Amazon Fire Tracker 2020: Over 500 Illegal Major Fires In Brazilian Amazon

The Brazilian Amazon just passed a grim milestone: Over 500 illegal major fires thus far in 2020.

The other major headline is that, although most fires continue to burn recently deforested areas, we are now seeing an increase in forest fires.

Our Real-time Amazon Fire Monitoring app has detected 504 major fires in the Brazilian Amazon as of August 24 (starting from the first major fire detected on May 28).

Brazilian Amazon Fire #449, burning both recently deforested area (center) and forest fire (upper center). Data: Planetscope (Planet), MAAP. Click to enlarge.
Brazilian Amazon Fire #449, burning both recently deforested area (center) and forest fire (upper center). Data: Planetscope (Planet), MAAP. Click to enlarge.

Some striking stats about this year’s major fires:

  • 97% have occured after the burning moratoriums established in July and are illegal.
  • 85% have occurred in August. Thus, the fire season has been accelerating.
  • 83% have burned recently deforested areas. Thus, the fires are actually a smoking indicator of the rampant deforestation.
  • 12% have been Forest Fires. This number marks a major increase from previous estimate as the fire season intensifies.
  • 4% have occurred in Protected Areas or Indigenous Territories.
  • 856,000 acres (353,000 hectares) have burned in the recently deforested areas fires.
  • 165,000 acres (66,000) have burned in the forest fires.

 

Base Map

The Base Map is a screen shot of the app’s “Major Amazon Fires 2020” layer. The majority of the major fires in the Brazilian Amazon have been in the states of Pará (42%) and Amazonas(31%), followed by Mato Grosso (13%) and Rondônia (13%). Note we are also now seeing an increase in major fires in the Bolivian Amazon, particularly savannah fires, and that will be the subject of a future report.

Base Map. Major Amazon Fires 2020. Data: MAAP.
Base Map. Major Amazon Fires 2020. Data: MAAP.

 

*Notes and Methodology

The app specializes in filtering out thousands of the traditional heat-based fire alerts to prioritize only those burning large amounts of biomass (defined here as a major fire).

In a novel approach, the app combines data from the atmosphere (aerosol emissions in smoke) and the ground (heat anomaly alerts) to effectively detect and visualize major Amazon fires.

When fires burn, they emit gases and aerosols. A new satellite (Sentinel-5P from the European Space Agency) detects these aerosol emissions. Thus, the major feature of the app is detecting elevated aerosol emissions which in turn indicate the burning of large amounts of biomass. For example, the app distinguishes small fires clearing old fields (and burning little biomass) from larger fires burning recently deforested areas or standing forest (and burning lots of biomass).

We define “major fire” as one showing elevated aerosol emission levels on the app, thus indicating the burning of elevated levels of biomass. This typically translates to an aerosol index of >1 (or cyan-green to red on the app). To identify the exact source of the elevated emissions, we reduce the intensity of aerosol data in order to see the underlying terrestrial heat-based fire alerts. Typically for major fires, there is a large cluster of alerts. The major fires are then confirmed, and burn areas estimated, using high-resolution satellite imagery from Planet Explorer.

See MAAP #118 for additional details on how to use the app.

No fires permitted in the Brazilian state of Mato Grosso after July 1, 2020. No fires permitted in all of Brazilian Amazon after July 15, 2020. Thus, we defined “illegal” as any major fires detected after these respective dates.

A major fire may be classified as burning across multiple land categories (for example, both recently deforested area and surrounding forest fire) so those percentages do not total 100%.

There was no available Sentinel-5 aerosol data on July 4, 15, and 26.

Acknowledgements

This analysis was done by Amazon Conservation in collaboration with SERVIR Amazonia.

Citation

Finer M, Vale H, Villa L, Nicolau A (2020) Over 500 Illegal Major Fires in Brazilian Amazon. MAAP.

MAAP #123: Detecting Illegal Logging In The Peruvian Amazon

Image 1. Example of a 2019 logging road with signs of illegality. Data: Planet.
Image 1. Example of a 2019 logging road with signs of illegality. Data: Planet.

In the Peruvian Amazon, the widespread illegal logging is difficult to detect with satellites because it is selective for high-value species (not clearcutting).

It is possible, however, to detect the associated logging roads.

In this report, we present a novel technique to identify illegal logging: analyze new logging roads in relation to detailed land use data available from government agencies.

Thus, our new method detects the crime in real-time and preventive action is still possible. This is important because when an intervention against illegal logging normally occurs, stopping a boat or truck with illegal timber, the damage is done.

This analysis has two parts. First, we identified the new logging roads built in the Peruvian Amazon during 2019, updating our previous work for 2015-18 (see Base Map).

Second, we analyzed the new logging road data in relation to governmental land use information in order to identify possible illegality.

This data is from 2019, but we are now applying this technique in real time during 2020.

 

Base Map. 2019 Logging roads, in relation to 2015-18 logging roads. Data: MAAP.
Base Map. 2019 Logging roads, in relation to 2015-18 logging roads. Data: MAAP.

Logging Roads 2019

The Base Map illustrates the location of logging roads built in the Peruvian Amazon during the last 5 years.

Previously (MAAP #99), we documented the construction of 3,300 kilometers of logging roads between 2015 and 2018.

Here, we estimate the construction of an additional 1,500 kilometers in 2019 (see red).

Note that forest roads are mainly concentrated in the Ucayali, Madre de Dios and Loreto regions.

Below, we show three types of possible illegality that detected in 2019:

  • Logging roads in areas without forestry concessions or permits (Cases 1-2).
    .
  • Logging roads in existing forest concessions, but whose current status is defined as “Non-Active or Undefined” (Cases 3-5).
    .
  • Logging roads in Native Communities (Case 6).

 

Cases of Possible Illegality

Logging roads in areas without forestry concessions or permits

Case 1. We detected the opening of a new logging road network (55 km) in an area without forestry concessions or permits, between the limits of the Loreto and San Martín regions. The image shows the digitized logging roads (red, left panel), and non-digitized satellite image (right panel). The arrows provide reference points between panels.

