<p>Join our webinar for an overview of the satellite imagery and resources available on GFW. We will be joined by partners from KSAT and Planet to share how this satellite imagery can help users investigate what’s happening in the world’s forests. <a href="https://wri.zoom.us/webinar/register/7617086143771/WN_8_Q9dGGwQT2mqRpd4LgRZA" rel="noopener" target="_blank">Register here!</a></p>


Investigating Deforestation with Satellite Imagery

Data and Research

GFW is first and foremost a data-driven platform. Read the latest forest research and analysis.

Data and Research | Global Forest Watch Blog

amazon

amazon | Global Forest Watch Blog

Bolivia

Bolivia | Global Forest Watch Blog

boreal

boreal | Global Forest Watch Blog

Brazil

Brazil | Global Forest Watch Blog

DRC | Global Forest Watch Blog

fires

fires | Global Forest Watch Blog

Indonesia

Indonesia | Global Forest Watch Blog

tree cover loss

tree cover loss | Global Forest Watch Blog

Tropical Deforestation

Tropical Deforestation | Global Forest Watch Blog

Mikaela Weisse

Liz Goldman

Forest Loss Remained Stubbornly High in 2021

La pérdida de bosques se mantuvo obstinadamente elevada en 2021

Perda florestal permanece resistentemente alta em 2021

Kehilangan Hutan Tetap Tinggi di Tahun 2021

La perte des forêts est restée chroniquement élevée en 2021

Forest loss, including tropical deforestation, remained stubbornly high in 2021 according to new data on the Global Forest Watch.

What Happened to Forests in 2021? | Global Forest Watch Blog

Comment

Global Forest Watch Content

https://dev-global-forest-watch-blog.pantheonsite.io/wp-content/uploads/2018/06/gfwGlobal.png

https://secure.gravatar.com/avatar/9be466a38cf8a59795eb5c3ed715baa9?s=96&d=mm&r=g

Amanda Dimartini

Planet satellite imagery of primary forest removal in Kalimantan Barat, Indonesia on GFW map

Deforestation Alerts

Deforestation Alerts | Global Forest Watch Blog

forest monitoring

forest monitoring | Global Forest Watch Blog

GLAD alerts

GLAD alerts | Global Forest Watch Blog

Radar

Radar | Global Forest Watch Blog

Anika Berger

Teresa Schofield

Amy Pickens

Johannes Reiche

Yaqing Gou

<p><em>Languages:&nbsp;<a href="https://www.globalforestwatch.org/blog/es/data-and-research/alertas-de-deforestacion-integradas/">Leerlo en español</a>, <a href="https://www.globalforestwatch.org/blog/pt/data-and-research/alertas-de-desmatamento-integrados/">Leia em portugues</a>, <a href="https://www.globalforestwatch.org/blog/id/data-and-research/peringatan-deforestasi-terintegrasi/">Baca dalam bahasa indonesia</a>, <a href="https://www.globalforestwatch.org/blog/fr/data-and-research/alertes-integrees-de-deforestation/">Lire en français</a></em></p>



<p>Global Forest Watch (GFW) is committed to bringing the most accurate forest data to users as quickly as possible. We are continually expanding the near-real-time deforestation alerts available on the GFW platform and are excited to introduce an updated layer. The integrated deforestation alerts layer combines the analytical power of GLAD, GLAD-S2 and RADD deforestation alerts to provide a faster, more confident view of forest disturbances than any one individual system. Prompted by user feedback, this update simplifies workflows and harnesses the best of each alert type to support a variety of monitoring purposes.</p>



<h2 class="wp-block-heading"><strong>Which Alert Systems are Being Integrated?</strong></h2>



<p>GFW currently offers three individual deforestation alert systems. <a href="https://www.globalforestwatch.org/blog/data-and-research/glad-deforestation-alerts/" target="_blank" rel="noreferrer noopener">GLAD-L</a> (often called “GLAD”) is GFW’s longest-standing alert product from the University of Maryland’s (UMD) Global Land Analysis and Discovery (GLAD) lab and uses imagery from NASA’s Landsat satellites. UMD introduced the higher resolution <a href="https://www.globalforestwatch.org/blog/data-and-research/glad-s2-offers-high-resolution-deforestation-alerts/" target="_blank" rel="noreferrer noopener">GLAD-S2</a> alerts in May 2021, which use a similar methodology to GLAD-L, but are sourced from the European Space Agency’s Sentinel-2 satellites. And finally, Radar for Detecting Deforestation (<a href="https://www.globalforestwatch.org/blog/data-and-research/radd-radar-alerts/" target="_blank" rel="noreferrer noopener">RADD</a>) alerts produced by Wageningen University use Sentinel-1 radar data. These alerts have the additional capability of detecting forest change through cloud cover that often blocks the view of the other satellites.</p>



<div class="imagecenter"><img decoding="async" class=" wp-image-3569 alignright" style="padding-bottom: 10px;" alt="Deforestation alert layers offered on GFW" src="https://dev-global-forest-watch-blog.pantheonsite.io/wp-content/uploads/2023/01/23.01.17-Integrated-Alerts-Blog.png" width="100%"></div>



