Country: Guayana

MAAP #226: AI to detect Amazon gold mining deforestation – 2024 update

Posted on by Matt Finer
Intro Image. Amazon Mining Watch interactive map.

As gold prices continue to increase, small-scale gold mining activity also continues to be one of the major deforestation drivers across the Amazon. 

It often targets remote areas, thus impacting carbon-rich primary forests. Moreover, in many cases, we presume that this mining is illegal based on its location within conservation areas (such as protected areas and Indigenous territories) and outside mining concessions.

Given the vastness of the Amazon, however, it has been a challenge to accurately and regularly monitor mining deforestation across all nine countries of the biome, in order to better inform related policies in a timely manner.

In a previous report (MAAP #212) we presented the initial results of the new AI-based dashboard (known as Amazon Mining Watch) designed to address the issue of gold mining and related policy implications. Amazon Mining Watch (AMW) is a partnership between Earth Genome, the Pulitzer Center’s Rainforest Investigations Network, and Amazon Conservation.

This online tool (see Intro Image) analyzes satellite imagery archives to estimate annual mining deforestation footprints across the entire Amazon, from 2018 to 2024 (Note 1). Although the data is not designed for precise area measurements,  it can be used to give timely estimates needed for management and conservation purposes.  

For example, the cumulative data can be used to estimate and visualize the overall Amazon-wide mining deforestation footprint, and the annual data can be used to identify trends and emerging new mining areas. The algorithm is based on 10-meter resolution imagery from the European Space Agency’s Sentinel-2 satellite and produces 480-meter resolution pixelated mining deforestation alerts.

The only tool of this kind to be truly regional (Amazon-wide) in coverage, AMW can also help foster regional cooperation, in particular in transfrontier areas where a lack of interoperability between official monitoring systems might hamper interventions.

The Amazon Mining Watch partnership is currently working to enhance the functionality and conservation impact of the dashboard, AMW will be a one-stop shop platform including real-time visualization of: 1) AI-based detection of mining deforestation across all nine Amazonian countries, with quarterly updates; 2) Hotspots of urgent mining cases, including river-based mining; and 3) the socio-environmental costs of illegal gold mining with the Conservation Strategy Fund (CSF) Mining Impacts Calculator.

Here, we present an update focused on the newly added 2024 data and its context within the cumulative dataset (since 2018).

MAJOR FINDINGS

In  the following sections, we highlight several major findings:

  • Gold mining is actively causing deforestation in all nine countries of the Amazon. This impact is concentrated in three major areas: southeast Brazil, the Guyana Shield, and southern Peru. In addition, mining in Ecuador is escalating.
  • The cumulative mining deforestation footprint in 2024 was over 2 million hectares (nearly 5 million acres) and has increased by over 50% in the past six years.
  • Over half of all Amazon mining deforestation occurred in Brazil, followed by Guyana, Suriname, Venezuela, and Peru.
  • While the cumulative footprint continues to grow, the rate of increase slowed in 2023 and 2024 after peaking in 2022, likely due to increased enforcement in Brazil.
  • Over one-third of the mining deforestation has occurred within protected areas and Indigenous territories, where much of it is likely illegal. We highlight the most impacted areas.
  • These results have important policy implications.
Base Map. Mining deforestation footprints, 2018-2024. Data: AMW, Amazon Conservation/MAAP.

Amazon & National Scale Patterns

The Base Map presents the gold mining footprint across the Amazon, as detected by the AMW algorithm. This data serves as our estimate of gold mining deforestation.

Yellow indicates the accumulated mining deforestation footprint for the years 2018- 2023; that is, all areas that the algorithm classified as a mining site vs other types of terrain, such as forest or agriculture. Red indicates the new mining areas detected in 2024.

Three major Amazon gold mining regions stand out: southeast Brazil (between the Tapajos, Xingu, and Tocantis Rivers), Guyana Shield (Venezuela, Guyana, Suriname, and French Guiana), and southern Peru (Madre de Dios).  In addition, Ecuador has emerged as an important mining deforestation front.

 

 

 

 

Graph 1. Amazon mining deforestation footprint. Data: AMW

Graph 1 quantifies the spatial data detected by the AMW algorithm. 

The cumulative mining deforestation footprint in 2024 was 2.02 million hectares (4.99 million acres)

For context, the initial mining deforestation footprint was around 970,000 hectares in 2018, the first year of Amazon Mining Watch data.

Between 2019 and 2024, we estimate that the gold mining deforestation grew by 1.06 million hectares (2.61 million acres).

Thus, over half (52.3%) of the cumulative footprint has occurred in just the past six years.

Note that while the cumulative footprint continues to grow, the rate of increase slowed in 2023 and 2024 after peaking in 2022.

 

 

 

Graph 2 shows that, of the total accumulated mining (2.02 million hectares), over half has occurred in Brazil (55.3%), followed by Guyana (15.4%), Suriname (12.4%), Venezuela (7.3%), and Peru (7.0%).

Graph 2. Gold mining deforestation across the Amazon, by country. Data: AMW, Amazon Conservation/MAAP

Graph 3 digs deeper into the AMW data, revealing additional trends between years. This data highlights the annual changes in detected mining deforestation. Note the trend across the entire Amazon at the top in green for overall context, followed by each country. Note that Brazil (orange line) accounts for much of the annual mining (over 50%).

In 2024, we documented the new gold mining deforestation of 111,603 hectares (275,777 acres). This total represents a decrease of 35% relative to the previous year 2023 and 45% relative to the peak year 2022.

The countries with the highest levels of new gold mining deforestation in 2024 were 1) Brazil (57,240 ha), 2) Guyana (19,372 ha), 3) Suriname (15,323 ha), 4) Venezuela (9,531 ha), and 5) Peru (6,020 ha). However, all five of these countries saw a major decrease in 2024, between 33% (Brazil and Suriname) and 46% (Peru).

Graph 3. Annual changes in new mining deforestation. Data: AMW
Figure 1. Protected areas & Indigenous territories impacted by mining deforestation. Data: AMW, ACA/MAAP.

Protected Areas & Indigenous Territories

We estimate that 36% of the accumulated mining deforestation in 2024 (over 725,000 hectares) occurred within protected areas and Indigenous territories (Figure 1; Note 2), where much of it is likely illegal.

Notably, the vast majority of this overall mining deforestation in protected areas and Indigenous territories has occurred in Brazil (88%).

 

 

 

 

 

 

 

 

 

Figure 2a. Top 10 impacted protected areas & Indigenous territories. Data: AMW, ACA/MAAP.

Figure 2a illustrates the top ten for both protected areas and Indigenous territories, in terms of both accumulated mining deforestation footprint and new mining deforestation in 2024. Figures 2b-d show zooms of the three main mining areas: southeast Brazil (2b), Guyana Shield (2c), and southern Peru (2d).

The top nine most impacted protected areas (in terms of accumulated footprint) are all in Brazil, led by Tapajós Environmental Protection Area. This area has lost over 377,000 hectares, followed by Amanã and Crepori National Forests, Rio Novo National Park, Urupadi, Jamanxim, and Itaituba National Forests, Jamanxim National Park, and Altamira National Forest. The top ten is rounded out by Yapacana National Park in Venezuela.

The three most impacted Indigenous territories are also in Brazil: Kayapó, Mundurucu, and Yanomami. Together, these three territories had a mining footprint of nearly 120,000 hectares. Fourth on the list is Ikabaru in Venezuela, followed by three in southern Peru (San Jose de Karene, Barranco Chico, and Kotsimba) with mining impact of over 17,000 hectares. Rounding out the top ten are Sai Cinza and Trincheira/Bacajá in Brazil, and San Jacinto in Peru.

We also estimate the expansion of over 38,000 hectares of new mining deforestation in protected areas and Indigenous territories in 2024. The protected area with the highest levels of new mining deforestation in 2024 was Tapajós Environmental Protection Area (nearly 19,000 hectares), followed by Amanã and Urupadi National Forests in Brazil, Rio Novo and Jamanxim National Parks in Brazil, Crepori National Forest in Brazil, Campos Amazonicos National Park in Brazil, Yapacan National Park in Venezuela, Guyane Regional Park in French Guiana, and Brownsberg Nature Reserve in Suriname.