Case 1. Data: MAAP, Planet. Click to enlarge.
Case 1. Data: MAAP, Planet. Click to enlarge.

 

Case 2. We detected the construction of a new logging road network (5.8 km) in the buffer zone of Asháninka Communal Reserve, reaching only 300 meters from the protected area. The image shows the digitized logging roads (red, left panel), and non-digitized satellite image (right panel). The arrows provide reference points between panels.

Case 2. Data: MAAP, Planet, IBC, SERNANP. Click to enlarge.
Case 2. Data: MAAP, Planet, IBC, SERNANP. Click to enlarge.

 

Logging roads in existing forest concessions, but whose current state is labelled as “Non-Active or Undefined” 

Case 3. We detected the construction of a new logging road (45.3 km) that crosses a native community and reaches a forest concession whose status is defined as “Undefined,” in the Loreto region just north of Pacaya Samiria National Reserve. The image shows the digitized logging roads (red, left panel), and non-digitized satellite image (right panel). The arrows provide reference points between panels.

Case 3. Data: MAAP, ESA, IBC, SERFOR. Click to enlarge.
Case 3. Data: MAAP, ESA, IBC, SERFOR. Click to enlarge.

 

Case 4. We detected the construction of a new logging road network (53.2 km), of which nearly half (21.4 km) crosses a forest concession whose status is defined is “Non Active”, near the town of Sepahua in the Ucayali region. The image shows the digitized logging roads (red, left panel), and non-digitized satellite image (right panel). The arrows provide reference points between panels.

Case 4. Data: MAAP, Planet, IBC, SERFOR. Click to enlarge.
Case 4. Data: MAAP, Planet, IBC, SERFOR. Click to enlarge.

 

Case 5. We detected the construction of a new logging road (17.7 km) in a forestry concession whose current status is defined as “Non Active,” in the Madre de Dios region. The image shows the digitized logging roads (red, left panel), and non-digitized satellite image (right panel). The arrows provide reference points between panels.

Case 5. Data: MAAP, ESA, IBC, SERFOR. Click to enlarge.
Case 5. Data: MAAP, ESA, IBC, SERFOR. Click to enlarge.

 

Logging roads in Native Communities

Case 6. We detected the construction of a logging road (23.4 km) within an indigenous community in the Ucayali region. We did not find evidence of a permit for this activity. The image shows the digitized logging roads (red, left panel), and non-digitized satellite image (right panel). The arrows provide reference points between panels.

Case 6. Data: MAAP, Planet, SERNANP, IBC, SERFOR. Click to enlarge.
Case 6. Data: MAAP, Planet, SERNANP, IBC, SERFOR. Click to enlarge.

 

Methodology

Our analysis included two main steps:

The first step consisted of evaluating linear patterns in the 2019 early warning and final forest loss data, available from Global Forest Watch (data from the University of Maryland) and Geobosques (data from the Peruvian Ministry of the Environment). From the linear patterns, we distinguished between logging access roads and other types of roads and highways. Logging roads tend to have linear patterns that branch into the interior of the forest where the commercial timber is found. Other types of roads have a more defined destination, such as towns or farms. Once logging roads were identified, we downloaded the associated high-resolution imagery (3 meters) from Planet Explorer and digitized the roads in ArcGIS. During this process, additional logging roads detected in the high resolution images were also digitized.

The second step focused on the legality analysis. The new logging road data were overlaid with other types of land use information, such as forestry concessions on the GeoSERFOR portal (SERFOR), permits and concessions on the SISFOR portal (OSINFOR), indigenous communities (IBC 2019), protected areas (SERNANP), population centers (INEI 2019), and official road networks (MTC 2018). For example, as shown above, this process identified logging roads near protected areas, within indigenous communities, and within non-active forest concessions.

We analyzed information on several websites now available from national and regional authorities, such as SISFOR (OSINFOR), GEOSERFOR (SERFOR), and IDERs (Spatial Data Infrastructure of Regional governments). These new resources provide valuable information, however may have limitations in ability to constantly update information on the status of concessions and forest permits, especially from regional governments.

 

References

Planet Team (2017). Planet Application Program Interface: In Space for Life on Earth. San Francisco, CA. https://api.planet.com

 

Acknowledgments

We thank R. Valle (OSINFOR), A. Felix (DAI), D. Suarez (ACCA), and G. Palacios for their helpful comments on this report.

This report was conducted with technical assistance from USAID, via the Prevent project. Prevent is an initiative that, over the next 5 years, will work with the Government of Peru, civil society, and the private sector to prevent and combat environmental crimes in Loreto, Ucayali and Madre de Dios, in order to conserve the Peruvian Amazon.

This publication is made possible with the support of the American people through USAID. Its content is the sole responsibility of the authors and does not necessarily reflect the views of USAID or the US government.

Citation

Finer M, Paz L, Novoa S, Villa L (2020) Detecting Illegal Logging in the Peruvian Amazon. MAAP: 123.

Amazon Fire Tracker 2020: Images Of The Brazilian Amazon Fires

Images of the 2020 Brazilian Amazon FireOur innovative new app for Real-time Amazon Fire Monitoring has now detected over 350 major fires in the Brazilian Amazon this season.*

Specifically, we have detected 365 major fires as of August 17, since the first major fire detected on May 28.

The fire season is accelerating, as 79% of the major fires have occured in August.

Below, we present a series of satellite images showing key examples from August 2020.

We highlight our key finding that the vast majority of major fires (88%burned recently deforested areas covering 557,000 acres (226,000 hectares). Thus, the fires are actually a striking indicator of the rampant deforestation currently threatening the the Brazilian Amazon.

We have detected 4 Forest fires (1% of the major fires) covering 2,790 acres (1,130 hectares) and 3 savanna fires covering 38,000 acres (15,000 hectares). The rest of the major fires are burning older agricultural areas.

Other key findings include:

  • The vast majority of the fires (96%) are illegal, occuring past the 120 day moratorium established in July.
  • At least 18 of the major fires have been in protected areas or indigenous territories.
  • Most of the fires (70%) have occurred in two departments: Amazonas and Para. Mato Grosso and Rondonia each account for 15%.

