Comment

Carajás, Brazil · Impact review

Carajás Iron Mine Impacts: 7 Effects Measured From Space

Mining removed 19.6% of the original canga vegetation at Carajás by 2016, a satellite study found. Across the Amazon, mining-linked forest loss was 12 times greater than the clearing inside leases.

Authored by Farmonaut Research & Content Team
Reviewed and verified by Farmonaut Mining & Mineral Exploration Team
Updated 5 November 2026 · 15 min read
Sentinel-2 Satellite View Of S11D Iron Ore Mine, Serra Sul, Carajás, Pará, Brazil
Sentinel-2 true-colour composite of S11D iron ore mine, Serra Sul, Carajás, Pará, Brazil (6.39°S, 50.33°W). Contains modified Copernicus Sentinel data, 2022 to 2026, processed in Google Earth Engine.

Key takeaways

  • Canga vegetation fell from 144.2 km² to 115.9 km² by 2016 (PLOS ONE, 2019).
  • Across 50 Amazon leases, mining-linked forest loss reached 11,670 km² from 2005 to 2015, up to 70 km from leases (Nature Communications, 2017).
  • Two canga areas at Carajás hold 856 seed plant species, 24 of them endemic; surveys there cover 844 iron caves.
  • S11D’s 93% water saving and no-tailings-dam design are Vale’s 2016 design claims, not audited results.
  • Vale’s Northern System produced 170.2 million tonnes of iron ore in 2025 (Vale, January 2026).
In this article
  1. Carajás at a glance: the mine and its setting
  2. 1. Lost canga vegetation and soil on the plateaus
  3. 2. Water use, dry processing and downstream quality
  4. 3. Rare plants, invasive species and iron caves
  5. 4. Forest loss beyond the lease boundary
  6. 5. The railway, roads and new corridors
  7. 6. Towns, migration and rural land use
  8. 7. Economic weight: output, investment and what it pays for
  9. How to run a Carajás-style impact review on your own licence
  10. Frequently asked questions

The Carajás iron mine has seven main impacts: lost ironstone vegetation and soil, changed water use, pressure on rare species and caves, forest clearing beyond the mine, a long rail corridor, growing towns and migration, and large economic weight for the region. The measured numbers are specific. A satellite study published in PLOS ONE in 2019 found that mining removed 28.3 square kilometres (km²) of canga vegetation by 2016, which is 19.6% of the original 144.2 km². A 2017 study in Nature Communications found that across the Brazilian Amazon, mining-linked forest loss was 12 times larger than the clearing inside mining leases.

Those two results point in different directions, and that is why this mine is studied so closely. The pits themselves take a small area. The wider effects, such as new roads, growing towns and supply chains, reach much further. This page sets out each impact with the figure on record, says who reported it, and ends with what a mining consultant can take from Carajás to a licence area of their own. A calculator lets you test the off-lease effect on your own numbers.

Carajás at a glance: the mine and its setting

Carajás is a group of open-pit iron ore mines in Pará state, northern Brazil, owned and run by Vale S.A. Mining Technology calls it the world’s largest iron ore mine, and the Nature Communications authors describe it the same way. For the production detail, grades and expansion plans, see our separate guide to Carajás output, grade and expansion. This page deals with what the mine does to the land, water and people around it.

Three facts set the scene. First, most of the pits sit inside the Carajás National Forest, a protected unit of 391,263 hectares created by Decree 2,486 of February 1998, according to Brazil’s biodiversity agency ICMBio. Second, this type of protected area allows sustainable use, including mining under licence, as the PLOS ONE authors note. Third, the ore sits under a rare habitat. Canga is a hard iron-rich crust on plateau tops, with low savanna-like plants growing on it, surrounded by tall evergreen rainforest. Because the ore lies under the canga, mining and this habitat compete for the same ground.

Impact Measured figure Who measured it, and when
Canga vegetation lost 28.3 km², or 19.6% of 144.2 km² Souza-Filho and others, Landsat images 1973 to 2016
Land converted to mines 26.6 km² from forest and 24.3 km² from canga Same study, 1973 to 2016
Water use at S11D 93% lower by design, using dry processing Vale design claim, reported in 2016
Rare plants 856 seed plant species, 24 endemic Reported in the same study, for two canga areas
Caves 844 iron caves in the surveys reviewed Jaffé and others, 2016
Forest beyond the lease Mining effect detected up to 70 km out Sonter and others, Amazon-wide, 2005 to 2015
Rail corridor About 890 km to the port near São Luís Sonter and others, 2017
Output 170.2 million tonnes of iron ore in 2025 Vale production report, January 2026
The seven environmental impacts of mining, explained in plain words

▶ The seven environmental impacts of mining, explained in plain words · 2:31

1. Lost canga vegetation and soil on the plateaus

The most direct impact is simple: an open pit removes what grows above the ore. Researchers mapped this using Landsat satellite images from 1973, 1984, 2001 and 2016, which gives a record from before the mine was built. The canga originally covered 144.2 km² of the Carajás region. By 2016, 28.3 km² had been removed, leaving 115.9 km², or 80.4% of the original area.