<h2 class="wp-block-heading"><strong>What Does the Integrated Alerts Layer Offer?</strong></h2>



<div class="imagecenter"><img decoding="async" class=" wp-image-3569 alignright" style="padding-bottom: 10px;" alt="integrated deforestation alerts layer combines the analytical power of GLAD, GLAD-S2 and RADD deforestation alerts" src="https://dev-global-forest-watch-blog.pantheonsite.io/wp-content/uploads/2021/12/integrated-deforestation-alerts-layer-2021.png" width="100%"></div>



<p>These three systems alert us of forest change happening in different parts of the world, gather different levels of resolution, and are updated at varying frequencies. The integrated option enables users to derive meaningful summary statistics that incorporate each of the individual systems. Using a common grid for all three alert products, the integrated layer understands that multiple systems have reported an alert in any given location and displays it through confidence levels:</p>



<div class="imagecenter"><img decoding="async" class=" wp-image-3569 alignright" style="padding-bottom: 10px;" alt="Confidence levels of integrated deforestation alerts" src="https://dev-global-forest-watch-blog.pantheonsite.io/wp-content/uploads/2021/12/21.12.01-Integrated-Alerts-Blog_Graphic-2-2-e1641415667423.png" width="100%"></div>



<p>This new layer comes with an important change: we are now reporting on the <em>area </em>affected by alerts as opposed to the <em>count of alert pixels.</em> As the integrated alert comes from systems with varying resolution, it is necessary to report areas rather than the number of alert pixels. However, be aware that the alerts are designed to quickly identify forest disturbances rather than precisely delineate forest loss – as a result, the areas presented on GFW may under or overestimate the actual area of forest loss.</p>



<p>Combining the alerts offers several benefits:</p>



<p><strong>1. Users Will Always Have the Earliest Information Available</strong></p>



<p>By combining all three alert systems into one layer, we are harnessing the different timing of the satellites to increase the chances of getting an unobscured look at the forest and therefore detect change faster. Users looking at integrated alerts in Peru from 2020 would have been alerted to forest loss on average 11 days earlier than by referring to any one system alone. This improvement in timeliness varies by region and across seasons, as the alert systems complement each other by providing their own strengths. For example, in consistently cloud-covered areas, the radar-based alert system provides the advantage of detecting forest loss through clouds. In areas with a regular dry season, however, the higher revisit time of the optical alert systems gives a more frequent opportunity to detect disturbances every 5-8 days where the radar system may take up to 12 days to detect change.</p>



<p><strong>2. The Overlap of Multiple Systems Increases Confidence to Prioritize Alerts</strong></p>



<p>If two or more alert systems detect a change in the same location, we have the highest confidence that these alerts indicate real disturbance. For individual systems, there is a delay before a first detection can be verified by additional satellite passes and thus reach “high confidence”. The integrated layer displays where multiple systems overlap, providing increased confidence faster than waiting for individual systems to reach high confidence through additional satellite images, which can take weeks or months. For forest monitors who rely on high confidence alerts to respond to deforestation events, we hope this new layer facilitates using the various alert systems in conjunction with one another and can promote faster follow up by identifying these highest confidence alerts as soon as a second system detects the same disturbance.</p>



<p><strong>3. Different Tree Cover Loss Types Are Captured Across a Variety of Landscapes </strong></p>



<p>The integrated alerts option helps to fill the gaps that the use of any one system alone might present. One of the greatest strengths of higher resolution RADD and GLAD-S2 alerts is their ability to capture smaller changes in the forest than the 30-meter GLAD-L product. This presents promising developments for detecting small canopy gaps from logging activities, which are widespread across the Amazon Basin. Now that all three systems are available in the Amazon basin, they can be used in concert to identify illegal activities quickly and with the highest confidence possible, a critical tool for both law enforcement and Indigenous Peoples and local communities working to combat deforestation.</p>



<iframe loading="lazy" width="670" height="490" frameborder="0" src="https://www.globalforestwatch.org/embed/map/?mainMap=eyJoaWRlTGVnZW5kIjp0cnVlfQ%3D%3D&map=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%3D"></iframe>



<div class="attribute"> The integrated deforestation alerts layer demonstrates the ability of multiple alert systems to detect small changes, such as tree removal along logging roads in the Peruvian Amazon.</div>



<h2 class="wp-block-heading"><strong>Get Started with Integrated Alerts</strong></h2>



<p>With three alert systems to choose from, we aim to simplify users’ workflow in identifying priority alerts. The integrated layer provides the most information on recent forest change for users who want to be notified of forest loss with the highest confidence possible. Though we see these systems as complimentary to each other, users may still opt to select individual systems for alert notifications and analysis if these suit their needs.</p>



<h2 class="wp-block-heading"><strong>What’s Next?</strong></h2>



<p>This integrated deforestation alerts layer is just the first step in combining the potential of multiple alert systems. A more comprehensive study is coming that will identify how to further integrate the systems into one downloadable dataset, assess how the different systems vary across regions, and factor in nearby pixels that detect the same disturbance. Stay tuned for future advancements in the field of forest monitoring.</p>