Finally, the Indigenous territory with the highest levels of new mining deforestation in 2024 was Kayapó in Brazil (over 2,100 hectares), followed by Ikabaru in Venezuela, Yanomami, Aripuana, and Mundurucu in Brazil, Baramita in Guyana, Kuruáya in Brazil, Isseneru and Kamarang Keng, San Jose de Karene in Peu. It is worth noting that Kayapó, Mundurucu, and Yanomami territories in Brazil all experienced declines in the mining deforestation rate in 2024. For example, Yanomami went from its peak in 2021 to the lowest on record in 2024.

Most impacted areas in eastern Brazilian Amazon

Figure 2b. Most impacted areas in eastern Brazilian Amazon. Data: AMW, Amazon Conservation/MAAP.

Most impacted areas in the Guyana Shield

Figure 2c. Most impacted areas in the Guyana Shield. Data: AMW, Amazon Conservation/MAAP.

Most impacted areas in the southern Peruvian Amazon

Figure 2d. Most impacted areas in the southern Peruvian Amazon. Data: AMW, Amazon Conservation/MAAP.

Conclusion & Policy Implications

Despite a recent downward trend in the rate of gold mining deforestation, the cumulative gold mining deforestation footprint continues to grow across the Amazon.

Our analysis shows that over one-third of this mining occurs within protected areas and Indigenous territories, the vast majority in Brazil. However, since the return of the Lula administration in 2023, Brazil has been ramping up enforcement efforts. This has contributed to the rapid decrease in area lost to mining across the Amazon, given Brazil’s outsized contribution to regional figures. This highlights again the importance of protected areas and Indigenous territories as a crucial policy instrument for the protection of the region’s ecosystems.

Although advances have been made in reducing illegal mining from protected areas in southern Peru, it continues to impact several Indigenous territories (MAAP #208, MAAP #196), particularly those surrounding the government-designated Mining Corridor. In fact, the most affected Indigenous territory in Peru, San Jose de Karene, has already lost over a third of its total area to illegal gold mining.  These territories are part of a regional organization known as FENAMAD, which has been supporting legal actions to help the government make decisions for a rapid response to illicit activities (such as illegal mining) that affect indigenous territories. This process led to five government-led operations between 2022 and 2024, in three communities: Barranco Chico, Kotsimba and San José de Karene (MAAP #208).

In Ecuador, mining deforestation continues to threaten numerous sites, including protected areas and Indigenous territories, along the Andes-Amazon transition zone (MAAP #206, MAAP #221, MAAP #219). An upcoming series of reports will detail these threats.

AMW is an emerging and powerful new tool, but it does have some caveats. One is that any mining activity less than 500 square meters may not be accurately detected. For example, we have been monitoring small-scale mining in several protected areas, such as Madidi National Park in Bolivia and Puinawai National Park in Colombia, that are not yet detected by the algorithm. In these cases, direct real-time monitoring with satellites is still needed. These areas will soon be added to the AMW as mining “Hotspots” (MAAP#197).

This is also the case for river-based mining that does not cause a large footprint on the ground. Imagery with very high resolution has revealed active river barge mining in northern Peru (MAAP #189) and along the Colombia/Brazil border (MAAP#197). These areas will also soon be added to the AMW as mining “Hotspots.”

Gold mining in the Amazon is certain to stay a major issue in the coming years as gold prices continue to skyrocket, reaching over $3,000 an ounce in April 2025, driven by global economic uncertainty. While there are encouraging signs of effective enforcement in Brazil, governments here and across the region will have to compete with this rising financial incentive for mining activities.

Tools such as the Amazon Mining Watch, which will eventually publish quarterly updates of newly detected mining deforestation areas, can help governments, civil society, and local community defenders spot new fronts of gold mining and take action in near real-time. In a feature developed by Conservation Strategy Fund (CSF), it will also evaluate the economic costs of socio-environmental mining damages necessary for communities and managers to declare punitive damages.

The only dashboard of this kind to be fully regional in coverage, the AMW can also help foster regional cooperation, in particular in transfrontier areas where a lack of interoperability between official monitoring systems might hamper interventions that are aimed at combating a phenomenon that is linked to other nature crimes and is mostly controlled by international organized crime. 

In the coming years, the MAAP and AMW teams will continue to publish both quarterly and annual reports of the dynamic mining situation in each country and across the Amazon, in addition to confidential reports directly to governments and community leaders on the most urgent cases.

Notes

1. Note that in this report, we focus on mining activity that causes deforestation. The vast majority is artisanal or small-scale gold mining, but other mining activities have also been detected, such as iron, aluminum, and nickel mines in Brazil and Colombia. Additional critical gold mining areas in rivers that are not yet causing deforestation (such as in northern Peru, southeast Colombia, and northwest Brazil; see MAAP #197), are not included in this report. This information is not yet displayed in Amazon Mining Watch, but future updates will include river-based mining hotspots. 

2. Our data source for protected areas and Indigenous territories is from RAISG (Amazon Network of Georeferenced Socio-Environmental Information), a consortium of civil society organizations in the Amazon countries. This source (accessed in December 2024) contains spatial data for 5,943 protected areas and Indigenous territories, covering 414.9 million hectares across the Amazon.

Acknowledgments

We thank colleagues from partner organizations around the Amazon for helpful comments on the report, including: Earth Genome, Conservación Amazónica (ACCA & ACEAA) & Federación Nativa del Río Madre de Dios y Afluentes (FENAMAD), Fundación EcoCiencia, Fundación para la Conservación y el Desarrollo Sostenible (FCDS), and Instituto Centro de Vida (ICV) & Instituto Socioambiental (ISA).

This report was made possible by the generous support of the Gordon and Betty Moore Foundation.

MAAP #225: Carbon in the Amazon (part 4): Protected Areas & Indigenous Territories

Posted on by Matt Finer
Figure 1. Total aboveground carbon change, Amazon protected areas & Indigenous territories 2013-2022. Data: Planet, ACA/MAAP.

We continue our ongoing series about carbon in the Amazon.

In part 1 (MAAP #215), we introduced a new dataset (Planet’s Forest Carbon Diligence) with wall-to-wall estimates for aboveground carbon at an unprecedented 30-meter resolution between 2013 and 2022. In part 2 (MAAP #217), we highlighted which parts of the Amazon are currently home to the highest (peak) carbon stocks. In part 3 (MAAP #220), we showed key cases of carbon loss (deforestation) and gain across the Amazon.

A key finding from this series is that the Amazon biome is teetering between a carbon source and sink. That is, historically the Amazon has functioned as a critical sink, with its forests accumulating carbon if left undisturbed. However, relative to the 2013 baseline, the Amazon flipped to a source during the high deforestation, drought, and fire seasons of 2015-2017. It then rebounded as a narrow carbon sink in 2022.

Here, in part 4, we focus on the importance of aboveground carbon in protected areas and Indigenous territories, which together cover 49.5% (414.9 million hectares) of the Amazon biome (see Figure 1).

We find that, as of 2022, Amazonian protected areas and Indigenous territories contained 34.1 billion metric tons of aboveground carbon (60% of the Amazon’s total). Importantly, in the ten years between 2013 and 2022, they functioned as a significant carbon sink, gaining 257 million metric tons.

With this data, we can also analyze aboveground carbon for each protected area and Indigenous territory. For example, Figure 1 illustrates aboveground carbon loss vs. gain for each protected area and Indigenous territory during the 10-year period of 2013 – 2022 (see details below).

Below, we further explain and illustrate the key findings.

Amazon-wide & Country-level Results

Amazonian protected areas and Indigenous territories currently cover nearly half (49.5%) of the Amazon biome, but contain 60% of the aboveground carbon. Together they contained 34.1 billion metric tons of aboveground carbon as of 2022, gaining 257 million metric tons since 2013, thus functioning as a carbon sink (Figure 2).1,2 

In contrast, areas outside of protected areas and Indigenous territories (424 million hectares) contained 22.6 billion metric tons of aboveground carbon as of 2022, losing 255 million metric tons since 2013, thus functioning as an overall carbon source.

Thus, the carbon sink function of protected areas and Indigenous territories narrowly offsets the emissions in the rest of the Amazon.

We emphasize that the protected areas and Indigenous territories functioned as a significant carbon sink (p-value = 0.01), while the outside areas were not a significant source (p-value= 0.15).

Regarding results by country, protected areas and Indigenous territories were significant carbon sinks in Colombia, Brazil, Suriname, and French Guiana (Guyana gained carbon but not significantly). In contrast, they were significant carbon sources in Bolivia and Venezuela (Peru and Ecuador lost carbon but not significantly).

