Ten years of land opened: mine land opening and recovery across 124 coal permits in South Sumatra
Produced with the support of the SORE team. Attribution and author list to be confirmed before launch.
An annual, wall-to-wall record of how much land was opened by coal mining across South Sumatra between 2016 and 2025, and how much of it grew back. Built to supply a baseline that post-mining land planning currently lacks.
Why we did this
South Sumatra is planning what to do with land after coal. Coffee, solar, oil palm and tourism are all named as intended uses, and the economics have been modelled.
The planning is proceeding without a measured answer to a prior question: how much post-mining land is there, where is it, and what condition is it in. Land reuse cannot be sized without it. We produced the baseline.
Method
We traced the footprint of 124 coal mining permits across South Sumatra, one year at a time, from 2016 to 2025. Every patch of open ground was delineated by hand on 30 m Harmonized Landsat and Sentinel imagery, then cross-checked against 3 m imagery and base maps. Overlapping polygons were removed first, so no hectare is counted twice. 2015 was dropped because the imagery that year was too cloudy to read.
The result is an annual record of two things usually reported separately: how much land was opened, and how much actually grew back.
What the record shows
- Open land grew 3.6 times over, from 6,291 ha in 2016 to 22,373 ha in 2025.
- 29,178 ha were opened at least once over the decade. 6,805 ha, or 23 per cent, are now vegetated again.
- Between 2017 and 2025, 25,589 ha were opened against 9,363 ha recovered, a ratio of 1 to 2.7. In nine consecutive years, recovery never exceeded clearing.
- 2,464 ha have stayed bare for the entire ten years, and a third of the currently open land has been open for five years or more.
- 41 per cent of land open in 2025 sits within 1 km of housing, and 77 of the 124 permits have a main river within 500 m of their boundary.
Where recovery did happen, it holds up. NDVI on recovered ground climbs steadily for four to five years, then settles near 0.74, well above the 0.33 threshold for healthy cover. That measures greenness, not species composition, which is a limit worth stating plainly in any rehabilitation report.
What it means for land reuse planning
Three findings bear directly on the transformation agenda:
- The available area is smaller than the opened area. Of 29,178 ha opened, a substantial share is still active workings rather than land awaiting a new use. Reuse proposals sized against total disturbance will overstate what is actually free.
- Recovery is not keeping pace with clearing, and has not in any year of the record. Any plan that assumes reclamation proceeds in step with extraction is assuming something the last decade did not deliver.
- Much of this land has neighbours. With 41 per cent of open land within a kilometre of housing, decisions about what happens next are not only agronomic or financial.
Limits, and what this is a first step toward
This is a preliminary study. It covers one province and one commodity, and it answers the spatial question only: how much land, where, and in what condition. The economic, energy, social and environmental questions that follow from it are the ones that matter for policy, and they are the direction of our work from here. This baseline exists so those questions can be answered against a measured quantity rather than an assumed one.
The method is built to be repeated. We intend to run the same annual, wall-to-wall measurement across other Indonesian regions and other coal-producing countries, because one province on its own shows what happened there, while a comparison separates what is structural in coal mining from what is particular to South Sumatra.
The analysis measures land cover, not land status. It does not establish tenure, permit condition, reclamation liability, or soil suitability for any particular crop. It is a baseline for planning, not a substitute for site assessment.
The full analysis
The findings are published as a self-contained interactive deck: animated maps, a year-by-year opened-against-recovered chart, a cumulative disturbance and recovery footprint, duration histograms, NDVI recovery curves, and proximity analysis for schools, villages, places of worship and rivers. Every figure traces back to the table it came from.