Physical land disturbance mapped from imagery (2019). Pollution is measured separately and is not included.
No mining footprint mapped here in the 2019 imagery.[3]
Outside known permit boundaries is not a finding of illegality: it can be a gap in the ESDM permit layer, an older or expired permit, or another land use.
Forest loss
Tree-cover loss 2001–2025, where tree cover in 2000 was at least 30 %. Annual satellite data (Hansen / University of Maryland).
Loss means tree cover was removed: clearing, but also plantation harvest or fire. It is not a finding of illegal deforestation, and the data is yearly, not monthly.
Permits
Visible with a free account · Permit list and numbersSign in
Kementerian ESDM geoportal (Minerba / Monaresia Wilayah Pertambangan), via legacy Eco Mining Data harvest · geoportal.esdm.go.id · retrieved 18 Oct 2026 · signal · unknown
Eco Mining Data computation: current permit geometries (ESDM) × area; overlap with WDPA protected areas · computed by Eco Mining Data · as of 18 Oct 2026 · signal · derived from the ESDM permit layer and WDPA (WDPA terms apply)Union of current permit geometries clipped to the subject; intersection with protected-area polygons; geodesic area
Cheng et al. 2026 · Mining land-use classes inside global mining polygons (South-Eastern Asia) · doi.org · as of 31 Dec 2022 · signal · CC BY-SA 4.0
Indonesia administrative regions (seed) · github.com · as of 1 Jan 2025 · confirmed · geometry: CC BY 3.0 IGO; codes: BPS public data (open; attribution: Badan Pusat Statistik)
Eco Mining Data computation: mining footprint (Cheng et al. 2026 · Maus et al. 2022 · ESMI digitised) × subject; inside and outside current ESDM permit geometries · doi.org · computed by Eco Mining Data · as of 31 Dec 2019 · signal · derived from Cheng et al. 2026 and Maus et al. 2022 (CC BY-SA 4.0), ESMI digitised land and the ESDM permit layerFootprint polygons clipped to the subject (geodesic area), counted once where sources overlap (Cheng > Maus > ESMI); the part inside the union of current permit geometries vs outside known permit boundaries (ESDM). Outside is not a finding of illegality: it can be a data gap or an older permit.
Hansen Global Forest Change GFC-2025-v1.13 · glad.earthengine.app · as of 31 Dec 2025 · confirmed · CC BY 4.0Hansen, M. C. et al. 2013. High-Resolution Global Maps of 21st-Century Forest Cover Change. Science 342: 850-53. Data: Global Forest Change 2000-2025 (v1.13), University of Maryland.
Model: comparable agriculture land Rp 9,108/m² minus expected reclamation ≈ Rp 15,054/m² (tin). Reclamation costs more than the land is worth: net −Rp 6,862/m², so its value is 0 (74% of model runs negative).
NJOP and compensation values are converted to market-equivalent and older years CPI-indexed. A fallback value is borrowed from a wider area and has a wider band and a lower grade. Not an appraisal. · landvalue.v1 · API (observations: Tier 1)
Value after restoration
What restoring abandoned mining land that is still bare or water could be worth, in rupiah. Figures for Kabupaten Bangka Barat.
Model estimate, not an appraisal.
MapBiomas mining-class measure
Every 30 m pixel that MapBiomas ever labelled as a mining pit and that is now abandoned and still bare or water.
Net gain
Rp 21.7 bnP10–P90 Rp 8.74–36.2 bngrade D
Net present value over 25 years at 10 % a year.
Land counted
464 ha · abandoned, still bare or water
Payback
2 yrs
Gain per rupiah of cost
0.66×
Footprint measure
Land inside the mapped mining footprint polygons that is abandoned and still bare or water.
No abandoned bare or water land in this measure here.
The two measures overlap. They are shown side by side and never added together. tn = trillion, bn = billion, m = million (rupiah). The letter after a figure is its grade (A best, D a rough model estimate); the range is P10–P90.
MapBiomas mining-class measure · Every 30 m pixel that MapBiomas ever labelled as a mining pit and that is now abandoned and still bare or water. 164 ha still bare and 300 ha water (pit lakes).
From value now to net gain
Value now → value after restoration, plus what restoration adds for everyone, minus the costs, gives the net gain.
Value now (degraded land)model value−Rp 31.9 bnP10–P90 −Rp 81.2 bn – Rp 54.9 bngrade D
Value after restorationmodel valueRp 38.3 bnP10–P90 Rp 22.4–90.4 bngrade D
The value now is negative: the land is worth less than nothing until it is reclaimed. That is shown but not counted as gain; the restoration costs are counted instead.