Figure 2. Amazon aboveground carbon 2013-2022, within vs. outside protected areas and Indigenous territories. Data: Planet, ACA/MAAP.

Individual Protected Area & Indigenous Territory Results

Figure 1 (see above) illustrates total aboveground carbon loss vs. gain for each protected area and Indigenous territory during the 10-year period of 2013 – 2022. 

Overall, we found 1,103 areas that served as significant carbon sinks (dark green) during this period (238 protected areas and 865 Indigenous territories). These areas are concentrated in the northern and central Amazon. See Annex 1 for a list of specific areas that were significant carbon sinks.

It is important to note that deforestation pressures currently threaten several of these significant carbon sinks, including Chiribiquete National Park and Nukak-Maku Indigenous Reserve in Colombia, Sierra del Divisor National Park in Peru, and Canaima National Park in Venezuela.

In contrast, we found 1,439 areas (156 protected areas and 1,283 Indigenous territories) that served as significant carbon sources. It is important to note that some areas with little documented deforestation, such as Alto Purus National Park, may have carbon loss from natural causes.

Figure 3. Total aboveground carbon stocks in each protected area and Indigenous territory. Data: Planet, ACA/MAAP.

Figure 3 offers the most recent snapshot of total aboveground carbon stocks in each protected area and Indigenous territory.

It presents data for 2022 categorized into three groups of High, Medium, and Low. Note that the highest carbon totals (over 330 million metric tons) are concentrated across the large designated areas of the northern Amazon.

These High and Medium carbon areas may be considered to have the highest overall conservation value purely in terms of total carbon.

See Annex 1 for specific areas with the highest carbon stocks as of 2022.

 

 

 

 

 

 

 

Figure 4. Aboveground carbon density in each protected area and Indigenous territory (2022). Data: Planet, ACA/MAAP

Finally, Figure 4 also displays the most recent data (2022) in each protected area and Indigenous territory, but standardized for area (aboveground carbon/hectare).

Note that the highest carbon totals (over 50 metric tons per hectare) are more evenly concentrated across the Amazon.

These High and Medium carbon areas may be considered to have the highest carbon conservation value per hectare.

 

 

 

 

 

 

 

 

 

Policy Implications:
Unlocking the Climate Value of Protected Areas and Indigenous Territories in the Amazon

Policy and finance for tropical forests as a climate solution have largely focused on reducing emissions from deforestation and forest degradation (REDD+). These efforts have made important strides in slowing and directing finance to tackle forest loss, particularly in high-deforestation regions. However, this emphasis on avoided emissions overlooks a critical component of the global carbon cycle: the carbon sink function (gaining of carbon over time) of primary tropical forests — which this analysis using Planet’s Forest Carbon Diligence data show is both measurable and significant.

This omission leaves a major flux in the carbon system—ongoing carbon sequestration in old-growth forests—outside the scope of existing market or non-market incentives. Critically, many of these carbon-absorbing forests are already located within established protected areas and indigenous territories. These areas are globally recognized for their importance in biodiversity conservation and for the stewardship provided by Indigenous Peoples and local communities. 

As global attention increasingly turns to engineered carbon removal strategies such as BECCS (Bioenergy with carbon capture and storage) and Direct Air Capture, there is an urgent need to recognize that Amazonian forests are already performing this function—naturally and at scale. Yet the value of Protected Areas and Indigenous territories as a potent carbon sink is neither monetized nor rewarded under current frameworks, unless they can demonstrate that they are under threat from deforestation or degradation in order to access REDD+ finance. An emerging exception is the High Integrity Forests Investment Initiative (HIFOR), which recognizes the value of carbon sequestration in old-growth forests, but does not generate tradable credits for each ton absorbed.5 The Tropical Forests Forever Fund (TFFF) proposed by Brazil for adoption at COP 30, would also reward forest countries at a rate of approximately US$ 4.00/year for every hectare of tropical forest they protect, regardless of whether they are under threat.6

To date, however, protected areas and Indigenous territories, despite their proven climate contribution, often lack the financial support necessary to ensure long-term effectiveness and resilience. As a result, they often face chronic underfunding,7 limiting their long-term effectiveness and resilience. Policy innovation is needed to close this gap and integrate the carbon sink function of mature forests into funding mechanisms for forest protection. Doing so would unlock meaningful incentives for the continued, long-term stewardship of these high-carbon ecosystems and would ensure that one of the planet’s most effective natural climate solutions receives the attention and resources it deserves.

Annex 1

Specific areas that were significant carbon sinks include:

Otishi, Sierra del Divisor, Güeppí-Sekime and Yaguas National Parks, Matsés, and Pucacuro National Reserves, Ashaninka Communal Reserve, and Cordillera Escalera and Alto Nanay- Pintuyacu Chambira Regional Conservation Area, Matses, Pampa Hermosa, and Yavarí – Tapiche Indigenous Reserves, and Kugapakori, Nahua, Nanti Territorial Reserve in Peru;

Amacayacu, Chiribiquete, Cahuinari, Rio Pure, and Yaigoje Apaporis National Parks, Nukak Natural Reserve, Amazonas Forest Reserve, and Putumayo and Nukak-Maku, Yaigoje Rio Apaporis and Vaupes Indigenous Reserve in Colombia;

Campos Amazônicos, Juruena, Mapinguari, Nascentes do Lago Jari, Serra do Divisor, and Montanhas do Tumucumaque National Parks, Amanã, Aripuanã, Crepori, Tapajós, and Tefé National Forests in Brazil, Itaituba and Jatuarana National Forests, and Alto Rio Negro, Baú, Aripuanã, Aripuanã, Apyterewa, Mundurucu, and Vale do Javari Indigenous Territories in Brazil.

Achuar Indigenous Territory and Zona Intangible Tagaeri – Taromenane in Ecuador; Manuripi Heath National Reserve and Takana, Takana II, and Yuracare Indigenous Reserves in Bolivia; Central Suriname and Sipaliwini Nature Reserves in Suriname; Canaima National Park in Venezuela; and Parc Amazonien de Guyane National Park in French Guiana, 

Specific areas with the highest carbon stocks, as of 2022, include:

Alto Purús, Manu, Sierra del Divisor, and Cordillera National Parks in Peru; Chiribiquete National Park in Colombia; Montanhas do Tumucumaque, Pico da Neblina, Jaú, and Juruena National Parks and Yanomami, Menkragnoti, Kayapó, Mundurucu, and Vale do Javari Indigenous Territories in Brazil; Caura and Canaima National Parks in Venezuela; and Parc Amazonien de Guyane National Park in French Guiana;

Methodology

We analyzed Planet Forest Carbon Diligence, a cutting-edge new dataset from the satellite-based company Planet, featuring a 10-year historical time series (2013 – 2022) with wall-to-wall estimates for aboveground carbon density at 30-meter resolution.3,4

One notable caveat of this data is that it does not distinguish aboveground carbon loss from natural vs human-caused drivers, so additional information may be incorporated to understand the context of each area. 

Based on these data, annual aboveground carbon values ​​were estimated in Amazonian protected areas and Indigenous territories to obtain a time series for 2013-2022. In addition, the Mann-Kendall test was used to analyze trends in the generated time series.

Our data source for protected areas and Indigenous territories is from RAISG (Amazon Network of Georeferenced Socio-Environmental Information), a consortium of civil society organizations in the Amazon countries. This source (accessed in December 2024) contains spatial data for 5,943 protected areas and Indigenous territories, covering 414.9 million hectares across the Amazon.

We determined that many of these areas (4,000) did not include creation date metadata, prohibiting any time-series control for that variable. Instead, we used the most current extent of protected areas and Indigenous territories as a proxy for those that existed from 2013 to 2022.

There was substantial overlap between protected areas and Indigenous territories, but we accounted for this to avoid double counting of the overlapping areas.

The aboveground carbon values for protected areas and Indigenous territories were calculated for each country and then summed across the Amazon.

The remaining areas were combined into the category of “Outside protected areas and Indigenous territories” and also calculated for each country and summed across the Amazon.

Our geographic range for the Amazon is a hybrid designed for maximum inclusion: biogeographic boundary (as defined by RAISG) for all countries, except for Bolivia and Peru, where we use the watershed boundary, and Brazil, where we use the Legal Amazon boundary. Our area estimate for this definition of the Amazon biome is 839.2 million hectares.