<p><em><em><em>For an in-depth look at how the integrated layer was designed and how to use it, see the <a href="https://www.globalforestwatch.org/help/forest-watcher/webinars/visualizing-analyzing-integrated-deforestation-alerts/">Help Center</a>.</em></em></em></p>



<hr> 

<div class="attribute" style="”padding-bottom:" 10px;”=""> 
Amy Pickens is a postdoc at University of Maryland and operates the GLAD alert systems.<br><br>

Johannes Reiche is an associate professor at Wageningen University and operates the RADD alert system.<br><br>

Yaqing Gou is a postdoc at Wageningen University and researches the various alert systems.<br>
</div>



<style type="text/css"></style>

Looking for the Quickest Signal of Deforestation? Turn to GFW’s Integrated Alerts

<p><em>Languages:&nbsp;<a href="https://www.globalforestwatch.org/blog/es/data-and-research/alertas-de-deforestacion-integradas/">Leerlo en español</a>, <a href="https://www.globalforestwatch.org/blog/pt/data-and-research/alertas-de-desmatamento-integrados/">Leia em portugues</a>, <a href="https://www.globalforestwatch.org/blog/id/data-and-research/peringatan-deforestasi-terintegrasi/">Baca dalam bahasa indonesia</a>, <a href="https://www.globalforestwatch.org/blog/fr/data-and-research/alertes-integrees-de-deforestation/">Lire en français</a></em></p>



<p>Global Forest Watch (GFW) is committed to bringing the most accurate forest data to users as quickly as possible. We are continually expanding the near-real-time deforestation alerts available on the GFW platform and are excited to introduce an updated layer. The integrated deforestation alerts layer combines the analytical power of GLAD, GLAD-S2 and RADD deforestation alerts to provide a faster, more confident view of forest disturbances than any one individual system. Prompted by user feedback, this update simplifies workflows and harnesses the best of each alert type to support a variety of monitoring purposes.</p>



<h2><strong>Which Alert Systems are Being Integrated?</strong></h2>



<p>GFW currently offers three individual deforestation alert systems. <a href="https://www.globalforestwatch.org/blog/data-and-research/glad-deforestation-alerts/" target="_blank" rel="noreferrer noopener">GLAD-L</a> (often called “GLAD”) is GFW’s longest-standing alert product from the University of Maryland's (UMD) Global Land Analysis and Discovery (GLAD) lab and uses imagery from NASA's Landsat satellites. UMD introduced the higher resolution <a href="https://www.globalforestwatch.org/blog/data-and-research/glad-s2-offers-high-resolution-deforestation-alerts/" target="_blank" rel="noreferrer noopener">GLAD-S2</a> alerts in May 2021, which use a similar methodology to GLAD-L, but are sourced from the European Space Agency's Sentinel-2 satellites. And finally, Radar for Detecting Deforestation (<a href="https://www.globalforestwatch.org/blog/data-and-research/radd-radar-alerts/" target="_blank" rel="noreferrer noopener">RADD</a>) alerts produced by Wageningen University use Sentinel-1 radar data. These alerts have the additional capability of detecting forest change through cloud cover that often blocks the view of the other satellites.</p>



<div class="imagecenter"><img class=" wp-image-3569 alignright" style="padding-bottom: 10px;" alt="Deforestation alert layers offered on GFW" src="https://dev-global-forest-watch-blog.pantheonsite.io/wp-content/uploads/2023/01/23.01.17-Integrated-Alerts-Blog.png" width="100%"></div>



<h2><strong>What Does the Integrated Alerts Layer Offer?</strong></h2>



<div class="imagecenter"><img class=" wp-image-3569 alignright" style="padding-bottom: 10px;" alt="integrated deforestation alerts layer combines the analytical power of GLAD, GLAD-S2 and RADD deforestation alerts" src="https://dev-global-forest-watch-blog.pantheonsite.io/wp-content/uploads/2021/12/integrated-deforestation-alerts-layer-2021.png" width="100%"></div>



<p>These three systems alert us of forest change happening in different parts of the world, gather different levels of resolution, and are updated at varying frequencies. The integrated option enables users to derive meaningful summary statistics that incorporate each of the individual systems. Using a common grid for all three alert products, the integrated layer understands that multiple systems have reported an alert in any given location and displays it through confidence levels:</p>



<div class="imagecenter"><img class=" wp-image-3569 alignright" style="padding-bottom: 10px;" alt="Confidence levels of integrated deforestation alerts" src="https://dev-global-forest-watch-blog.pantheonsite.io/wp-content/uploads/2021/12/21.12.01-Integrated-Alerts-Blog_Graphic-2-2-e1641415667423.png" width="100%"></div>



<p>This new layer comes with an important change: we are now reporting on the <em>area </em>affected by alerts as opposed to the <em>count of alert pixels.</em> As the integrated alert comes from systems with varying resolution, it is necessary to report areas rather than the number of alert pixels. However, be aware that the alerts are designed to quickly identify forest disturbances rather than precisely delineate forest loss – as a result, the areas presented on GFW may under or overestimate the actual area of forest loss.</p>