We have detected an additional 10 major fires in the Bolivian Amazon, and that will be the feature of a future report.

 

Images of the 2020 Brazilian Amazon Fires

1) Fires burning recently deforested areas

Brazilian Amazon Fire #338 (August 16, 2020)

Brazilian Amazon Fire #338 (August 16, 2020)

 

Brazilian Amazon Fire #335 (August 16, 2020)

Brazilian Amazon Fire #335 (August 16, 2020)

 

Brazilian Amazon Fire #233 (August 11, 2020)Brazilian Amazon Fire #233 (August 11, 2020)

 

Brazilian Amazon Fire #230 (August 11, 2020)

Brazilian Amazon Fire #230 (August 11, 2020)

 

Brazilian Amazon Fire #221 (August 11, 2020)

Brazilian Amazon Fire #221 (August 11, 2020)

Brazilian Amazon Fire #202 (August 10, 2020)

Brazilian Amazon Fire #202 (August 10, 2020)

 

Brazilian Amazon Fire #188 (August 9, 2020)

Brazilian Amazon Fire #188 (August 9, 2020)

 

Brazilian Amazon Fire #124 (August 6, 2020)

Brazilian Amazon Fire #124 (August 6, 2020)

 

Brazilian Amazon Fire #110 (August 4, 2020)

Brazilian Amazon Fire #110 (August 4, 2020)

 

Brazilian Amazon Fire #109 (August 4, 2020)

Brazilian Amazon Fire #109 (August 4, 2020)

 

Brazilian Amazon Fire #76 (August 1, 2020)

Brazilian Amazon Fire #76 (August 1, 2020)

 

2) Forest Fires 

Brazilian Amazon Fire #218, August 2020

Brazilian Amazon Fire #218, August 2020

 

Brazilian Amazon Fire #195, August 2020

Brazilian Amazon Fire #195, August 2020

 

3) Grassland (Savanna) Fires 

Brazilian Amazon Fire #219, August 2020

Brazilian Amazon Fire #219, August 2020

 

*Notes and Methodology

The app specializes in filtering out thousands of the traditional heat-based fire alerts to prioritize only those burning large amounts of biomass (defined here as a major fire).

In a novel approach, the app combines data from the atmosphere (aerosol emissions in smoke) and the ground (heat anomaly alerts) to effectively detect and visualize major Amazon fires.

When fires burn, they emit gases and aerosols. A new satellite (Sentinel-5P from the European Space Agency) detects these aerosol emissions. Thus, the major feature of the app is detecting elevated aerosol emissions which in turn indicate the burning of large amounts of biomass. For example, the app distinguishes small fires clearing old fields (and burning little biomass) from larger fires burning recently deforested areas or standing forest (and burning lots of biomass).

We define “major fire” as one showing elevated aerosol emission levels on the app, thus indicating the burning of elevated levels of biomass. This typically translates to an aerosol index of >1 (or cyan-green to red on the app). To identify the exact source of the elevated emissions, we reduce the intensity of aerosol data in order to see the underlying terrestrial heat-based fire alerts. Typically for major fires, there is a large cluster of alerts. The major fires are then confirmed, and burn areas estimated, using high-resolution satellite imagery from Planet Explorer.

See MAAP #118 for additional details.

No fires permitted in the Brazilian state of Mato Grosso after July 1, 2020. No fires permitted in all of Brazilian Amazon after July 15, 2020. Thus, we defined “illegal” as any major fires detected after these respective dates.

There was no available Sentinel-5 aerosol data on July 4, 15, and 26.

 

Acknowledgements

This analysis was done by Amazon Conservation in collaboration with SERVIR Amazonia.

Citation

Finer M, Nicolau A, Vale H, Villa L, Mamani N (2020) Amazon Fire Tracker 2020: Images of the Brazilian Amazon Fires. MAAP.

Amazon Fire Tracker 2020: Over 200 Major Fires As Of Aug 10

Photo of Brazil Fire 76 on July 31 2020Our innovative new app for Real-time Amazon Fire Monitoring has detected over 200 major fires in 2020.

The app specializes in filtering out thousands of the traditional heat-based fire alerts to prioritize only those burning large amounts of biomass (defined here as a major fire).*

Our key findings include:

  • We have detected 227 major Amazon fires (Brazil 220, Bolivia 6; Peru 1), as of August 10.
    ,
  • The vast majority of major fires have been in the Brazilian Amazon, where a strikingly high number (85%) have burned recently deforested areas. Thus, the fires are actually a smoking indicator of the rampant deforestation now in Brazil.
    k
  • In Brazil, we have detected two forest fires, but this risk increases as we get deeper into the dry season. The rest of the fires have been on older fields.
    l
  • In Brazil, the vast majority (94%) of the major fires have been illegal, in violation of the state and national fire moratoriums established in July. In fact, despite the moratoriums, the number of major fires is accelerating: 143 so far in August following 77 in May through July.
    m
  • In Brazil, 14 of the fires have been in Protected Areas.
    k
  • In the Bolivian and Peruvian Amazon, we have recently started detecting fires in the drier ecosystems (savannahs and grasslands).

See below for a more detailed breakdown of the results.

 

Additional Results

The Base Map is a screen shot of the app’s “Major Amazon Fires 2020” layer.

The vast majority of the fires have been in the Brazilian AmazonPará (37%) and Amazonas (39%), followed by Mato Grosso (17%) and Rondônia (8%).

Importantly, the vast majority of the major fires in the Brazilian Amazon (85%) have burned recently deforested areas (cleared between 2018 and 2020) covering 280,000 acres (113,000 hectares). Thus, we argue that the central issue is actually deforestation and the fires are actually a smoking indicator of this forest loss.

We have detected the first two forest fires, burning 388 acres (1,447 hectares) in Mato Grosso and Para.

The rest of the major fires have been on older cattle or agricultural lands (deforested prior to 2018).