+Gain in land valuemodel valueRp 28.4 bnP10–P90 Rp 18.5–39.5 bngrade D
+Ecosystem servicesmodel valueRp 16.9 bnP10–P90 Rp 11.4–24 bngrade D
+CarbonRp 633 mP10–P90 Rp 161 m – Rp 1.34 bngrade D
+Avoided downstream impactsmodel valueRp 7.55 bnP10–P90 Rp 4.05–12.9 bngrade D
−Restoration costsRp 32.6 bnP10–P90 Rp 27.9–38.4 bngrade D
Shown, not added
Plan income (capitalised)model valueRp 23.2 bnP10–P90 Rp 15.8–34.5 bngrade D
Reclamation duty dischargedRp 78.2 bnP10–P90 Rp 50.4–114 bngrade D
By permit zone
Permit zone
Land
Net gain
Inside an expired or revoked permit: reclamation duty applies
–
–
Inside an active permit
267 ha
Rp 13 bnP10–P90 Rp 5.42–21.4 bngrade D
Inside a permit with another statusPending, suspended or unknown status.
–
–
Outside known permit boundaries (ESDM)This can be a gap in the ESDM permit layer or an older permit: the layer lists current permits only.
197 ha
Rp 8.64 bnP10–P90 Rp 3.27–14.8 bngrade D
Who benefits
How the benefits are shared, before costs (median path). Benefit-sharing defaults from the Restoration Studio: 80 % cooperative, 20 % village fund.
LandownerThe gain in land value.Rp 23.6 bngrade D45 %
CooperativePlan income above the land value and the community share of carbon (80 % stays in the cooperative).Rp 3.37 bngrade D6 %
Village fund20 % of the community share goes to the village fund.Rp 842 mgrade D2 %
Government and the publicEcosystem services and avoided downstream impacts.Rp 24.4 bngrade D46 %
Project developerThe rest of the carbon value.Rp 360 mgrade D1 %
Source and grade of every part
Part
Per hectare
Source
Value now (degraded land)model valueToday's use value of the degraded land: comparable land value minus the cost of reclamation. It can be negative. A negative value is shown but not counted as gain, because the restoration costs are counted below.
−Rp 68.7 m/haP10–P90 −Rp 175 m/ha – Rp 118 m/hagrade D
Eco Mining Data land value: degraded mining land = comparable value − reclamation cost (model value, not floored at zero)
Value after restorationmodel valueThe larger of the comparable land value and the capitalised plan income, never both.
Rp 82.5 m/haP10–P90 Rp 48.2–195 m/hagrade D
The larger of (b) the comparable land value (agriculture/plantation; pit lakes × water factor) and (a) the capitalised plan income
Plan income (capitalised)model valueIncome of a default restoration plan (forest, agroforestry, cassava fields, fish cages in pit lakes) over 25 years, discounted. Used in the value after restoration only when it is the larger of the two.
Rp 50.1 m/haP10–P90 Rp 34–74.3 m/hagrade D
Restoration Studio option economics (Sulawesi Utara price table applied nationally), default plan mix per commodity; present value of operating income over the plan period
=Net gainmodel valueRp 21.7 bnP10–P90 Rp 8.74–36.2 bngrade D
Value after restoration minus today's use value (a negative value now counts as zero).
Rp 61.1 m/haP10–P90 Rp 39.8–85.2 m/hagrade D
(value after − today's use value, a negative value counted as 0), realised in year 2 and discounted
Ecosystem servicesmodel valueNon-market value of the restored land, such as water regulation and habitat (benefit transfer from de Groot et al. 2012).
Rp 36.5 m/haP10–P90 Rp 24.5–51.8 m/hagrade D
de Groot, R. et al. 2012. Global estimates of the value of ecosystems and their services in monetary units. Ecosystem Services 1: 50-61 (biome means). Benefits transferred at 0.25-1× the mean (most likely 0.5). Cultivated land: no extra non-market value counted.
CarbonCarbon credits from replanting, after buffer, leakage and uncertainty deductions, at a placeholder price.
Rp 1.36 m/haP10–P90 Rp 346,287/ha – Rp 2.89 m/hagrade D
Restoration Studio carbon profiles × (1 − buffer − leakage − uncertainty) × carbon price, minus measurement and issuance costs
Avoided downstream impactsmodel valueAvoided sediment, flooding and water-treatment costs downstream. A model value, not a measurement.
Rp 16.3 m/haP10–P90 Rp 8.73–27.7 m/hagrade D
MODEL value: avoided sediment, flooding and water-treatment costs downstream of restored mining land, Rp/ha/yr. No Indonesian per-hectare study found; ESMI assumption, to be replaced by causal estimates.
RemediationRegrading, soil and water testing, labour, and POCO compost with microbes where metals are expected.
Rp 35.4 m/haP10–P90 Rp 30.4–41.3 m/hagrade D
Restoration Studio remediation costs: regrading, XRF and lab sampling, labour, POCO + microbes on the share of land with metals
PlantingSeedlings, compost and planting for the plan mix.
Rp 29.6 m/haP10–P90 Rp 24.6–36 m/hagrade D
Restoration Studio option costs (planting, seedlings, compost) of the plan mix
Restoration costsRemediation and planting, plus a contingency.
Rp 70.3 m/haP10–P90 Rp 60.1–82.7 m/hagrade D
Remediation + planting + contingency
Reclamation duty dischargedThe reclamation duty that restoration would discharge. It is a transfer between parties (who pays): shown, not added to the gain.