<p>Combining the alerts offers several benefits:</p>



<p><strong>1. Users Will Always Have the Earliest Information Available</strong></p>



<p>By combining all three alert systems into one layer, we are harnessing the different timing of the satellites to increase the chances of getting an unobscured look at the forest and therefore detect change faster. Users looking at integrated alerts in Peru from 2020 would have been alerted to forest loss on average 11 days earlier than by referring to any one system alone. This improvement in timeliness varies by region and across seasons, as the alert systems complement each other by providing their own strengths. For example, in consistently cloud-covered areas, the radar-based alert system provides the advantage of detecting forest loss through clouds. In areas with a regular dry season, however, the higher revisit time of the optical alert systems gives a more frequent opportunity to detect disturbances every 5-8 days where the radar system may take up to 12 days to detect change.</p>



<p><strong>2. The Overlap of Multiple Systems Increases Confidence to Prioritize Alerts</strong></p>



<p>If two or more alert systems detect a change in the same location, we have the highest confidence that these alerts indicate real disturbance. For individual systems, there is a delay before a first detection can be verified by additional satellite passes and thus reach "high confidence". The integrated layer displays where multiple systems overlap, providing increased confidence faster than waiting for individual systems to reach high confidence through additional satellite images, which can take weeks or months. For forest monitors who rely on high confidence alerts to respond to deforestation events, we hope this new layer facilitates using the various alert systems in conjunction with one another and can promote faster follow up by identifying these highest confidence alerts as soon as a second system detects the same disturbance.</p>



<p><strong>3. Different Tree Cover Loss Types Are Captured Across a Variety of Landscapes </strong></p>



<p>The integrated alerts option helps to fill the gaps that the use of any one system alone might present. One of the greatest strengths of higher resolution RADD and GLAD-S2 alerts is their ability to capture smaller changes in the forest than the 30-meter GLAD-L product. This presents promising developments for detecting small canopy gaps from logging activities, which are widespread across the Amazon Basin. Now that all three systems are available in the Amazon basin, they can be used in concert to identify illegal activities quickly and with the highest confidence possible, a critical tool for both law enforcement and Indigenous Peoples and local communities working to combat deforestation.</p>



<iframe width="670" height="490" frameborder="0" src="https://www.globalforestwatch.org/embed/map/?mainMap=eyJoaWRlTGVnZW5kIjp0cnVlfQ%3D%3D&amp;map=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%3D"></iframe>



<div class="attribute"> The integrated deforestation alerts layer demonstrates the ability of multiple alert systems to detect small changes, such as tree removal along logging roads in the Peruvian Amazon.</div>



<h2><strong>Get Started with Integrated Alerts</strong></h2>



<p>With three alert systems to choose from, we aim to simplify users’ workflow in identifying priority alerts. The integrated layer provides the most information on recent forest change for users who want to be notified of forest loss with the highest confidence possible. Though we see these systems as complimentary to each other, users may still opt to select individual systems for alert notifications and analysis if these suit their needs.</p>



<h2><strong>What's Next?</strong></h2>



<p>This integrated deforestation alerts layer is just the first step in combining the potential of multiple alert systems. A more comprehensive study is coming that will identify how to further integrate the systems into one downloadable dataset, assess how the different systems vary across regions, and factor in nearby pixels that detect the same disturbance. Stay tuned for future advancements in the field of forest monitoring.</p>



<p><em><em><em>For an in-depth look at how the integrated layer was designed and how to use it, see the <a href="https://www.globalforestwatch.org/help/forest-watcher/webinars/visualizing-analyzing-integrated-deforestation-alerts/">Help Center</a>.</em></em></em></p>



<hr> 

<div class="attribute" style="”padding-bottom:" 10px;”=""> 
Amy Pickens is a postdoc at University of Maryland and operates the GLAD alert systems.<br><br>

Johannes Reiche is an associate professor at Wageningen University and operates the RADD alert system.<br><br>

Yaqing Gou is a postdoc at Wageningen University and researches the various alert systems.<br>
</div>



<style type="text/css"></style>

New integrated deforestation alerts layer combines the analytical power of multiple alert systems for an impoved view of forest disturbances.

Explore GFW's New Integrated Deforestation Alerts

https://secure.gravatar.com/avatar/be19f67e91ba81d93d32ae47706dcbb8?s=96&d=mm&r=g

Paige Johnson

GCS REDD+, Madre de Dios, Peru.

data transparency

data transparency | Global Forest Watch Blog

University of Maryland

University of Maryland | Global Forest Watch Blog

Jessica Webb

<p>There’s a new deforestation alert system in town, and it’s monitoring forest loss across the Amazon Basin in greater detail than ever before. The new 10-meter resolution GLAD-S2 alerts, based on the European Space Agency’s Sentinel-2 satellite program and developed by the University of Maryland (UMD), are now available on Global Forest Watch (GFW) and offer another tool to track changes in the Amazon (and soon the entire humid tropics) in near-real-time.</p>



<p>Here’s what you need to know about the new system:</p>



<p><strong>1. GLAD-S2 alerts are the highest resolution alert system for the Amazon Basin</strong></p>