The most impacted protected areas are Jamanxim and Altamira National Forests in Pará. We emphasize, however, that these fires were burning recently deforested areas (not forest fires) and so, again, the primary issue is deforestation.

In Brazil, the vast majority of the major fires (94%) appear to be illegal as they violate the state and national government mandated fire moratoriums established in July. In fact, despite the moratoriums, the number of major fires is accelerating: 143 so far in August, following 64 in July, 12 in June, and the first one in May.

In the Bolivian Amazon, we have recently started detecting fires in the savannahs in the department of Beni. We also detected one fire in a recently deforested area in the Santa Cruz department.

In the Peruvian Amazon, we have recently started detecting fires in the upper elevation grasslands. The biggest one was actually within a protected area (Otishi National Park). There have also been smaller grassland fires near the buffer zone of upper Manu National Park.

 

Key Examples of 2020 Fires

Overall our key finding is that most major Brazilian Amazon fires are burning recently deforested areas, and not raging forest fires. Below is a series of satellite image time-lapse videos showing examples of recent deforestation followed by a major 2020 fire.

 

Brazilian Amazon Fire #54, July 2020

 

Brazilian Amazon Fire #59, July 2020

 

Brazilian Amazon Fire #76, July 2020

 

Brazilian Amazon Fire #110, August 2020

 

*Notes and Methodology

In a novel approach, the app combines data from the atmosphere (aerosol emissions in smoke) and the ground (heat anomaly alerts) to effectively detect and visualize major Amazon fires.

When fires burn, they emit gases and aerosols. A new satellite (Sentinel-5P from the European Space Agency) detects these aerosol emissions. Thus, the major feature of the app is detecting elevated aerosol emissions which in turn indicate the burning of large amounts of biomass. For example, the app distinguishes small fires clearing old fields (and burning little biomass) from larger fires burning recently deforested areas or standing forest (and burning lots of biomass).

We define “major fire” as one showing elevated aerosol emission levels on the app, thus indicating the burning of elevated levels of biomass. This typically translates to an aerosol index of >1 (or cyan-green to red on the app). To identify the exact source of the elevated emissions, we reduce the intensity of aerosol data in order to see the underlying terrestrial heat-based fire alerts. Typically for major fires, there is a large cluster of alerts. The major fires are then confirmed, and burn areas estimated, using high-resolution satellite imagery from Planet Explorer.

See MAAP #118 for additional details.

No fires permitted in the Brazilian state of Mato Grosso after July 1, 2020. No fires permitted in all of Brazilian Amazon after July 15, 2020. Thus, we defined “illegal” as any major fires detected after these respective dates.

There was no available Sentinel-5 aerosol data on July 4, 15, and 26.

 

Acknowledgements

This analysis was done by Amazon Conservation in collaboration with SERVIR Amazonia.

Citation

Finer M, Nicolau A, Villa L (2020) 200 Major Amazon Fires in 2020: Tracker Analysis. MAAP.

Amazon Fire Tracker 2020 – July Update

Recall we recently launched an innovative new app for Real-time Amazon Fire Monitoring (see MAAP #118 for details).

In a novel approach, the app combines data from the atmosphere (aerosol emissions in smoke) and the ground (heat anomaly alerts) to effectively detect and visualize major Amazon fires.

The app specializes in filtering out thousands of the heat-based fire alerts to prioritize only those burning large amounts of biomass and thus emitting elevated levels of aerosol (defined here as a major fire).*

Example of a major fire in the Brazilian Amazon burning recently deforested area. Brazil major fire #54, July 30, Mato Grosso. Image: Planetscope (Planet).
Example of a major fire in the Brazilian Amazon burning recently deforested area. Brazil major fire #54, July 30, Mato Grosso. Image: Planetscope (Planet).

As of the end of July, we have detected 77 major Amazon fires thus far in 2020, all in Brazil.

In summary, 84% of the major fires are burned recently deforested areas and 83% were illegal (in violation of fire moratoriums). We detected the first forest fire on the last day of the month.

We have started detecting large and uncontrolled fires in the drier ecosystems of Bolivia, but outside the Amazon watershed.

See below for a more detailed overview of the 2020 Amazon fire season thru the end of July.

 

Key Results

The Base Map is a screen shot of the app’s “Major Amazon Fires 2020” layer.

Base Map. Screen shot of the app’s “Major Amazon Fires 2020” layer.
Base Map. Screen shot of the app’s “Major Amazon Fires 2020” layer.

As noted above, we have detected 77 major Amazon fires thus far in 2020, all in Brazil.

The first major fire was detected on May 28 in the state of Mato Grosso in southeastern Brazilian Amazon (see MAAP #118). This event was followed by 12 major fires in June, all in Mato Grosso (see Fire Tracker #12).

The number of major fires in Mato Grosso decreased in July, suggesting the state’s new fire moratorium (starting July 1) may be working.

Starting in mid-July, the major fire activity shifted to the surrounding Brazilian states of Amazonas, Rondônia and Pará. This shift coincided with national fire moratorium (starting July 15), indicating it has not been as effective.

Overall, most of the major fires (83%) appear to be illegal as they violate the state and national government mandated fire moratoriums established in July.

Importantly, most of the major fires (84%) have burned recently deforested areas (deforested 2018-20) covering 108,000 acres (44,000 hectares). See MAAP #113 for more on this important point in regards to the 2019 fires.

We detected the first forest fire on the last day of the month. It burned 388 acres (157 hectares).

The other major fires have been in older cattle or agricultural areas (deforested pre 2018).

We have started detecting large and uncontrolled fires in the drier ecosystems of Bolivia, but outside the Amazon watershed.

 

Key Examples of 2020 Fires

Overall our key finding is that most major Brazilian Amazon fires are burning recently deforested areas, and not raging forest fires. Below is a series of four satellite images time-lapse videos showing examples of recent deforestation (2019) followed by a major 2020 fire burning lots of biomass that was detected by the app.