Rp 169 m/haP10–P90 Rp 109–245 m/hagrade D
Reclamation standard (Kepmen ESDM 344.K/MB.01/MEM.B/2025, IKK-adjusted): the duty restoration discharges. A TRANSFER (who pays), not added to the gain
Net gainmodel valueGain in land value + ecosystem services + carbon + avoided downstream impacts − restoration costs.
Rp 46.8 m/haP10–P90 Rp 18.8–77.9 m/hagrade D
Net gain = (value after − today's use value) + ecosystem services + carbon + avoided downstream impacts − costs; present value, model estimate, not an appraisal
Mining footprint: Cheng et al. 2026, Maus et al. 2022 (v2), ESMI digitised; counted once · 31 Dec 2022
Values in Rp (ecosystem services) are a data gap until ESVD valuation is connected. · il.v1
What is the EAI (Environmental Affected Index) of this area, and which component weighs most?
Answered
10.9EAI
EAI 10.9 (low). Heaviest component: Permits in protected areas. 56% of components have data.
For Kabupaten Bangka Barat
Grade C
Sources · Method
Eco Mining Data EAI v1 (Environmental Affected Index; method in STATUS-answers.md)
Mining footprint: Cheng et al. 2026, Maus et al. 2022 (v2), ESMI digitised; counted once · Maus, V. et al. (2022) Scientific Data 9, 289; Cheng et al. (2026) · CC BY 4.0 / CC BY-SA 4.0
Method: Sentinel-5P tropospheric NO2 column enhancement over a local background, mean of the newest 12 months (strongest of the regency mean and its industrial zones). Satellite column density, not a ground concentration and not a health exposure value
Partial answer: ISPU stations in the region; Sentinel-5P, CAMS and NEA are not connected yet.
Missing: CAMS (Copernicus atmosfeer: analyses en voorspellingen) · NEA Singapore (PSI/PM2.5 per uur, data.gov.sg) + ASMC (ASEAN-nevelkaarten)
Where do most new alerts come from this week/month?
Answered
0alerts
0 confirmed alerts in the last 30 days (0 in the last 7 days). Newest alert: 10 Aug 2026.
For Kabupaten Bangka Barat
Grade C
Data as of 10 Aug 2026 outdated
Details to verified organisations
Sources · Method · Partial answer
Eco Watch alerts (Eco Mining Data; confirmed after four-eyes review)
Method: Eco Watch alerts in the last 7 / 30 days; public count = confirmed only, unconfirmed signals counted for T3 in region scope
Partial answer: Eco Watch alerts (confirmed in public; unconfirmed signals counted for verified organisations only). RADD/GLAD and own detection are not live yet.
Social: Conflicts, displacement, safety (accidents, collapses at PETI sites)
Answered
0events
0 reported conflicts, accidents or disasters in 24 months (0 confirmed). Most frequent: –.
For Kabupaten Bangka Barat
Grade D
Data as of 18 Oct 2026
Sources · Method · Partial answer
Eco Mining Data news engine (headlines as published; events extracted, counted) · 18 Oct 2026
Method: Media-reported conflicts, accidents, disasters and spills (24 months); counted, mostly unconfirmed, no organisation named
Partial answer: Media-reported conflicts, accidents and disasters (counted, mostly unconfirmed, no organisations named); police and field data not yet.
Missing: GDELT (wereldwijd nieuws, geotagged) · Field reports
Mining footprint: Cheng et al. 2026, Maus et al. 2022 (v2), ESMI digitised; counted once · Maus, V. et al. (2022) Scientific Data 9, 289; Cheng et al. (2026) · CC BY 4.0 / CC BY-SA 4.0
Method: Hansen GFC loss inside Maus 2022 mining polygons, per province; all tree-cover loss (GFW) for comparison
Partial answer: Hansen forest loss inside vs outside permits for the Tijdmachine areas, provincial mining deforestation (Hansen × Maus) as context.
Missing: MapBiomas Indonesia (Auriga Nusantara): jaarlijkse landbedekking incl. mijnbouw
Regency mean compared with the background of its island group
Grade C
Satellite column density, not air at ground level. Sentinel-5P (TROPOMI) measures the amount of NO₂ and SO₂ in the whole air column. The figure is the ENHANCEMENT over the surrounding background, not a ground-level concentration and not a health exposure value. The zone is named by place; this is not a finding about any company.
0.05× 10¹⁵ molecules/cm² NO₂ above backgroundmean of the newest 12 months (to 2026-09), ± 0.01 (1 SE, statistical)No significant trend (-0.00 per year, p = 0.466, 9 years)
Monthly NO₂ enhancement (thin), 12-month running mean (bold), ± 1 SE band; 0 = background
SO₂ for this place is not available yet (sampled orbits still being processed).
The ±SE is statistical only; the retrieval (air-mass factor) uncertainty is 25–50 % for NO₂ and larger for small SO₂ columns. No wind data is used yet: the background ring covers all directions. Sources: KNMI TEMIS monthly TROPOMI NO₂ v1.0; Copernicus Sentinel-5P OFFL L2 SO₂ via Microsoft Planetary Computer.