<p>The new alerts are based on the European Space Agency’s Sentinel-2 satellites with a resolution of 10 meters, which make them the highest resolution alert system available across the Amazon. The GLAD-S2 system offers much more detail than the Landsat-based GLAD alerts, with 6.25 GLAD-S2 alert pixels fitting inside a single GLAD pixel.</p>



<p>Because of their higher resolution, the new alerts pick up much more forest change than GLAD alerts. For example, in 2019, GLAD-S2 alerts detected about double the area of primary forest loss in Peru as GLAD. &nbsp;</p>



<p>The system brings improved resolution that is especially important for the detection of logging activities. The GLAD-S2 alerts pick up logging roads more consistently than GLAD, and also show small canopy gaps along those roads that are indicative of selective logging. Illegal logging is widespread <a href="https://news.mongabay.com/2020/03/brazil-drastically-reduces-controls-over-suspicious-amazon-timber-exports/" target="_blank" rel="noreferrer noopener">across</a> <a href="https://gfintegrity.org/press-release/new-trade-brief-finds-large-scale-trade-misinvoicing-in-colombian-timber-exports/" target="_blank" rel="noreferrer noopener">the</a> <a href="https://news.mongabay.com/2019/02/peru-report-reveals-high-rates-of-illegality-in-timber-extraction/" target="_blank" rel="noreferrer noopener">Amazon</a>, so being able to better track where logging is happening will provide a critical tool for law enforcement.</p>



<iframe loading="lazy" width="670" height="490" frameborder="0" src="https://www.globalforestwatch.org/embed/map/?mainMap=eyJoaWRlTGVnZW5kIjp0cnVlfQ%3D%3D&map=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%3D"></iframe>



<div class="attribute"> 

GLAD-S2 alerts detect a logging road and associated selective logging in the buffer zone of Cordillera Azul National Park, Peru.
</div>



<p><strong>2. GLAD-S2 can detect changes faster than existing alerts in the Amazon</strong></p>



<p>The Sentinel-2 satellites also have another advantage over Landsat, in that they cover every place on Earth every 5 days, compared to Landsat’s more than 8 days. That means we’re more likely to get a cloud-free image for a newly cleared area sooner than with Landsat, and can therefore detect the change and get it on GFW faster. In an assessment of the 2019 alerts in Peru, we found that GLAD-S2 detected new alerts an average of 11 days earlier than GLAD in 2019 – that difference rose to 33 days during the wet season.</p>



<p>Since <a href="https://www.globalforestwatch.org/blog/data-and-research/early-warning-systems-for-deforestation-an-explainer/" target="_blank" rel="noreferrer noopener">the purpose of alert systems</a> is to identify change as quickly as possible in order to promote action, the reduced delay from when a deforestation event happens to when it is detected is critical. It allows for an even more rapid response in the case of illegal or undesirable activity, which can ideally prevent more deforestation from happening.</p>



<iframe loading="lazy" width="670" height="490" frameborder="0" src="https://www.globalforestwatch.org/embed/map/wdpa/19995/?map=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%3D"></iframe>



<div class="attribute"> 

GLAD-S2 alerts detect deforestation in Colombia’s Tinigua National Park, sometimes weeks faster than GLAD alerts.
</div>



<p><strong>3. The new alerts cover primary forests in the Amazon Basin – for now</strong></p>



<p>Right now, the system covers humid tropical primary forests in the Amazon Basin, which is the focus of <a href="https://hivos.org/program/all-eyes-on-the-amazon/" target="_blank" rel="noreferrer noopener">All Eyes on the Amazon</a> — a project that supports Indigenous Peoples and local communities in combatting deforestation and protecting human rights, and which funded the alert development. However, UMD plans to expand the GLAD-S2 system to eventually cover the entire tropics. The Sentinel-2 satellites have global coverage and UMD has overcome many of the challenges of processing these images, which will enable rapid scaling. We hope to have the system operational across the tropics by the end of 2021.</p>



<iframe loading="lazy" width="670" height="490" frameborder="0" src="https://www.globalforestwatch.org/embed/map/?map=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"></iframe>



<div class="attribute"> 

GLAD-S2 alerts currently cover primary forests in the Amazon Basin.
</div>



<p><strong>4. We’re working to integrate GLAD-S2 and other alert systems on Global Forest Watch</strong></p>



<p>With the new GLAD-S2 system, GLAD alerts, and the radar-based <a href="https://www.globalforestwatch.org/blog/data-and-research/radd-radar-alerts/" target="_blank" rel="noreferrer noopener">RADD alerts</a> recently released for tropical Africa and Asia, there are likely to be a lot of questions about the differences in these systems. We do not recommend a single “best” system, but instead view these systems as complimentary rather than competing. When put together, these systems can provide users with even more actionable information to halt deforestation.</p>



<p>We’re currently working on an integrated system that will allow users to view all of the alerts together on GFW rather than have to pick a single system. Since the systems are completely independent from each other, if two or more detect an alert in the same location, it gives us a lot of confidence that there is a real change. And having three different systems allows the detection and confirmation of changes faster than any one of them alone. More information about this integrated system will be available soon.</p>



<p>Check out the GLAD-S2 alerts on the GFW platform <a href="https://gfw.global/3xMIUe4" target="_blank" rel="noreferrer noopener">here</a> or download the spatial data on UMD’s platform <a href="https://glad.earthengine.app/view/s2-forest-alerts" target="_blank" rel="noreferrer noopener">here</a>. We’re working to fully incorporate the GLAD-S2 alerts into other GFW tools, including email subscriptions, Forest Watcher, and MapBuilder – stay tuned!</p>



<hr> 

<div class="attribute" style="”padding-bottom:" 10px;”=""> 

 

Amy Pickens is a Faculty Specialist at University of Maryland and operates the GLAD-S2 Alert System.

