the same solution: recompute the earlier period on the current basis, and document the restatement.

The alternative — comparing two independently-produced national maps — yields a change figure dominated by methodology.

The Classification Problem

Indonesia's national classification is anchored in the natural ecosystem types identified in its biodiversity strategy and status reporting, spanning marine, freshwater and terrestrial realms, with anthropogenic types identified separately. A working compilation extends this to a larger set for mapping purposes.

That national scheme has to be crosswalked to Level 3 of the IUCN Global Ecosystem Typology — Ecosystem Functional Groups — because that is the level at which types must be reported for international comparability.

Crosswalks are never clean. National types are defined by what the country needs to manage; international types are defined by ecological function. A national type can map to several functional groups, or several national types can collapse into one. Every one of those relationships is a judgement that has to be recorded, or the account cannot be reproduced.

What Actually Consumes the Effort

Three things, none of them glamorous:

Source reconciliation. Multiple national layers cover overlapping ground with different vintages, resolutions and definitions. Deciding which takes precedence where is the substantive methodological work.

Boundary arithmetic. Areas must sum to the national territory. Overlaps and gaps between thematic layers are normal in cartographic products and fatal in an account.

Provenance. Every figure in the final table must trace to a named source layer with a version and a date. An account nobody can reproduce is an assertion.

Why This Matters Beyond Statistics

Ecosystem extent accounting is the substrate that several other things sit on. A parallel indicator on services provided by ecosystems needs the same spatial reference. Carbon accounting at jurisdictional scale needs consistent land classification. Any jurisdictional REDD+ baseline allocation depends on knowing what is where, consistently, over time.

Countries that build this layer once, properly, stop rebuilding it for every new framework. Countries that do not end up with a different land classification for each international obligation, none of which reconcile — which is where a great deal of national reporting effort currently goes.

The Transferable Lesson

The instinct when facing a mapping question is to make a map. The higher-value move is usually to make the existing maps agree with each other and to write down how. That output is less visually impressive and far more durable, and it is what turns a collection of spatial data into something an accountant, a statistician and a verifier can all use.

Selling into a compliance market?

Compliance buyers examine sampling design, uncertainty and data traceability directly — and a later finding of ineligibility returns the liability to them. TREEO builds the evidence file that answers it.

Related Articles