 

Brazilian Amazon Fire #1, May 2020

Brazilian Amazon Fire #4, June 2020

 

Brazilian Amazon Fire #12, June 2020

Brazilian Amazon Fire #18, July 2020

Brazilian Amazon Fire #54, July 2020

*Notes and Methodology

When fires burn, they emit gases and aerosols. A new satellite (Sentinel-5P from the European Space Agency) detects these aerosol emissions. Thus, the major feature of the app is detecting elevated aerosol emissions which in turn indicate the burning of large amounts of biomass. For example, the app distinguishes small fires clearing old fields (and burning little biomass) from larger fires burning recently deforested areas or standing forest (and burning lots of biomass).

We define “major fire” as one showing elevated aerosol emission levels on the app, thus indicating the burning of elevated levels of biomass. This typically translates to an aerosol index of >1 (or cyan-green to red on the app). To identify the exact source of the elevated emissions, we reduce the intensity of aerosol data in order to see the underlying terrestrial heat-based fire alerts. Typically for major fires, there is a large cluster of alerts. The major fires are then confirmed, and burn areas estimated, using high-resolution satellite imagery from Planet Explorer.

No fires permitted in the Brazilian state of Mato Grosso after July 1, 2020. No fires permitted in all of Brazilian Amazon after July 15, 2020. Thus, we defined “illegal” as any major fires detected after these respective dates.

There was no available Sentinel-5 aerosol data on July 4, 15, and 26.

Acknowledgements

This analysis was done by Amazon Conservation in collaboration with SERVIR Amazonia.

Citation

Finer M, Nicolau A, Villa L (2020) Amazon Fire Tracker 2020 – July Update. MAAP.

MAAP #122: Amazon Deforestation 2019

Newly released data for 2019 reveals the loss of over 1.7 million hectares (4.3 million acres) of primary Amazon forest in our 5 country study area (Bolivia, Brazil, Colombia, Ecuador, and Peru).* That is twice the size of Yellowstone National Park.

Table 1 shows 2019 deforestation (red) in relation to 2018 (orange).

Table 1. Amazon 2019 primary forest loss for 2019 (red) compared to 2018 (orange). Data: Hansen/UMD/Google/USGS/NASA, MAAP.
Table 1. Amazon 2019 primary forest loss for 2019 (red) compared to 2018 (orange). Data: Hansen/UMD/Google/USGS/NASA, MAAP.

Primary forest loss in the Brazilian Amazon (1.29 million hectares) was over 3.5 times higher than the other four countries combined, with a slight increase in 2019 relative to 2018. Many of these areas were cleared in the first half of the year and then burned in August, generating international attention.

Primary forest loss rose sharply in the Bolivian Amazon (222,834 hectares), largely due to uncontrolled fires escaping into the dry forests of the southern Amazon.

Primary forest loss rose slightly in the Peruvian Amazon (161,625 hectares) despite a relatively successful crackdown on illegal gold mining, pointing to small-scale agriculture (and cattle) as the main driver.

On the positive side, primary forest loss decreased in the Colombian Amazon (91,400 hectares) following a major spike following the 2016 peace accords (between the government and FARC). It is worth noting, however, that we have now documented the loss of 444,000 hectares (over a million acres) of primary forest in the Colombian Amazon in the past four years since the peace agreement (see Annex).

*Two important points about the data. First, we use annual forest loss from the University of Maryland to have a consistent source across all five countries. Second, we applied a filter to only include loss of primary forest (see Methodology).

 

2019 Deforestation Hotspots Map

The Base Map below shows the major 2019 deforestation hotspots across the Amazon.

2019 deforestation hotspots across the Amazon. Data: Hansen/UMD/Google/USGS/NASA, MAAP.
2019 deforestation hotspots across the Amazon. Data: Hansen/UMD/Google/USGS/NASA, MAAP.

Many of the major deforestation hotspots were in Brazil. Early in the year, in March, there were uncontrolled fires up north in the state of Roraima. Further south, along the Trans-Amazonian Highway, much of the deforestation occurred in the first half of the year, followed by the high profile fires starting in late July. Note that many of these fires were burning recently deforested areas, and were not uncontrolled forest fires (MAAP #113).

The Brazilian Amazon also experienced escalating gold mining deforestation in indigenous territories (MAAP #116).

Bolivia also had an intense 2019 fire season. Unlike Brazil, many were uncontrolled fires, particularly in the Beni grasslands and Chiquitano dry forests of the southern Bolivian Amazon (MAAP #108).

In Peru, although illegal gold mining deforestation decreased (MAAP #121), small-scale agriculture (including cattle) continues to be a major driver in the central Amazon (MAAP #112) and an emerging driver in the south.

In Colombia, there is an “arc of deforestation” in the northwestern Amazon. This arc includes four protected areas (Tinigua, Chiribiquete and Macarena National Parks, and Nukak National Reserve) and two Indigenous Reserves (Resguardos Indígenas Nukak-Maku and Llanos del Yari-Yaguara II) experiencing substantial deforestation (MAAP #120). One of the main deforestation drivers in the region is conversion to pasture for land grabbing or cattle ranching.

Annex – Colombia peace accord trend

Table 1. Deforestation of primary forest in the Colombian Amazon, 2015-20. Data: Hansen/UMD/Google/USGS/NASA, UMD/GLAD. *Until May 2020
Table 1. Deforestation of primary forest in the Colombian Amazon, 2015-20. Data: Hansen/UMD/Google/USGS/NASA, UMD/GLAD. *Until May 2020

Methodology

The baseline forest loss data presented in this report were generated by the Global Land Analysis and Discovery (GLAD) laboratory at the University of Maryland (Hansen et al 2013) and presented by Global Forest Watch. Our study area is strictly what is highlighted in the Base Map.

For our estimate of primary forest loss, we used the annual “forest cover loss” data with density >30% of the “tree cover” from the year 2001. Then we intersected the forest cover loss data with the additional dataset “primary humid tropical forests” as of 2001 (Turubanova et al 2018). For more details on this part of the methodology, see the Technical Blog from Global Forest Watch (Goldman and Weisse 2019).

For boundaries, we used the biogeographical limit (as defined by RAISG) for all countries except Bolivia, where we used the Amazon watershed limit (see Base Map).