 

</div> 

 

<hr>



<style type="text/css"></style>

Higher Resolution Alerts Offer More Detailed Picture of Forest Loss

<p>There's a new deforestation alert system in town, and it’s monitoring forest loss across the Amazon Basin in greater detail than ever before. The new 10-meter resolution GLAD-S2 alerts, based on the European Space Agency’s Sentinel-2 satellite program and developed by the University of Maryland (UMD), are now available on Global Forest Watch (GFW) and offer another tool to track changes in the Amazon (and soon the entire humid tropics) in near-real-time.</p>



<p>Here’s what you need to know about the new system:</p>



<p><strong>1. GLAD-S2 alerts are the highest resolution alert system for the Amazon Basin</strong></p>



<p>The new alerts are based on the European Space Agency’s Sentinel-2 satellites with a resolution of 10 meters, which make them the highest resolution alert system available across the Amazon. The GLAD-S2 system offers much more detail than the Landsat-based GLAD alerts, with 6.25 GLAD-S2 alert pixels fitting inside a single GLAD pixel.</p>



<p>Because of their higher resolution, the new alerts pick up much more forest change than GLAD alerts. For example, in 2019, GLAD-S2 alerts detected about double the area of primary forest loss in Peru as GLAD. &nbsp;</p>



<p>The system brings improved resolution that is especially important for the detection of logging activities. The GLAD-S2 alerts pick up logging roads more consistently than GLAD, and also show small canopy gaps along those roads that are indicative of selective logging. Illegal logging is widespread <a href="https://news.mongabay.com/2020/03/brazil-drastically-reduces-controls-over-suspicious-amazon-timber-exports/" target="_blank" rel="noreferrer noopener">across</a> <a href="https://gfintegrity.org/press-release/new-trade-brief-finds-large-scale-trade-misinvoicing-in-colombian-timber-exports/" target="_blank" rel="noreferrer noopener">the</a> <a href="https://news.mongabay.com/2019/02/peru-report-reveals-high-rates-of-illegality-in-timber-extraction/" target="_blank" rel="noreferrer noopener">Amazon</a>, so being able to better track where logging is happening will provide a critical tool for law enforcement.</p>



<iframe width="670" height="490" frameborder="0" src="https://www.globalforestwatch.org/embed/map/?mainMap=eyJoaWRlTGVnZW5kIjp0cnVlfQ%3D%3D&map=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%3D"></iframe>



<div class="attribute"> 

GLAD-S2 alerts detect a logging road and associated selective logging in the buffer zone of Cordillera Azul National Park, Peru.
</div>



<p><strong>2. GLAD-S2 can detect changes faster than existing alerts in the Amazon</strong></p>



<p>The Sentinel-2 satellites also have another advantage over Landsat, in that they cover every place on Earth every 5 days, compared to Landsat’s more than 8 days. That means we’re more likely to get a cloud-free image for a newly cleared area sooner than with Landsat, and can therefore detect the change and get it on GFW faster. In an assessment of the 2019 alerts in Peru, we found that GLAD-S2 detected new alerts an average of 11 days earlier than GLAD in 2019 – that difference rose to 33 days during the wet season.</p>



<p>Since <a href="https://www.globalforestwatch.org/blog/data-and-research/early-warning-systems-for-deforestation-an-explainer/" target="_blank" rel="noreferrer noopener">the purpose of alert systems</a> is to identify change as quickly as possible in order to promote action, the reduced delay from when a deforestation event happens to when it is detected is critical. It allows for an even more rapid response in the case of illegal or undesirable activity, which can ideally prevent more deforestation from happening.</p>



<iframe width="670" height="490" frameborder="0" src="https://www.globalforestwatch.org/embed/map/wdpa/19995/?map=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%3D"></iframe>



<div class="attribute"> 

GLAD-S2 alerts detect deforestation in Colombia’s Tinigua National Park, sometimes weeks faster than GLAD alerts.
</div>



<p><strong>3. The new alerts cover primary forests in the Amazon Basin – for now</strong></p>



<p>Right now, the system covers humid tropical primary forests in the Amazon Basin, which is the focus of <a href="https://hivos.org/program/all-eyes-on-the-amazon/" target="_blank" rel="noreferrer noopener">All Eyes on the Amazon</a> — a project that supports Indigenous Peoples and local communities in combatting deforestation and protecting human rights, and which funded the alert development. However, UMD plans to expand the GLAD-S2 system to eventually cover the entire tropics. The Sentinel-2 satellites have global coverage and UMD has overcome many of the challenges of processing these images, which will enable rapid scaling. We hope to have the system operational across the tropics by the end of 2021.</p>