Chart

Canga vegetation cleared by ridge, to July 2016

Square kilometres; the regional total was 28.3 km²

0.0 km²4.0 km²8.0 km²12.0 km²16.0 km²13.0 km²N48.9 km²N56.0 km²S11DSource: Souza-Filho and others, PLOS ONE, 2019.
0.0 km²4.0 km²8.0 km²12.0 km²16.0 km²13.0 km²N48.9 km²N56.0 km²S11DSource: Souza-Filho and others, PLOS ONE, 2019.
Three ridges account for nearly all of the clearing mapped up to 2016.

Inside the national forest the picture is sharper, because that is where the three largest iron ore mines are. Canga there fell from 105.2 km² to 77.2 km². The clearing was concentrated on three ridges: N4 lost 13.0 km², N5 lost 8.9 km² and S11D lost 6.0 km², all by July 2016. Readers should note that the study was funded by Instituto Tecnológico Vale, a non-profit research institute, and that one author worked for Vale. The paper declares both.

Why the timing of the clearing matters

Clearing is not steady over a mine’s life. At N4 and N5, the study found that canga loss from 1984 to 2001, the ramp-up period, ran at almost twice the pace of the later open-pit phase from 2001 to 2016. S11D lost its 6.0 km² in about three years of early construction. The authors expected the rate there to slow to about 0.2 km² a year over a mine life of about 30 years, but warned that faster demand for ore could shorten the mine life and raise the clearing rate. For a consultant, the lesson is that a baseline taken after construction starts has already missed the biggest change.

Slow recovery and land conversion

The canga plateaus are capped by hard ironstone, and the study’s authors say that knowledge of how canga plants grow and recover is still very limited. That matters for closure planning. Restoring a site is not the same task as stabilising it, as our guide to mining reclamation and restoration explains. Over the full 1973 to 2016 period, the mapped conversions to mine land were 26.6 km² from forest and 24.3 km² from canga, so the study area lost both habitats in similar amounts.

2. Water use, dry processing and downstream quality

Iron ore plants traditionally use a lot of water to wash and concentrate ore, and the fine waste goes to tailings dams (ponds that store the wet leftovers of processing). Vale’s answer at S11D was to process the ore at its natural moisture. Mining.com reported in December 2016 that this design was expected to cut water use by 93% and remove the need for tailings dams. Those were the company’s design figures at opening, not measured results from an independent audit, so they should be quoted that way.

On water quality, the 2019 PLOS ONE paper states that heavy-metal pollution in water from iron ore mining had not been detected in the Carajás region at that time, citing earlier work. That is a dated statement about the literature it reviewed. It does not replace site monitoring, because older pits and the newer dry plant run differently, and because rainfall in the eastern Amazon can move fine sediment off any bare ground. For a wider look at how mines affect streams and what studies measure, see our review of mining impacts on water.

A consultant reviewing a comparable site should ask for four things: the water balance by plant, not by company; sediment and turbidity readings upstream and downstream of each disturbed catchment; the year each monitoring point began; and which figures are design targets and which are measured.

3. Rare plants, invasive species and iron caves

Canga matters out of all proportion to its size. It covered just 2.5% of the national forest before mining, yet two locations in the study area, the Carajás National Forest and the Campos Ferruginosos National Park, hold 856 seed plant species. Of these, 24 are endemic, meaning they are found only in a restricted area. The same two places hold 17 exotic invasive plant species, most of them in the park. The paper names clearing for open pits, fire and invasive plants as the main threats to this vegetation.

Chart

Canga area before mining and in 2016

Square kilometres of canga vegetation

Before mining20160 km²37.5 km²75 km²112 km²150 km²Carajás region144 km²116 km²National Forest105 km²77.2 km²Source: Souza-Filho and others, PLOS ONE, 2019.
Before mining20160 km²37.5 km²75 km²112 km²150 km²Carajás region144 km²116 km²National Forest105 km²77.2 km²Source: Souza-Filho and others, PLOS ONE, 2019.
About four-fifths of the regional canga remained in 2016, the basis for the offset debate.