All data were processed under the geographical coordinate system WGS 1984. To calculate the areas in metric units, the projection was: Peru and Ecuador UTM 18 South, Bolivia UTM 20 South, Colombia MAGNA-Bogotá, and Brazil Eckert IV.

Lastly, to identify the deforestation hotspots, we conducted a kernel density estimate. This type of analysis calculates the magnitude per unit area of a particular phenomenon, in this case forest cover loss. We conducted this analysis using the Kernel Density tool from Spatial Analyst Tool Box of ArcGIS. We used the following parameters:

Search Radius: 15000 layer units (meters)
Kernel Density Function: Quartic kernel function
Cell Size in the map: 200 x 200 meters (4 hectares)
Everything else was left to the default setting.

For the Base Map, we used the following concentration percentages: Medium: 7%-10%; High: 11%-20%; Very High: >20%.

References

Goldman L, Weisse M (2019) Explicación de la Actualización de Datos de 2018 de Global Forest Watch. https://blog.globalforestwatch.org/data-and-research/blog-tecnico-explicacion-de-la-actualizacion-de-datos-de-2018-de-global-forest-watch

Hansen, M. C., P. V. Potapov, R. Moore, M. Hancher, S. A. Turubanova, A. Tyukavina, D. Thau, S. V. Stehman, S. J. Goetz, T. R. Loveland, A. Kommareddy, A. Egorov, L. Chini, C. O. Justice, and J. R. G. Townshend. 2013. “High-Resolution Global Maps of 21st-Century Forest Cover Change.” Science 342 (15 November): 850–53. Data available on-line from: http://earthenginepartners.appspot.com/science-2013-global-forest.

Turubanova S., Potapov P., Tyukavina, A., and Hansen M. (2018) Ongoing primary forest loss in Brazil, Democratic Republic of the Congo, and Indonesia. Environmental Research Letters  https://doi.org/10.1088/1748-9326/aacd1c 

Acknowledgements

We thank G. Palacios for helpful comments to earlier versions of this report.

This work was supported by the following major funders: Norwegian Agency for Development Cooperation (NORAD), Gordon and Betty Moore Foundation, International Conservation Fund of Canada (ICFC), Metabolic Studio, Erol Foundation, MacArthur Foundation, and Global Forest Watch Small Grants Fund (WRI).

Citation

Finer M, Mamani N (2020) 2019 Amazon Deforestation. MAAP: 122.

Amazon Fire Tracker 2020: Brazil Fire #12 (June 29)

As presented in MAAP #118, Amazon Conservation launched a real-time fire monitoring app that specializes in the rapid and user-friendly detection and visualization of major Amazon fires.

In a novel approach, the app combines data from the atmosphere (aerosol emissions in smoke) and the ground (heat anomalies) to effectively monitor large Amazon fires.

2020 Amazon major fire #12, in Mato Grosso Brazil, on June 29. Data: Planet. Analysis: MAAP/Amazon Conservation.
2020 Amazon major fire #12, in Mato Grosso Brazil, on June 29. Data: Planet. Analysis: MAAP/Amazon Conservation.

As detailed below, the app just detected the 12th major Amazon fire of the year on June 29, 2020 (see the high-resolution image to the right).

It burned 587 hectares (1,451 acres) of land deforested in 2019.

Thus far, all 12 major Amazon fires of 2020 have:

  • Occured in the state of Mato Grosso, in the southeastern Brazilian Amazon.
  • Burned recently deforested areas (that is, areas deforested in 2018, 2019, or 2020). In other words, these are not forest fires. See MAAP #113 for background on this important point.

Below, we describe the process of using the app to detect and confirm the fire on June 29.

 

Step 1.

Detection of elevated emissions in the southeastern Brazilian Amazon (Mato Grosso).

Step 1. Detection of elevated emissions in the southeastern Brazilian Amazon (Mato Grosso).

Step 2.

Zoom in to confirm the elevated aerosol emissions, indicating the burning of abundant biomass.

Step 2. Zoom in to confirm the elevated aerosol emissions, indicating the burning of abundant biomass.

 

Step 3.

Adjust the transparency to see the underlying fire alerts that indicate the exact location of the fires. Obtain coordinates of the source of the fires.

Step 3. Adjust the transparency to see the underlying fire alerts that indicate the exact location of the fires. Obtain coordinates of the source of the fires.

Step 4.

Confirm fire with satellite imagery. For example, here is a high resolution image from Planet Explorer showing the fire burning on June 29.

Step 4. Confirm fire with satellite imagery. For example, here is a high resolution image from Planet Explorer showing the fire burning on June 29.

Step 5.

Using Planet’s extensive imagery archive, we were able to determine that the fires were burning an area deforested in 2019 (and not a forest fire). In the timelapse below, see that the deforestation occurred between September and October 2019, and then burned on June 29, 2020. The final image shows the day after the fires, June 30, to see the full extent of the burn.

Coordinates

-10.99, -55.13

References

Gorelick, N., Hancher, M., Dixon, M., Ilyushchenko, S., Thau, D., & Moore, R. (2017). Google Earth Engine: Planetary-scale geospatial analysis for everyone. Remote Sensing of Environment.
Planet Team (2017). Planet Application Program Interface: In Space for Life on Earth. San Francisco, CA. https://api.planet.com

Acknowledgements

This work was supported by the following major funders: USAID/NASA (SERVIR), Global Forest Watch Small Grants Fund (WRI), Norwegian Agency for Development Cooperation (NORAD),  International Conservation Fund of Canada (ICFC), Metabolic Studio, and Erol Foundation.

 

Citation

Finer M, Villa L (2020) Amazon Fire Tracker 2020: Brazil #12 (June 29). MAAP.

MAAP#121: Reduction Of Illegal Gold Mining In The Peruvian Amazon

Thanks to the support of the USAID, via the Prevent Project, dedicated to the prevention and combat of environmental crimes in the Amazon, we conducted a detailed analysis of recent illegal gold mining deforestation in the southern Peruvian Amazon.