<iframe width="670" height="490" frameborder="0" src="https://www.globalforestwatch.org/embed/map/?map=eyJjZW50ZXIiOnsibGF0IjotNC4xNDcxNjY1MDQ5MzAzNDQsImxuZyI6LTY0LjEwMTIzMDEzMDAyODcyfSwiem9vbSI6NC4xOTIyMzU2Njc2Njc4MDcsImRhdGFzZXRzIjpbeyJkYXRhc2V0IjoiZ2xhZC1zMi1kZWZvcmVzdGF0aW9uLWFsZXJ0cyIsIm9wYWNpdHkiOjEsInZpc2liaWxpdHkiOnRydWUsImxheWVycyI6WyJnbGFkLXMyLWRlZm9yZXN0YXRpb24tYWxlcnRzLTIwMjEiLCJnbGFkLXBsdXMtZGVmb3Jlc3RhdGlvbi1hbGVydHMtZ2VvZ3JhcGhpYy1jb3ZlcmFnZSJdLCJ0aW1lbGluZVBhcmFtcyI6eyJzdGFydERhdGUiOiIyMDE5LTAxLTAxIiwiZW5kRGF0ZSI6IjIwMjEtMDQtMTQiLCJ0cmltRW5kRGF0ZSI6IjIwMjEtMDQtMTQifX0seyJkYXRhc2V0IjoicG9saXRpY2FsLWJvdW5kYXJpZXMiLCJsYXllcnMiOlsiZGlzcHV0ZWQtcG9saXRpY2FsLWJvdW5kYXJpZXMiLCJwb2xpdGljYWwtYm91bmRhcmllcyJdLCJvcGFjaXR5IjoxLCJ2aXNpYmlsaXR5Ijp0cnVlfSx7ImRhdGFzZXQiOiJpbnRlZ3JhdGVkLWRlZm9yZXN0YXRpb24tYWxlcnRzLThiaXQiLCJvcGFjaXR5IjoxLCJ2aXNpYmlsaXR5Ijp0cnVlLCJsYXllcnMiOlsiaW50ZWdyYXRlZC1kZWZvcmVzdGF0aW9uLWFsZXJ0cy04Yml0LWdsYWRTIiwiZ2xhZFNPbmx5Il19LHsiZGF0YXNldCI6InByaW1hcnktZm9yZXN0cyIsIm9wYWNpdHkiOjEsInZpc2liaWxpdHkiOnRydWUsImxheWVycyI6WyJwcmltYXJ5LWZvcmVzdHMtMjAwMSJdfV19"></iframe>



<div class="attribute"> 

GLAD-S2 alerts currently cover primary forests in the Amazon Basin.
</div>



<p><strong>4. We’re working to integrate GLAD-S2 and other alert systems on Global Forest Watch</strong></p>



<p>With the new GLAD-S2 system, GLAD alerts, and the radar-based <a href="https://www.globalforestwatch.org/blog/data-and-research/radd-radar-alerts/" target="_blank" rel="noreferrer noopener">RADD alerts</a> recently released for tropical Africa and Asia, there are likely to be a lot of questions about the differences in these systems. We do not recommend a single “best” system, but instead view these systems as complimentary rather than competing. When put together, these systems can provide users with even more actionable information to halt deforestation.</p>



<p>We’re currently working on an integrated system that will allow users to view all of the alerts together on GFW rather than have to pick a single system. Since the systems are completely independent from each other, if two or more detect an alert in the same location, it gives us a lot of confidence that there is a real change. And having three different systems allows the detection and confirmation of changes faster than any one of them alone. More information about this integrated system will be available soon.</p>



<p>Check out the GLAD-S2 alerts on the GFW platform <a href="https://gfw.global/3xMIUe4" target="_blank" rel="noreferrer noopener">here</a> or download the spatial data on UMD’s platform <a href="https://glad.earthengine.app/view/s2-forest-alerts" target="_blank" rel="noreferrer noopener">here</a>. We’re working to fully incorporate the GLAD-S2 alerts into other GFW tools, including email subscriptions, Forest Watcher, and MapBuilder – stay tuned!</p>



<hr> 

<div class="attribute" style="”padding-bottom:" 10px;”=""> 

 

Amy Pickens is a Faculty Specialist at University of Maryland and operates the GLAD-S2 Alert System.

 

</div> 

 

<hr>



<style type="text/css"></style>

New GLAD-S2 deforestation alerts, derived from the European Space Agency’s Sentinel-2 satellites, offer a high resolution system for detecting forest loss.