Notes

1 Breaking down the results by category, protected areas contained nearly 21.1 billion metric tons of aboveground carbon as of 2022, gaining over 204 million metric tons since 2013, while Indigenous territories contained over 16.8 billion metric tons of aboveground carbon as of 2022, gaining over 132 million metric tons since 2013. Note that protected areas and Indigenous territories overlap in many areas.

2 Standardizing for area (that is, calculating the results per hectare), protected areas and Indigenous territories contained 82.2 metric tons of aboveground carbon per hectare as of 2022, gaining a net 0.6 metric tons per hectare since 2013. In contrast, areas outside of protected areas and Indigenous territories contained 53.2 metric tons of aboveground carbon per hectare as of 2022, losing a net 0.6 metric tons per hectare since 2013.

3 Anderson C (2024) Forest Carbon Diligence: Breaking Down the Validation and Intercomparison Report. https://www.planet.com/pulse/forest-carbon-diligence-breaking-down-the-validation-and-intercomparison-report/

4 In terms of the limitations of Planet’s Forest Carbon Diligence data, Duncanson et al (2025) recently wrote a Letter in Science focused on spatial resolution for forest carbon maps. Given the natural constraint of the size of a tree, they discuss the challenge of pixel-level validation below 5 meters for forest carbon monitoring. The authors state that spatial resolution should at minimum exceed the crown diameter of a typical large tree, which is about 20 meters for tropical forests. In this sense, the 30-meter product exceeds this limitation.

Duncanson et al (2025) Spatial resolution for forest carbon maps. Science 387: 370-71.

5 WCS High Integrity Forest Investment Initiative (HIFOR): The Science Basis

6 https://www.bloomberg.com/news/newsletters/2025-04-04/too-big-to-fell-brazil-takes-trees-to-wall-street?cmpid=BBD040425_GR

7 UNEP-WCMC, IUCN, and NGS. (2022). Protected Planet Report 2022. Cambridge, UK: UNEP-WCMC.

Acknowledgments

Through a generous sharing agreement with the satellite company Planet, we have been granted access to this data across the entire Amazon biome for the analysis presented in this series.

We thank colleagues from the following organizations for helpful comments on this report: Planet, Conservación Amazónica – ACCA, Conservación Amazónica -ACEAA, Gaia Amazonas, Ecociencia, and Instituto del Bien Común.

We especially thank colleagues at Conservación Amazónica – ACCA for help with the 10-year data analysis.

This report was made possible by the generous support of the Norwegian Agency for Development Cooperation (NORAD)

Citation

Bodin B, Finer M, Castillo H, Mamani N (2025) Carbon in the Amazon (part 4): Protected Areas & Indigenous Territories. MAAP: 225.

MAAP #220: Carbon across the Amazon (part 3): Key Cases of Carbon Loss & Gain

Posted on by dcadmin
Graph 1. The Amazon biome functions as a narrow carbon sink from 2013 to 2022. Data: Planet, ACA/MAAP.

In part 1 of this series (MAAP #215), we introduced a critical new dataset (Planet’s Forest Carbon Diligence) with wall-to-wall estimates for aboveground carbon at an unprecedented 30-meter resolution between 2013 and 2022. This data uniquely merges machine learning, satellite imagery, airborne lasers, and a global biomass dataset from GEDI, a NASA mission.

In part 2 (MAAP #217), we highlighted which parts of the Amazon are currently home to the highest (peak) aboveground carbon levels and the importance of protecting these high-integrity forests (see Annex 1).

Here, in part 3, we focus on aboveground carbon loss and gain across the Amazon over the 10 years for which we have data (2013-22; see Base Map below).

The Amazon loses carbon to the atmosphere due to deforestation, logging, human-caused fires, and natural disturbances, while it gains carbon from forest regeneration and old-growth forests continuing to sequester atmospheric carbon.4

Overall, we find that the Amazon still narrowly functions as a carbon sink (meaning the carbon gain is greater than the loss) during this period, gaining 64.7 million metric tons of aboveground carbon between 2013 and 2022 (see Graph 1).

This finding underscores the importance of both primary and secondary forests in countering widespread deforestation. Moreover, it highlights the critical potential of primary forests to continue accumulating carbon if left undisturbed.

This gain, however, is quite small relative to the total 56.8 billion metric tons of aboveground carbon contained in the Amazon biome (that is, a gain of just +0.1%), reinforcing concerns that the Amazon could flip to a carbon source in the coming years (with carbon loss becoming greater than its gain) due to increasing deforestation, degradation, and fires.1  See Annex 2 for more details, including how the Amazon became a carbon sink following the 2015 drought, but since rebounded.

The countries with the largest carbon gain are 1) Brazil, 2) Colombia, 3) Suriname, 4) Guyana, and 5) French Guiana. In contrast, the countries with the greatest carbon loss are 1) Bolivia, 2) Venezuela, 3) Peru, and 4) Ecuador.

Zooming in to the site level yields additional insights. For example, we can now estimate the carbon loss from major deforestation events across the Amazon from 2013 to 2022. On the flip side, we can also calculate the carbon gain from both secondary and primary forests.

Areas with carbon gain in intact areas indicate excellent candidates for the High Integrity Forest (HIFOR) initiative, a new financing instrument uniquely focused on maintaining intact tropical forests.2 Importantly, a HIFOR unit represents a hectare of high-integrity tropical forest within a high-integrity landscape that has been “well-conserved” for over a decade.3  Intact areas with carbon gain between 2013-22 may indicate decadally “well-conserved” areas that can be overlapped with areas of high ecological integrity.

Below, we illustrate these findings with a series of novel maps zooming in on emblematic cases of large carbon loss and gain across the Amazon from 2013 – 2022. These cases include forest loss driven by agriculture, gold mining, and roads, as well as forest gain in remote primary forests.

Base Map – Amazon Carbon Loss & Gain (2013-2022)

The Base Map shows wall-to-wall estimates of aboveground carbon loss and gain across the Amazon between 2013 and 2022.

Carbon loss is indicated by yellow to red, indicating low to high carbon loss. Carbon gain is indicated by light to dark green, indicating low to high carbon gains.

Below, we present a series of notable cases of high carbon loss and gain indicated in Insets A-I.

Base Map. Areas of major carbon loss and gain across the Amazon between 2013 and 2022. Source: Amazon Conservation/MAAP, Planet.

Emblematic Cases of Carbon Loss & Gain

Figure 1 highlights emblematic cases of carbon loss (Insets A-F in red) and carbon gain (Insets G-I in green). Below we highlight a series of emblematic cases.

Figure 1. Emblematic cases of carbon loss and gain across the Amazon. Source: Amazon Conservation/MAAP, Planet.

Carbon Loss

We can now estimate the carbon loss from major deforestation events across the Amazon during the past ten years, directly from a single dataset. These cases include forest loss from agriculture, gold mining, and roads. Note that the presented values represent just the carbon loss featured in the selected area.

A. Colombia – Arc of Deforestation

Figure 1A. Carbon loss in the Colombian Amazon’s arc of deforestation. Source: Amazon Conservation/MAAP, Planet.

Figure 1A shows the extensive carbon emissions (39.5 million metric tons) associated with the major deforestation within and surrounding protected areas and Indigenous territories in the Colombian Amazon‘s arc of deforestation.

The carbon loss within the protected areas and Indigenous territories is likely from illegal deforestation.

See MAAP #211 for more details.

 

 

 

 

 

 

 

 

 

B. Peru – Mennonite Colonies

Figure 1B. Carbon loss by new Mennonite colonies in the Peruvian Amazon. Source: Amazon Conservation/MAAP, Planet.

Figure 1B shows the carbon emissions of 224,300 metric tons associated with the recent deforestation carried out by new Mennonite colonies arriving in the central Peruvian Amazon starting in 2017.

See MAAP #188 for more details, including information regarding the legality of  the deforestation causing the carbon loss.

 

 

 

 

 

 

 

 

 

 

C. Peru – Gold Mining

Figure 1C. Carbon loss associated with gold mining deforestation in  southern Peruvian Amazon. Source: ACA/MAAP, Planet.

Figure 1C shows the extensive carbon emissions (11.3 million metric tons) associated with gold mining deforestation in the southern Peruvian Amazon.

Most of the carbon loss within the protected areas (and their buffer zones) and Indigenous territories is likely from illegal deforestation.

See MAAP #208 for more information, including details regarding the legality of the deforestation causing the carbon loss.

 

 

 

 

 

 

 

 

 

D. Brazil – Road BR-364

Figure 1D. Carbon loss along BR-364 in the southwest Brazilian Amazon. Source: ACA/MAAP, Planet.