The objective is to understand the trends from early 2017 to June 2020 (which includes the first part of the mandatory quarantine issued by the Peruvian government as of March 16, 2020 due to the coronavirus pandemic).

Base Map. Illegal gold mining deforestacion in the protected area buffer zones of the southern Peruvian Amazon, 2017-2019. Data: MAAP. Click to enlarge image.
Base Map. Illegal gold mining deforestacion in the protected area buffer zones of the southern Peruvian Amazon, 2017-2019. Data: MAAP. Click to enlarge image.

We focus on the buffer zones of two protected areas in the Madre de Dios region: Tambopata National Reserve and Bahuaja Sonene National Park (see Base Map).*

This area includes La Pampa, the current highest intensity illegal mining zone in the country. In February 2019, the Peruvian government launched Operation Mercury  to confront the illegality in La Pampa and surrounding areas.

The Base Map shows that gold mining deforestation in La Pampa decreased over 90% following Operation Mercury.

However, illegal gold mining does continue after Operation Mercury (including during the coronavirus state of emergency), but at lower rates. Thus, current snapshots may be misleading and recent context is important.

On the Base Map, the red arrows indicate the areas with the most recent illegal activity (click the image to enlarge). See below for more details.

 

Main Results

Table 1. Illegal gold mining deforestation before (yellow) and after (red) Operation Mercury in the buffer zones of Madre de Dios. Data: MAAP.
Table 1. Illegal gold mining deforestation before (yellow) and after (red) Operation Mercury in the buffer zones of Madre de Dios. Data: MAAP.

The Base Map and Table 1 illustrate the following key results:

  • In La Pampa, we documented mining deforestation of 173 hectares (428 acres) per month before Operation Mercury (January 2018 – February 2019). After the intervention, deforestation was reduced to 14 hectares (36 acres) per month (March 2019 – May 2020), a decrease of 92%.
    .
  • Upstream, in the Alto Malinowski, we documented the mining deforestation of 61 hectares (150 acres) per month before Operation Mercury. After the intervention, deforestation was reduced to 28 hectares (69 acres) per month, a decrease of 53%.
    .
  • Downstream, in the Apaylon area, we documented the mining deforestation of 2.9 hectares (7 acres) per month, before Operation Mercury. After the intervention, deforestation increased to 4 hectares (10 acres) per month, an increase of 41%. Apaylon is main area in the buffer zone where deforestation has increased.
    .
  • Within Tambopata National Reserve, we documented the mining deforestation of 6.5 hectares (16 acres) per month, before Operation Mercury. After the intervention, deforestation was reduced to 0.5 hectares (1.2 acres) per month, a decrease of 93%.
    .
  • Overall, illegal gold mining does continue in the buffer zones of Madre de Dios, but at lower rates than the previous two years. We documented the gold mining deforestation of 797 hectares (1,670 acres) after Operation Mercury.
    .
  • Regarding the speculation that illegal activity would increase during the coronavirus pandemic, we have not documented any major increase or surge in the buffer zones of Madre de Dios.* Illegal mining does continue, however, we documented the deforestation of 80 hectares (198 acres) during the quarantine.

 

Reduction of 90% in La Pampa

The following images show the major decrease in gold mining deforestation in La Pampa after Operation Mercury. Image 1 shows the rapid deforestation before Operation Mercury, between January 2017 (left panel) and February 2019 (right panel). Image 2 shows how the deforestation decreased after Operation Mercury, between February 2019 (left panel) and May 2020 (right panel). The red dot represents a reference point between the images.

Image 1. Rapid gold mining deforestation in La Pampa before Operation Mercury, between January 2017 (left panel) and February 2019 (right panel). Data: Planet.
Image 1. Rapid gold mining deforestation in La Pampa before Operation Mercury, between January 2017 (left panel) and February 2019 (right panel). Data: Planet.
Image 2. Mining deforestation decreased in La Pampa after Operation Mercury, between February 2019 (left panel) and May 2020 (right panel). Data: Planet.
Image 2. Mining deforestation decreased in La Pampa after Operation Mercury, between February 2019 (left panel) and May 2020 (right panel). Data: Planet.

 

Displaced Miners?

There has also been speculation that the focus of Operation Mercury in La Pampa would lead to illegal miners moving to other areas.* Base Map 2 shows two of the most threatened areas: Camanti and Pariamanu.

Table 2. Deforestation by illegal gold mining before (yellow) and after (red) Operation Mercury in two other threatened areas. Data: MAAP.
Table 2. Deforestation by illegal gold mining before (yellow) and after (red) Operation Mercury in two other threatened areas. Data: MAAP.

These are the main results for these two areas:

  • In Camanti (located in the buffer zone of Amarakaeri Communal Reserve), we documented the gold mining deforestation of 13.3 hectares (33 acres) per month before Operation Mercury. After the intervention, deforestation was reduced to 6.1 hectares (15 acres) per month, a decrease of 54%.
    .
  • In Pariamanu, we documented  the mining deforestation of 2.5 hectares (6 acres) per month before Operation Mercury. After the intervention, it increased to 4.2 hectares (10 acres) per month, an increase of 70%.
  • In summary, illegal gold mining continues in these two areas outside La Pampa. We documented the mining deforestation of 175 hectares (432 acres) after Operation Mercury (including 22 hectares during the pandemic). There is some evidence that miners are being displaced to Pariamanu, but there has not been a surge in Camanti.
Base Map 2. Main mining areas in the south of the Peruvian Amazon. Click to enlarge image.
Base Map 2. Main mining areas in the south of the Peruvian Amazon. Click to enlarge image.