The answer so far has been to set aside other canga. In June 2017 Brazil created the Campos Ferruginosos National Park, which ICMBio lists at 79,086 hectares. The PLOS ONE paper says it covers the Bocaina and Tarzan ridges, holds about 24 km² of canga, or 21% of the regional total, and bars mining. It was financed through offsets for mining operations. An offset is a conservation measure elsewhere that makes up for damage at the mine site. Whether a separate ridge truly replaces a mined one depends on whether it holds the same species, and the authors treat the choice of offset areas as a research question.

Iron caves: the impact most outsiders miss

The ironstone is full of small caves, and these can stop a project. A 2016 study in PLOS ONE reviewed eight speleology reports covering 844 iron caves across Serra Norte, Serra Sul, Serra Leste and Serra da Bocaina. It found that larger caves held more species, and were more likely to hold troglobites (animals that live their whole lives underground) and resident bat colonies. Caves with guano, a sign of bat activity, gained species faster as they got larger.

The same study says Brazil has one of the strictest cave protection regimes in the world, requiring extensive surveys before licensing, and notes that troglobites have halted mining projects worldwide. The law it reviewed was Federal Decree 6640/2008, so check the current text with a Brazilian environmental lawyer before relying on it. For anyone planning work on ironstone, a cave survey belongs near the start of the schedule, not near the end.

4. Forest loss beyond the lease boundary

This is the largest measured effect, and the hardest to manage. Sonter and colleagues studied 50 large operating mining leases in the Brazilian Amazon, compared them with matched control areas, and looked at forest loss from 2005 to 2015. Their main results:

Chart

Forest cleared around 50 Amazon mining leases, 2005 to 2015

Square kilometres; all 50 leases, not Carajás alone

0 km²10,000 km²20,000 km²30,000 km²40,000 km²All clearing, lease + 70 km37,830 km²Linked to mining, in total11,670 km²Inside the leases983 km²Source: Sonter and others, Nature Communications, 2017.
0 km²10,000 km²20,000 km²30,000 km²40,000 km²All clearing, lease + 70 km37,830 km²Linked to mining, in total11,670 km²Inside the leases983 km²Source: Sonter and others, Nature Communications, 2017.
The clearing attributed to mining far exceeds what happened inside the leases.
  • Within the leases, 983 km² of forest was cleared, which is 14% of what stood there in 2005.
  • Mining-linked clearing in total reached 11,670 km², counting leases and zones up to 70 km around them.
  • That total is 12 times the on-lease figure, and 9% of all Amazon deforestation over the period.
  • Rates off the lease were highest in the 0 to 10 km and 40 to 50 km zones, at 4.4%.

These are Amazon-wide results across all 50 leases, not a Carajás-only measurement. The paper does treat Carajás as a case study, though, and says its operations drove clearing within 70 km of the leases through urban growth and infrastructure. It also says that iron ore creates demand for charcoal to make pig iron, which drives tree plantations and native forest loss.

The picture inside the protected area is different. The PLOS ONE paper says forests within the Carajás protected areas remain mostly undisturbed, helped by surveillance against fire, settlement and illegal gold digging. It argues that without this cover, the forest would likely have gone the way of the adjacent land, much of it now cropland or pasture. Both results can be true: strong protection inside the boundary, and heavy clearing pressure outside it.

Off-lease clearing calculator: what would the Amazon ratio mean for your site?

The calculator below applies a ratio to your own on-lease clearing. The default of 12 is the Amazon-wide figure from Sonter and colleagues. It is an average from one region and one decade, so treat the result as a scoping question for your impact study, not a forecast.

Calculator

How much clearing could a mine trigger beyond its lease?

Enter your planned on-lease clearing, a ratio and an offset multiple.

hectares

times

times
6,000 hatotal clearing to scope, at your ratio
  • Part beyond the lease (hectares): 5,500
  • Offset on on-lease clearing only (hectares): 1,000
  • Offset if off-lease clearing counts too (hectares): 12,000

Example values only, not a benchmark or forecast. The default ratio of 12 is the Amazon-wide figure from Sonter and others (Nature Communications, 2017) for 50 leases, 2005 to 2015; other regions differ. The offset multiple is your own assumption.

5. The railway, roads and new corridors

Ore leaves Carajás by rail, on a line of about 890 km to the deep-water terminal at Ponta da Madeira near São Luís, Maranhão. Railway Gazette records that the line opened in 1985 and is operated by Vale. Sonter and colleagues name the railway as a case where effects reach more than 70 km from the lease, and emerge only after a long period of operation. A long line is one reason lease-based impact studies can miss so much: the mine is in one state and the corridor runs on through another.