Figure 1D shows the carbon emissions along road BR-364 that crosses the state of Acre in the southwest Brazilian Amazon.

This road was opened in the 1960s and paved in the 1980s.

 

 

 

 

 

 

 

 

 

 

 

E. Brazil – Road BR-319

Figure 1E. Carbon loss along paved roads. Source: ACA/MAAP, Planet.

Figure 1E shows a controversial road paving project that would effectively link the arc of deforestation to the south with more intact forests to the north in Amazonas and Roraima states.

Note that the current carbon loss is concentrated along the paved roads.

The paving of road BR-319 has recently caused headlines as President Luiz Inácio Lula da Silva recently authorized the paving of 20 km of the road and plans to bid for an additional 32 km (thus, paving of 52 km in total).

Modeling studies predict extensive new deforestation from this road construction, and thus additional associated carbon loss.

 

 

 

 

 

 

 

 

F. Brazil – Road BR-163

Figure 1F. Carbon loss along BR-163 in the eastern Brazilian Amazon. Source: ACA/MAAP, Planet.

Figure 1F shows the extensive carbon emissions (71.4 million metric tons) along a recently paved stretch of road BR-163 which crosses the state of Pará in the eastern Brazilian Amazon.

Importantly, this stretch of road has been presented as a case study of what may happen along road BR-319 if it is paved.

 

 

 

 

 

 

 

 

 

 

 

Carbon Gain

We can also calculate the carbon gain from both secondary and primary forests. These cases include forest gain from remote primary forests that may be good candidates for the HIFOR initiative.

Figure 1G. Carbon gains in the southeast Colombian Amazon. Source: ACA/MAAP, Planet.

G. Southeast Colombia

Figure 1G shows the carbon gain of over 52.5 million metric tons in the remote southeast Colombian Amazon.

This area is anchored by three national parks and several large indigenous territories.

 

 

 

 

 

 

 

 

 

 

Figure 1H. Carbon gains along the border of eastern Ecuador and northern Peru. Source: ACA/MAAP, Planet.

H. Ecuador – Peru border

Figure 1H shows the carbon gain of nearly 40 million metric tons along the border in eastern Ecuador and northern Peru.

Note this area is anchored by numerous protected areas, including Yasuni National Park in Ecuador and Pucacuro National Reserve in Peru, and Indigenous territories.

 

 

 

 

 

 

 

 

 

Figure 1I. Carbon gains in the tri-border region of the northeast Amazon. Source: ACA/MAAP, Planet.

I. Northeast Amazon

Figure 1I shows the carbon gain of 164.7 million metric tons in the tri-border region of the northeast Amazon (northern Brazil, French Guiana, and Suriname).

For example, note the carbon gains in Montanhas do Tumucumaque National Park and Tumucumaque Indigenous territory in northeast Brazil.

Also note that this was an Amazonian “peak carbon area,” as described in MAAP #217.

 

 

 

 

 

 

 

 

 

Annex 1

Annex 1. Peak carbon areas in relation to the carbon loss and gain data. Source: Amazon Conservation/MAAP, Planet.

In part 2 of this series (MAAP #217), we highlighted which parts of the Amazon are currently home to the highest (peak) aboveground carbon levels.

Annex 1 shows these peak carbon areas in relation to the carbon loss and gain data presented above.

Note that both peak carbon areas (southeast and northeast Amazon) are largely characterized by carbon gain.

 

 

 

 

 

 

 

 

 

Annex 2

Annex 2. Amazon biome functions as a narrow carbon sink from 2013 to 2022, but became a source in between. Data: Planet, ACA/MAAP.

Annex 2 shows all ten years of aboveground carbon data grouped by two-year intervals (thus, it is an extension of Graph 1 above, adding data for the intermediate years).

In this context, black indicates our baseline of 2013-14, red indicates a decrease from the baseline (carbon source), and green indicates an increase from the baseline (carbon sink).

Importantly, there was a decrease in aboveground carbon from 2015-18, which likely reflects the severe droughts of 2015 and 2016 and subsequent severe fire seasons of 2016 and 2017. Aboveground carbon rebounded from 2019-22.

This trend supports the hypothesis that the Amazon biome is teetering on being an aboveground carbon source vs sink.

It also raises the possibility that the Amazon may return to being a carbon source following the intense drought and fires of 2024.

.

.

Notes

1 In part 1 of this series (MAAP #215), we found the Amazon “is still functioning as a critical carbon sink”. As pointed out in a companion blog by Planet, however, the net carbon sink of +64 million metric tons is quite small relative to the total estimate of 56.8 billion metric tons of aboveground carbon across the Amazon. That is a net positive change of just +0.1%. As the blog notes, that’s a “very small buffer” and there’s “reason to worry that the biome could flip from sink to source with ongoing deforestation.”

2 High Integrity Forest (HIFOR) units are a new, non-offset asset that recognizes and rewards the essential climate services and biodiversity conservation that intact tropical forests provide, including ongoing net removal of CO2 from the atmosphere. HIFOR rewards the climate services that intact tropical forests provide, including ongoing net carbon removal from the atmosphere, and complements existing instruments to reduce emissions from deforestation and degradation (REDD+) by focusing on tropical forests that are largely undegraded. A HIFOR unit represents a hectare of well-conserved, high-integrity tropical forest where ‘well-conserved’ means that high ecological integrity is maintained over a decade of monitoring as part of equitable, effective management of a site and ‘high ecological integrity’ means a score of >9.6 on the Forest Landscape Integrity Index. For more information see https://www.wcs.org/our-work/climate-change/forests-and-climate-change/hifor

3 Two additional important references regarding HIFOR methodology and application:

High Integrity Forest Investment Initiative, Methodology for HIFOR units, April 2024. Downloaded from https://www.wcs.org/our-work/climate-change/forests-and-climate-change/hifor

Forest Landscape Integrity Index metric used by HIFOR: www.forestintegrity.com

4 In Planet’s Forest Carbon Diligence product, carbon loss and gain are detected via changes in canopy cover and canopy height during the given periods (in this case, 2013 vs 2022).

Acknowledgments

Through a generous sharing agreement with the satellite company Planet, we have been granted access to this data across the entire Amazon biome for the analysis presented in this series.

We also thank D. Zarin (WCS) for helpful comments regarding the implications of our findings for the HIFOR initiative.

This report was made possible by the generous support of the Norwegian Agency for Development Cooperation (NORAD)

Citation

Finer M, Mamani N, Anderson C, Rosenthal A (2024) Carbon across the Amazon (part 3): Key Cases of Carbon Loss & Gain. MAAP: 220.

MAAP #217: Carbon across the Amazon (part 2): Peak Carbon Areas

Posted on by dcadmin
Figure 1. Example of peak carbon areas in southern Peru and adjacent western Brazil. Data: Planet.

In part 1 of this series (MAAP #215), we introduced a critical new resource (Planet Forest Carbon Diligence) that provides wall-to-wall estimates for aboveground carbon density at an unprecedented 30-meter resolution. This data uniquely merges machine learning, satellite imagery, airborne lasers, and a global biomass dataset from GEDI, a NASA mission.4

In that report, we showed that the Amazon contains 56.8 billion metric tons of aboveground carbon (as of 2022), and described key patterns across all nine countries of the Amazon biome over the past decade.

Here, in part 2, we focus on the peak carbon areas of the Amazon that are home to the highest aboveground carbon levels.

These peak carbon areas correspond to the upper one-third of aboveground carbon density levels (>140 metric tons per hectare).1

They likely have experienced minimal degradation (such as selective logging, fire, and edge/fragmentation effects)2 and are thus a good proxy for high-integrity forests.

Figure 1 shows an important example of peak carbon areas in southern Peru and adjacent western Brazil.

The peak carbon areas are often found in the remote primary forests of protected areas and Indigenous territories, but some are located in forestry concessions (specifically, logging concessions) or undesignated lands (also referred to as undesignated public forests).

Our goal in this report is to leverage unprecedented aboveground carbon data to reinforce the importance of these designated areas and draw attention to the remaining undesignated lands.

For example, peak carbon areas would be excellent candidates for the High Integrity Forest (HIFOR) initiative, a new financing instrument that uniquely focuses on maintaining intact tropical forests.3 HIFOR rewards the climate services that intact tropical forests provide, including ongoing net carbon removal from the atmosphere, and complements existing instruments to reduce emissions from deforestation and degradation (REDD+) by focusing on tropical forests that are largely undegraded.

