How do my points improve the map?
Each point marks a spot as degraded or forest. The model counts your points far above its starting labels and reruns on the same imagery release, so each correction is compared like for like.
Degraded land mapping
Where on this property is the land open, grass and bare ground, and where are trees and shrubs? Open land is where to look for planting sites. The analysis classifies recent high-resolution imagery, starting from canopy height, and you can refine it by placing points on the map marking land as degraded or forest.
Pilot, tested on properties in the Azuero region of Panama.
Who it is for
Carbon project developers
Estimate the land open for ARR planting from recent imagery, with outlines you can check in the field.
Planting and field teams
Outline the patches to plant, and correct the map point by point where you know the ground.
Restoration organizations and funders
Show where open land sits on a property before the work starts.
What goes in
A property boundary as a polygon, and these settings, which you can change before each run.
| Setting | Default | Range |
|---|---|---|
| Points a person placed on the map, each labelled degraded or forest. | None | — |
| Esri Wayback release to analyze. Reruns keep the release of the run they refine. | Set per run | — |
| Satellite embedding year for a regional starting map. Reruns keep the year of the run they refine. | Set per run | — |
| Start from canopy height before any points are added. | On | — |
| Smallest degraded patch kept on the map, in hectares. | 0.1 | 0 or more |
| Largest forest gap inside degraded land that is filled, in square meters. | 100 | 0 or more |
Data sources
- Esri World Imagery WaybackRead on tiles of about 0.6 m; the source imagery’s own resolution varies by place. The newest release unless you pick another; each run reports the release and, when available, the capture dates.
- Meta and WRI Global Canopy Height Map v2 (CHMv2)About 1.2 m pixels, read at native resolution. Imagery from 2017 to 2020, mostly 2018 to 2020.
- Google Satellite Embedding (annual)10 m, yearly. Context for the regional model in Azuero, Panama.
What you get
- Hectares and share of degraded land and forest on the property.
- Outlines of degraded land, with patches under 0.1 ha removed and small forest gaps inside filled.
- The imagery release and capture dates the map was made from.
- In Azuero, the land likely to be planted, with groups of trees taken out.
Limits
- A pilot. The regional model is trained in Azuero, Panama; elsewhere it starts from canopy height alone, so expect to place more points.
- Trees that drop their leaves in the dry season can read as degraded, and seasonal woodland such as miombo is not reliable yet.
- Capture dates vary across a property and can be older than the imagery release; each run reports them when the imagery service provides them.
- Up to about 500 ha of bounding box per run; long or fragmented boundaries can hit the limit sooner.
- It maps what the imagery shows. Whether land counts as degraded under a standard or national rule is a separate check.
Where it fits in your work
Run it once a property is a likely candidate, after forest cover history and land structure, and before planting design. Field teams correct it with points, and the outlines go to the map and the field app for planning.
Questions teams ask
Which imagery does it use?
The newest Esri World Imagery Wayback release, read on tiles of about 0.6 m, unless you pick an older one. Each run reports the release and, when available, the capture dates.
Does it run outside Panama?
Yes, from canopy height and your points. The regional model, which adds context from satellite embeddings, exists for Azuero only; other regions need their own training data.
Is degraded land the same as eligible land?
No. It shows open land on recent imagery. Eligibility also depends on the land’s history, your standard’s rules and what you can confirm in the field.
Run it on your land
Tell us about the properties you are looking at, and we will set the analyses up in your project.