Corridors keep growing. S11D alone needed a 101 km rail branch, according to Mining.com, along with expansions of the main line and the port. Inside the mine, its truckless system replaced haul trucks with mobile crushers and 68 km of conveyor belts. Vale said this would cut fuel costs by 70%, again a design claim from 2016. Conveyors reduce diesel use, but they also create long fixed lines on the ground, which is why route choice matters for wildlife movement and drainage.

The next round of work is already licensed. Brazil’s environmental regulator Ibama issued, in September 2025, an operating licence for the Serra Sul +20 project, InfoMoney reported, which adds 20 million tonnes a year at S11D through new mining areas, a new semi-mobile crusher and a second long-distance conveyor. At that time, commissioning was scheduled to begin in the second half of 2026. New mining areas mean new clearing, so the 2016 canga figures above are a floor for today, not a current total.

6. Towns, migration and rural land use

Large mines draw people. Sonter and colleagues describe how the 2007 expansion at Carajás, a US$2.48 billion project, brought an influx of workers and further deforestation as nearby towns and facilities grew. They list urban expansion, support industries and new roads among the routes by which mining drives clearing off the lease. So the social and land effects are linked: the same growth that brings work and services also pushes on the forest edge.

Land use outside the boundary also shows up in the satellite record. Much of the adjacent land outside the protected areas has become pasture and cropland, the PLOS ONE paper notes, while the area inside the boundary stayed forested. For rural residents, this means two very different kinds of land sit side by side, with different rules on each.

Community questions at a site like this are rarely answered by a single number. A consultant should separate jobs by place of residence, count in-migration against local hiring, track the growth of towns on imagery from year to year, and record which land claims and traditional uses sit along the access corridors. The satellite part of that list is the easiest to do independently, because the record goes back decades.

7. Economic weight: output, investment and what it pays for

The economic impact starts with scale. Vale’s production report gives 170.2 million tonnes of iron ore from the Northern System in 2025, about half of the company’s 336 million tonnes. S11D produced a record 86.0 million tonnes, while Serra Norte and Serra Leste fell to 84.1 million tonnes, from 92.7 million tonnes in 2024. For context, USGS estimates Brazil’s 2025 iron ore mine output at 420 million tonnes, out of about 2.6 billion tonnes worldwide, though the two figures are not on the same basis. The USGS counts usable ore.

Chart

Northern System iron ore output, 2024 and 2025

Million tonnes of iron ore

202420250 Mt50 Mt100 Mt150 Mt200 MtSerra Norte + Leste92.7 Mt84.1 MtS11D83 Mt86 MtNorthern total176 Mt170 MtSource: Vale, Production and sales report 4Q25 (January 2026).
202420250 Mt50 Mt100 Mt150 Mt200 MtSerra Norte + Leste92.7 Mt84.1 MtS11D83 Mt86 MtNorthern total176 Mt170 MtSource: Vale, Production and sales report 4Q25 (January 2026).
S11D grew while the older northern pits fell.

Investment is also planned at a large scale. InfoMoney reports that Vale plans to invest R$70 billion through its Novo Carajás programme to 2030, aiming for 200 million tonnes a year from the complex and 32% more copper. Serra Sul +20 alone is estimated at US$2.8 billion. For how output and spending feed through to an economy in general, see our overview of the economic benefits of mining.

On climate, Vale reports company-wide figures: Scope 1 and 2 emissions of 7.8 million tonnes of CO2 equivalent in 2025, down 25.3% from 2017. Scope 1 is a company’s own fuel burn, and Scope 2 is the power it buys. These cover all of Vale’s operations, so they cannot be read as a Carajás number.

What mining does for an economy: exports, jobs and revenue

▶ What mining does for an economy: exports, jobs and revenue · 5:03

How to run a Carajás-style impact review on your own licence

Carajás is unusually well documented, and the methods used there work at far smaller sites. These five steps follow what the studies above actually did.

  1. Build the baseline from old imagery. Landsat goes back to the 1970s, so you can map the land before your project, even if no one surveyed it then.
  2. Draw two boundaries. Assess your lease, then assess a wider zone. The Amazon result suggests looking well beyond 10 km where roads and towns are likely to grow.
  3. Find the rare habitat first. Ironstone, caves and narrow-range plants can decide a project, so map them before you fix pit shells and access routes.
  4. Label each figure. Mark every number as a design target, a company report or an independent measurement, and keep the year beside it.
  5. Repeat on a schedule. Sentinel-2, a European Space Agency satellite mission with 10 m detail and a 5-day revisit, can show cleared ground, new roads and growing waste dumps between site visits.