Below, we detail the major findings and then zoom in on the peak carbon areas in the northeast and southwest Amazon.

Peak Carbon Areas in the Amazon   

The Base Map below illustrates our major findings.

The peak carbon areas (>140 metric tons per hectare; indicated in pink) are concentrated in the southwest and northeast Amazon, covering 27.8 million hectares (11 million ha in the southwest and 16.8 million ha in the northeast).
k

Base Map. Planet Forest Carbon Diligence across the Amazon biome for the year 2022. Data: Planet.

In the southwest Amazon, peak carbon levels are found in southern & central Peru, and adjacent western Brazil.

In the northeast Amazon, peak carbon levels are found in northeast Brazil, much of French Guiana, and parts of Suriname.

By country, Brazil and Peru have the largest area of peak carbon (10.9 million and 10.1 million hectares respectively), followed by French Guiana (4.7 million ha), and Suriname (2.1 million ha).

Protected areas and Indigenous territories cover much (61%) of the peak carbon area (16.9 million hectares).

The remaining 39% remains unprotected, and arguably threatened, in undesignated lands (9.4 million hectares) and forestry concessions (1.5 million ha), respectively.

In addition, high carbon areas (>70 metric tons per hectare; indicated by the greenish-yellow coloration in the Base Map) are found in all nine countries of the Amazon biome, notably Colombia, Ecuador, Bolivia, Venezuela, and Guyana.

Southwest Amazon

­Southern Peru

Figure 2a. Peak carbon area in the southern Peruvian Amazon. Data: Planet, SERNANP, RAISG.

Figure 2a zooms in on the peak carbon area covering 7.9 million hectares in southern Peru (regions of Madre de Dios, Cusco, and Ucayali) and adjacent southwest Brazil (Acre).

Several protected areas (such as Manu and Alto Purús National Parks, and Machiguenga Communal Reserve) anchor this area.

It is also home to numerous Indigenous territories (such as Mashco Piro, Madre de Dios, and Kugapakori, Nahua, Nanti & Others Indigenous Reserves).

 

 

 

 

 

 

 

 

 

 

Figure 2b highlights the major land designations within the peak carbon area of southern Peru.

Figure 2b. Peak carbon areas (outlined in pink), categorized by land designation in southern Peru and adjacent western Brazil. Data: Planet, NICFI, SERNANP, SERFOR, RAISG.

Protected areas and Indigenous territories cover 77% of this area (green and brown, respectively).

The remaining 23% could be considered threatened, as they are located in forestry concessions or undesignated lands (orange and red, respectively). Thus, these areas are ideal candidates for increased protection to maintain their peak carbon levels.

 

 

 

 

 

 

 

 

 

 

 

Central Peru

Figure 3a. Peak carbon area in the central Peruvian Amazon. Data: Planet, SERNANP, RAISG.

Figure 3a zooms in on the peak carbon area in the central Peruvian Amazon, which covers 3.1 million hectares in the regions of Ucayali, Loreto, Huánuco, Pasco, and San Martin.

Several protected areas (including Sierra del Divisor, Cordillera Azul, Rio Abiseo, and Yanachaga–Chemillén National Parks, and El Sira Communal Reserve) anchor this area.

It is also home to numerous Indigenous territories (such as Kakataibo, Isconahua, and Yavarí Tapiche Indigenous Reserves).

 

 

 

 

 

 

 

 

 

 

Figure 3b. Peak carbon areas (outlined in pink), categorized by land designation in central Peru. Data: Planet, NICFI, SERNANP, SERFOR, RAISG.

Figure 3b highlights the major land designations within the peak carbon area of central Peru.

Protected areas and Indigenous territories cover 69% of this area (green and brown, respectively).

The remaining 31% could be considered threatened, as they are located in forestry concessions or undesignated lands (orange and red, respectively), and are ideal candidates for increased protection.

 

 

 

 

 

 

 

 

 

 

 

 

 

Northeast Amazon

Figure 4a. Peak carbon area in the tri-border region of the northeast Amazon. Data: Planet, RAISG.

Figure 4a zooms in on the peak carbon area in the tri-border region of the northeast Amazon, which covers 16.8 million hectares in northern Brazil, French Guiana, and Suriname.

Several protected areas (including Montanhas do Tumucumaque National Park in northeast Brazil, Amazonien de Guyane National Park in French Guiana, and Central Suriname Nature Reserve) anchor this area.

It is also home to numerous Indigenous territories (such as Tumucumaque, Rio Paru de Este, and Wayãpi in northeast Brazil).

 

 

 

 

 

 

Figure 4b. Peak carbon areas (outlined in pink), categorized by land designation in northeast Amazon. Data: Planet, NICFI, RAISG.

Figure 4b highlights the major land designations within the peak carbon area of the northeast Amazon.

Protected areas and Indigenous territories cover just over half (51%) of this area (green and brown, respectively).

The remaining 49% could be considered threatened, as they are located in undesignated lands, and are ideal candidates for increased protection.

 

 

 

 

 

 

 

 

 

Notes

1 We selected this value (upper 33%) to capture the highest aboveground carbon areas and include a range of high carbon areas. Additional analyses could target different values, such as the highest 10% or 20% of aboveground carbon.

2  A recent paper documented a strong relationship between selective logging and aboveground carbon loss (Csillik et al. 2024, PNAS). The link between forest edges and carbon is presented in Silva Junior et al, Science Advances.

3 High Integrity Forest (HIFOR) units are a new tradable asset that recognizes and rewards the essential climate services and biodiversity conservation that intact tropical forests provide, including ongoing net removal of CO2 from the atmosphere. For more information see https://www.wcs.org/our-work/climate-change/forests-and-climate-change/hifor

4 For more information, see the “What is Forest Carbon Diligence?” section in this recent blog from Planet.

Citation

Finer M, Mamani N, Anderson C, Rosenthal A (2024) Carbon across the Amazon (part 2): Peak Carbon Areas. MAAP #217.

Acknowledgments

This report was made possible by the generous support of the Norwegian Agency for Development Cooperation (NORAD)

MAAP #215: Unprecedented Look at Carbon across the Amazon (part 1)

Posted on by dcadmin
Figure 1. Example of Planet Forest Carbon Diligence, focused on southern Peru and adjacent western Brazil.

The Amazon biome has long been one of the world’s largest carbon sinks, helping stabilize the global climate.

Precisely estimating this carbon, however, has been a challenge. Fortunately, new satellite-based technologies are providing major advances, most notably NASA’s GEDI mission (see MAAP #213) and, most recently, Planet Forest Carbon Diligence.1

Here, we focus on the latter, analyzing Planet’s cutting-edge new dataset, featuring a 10-year historical time series (2013 – 2022) with wall-to-wall estimates for aboveground carbon density at 30-meter resolution.

As a result, we can produce high-resolution aboveground carbon maps and estimates for anywhere and everywhere across the vast Amazon (see Figure 1).

Through a generous sharing agreement with Planet, we have been granted access to this data across the entire Amazon biome for the analysis presented in the following three-part series:

  1. Estimate and illustrate total aboveground forest carbon across the Amazon biome in unprecedented detail (see results of this first report, below).
    j
  2. Highlight which parts of the Amazon are home to the highest aboveground carbon levels, including protected areas and Indigenous territories (see second report, MAAP #217).
    l
  3. Present emblematic deforestation cases that have resulted in the highest aboveground carbon emissions across the Amazon (see third report, MAAP #220).

Major Results

Carbon across the Amazon

Based on our analysis of Planet Forest Carbon Diligence, we estimate that the Amazon contained 56.8 billion metric tons of aboveground carbon, as of 2022 (see Base Map). Applying a standard root-to-shoot ratio conversion (26%), this estimate increases to 71.5 billion metric tons of above and belowground carbon. This total is equivalent to nearly two years of global carbon dioxide emissions at the peak 2022 level (37.15 billion metric tons).5

The peak carbon levels are largely concentrated in the southwest Amazon (southern Peru and adjacent western Brazil) and northeast Amazon (northeast Brazil, French Guiana, and Suriname).

Base Map. Planet Forest Carbon Diligence across the Amazon biome.

Total Carbon by Country

As shown in Graph 1, countries with the most aboveground carbon are 1) Brazil (57%; 32.1 billion metric tons), 2) Peru (15%; 8.3 billion metric tons), 3) Colombia (7%; 4 billion metric tons), 4) Venezuela (6%; 3.3 billion metric tons), and 5) Bolivia (6%; 3.2 billion metric tons). These countries are followed by Guyana (3%; 2 billion metric tons), Suriname (3%; 1.6 billion metric tons), Ecuador (2%; 1.2 billion metric tons), and French Guiana (2%; 1.1 billion metric tons).