Statement of the Peruvian Protected Area Agency (SERNANP)

El Servicio Nacional de Áreas Naturales Protegidas por el Estado (SERNANP) nos ha comunicado lo siguiente:

  • La actividad de control y vigilancia en la Reserva Nacional Tambopata es permanente y las autoridades (SERNANP, Policía Nacional del Perú, Fiscalías Especializadas en Materia Ambiental, y Marina de Guerra del Perú) continúan interviniendo a todas las actividades de minería ilegal, manteniendo el 100%.
  • Las zonas de amortiguamiento son espacios que están sujetos a la intervención de las autoridades de la Operación Mercurio (no del SERNANP). Se han realizado intervenciones continuas e interdicciones tanto en  las zonas indicadas en el reporte, como en Apaylon y Camanti.
    ,
  • Cabe mencionar que la Operación Mercurio, durante el 2019 y sobre todo en el 2020 (Incluyendo el período de cuarentena) ha ampliado sus operativos mas allá de la Pampa, lo cual explica porque en Camanti las cifras también se ha reducido.  En el segundo semestre de 2020 y en el 2021, se espera que los operativos es amplíen a otras zonas de Madre de Dios.

 

*Notes

 

Acknowledgments

We thank R. Segura, M. Castro, E. Ortiz, M. Silman, M. E. Gutierrez, S. Novoa, H. Balbuena, M. Allemant, and G. Palacios for their helpful comments on this report.

This report was conducted with technical assistance from USAID, via the Prevent project. Prevent is an initiative that, over the next 5 years, will work with the Government of Peru, civil society, and the private sector to prevent and combat environmental crimes in Loreto, Ucayali and Madre de Dios, in order to conserve the Peruvian Amazon.

This publication is made possible with the support of the American people through USAID. Its content is the sole responsibility of the authors and does not necessarily reflect the views of USAID or the US government.

USAID logo

Citation

Finer M, Mamani N (2020) Reduction of Illegal Gold Mining in the Peruvian Amazon. MAAP:

Amazon Fire Tracker 2020: Brazil #4 (June 17, 2020)

As presented in MAAP #118, Amazon Conservation launched a real-time fire monitoring app that specializes in detection of elevated aerosol emissions in the smoke coming from burning Amazon fires. As detailed below, the app just detected the fourth major Amazon fire of 2020 on June 17. All four fires thus far have been in the state of Mato Grosso and burning recently deforested areas (see MAAP #113 for background).

Step 1.

Detection of elevated emissions in the southeastern Brazilian Amazon (Mato Grosso).

Step 1 of Amazon Fire App. Detection of elevated emissions in the southeastern Brazilian Amazon (Mato Grosso).

Step 2.

Zoom in on the emissions.

Step 3.

Adjust the transparency to see the underlying fire alerts that indicate the exact location of the fires. Obtain coordinates of the source of the fires.

 

Step 4.

Check the satellite imagery in Planet Explorer. Here is a high resolution Planet image showing the fire burning on June 17. Also see the comparisons below, comparing the the June 17 fires with a pre-fire image from June 10.

Pre-fire image, June 10:

pre-fire image, June 10

 

June 17 fires:

Imagery source: Planet.

Step 5.

Using Planet’s extensive imagery archive, we were able to determine that the fires were burning an area deforested in 2019 (and not a forest fire).

Coordinates: -10.45, -53.55

 

Annex – Fire Alert vs. Aerosol Emission Data

This two images show us how aerosol emission data allows users to prioritize hundreds (or thousands) of heat-based fire alerts (photo 1). In other words, the aerosol data indicates just the fires that are  actually burning lots of biomass and putting out abundant smoke.

References

Gorelick, N., Hancher, M., Dixon, M., Ilyushchenko, S., Thau, D., & Moore, R. (2017). Google Earth Engine: Planetary-scale geospatial analysis for everyone. Remote Sensing of Environment.”
https://earthengine.google.com/faq/
Planet Team (2017). Planet Application Program Interface: In Space for Life on Earth. San Francisco, CA. https://api.planet.com

Acknowledgements

This work was supported by the following major funders: USAID/NASA (SERVIR), Global Forest Watch Small Grants Fund (WRI), Norwegian Agency for Development Cooperation (NORAD),  International Conservation Fund of Canada (ICFC), Metabolic Studio, and Erol Foundation.

 

Citation

Finer M, Villa L (2020) Amazon Fire Tracker 2020: Brazil #4 (June 17, 2020). MAAP.

MAAP: Amazon Fire Tracker #2 – Brazil, June 8 2020

As presented in MAAP #118, Amazon Conservation launched a real-time fire monitoring app that specializes in detection of elevated aerosol emissions from burning Amazon fires. As detailed below, the app detected the second major 2020 fire on June 8, 2020 in Mato Grosso, Brazil.

Step 1. Detection of elevated emissions in the southeastern Brazilian Amazon (Mato Grosso).

Step 4. Check the satellite imagery archive in Planet Explorer.

Step 2. Zoom in on the emissions, adjust the transparency to see the underlying fire alerts that indicate the fire location.

Step 2. Zoom in on the emissions, adjust the transparency to see the underlying fire alerts that indicate the fire location.

Step 3. Zoom in again to see precisely the fire location and obtain coordinates.

Step 3. Zoom in again to see precisely the fire location and obtain coordinates.

Step 4. Check the satellite imagery archive in Planet Explorer. Here is a Landsat image (30 meter resolution) showing the fire burned around 3,000 hectares (7,400 acres) of an area deforested in July 2018. Note that MAAP #113 made the important discovery that most of the 2019 Brazilian Amazon fires were burning recently deforested areas (and not uncontrolled forest fires).

Step 1. Detection of elevated emissions in the southeastern Brazilian Amazon (Mato Grosso).

Coordinates

lat: -12.57, lon: -54.06

 

References

Gorelick, N., Hancher, M., Dixon, M., Ilyushchenko, S., Thau, D., & Moore, R. (2017). Google Earth Engine: Planetary-scale geospatial analysis for everyone. Remote Sensing of Environment.”
https://earthengine.google.com/faq/

Acknowledgements

This work was supported by the following major funders: USAID/NASA (SERVIR), Global Forest Watch Small Grants Fund (WRI), Norwegian Agency for Development Cooperation (NORAD),  International Conservation Fund of Canada (ICFC), Metabolic Studio, and Erol Foundation.

 

Citation

Finer M, Villa L (2020) Amazon 2020 Fire Tracker #2 – Brazil, June 8. MAAP.