We measure the land.
Independent reviewers certify it.
Kabon.Africa runs its own Earth Observation and MRV (Measurement, Reporting & Verification) pipeline in production — satellite band-math and ground IoT sensors, not a third-party dashboard bolted on for show.
Measurement — Kabon
Our own pipeline processes Sentinel-2 satellite bands into an NDVI time series for the project's exact parcel, and combines it with IoT ground readings — soil, gas flow, temperature — where devices are deployed. This is the evidence a reviewer sees, generated by our code, not re-published from someone else's report.
Certification — independent reviewers
A Kabon-assigned or client-nominated reviewer assesses that evidence against the published Kabon Carbon Standard methodology and records the carbon figure. A second reviewer — never the one who assessed it — must sign off before any credit is issued.
Full issuance chain, buffer pool, and methodology library on the Kabon Carbon Standard page.
From orbit to NDVI reading
Every step below runs against live Sentinel-2 imagery for the project's own coordinates — not a static basemap tile.
- STEP 1
Authenticate & task
Each check-in authenticates to the Copernicus Data Space Ecosystem via OAuth client-credentials, with the token cached until it expires.
- STEP 2
Build the parcel window
A bounding box is built from the project's GPS centroid and hectares, capped at a 5 km half-side to stay within Copernicus's sample-size limit — with ground resolution chosen adaptively: 10 m up to 100 ha, 20 m up to 2,500 ha, 30 m above that, so large projects stay tractable.
- STEP 3
Run the evalscript
A custom Sentinel-2 L2A evalscript reads the red (B04) and near-infrared (B08) bands to compute NDVI, then masks cloud and shadow pixels using the Scene Classification Layer (SCL 4–6: vegetation, soil, water).
- STEP 4
Aggregate the time series
A 60-day daily NDVI series is pulled and reduced to 25th / 50th / 75th percentile statistics plus a derived cloud-cover figure, taking the most recent valid interval as the reading.
- STEP 5
Fall back when needed
If Copernicus is unavailable, the pipeline falls back to the Agromonitoring NDVI history API over the same 60-day window, so a reading is still produced.
IoT sensors, on-site
For clean-energy projects — solar, biogas, cookstoves — Kabon pulls country-specific grid-intensity and fuel emission factors and caches them per ISO country code, underpinning the avoided-emissions maths for the project. Soil and gas-flow sensors on land-restoration sites report continuous, tamper-proof readings from the field: no self-reporting, no consultant with a clipboard.
What's live, what's next
We name our sources and their availability plainly — free, integrated, or still to be negotiated.
Primary source. 10 m multispectral optical imagery, atmospherically corrected, accessed via the Copernicus Data Space Ecosystem Statistics API.
Secondary source used automatically when the Copernicus API is unreachable, over the same 60-day window.
Free radar imagery that sees through cloud cover — a natural next input for the rainy-season gaps optical NDVI can't fill.
Free, longer-baseline optical archive (since 2013) — useful for pre-project baseline and long-run change detection.
High-resolution (≤5 m) tropical basemaps, free for approved non-commercial monitoring use under the Norway's International Climate & Forests Initiative programme.
Kabon's own processing layer
Sentinel-2 imagery is open Copernicus data — anyone can request it. The processing layer above it (adaptive bounding-box and resolution logic, the NDVI evalscript, cloud/shadow masking, time-series aggregation, and the IoT emissions-factor pipeline) is built and maintained in-house, and is Kabon.Africa's own intellectual property.
Satellite NDVI and IoT readings inform the reviewer's assessment; the reviewer currently records the final carbon figure by hand against that evidence, rather than the pipeline deriving it automatically. Fusing satellite and IoT evidence into an automated carbon estimate is on our near-term roadmap — we'd rather tell you where the line sits today than round up.