Overall, we documented the total gain of 64.7 million metric tons of aboveground carbon across the Amazon during the ten years between 2013 and 2022.2 In other words, the Amazon is still functioning as a critical carbon sink.

The countries with the most aboveground carbon gain over the past ten years are 1) Brazil, 2) Colombia, 3) Suriname, 4) Guyana, and 5) French Guiana. Note that we show Brazil as a carbon sink (gain of 102.8 million metric tons), despite other recent studies showing it as a carbon source.3 Also note the important gains in aboveground carbon across several key High Forest cover, Low Deforestation (HFLD) countries, namely Colombia, Suriname, Guyana, and French Guiana.4

In contrast, the countries with the most aboveground carbon loss over the past ten years are 1) Bolivia, 2) Venezuela, 3) Peru, and 4) Ecuador.

Graph 1. Planet Forest Carbon Diligence data across the Amazon biome, comparing 2013-14 with 2021-22. Note that a “+” symbol indicates that the country gained aboveground carbon, while a “-“ symbol indicates that the country lost aboveground carbon.

Carbon Density by Country

Standardizing for area, Graph 2 shows that countries with the highest aboveground carbon density (that is, aboveground carbon per hectare as of 2021-22) are located in the northeast Amazon: French Guiana (134 metric tons/hectare), Suriname (122 metric tons/hectare), and Guyana (85 metric tons/hectare). Ecuador is also high (94 metric tons/hectare).

Note that countries in the northeast Amazon (French Guiana, Suriname, and Guyana) have lower total aboveground carbon due to their smaller size (Graph 1), but high aboveground carbon density per hectare (Graph 2). This also applies to Ecuador.

Graph 2. Planet Forest Carbon Diligence data for aboveground carbon density by country across the Amazon, comparing 2013-14 with 2021-22. Note that a “+” symbol indicates that the country gained aboveground carbon, while a “-“ symbol indicates that the country lost aboveground carbon.

Notes & Citations

1 Anderson C (2024) Forest Carbon Diligence: Breaking Down The Validation And Intercomparison Report. https://www.planet.com/pulse/forest-carbon-diligence-breaking-down-the-validation-and-intercomparison-report/

2 In terms of uncertainty, the data contains pixel-level estimates, but not yet at national levels. To minimize annual uncertainty at the country level, we averaged 2013 and 2014 for the baseline and 2021 and 2022 for the current state.

3 Recently, in MAAP #144, we showed Brazil as a carbon source, based on data from 2001 to 2020. In contrast, Planet Forest Carbon Diligence is based on data from 2013 to 2022. Thus, one interpretation of the difference is that most carbon loss occurred in the first decade of the 2000s, which is consistent with historical deforestation data showing peaks in the early 2000s. It also highlights the likely importance of the interplay between forest loss/degradation (carbon loss) and forest regeneration (carbon gain) in terms of whether a country is a carbon source or sink during a given timeframe.

4 HFDL, or “High Forest cover, Low Deforestation” describes countries with both a) high forest cover (>50%) and low deforestation rates (<0.22% per year). For more information on HFDL, see https://www.conservation.org/blog/what-on-earth-is-hfld-hint-its-about-forests

5 Annual carbon dioxide (CO₂) emissions worldwide from 1940 to 2023

Citation

Finer M, Mamani N, Anderson C, Rosenthal A (2024) Unprecedented Look at Carbon across the Amazon. MAAP  #215.

Acknowledgments

This report was made possible by the generous support of the Norwegian Agency for Development Cooperation (NORAD)

 

MAAP #214: Agriculture in the Amazon: New data reveals key patterns of crops & cattle pasture

Posted on by dcadmin
Figure 1. Agricultural and pasture data in a section of the Brazilian Amazon.

A burst of new data and online visualization tools are revealing key land use patterns across the Amazon, particularly regarding the critical topic of agriculture. This type of data is particularly important because agriculture is the leading cause of overall Amazonian deforestation.

These new datasets include:

  • Crops. The International Food Policy Research Institute (IFPRI), a leading agriculture and food systems research authority, recently launched the latest version of their innovative crop monitoring product, the Spatial Production Allocation Model (SPAM).1 This latest version, developed with support from WRI’s Land & Carbon Lab, features spatial data for 46 crops, including soybean, oil palm, coffee, and cocoa. This data is mapped at 10-kilometer resolution across the Amazon and updated through 2020.2
    j
  • Cattle pasture. The Atlas of Pastures,3 developed by the Federal University of Goiás, facilitates access to data regarding Brazilian cattle pastures generated by MapBiomas. This data is mapped at 30-kilometer resolution and updated through 2022. We use Collection 5 from Mapbiomas for the rest of the Amazonian countries.4
    j
  • Gold mining. New mining data is included for additional context. Amazon Mining Watch uses machine learning to map open-pit gold mining.5 This data is mapped at 10-kilometer resolution across the Amazon and updated through 2023.

We merged and analyzed these new datasets to provide our first overall estimate of Amazonian land use, the most detailed effort to date across all nine countries of the biome. Figure 1 shows an example of this merged data in a section of the Brazilian Amazon.

Below, we present and illustrate the following major findings across the Amazon, and then zoom in on several regions across the Amazon to show the data in greater detail.

Major Findings

The Base Map illustrates several major findings detailed below.

Base Map. Overview of the merged datasets noted above for crops, pasture, and gold mining. Double-click to enlarge. Data: IFRI/SPAM, Lapig/UFG, Mapbiomas, AMW, ACA/MAAP.

1) Crops
We found that 40 crops in the SPAM dataset overlap with the Amazon, covering over 106 million hectares (13% of the Amazon biome).

Soybean covers over 67.5 million hectares, mostly in southern Brazil and Bolivia. Maize covers slightly more area (70 million hectares) but we consider this a secondary rotational crop with soy (thus, there is considerable overlap between these two crops).

Oil palm covers nearly 8 million hectares, concentrated in eastern Brazil, central Peru, northern Ecuador, and northern Colombia.

In the Andean Amazon zones of Peru, Ecuador, and Colombia, cocoa covers over 8 million hectares and the two types of coffee (Arabica and Robusta) cover 6.7 million hectares.

Other major crops across the Amazon include rice (13.8 million hectares), sorghum (10.9 million hectares), cassava (9.8 million hectares), sugarcane (9.6 million hectares), and wheat (5.8 million hectares).

2) Cattle Pasture
Cattle Pasture covers 76.3 million hectares (9% of the Amazon biome). The vast majority (92%) of the pasture is in Brazil, followed by Colombia and Bolivia.

3) Crops & Cattle Pasture
Overall, accounting for overlaps between the data, we estimate that crops and pasture combined cover 115.8 million hectares. This total is the equivalent of 19% of the Amazon biome.

In comparison, open-pit gold mining covered 1.9 million hectares (0.23% of the Amazon biome).

Zooms across the Amazon

Eastern Brazilian Amazon

Figure 2 shows the transition from the soy frontier to the cattle pasture frontier in the eastern Brazilian Amazon. Also note a mix of other crops, such as oil palm, sugarcane, and cassava, and some gold mining.

Figure 2. Eastern Brazilian Amazon. Data: IFRI/SPAM, Lapig/UFG, Mapbiomas, AMW, ACA/MAAP.

Andean Amazon (Peru and Ecuador)

Figure 3. Andean Amazon. Data: IFRI/SPAM, Lapig/UFG, Mapbiomas, AMW, ACA/MAAP.

The land use patterns are quite different in the Andean Amazon regions of Peru and Ecuador.

Figure 3 shows, that instead of soy and cattle pasture, there is instead oil palm, rice, coffee, and cocoa.

Also note the extension of the cattle pasture frontier in the western Brazilian Amazon, towards Peru and Bolivia.

 

 

 

 

 

 

 

 

 

 

 

 

Northeast Amazon (Venezuela, Guyana, Suriname, French Guiana)

Figure 4 shows the general lack of crops in the core Amazon regions Guyana, Suriname, and French Guiana, which is surely a major factor they are all considered High Forest cover, Low Deforestation countries (HFLD). In contrast, note there is abundant gold mining activity throughout this region.