Farmonaut’s satellite-based mineral detection works from this kind of imagery to give exploration targets, which sit ahead of a resource estimate and are not an environmental clearance. To discuss a licence area, use the mining query form or contact us. The image at the top of this page is a Sentinel-2 view of the S11D area.

Remote sensing for mining, explained in about two minutes

▶ Remote sensing for mining, explained in about two minutes · 2:01

Frequently asked questions

What are the main environmental impacts of the Carajás mine?

The main measured ones are loss of canga vegetation (28.3 km² by 2016), conversion of forest to mine land (26.6 km² from 1973 to 2016), pressure on rare plants and iron caves, and clearing linked to towns and infrastructure around the leases.

Is the Carajás mine inside a protected area?

Yes. The three largest iron ore mines are inside the Carajás National Forest, a 391,263 hectare unit where licensed mining is allowed under sustainable use rules.

How much vegetation has the Carajás mine removed?

About 19.6% of the region’s original canga by 2016, according to the satellite study in PLOS ONE. The figure predates recent expansion work, so today’s total will be higher.

What is canga?

Canga is a hard crust rich in iron that caps plateaus, with low savanna-like plants on top. At Carajás it forms islands in the rainforest and lies over the iron ore.

Does the Carajás iron mine cause deforestation in the Amazon?

Yes, on and off the lease. A 2017 Nature Communications study found mining-linked forest loss up to 70 km from leases across the Brazilian Amazon, and named Carajás among its case studies. Forest inside the protected areas stays mostly intact.

Does S11D use tailings dams?

S11D was designed to process ore at its natural moisture, which Vale said would remove the need for tailings dams and cut water use by 93%. Those were design figures reported in 2016.

How long is the Carajás railway?

About 890 km from the mine to Ponta da Madeira near São Luís, according to the Nature Communications paper. Railway Gazette lists a network length of 1,089 km and also mentions the 226 km Ferrovia Norte-Sul section that Vale runs.

What is the Campos Ferruginosos National Park?

It is a park created in 2017 and listed by ICMBio at 79,086 hectares. It covers the Bocaina and Tarzan ridges, holds about 21% of the region’s canga, and does not allow mining.

Why do caves matter for iron ore mining in Brazil?

Iron caves can hold species found nowhere else, and Brazilian law requires surveys and protection based on how important each cave is. A 2016 study at Carajás reviewed data on 844 caves.

Can satellites track the impact of a mine?

They can track what changes on the surface, such as cleared land, pits, dumps and roads. They cannot by themselves show the cause of a change or the quality of the water, so field monitoring is still needed.

Authored by: Farmonaut Research & Content Team

Reviewed and verified by: Farmonaut Mining & Mineral Exploration Team

Reviewed November 2026 against Souza-Filho and others (2019), Sonter and others (2017), Jaffé and others (2016), ICMBio unit pages, Vale’s 4Q25 report and USGS MCS 2026.

Farmonaut mining

Map your own project area from space

Mining consultants can send a licence boundary and get target zones from satellite data, ahead of a resource estimate.

Farmonaut Mining Pro 3D Target View (Sample, Synthetic Data)

Mining Pro sample view, synthetic data

Sources

  1. Souza-Filho and others, canga mapping, PLOS ONE (2019): journals.plos.org
  2. Sonter and others, Mining drives extensive deforestation in the Brazilian Amazon, Nature Communications (2017): nature.com
  3. Jaffé and others, cave biodiversity at Carajás, PLOS ONE (2016): journals.plos.org
  4. ICMBio, Floresta Nacional de Carajás: gov.br
  5. ICMBio, Parque Nacional dos Campos Ferruginosos: gov.br
  6. Vale, Production report 4Q25 (January 2026): filemanager-cdn.mziq.com
  7. Vale, Climate page: vale.com
  8. USGS, Mineral Commodity Summaries 2026, Iron Ore: pubs.usgs.gov
  9. Mining.com, S11D opening report (December 2016): mining.com
  10. Railway Gazette, Estrada de Ferro Carajás: railwaygazette.com
  11. InfoMoney, Vale licence for Serra Sul +20: infomoney.com.br
  12. ESA, Sentinel-2: esa.int
  13. Mining Technology, Carajás project profile: mining-technology.com



Discussion

Questions, corrections or field experience? Share them below. First-time comments are reviewed before they appear, and your email is never published.

No comments yet. Be the first to add a question or your experience.

Leave a comment

Farmonaut Farmonaut Trusted by 200,000+ users and 100+ businesses 200,000+ users trust us Get started