The files
| File | Format | What's in it | Notes |
|---|---|---|---|
| Annual land-cover classification | GeoTIFF | One raster per year over your boundary, every pixel assigned a class from the agreed scheme. The same scheme applies in every year, so year-on-year differences are change on the ground and not a change of legend. | Refreshed once a year. |
| Change statistics table | CSV | Area in hectares by land cover class and by administrative unit, for each year, plus a change matrix between any two years. Opening stock, closing stock and the movement between classes sit in the same table. | It is meant to be pasted into your indicator workbook. |
| Habitat loss and gain layer | GeoPackage | Polygons of area that changed class between two years, tagged with the class it left and the class it entered. Loss and gain are flagged so a natural or semi-natural class shrinking or growing can be pulled out with one filter. | Opens in QGIS or ArcGIS. |
| Reporting-ready export | CSV or spreadsheet | The change figures laid out for the national indicator submission you file. | Send us the template and we fit the columns to it. |
| Method note | A short note for each release: the classification scheme, the years and scenes used, and the places where cloud left a thinner record. | It travels with the figures into your submission file. |
How it works
Pipeline: what you supply, what we do, what comes back.
- 01 / You supply
Supply the boundary and the scheme
A national or sub-national boundary as a shapefile or GeoPackage. Your baseline year. Your preferred classification scheme if the ministry already has one, or we propose one to agree.
- 02 / We do
We composite the year
Wide-swath multispectral passes across the territory are collected for the year and combined into a cloud-reduced composite for each reporting year.
- 03 / We do
We classify and difference
Each year is classified on the agreed scheme and compared with the years before it. Class statistics and a change matrix come out of the same run, so the tables and the layer agree.
- 04 / Comes back
You receive the release
The classified rasters, the statistics table, the loss and gain layer, the reporting export and the method note. Each year after that, one new year is added to the same record.
Operating envelope
| Condition | Effect on the output | What the delivery does |
|---|---|---|
| 10–30 m pixels | A pixel here covers well under a hectare of ground, depending on the source. Change at landscape scale shows. Small patches below roughly a hectare, narrow hedgerows, single buildings and small wetlands do not resolve reliably. | For small, high-value sites pair this with local survey. |
| Cloud cover | Compositing a full year of passes helps with cloud. In persistently cloudy regions, some years have thinner data. | The method note lists where. |
| Land cover, not carbon | The figures are class and area change. | They are not carbon stocks, emissions or biodiversity measures. |
| Class separation | Classes that look alike from orbit, such as some grassland against low cropland, can be confused. | The scheme is kept to classes that multispectral imagery can tell apart. |
| Baseline and back-fill | The record is built forward from the baseline year you agree. | Years before that depend on what archive imagery supports for your territory. |
| Cadence | Annual is a target. | It follows when a year's usable imagery is complete. |
Why this exists
Each cycle the land-cover figures get rebuilt from whatever exists. There is an older national map, a regional product on a different legend, and the last consultant's classification study, which stopped when its contract did.
You reconcile class definitions by hand, argue about whether a change is real or a legend change, and add a caveat to the submission. A fresh classification commission means a tender and a legend workshop, and the deadline comes first. Habitat Loss Mapping replaces that with one scheme, refreshed every year.
Who it's for
For
- Statistics analysts at an environment ministry or national statistics office compiling land-cover change for SDG and national reporting.
- Evaluators who want one classification scheme and one repeatable method across reporting cycles.
- Teams working at national or sub-national scale who need area by class, by administrative unit.
Not for
- Teams that need carbon stocks, emissions or biodiversity scores. This is land cover only.
- Site-level habitat surveys of small patches. Field survey or finer imagery does that better.
- Real-time alerts for clearing. This is an annual record.
- A land-cover legend with many fine classes that need field-level separation.
Technical and procurement questions
Does 10–30 m resolve small habitat patches?
It resolves landscape-scale class change. Patches under about a hectare do not resolve reliably, so pair this with local survey for small, high-value sites.
Is the classification scheme fixed from year to year?
Yes. Consistency across years is the point. The scheme is set once with you and held.
Does this measure carbon?
No. It reports land cover class and area change. It gives no carbon stock or emissions figure.
What about cloud over a whole country?
Compositing a full year of passes helps with cloud. Persistently cloudy regions can have thinner data in some years, and the method note says where.
Can you back-fill historical years?
The record is built forward from an agreed baseline year. Back-fill depends on the archive imagery for your territory, and we check it before anything is agreed.
How do we validate against our own data?
Send your reference data, such as an existing national map or field plots. It can be set beside the classification so you can see where they agree and where they differ.
Will it open in our GIS and workbook?
The layers open in QGIS or ArcGIS, and the tables are plain CSV for your workbook.
Who owns the outputs?
Use and ownership terms for the delivered files are written into the contract before work starts. The public satellite archives keep their own licences.
How is it priced?
Each territory is scoped and quoted on its own. You get a written scope for procurement.