Figure 4. Northeastern Amazon. Data: IFRI/SPAM, Lapig/UFG, Mapbiomas, AMW, ACA/MAAP.

Methods

For the SPAM data, we used the physical area, which is measured in a hectare and represents the actual area where a crop is grown (not counting how often production was harvested from it). We only considered values ​​greater than or equal to 100 ha per pixel.

For the Base Map, due to their importance as primary economic crops, we layered soybean and oil palm as the top two layers, respectively. From there, crops were layered in order of their total physical area across the Amazon. Thus, the full extensions of some crops are not shown if they overlap pixels with other crops that have greater physical area. For overlaps with crops and pasture, we favored the crops.

Notes & Data Sources

1 International Food Policy Research Institute (IFPRI), 2024, “Global Spatially-Disaggregated Crop Production Statistics Data for 2020 Version 1.0” https://doi.org/10.7910/DVN/SWPENT, Harvard Dataverse, V1

Spatial Production Allocation Model (SPAM)
SPAM 2020 v1.0 Global data (Updated 2024-04-16)

2 Note that the spatial resolution is rather low (10-kilometers) so all crop coverage data above should be interpreted as referential only.

3 The Atlas of Pastures (Atlas das Pastagens), open to the public, was developed by the Image Processing and Geoprocessing Laboratory of the Federal University of Goiás (Lapig/UFG), to facilitate access to results and products generated within the MapBiomas initiative, regarding Brazilian pastures.

https://atlasdaspastagens.ufg.br/

4 MapBiomas Collection 5;  https://amazonia.mapbiomas.org/en/

5 See MAAP #212 for more information on Amazon Mining Watch.

Citation

Finer M, Ariñez A (2024) Agriculture in the Amazon: New data reveals key patterns of crops & cattle pasture. MAAP: 214.

MAAP #199: Amazon Carbon Update, based on NASA’s GEDI Mission

Posted on by Lauren deVilla

As we approach the COP28 climate summit, starting in Dubai in late November, we provide here a concise update on the current state of remaining Amazon carbon reserves.

We present the newly updated version of NASA’s GEDI data1, which uses lasers aboard the International Space Station to provide cutting-edge estimates of aboveground biomass density on a global scale.

Here, we zoom in on the Amazon and take a first look at the newly updated data, which covers the time period of April 2019 – March 2023.2

This data, which is measured in megagrams of aboveground biomass per hectare (Mg/ha) at a 1-kilometer resolution, serves as our estimate for aboveground carbon reserves.

Figure 1 displays aboveground biomass across the Amazon biome. Note the highest carbon densities (indicated in bright yellow) are located in both the northeast Amazon and southwest Amazon.

Aboveground Biomass across the Amazon

Figure 2 also displays aboveground biomass across the Amazon biome, but this time with country boundaries and labels added.

Note that the peak biomass concentrations in the northeast Amazon include Suriname, French Guiana, and the northeast corner of Brazil. The peak biomass concentrations in the southwest Amazon are centered in southern Peru. Also note that many parts of Ecuador, Colombia, Venezuela, Guyana, Bolivia, Brazil, and northern Peru have high carbon densities as well.

Figure 2. Aboveground biomass density (carbon estimate) across the Amazon biome, with country boundaries. Data: NASA/GEDI, NICFI.

Carbon Estimates

We calculated over 78 billion metric tons of aboveground biomass across the Amazon biome (78,184,161,090 metric tons to be exact). Using a general assumption that 48% of this biomass is carbon3, we estimate over 37 billion metric tons of carbon across the Amazon (37,528,397,323 metric tons).

Note that these totals are likely underestimates given that the laser-based data has not yet achieved full coverage across the Amazon (that is, there are many areas where the lasers have not yet recorded data, leaving visible blanks in the maps above).

This is consistent with a previous study based on another independent dataset, where we estimated 6.7 billion metric tons of carbon in the Peruvian Amazon as of 2013 (MAAP #148). The current GEDI data estimates at least 5.3 billion metric tons in the Peruvian Amazon.

Carbon Sink

In a previous report, we showed that the Brazilian Amazon has become a net carbon source, whereas the total Amazon is still a net carbon sink (MAAP #144). Our current report goes one step further in terms of showing just how much carbon is left in that sink.

Notes

1GEDI L4B Gridded Aboveground Biomass Density, Version 2.1. https://daac.ornl.gov/cgi-bin/dsviewer.pl?ds_id=2299

2Note that we previously reported on the initial data release, which covered the time period of April 2019 – August 2021 (see MAAP #160).

3Domke et al (2022) How Much Carbon is in Tree Biomass?. USDA/Forest Service.

Acknowledgements

This work was supported by NORAD (Norwegian Agency for Development Cooperation) and ICFC (International Conservation Fund of Canada).

Citation

Mamani N, Finer M, Ariñez A (2022) Amazon Carbon Update, based on NASA’s GEDI Mission. MAAP: 199.

MAAP #197: Illegal Gold Mining Across the Amazon

Posted on by Matt Finer
Example of major gold mining zone in the Peruvian Amazon. Data: Planet.

Illegal Gold Mining continues to be one of the major issues facing nearly all Amazonian countries.

In fact, following the recent high-level summit of the Amazon Cooperation Treaty Organization, the nations’ leaders signed the Belém Declaration, which contains a commitment to prevent and combat illegal mining, including strengthened regional and international cooperation (Objective 32).

Illegal gold mining is a major threat to the Amazon because it impacts both primary forests and rivers, often in remote and critical areas such as protected areas & indigenous territories.

That is, illegal gold mining is both a major deforestation driver and a source of water contamination (especially mercury) across the Amazon.

Previously, in MAAP #178, we presented a large-scale overview of the major gold mining deforestation hotspots across the entire Amazon biome. We found that gold mining is actively causing deforestation in nearly all nine countries of the Amazon.

Here, we update this analysis with two important additions. First, we add to the overview major gold mining operations taking place in rivers, in addition to those causing deforestation (see Figure 1).

Second, we present a new map of likely illegal gold mining sites, based on information from partners and location with protected areas and indigenous territories (see Figure 2).

Finally, we show a series of high-resolution satellite images of key examples of illegal Amazon gold mining.

Updated Amazon Gold Mining Map

Figure 1 is our updated Amazon gold mining map.

The orange dots indicate areas where gold mining is currently causing deforestation of primary forests. The blue dots indicate areas where gold mining is occurring in rivers. Combined, we documented 58 active forest and river-based mining sites across the Amazon.

The dots outlined in red indicate the mining sites that are likely illegal, for both forest and river-based mining. We found at least 49 cases of illegal mining across the Amazon, the vast majority of the active mining sites noted above.

Note the concentrations of illegal mining causing deforestation in southern Peru, across eastern Brazil, and across Ecuador. Similarly, note the concentrations of illegal mining in rivers in northern Peru and adjacent Colombia and Brazil.

Figure 1. Updated Amazon gold mining map. Data: ACA/MAAP. Click to enlarge.

Protected Areas & Indigenous Territories

Figure 2 adds protected areas and indigenous territories. We found at least 36 conflictive overlaps: 16 in protected areas and 20 in indigenous territories. We also found an additional two conflicts with Brazilian National Forests.

We highlight a number of high-conflict zones. For protected areas: Podocarpus National Park in Ecuador; Madidi National Park in Bolivia; Canaima, Caura, and Yapacana National Parks in Venezuela. We note that the Peruvian government has been effectively minimizing invasions in protected areas in the southern region of Madre de Dios (Tambopata National Reserve and Amarakaeri Communal Reserve).

For indigenous territories: Kayapo, Menkragnoti, Yanomami, and Mundurucu in Brazil; Pueblo Shuar Arutam in Ecuador, and a number of communities in southern Peru.

Figure 2. Amazon gold mining map., with protected areas and indigenous territories. Data: ACA/MAAP, RAISG. Click to enlarge.

Methods

The forest-based mining sites displayed in Figure 1 are largely based on information obtained over the last several years of our deforestation monitoring work. The river-based sites are largely based on information obtained from partners in country and on the ground.

We complemented this information with automated, machine-based data from Amazon Mining Watch, and data from RAISG.

For these sources, we checked recent imagery and only included sites that appeared to still be active.

Classification as an illegal mining site is largely based on location within protected areas or indigenous territories, or clearly
outside of an authorized mining zone

Citation

Finer M, Mamani N, Arinez A, Novoa S, Larrea-Alcázar D, Villa J (2023) Illegal Gold Mining Across the Amazon. MAAP: 197.