Higher Resolution Alerts Offer More Detailed Picture of Forest Loss | Global Forest Watch Blog

https://secure.gravatar.com/avatar/0fa0ee27160b46f12c1f3e6cc973f07a?s=96&d=mm&r=g

Katie Fletcher

satellite data image of amazon rainforest

Satellite Imagery

Satellite Imagery | Global Forest Watch Blog

<p>The tree cover loss data from Hansen et al. 2013, featured on Global Forest Watch, represents the best available spatial data on how forests are changing around the world. It is unparalleled in its global coverage, multidecadal time span, annual update frequency and 30-meter pixel resolution. As such, it is a powerful tool for mapping recent tree cover loss  and has been widely used by governments, academia, private sector and civil society alike to identify and address deforestation.</p>



<p>The data remains so widely used in part because it continues to improve as relevant methods and technologies evolve. Here we describe the changes to the data over time, explore the implications for how this data should be interpreted and used, and offer a glimpse of our plans to improve the consistency of this data in the future.</p>



<h3 class="wp-block-heading">How and why has the tree cover loss data changed over time?</h3>



<p>There are a few factors that cause inconsistencies in the data over its 20-year span (read more on the <a href="https://storage.googleapis.com/earthenginepartners-hansen/GFC-2020-v1.8/download.html" target="_blank" rel="noreferrer noopener">UMD download page</a>):</p>



<ul type="1"><li><strong>Algorithm adjustments</strong>: The original algorithm used to map tree cover loss from satellite images for 2001-2012 has been improved upon in subsequent updates. These improvements allowed for annual updates starting in 2013 (coupled with <a href="https://blog.globalforestwatch.org/data-and-research/a-fresh-look-at-forests-2011-2013/?utm_campaign=BLOG:+Trend+Analysis+Rebuttle&utm_medium=bitly&utm_source=GFWBlog" target="_blank" rel="noreferrer noopener">small additions</a> to the 2011 and 2012 data) and have resulted in enhanced detection of loss— particularly from 2015 onwards.</li><li><strong>Improved satellite data: </strong>The algorithm uses every available Landsat image (from the Landsat 5, 7 and, since 2013, Landsat 8 satellites) to detect tree cover loss. The Landsat 8 satellite has the same resolution as previous Landsat missions, but has an improved sensor that can better resolve features on the ground. Incorporation of Landsat 8 data into the loss algorithm resulted in better detections of tree cover loss smaller than an individual pixel (e.g. selective logging) starting in 2013.</li><li><strong>Variations in satellite image availability: </strong>The number of Landsat images available per year has generally increased over time, allowing us more opportunities to “see” changes on the ground. The number of images in 2012 was particularly low given the gap between the decommissioning of Landsat 5 in 2011 and the launch of Landsat 8 in 2013. Clouds and smoke can also delay the detection of tree cover loss if there are no clear images available between when the tree cover loss event happened and the end of the year. This is particularly relevant in the case of <a href="https://blog.globalforestwatch.org/data-and-research/technical-blog-caveats-to-the-2016-tree-cover-loss-data/?utm_campaign=BLOG:+Trend+Analysis+Rebuttle&utm_medium=bitly&utm_source=GFWBlog" target="_blank" rel="noreferrer noopener">late-year fires</a>, which are often not detected until the following calendar year.</li></ul>



<p>In general, the first two factors have resulted in increased sensitivity in detecting tree cover loss, giving us a better understanding of changes that are typically difficult to map, including fires, selective logging and shifting agriculture. Certain areas benefitted more than others from these improvements— the Brazilian Amazon and Indonesia are dominated by large-scale clearing and so loss was already well captured by the initial algorithm, while West and Central Africa, Southeast Asia, and other areas with smaller-scale forest clearing were more affected by the improvements. In European and North American temperate forests, the improved forest loss data after the year 2015 is more sensitive to selective logging. Surface fires that partially degrade the tree canopy are also detected better in the recent data compared to the 2000-2012 original product.</p>



<p>The tradeoff for improved loss detection is the introduction of inconsistencies into the time series. To some extent, this is inevitable – science and data improve over time, and it’s not always possible or desirable to maintain 100% consistency. As an example, consider COVID-19 case numbers: testing is more available and well understood than it was in early 2020, but that doesn’t mean that case counts from early in the pandemic are useless.</p>



<h3 class="wp-block-heading">How can the data be used to analyze trends?</h3>



<p>The most robust way to analyze loss trends for a specific area and time period of interest is to perform a sampling exercise using <a href="https://www.sciencedirect.com/science/article/pii/S0034425714000704" target="_blank" rel="noreferrer noopener">appropriate statistical models</a>. The tree cover loss map can be used to identify where change likely happened each year, and a sample of points within and outside such areas can be inspected using Landsat time-series data (e.g., analysis ready data provided by the <a href="https://www.glad.umd.edu/ard" target="_blank" rel="noreferrer noopener">UMD GLAD lab</a>) and high resolution imagery. The reference data collected through such an exercise can be used to develop an unbiased estimate of the area of tree cover loss that occurred, as well as a measure of uncertainty around the estimate. These sample-based estimates are independent of the method changes in the mapped data.</p>



<p>However, that approach is not feasible to implement for every project or analysis when considering the needs for timeliness, the number of areas to assess, or budget. Here are some tips to keep in mind when using the tree cover loss data directly to assess trends:</p>



<ul><li>There have been minimal changes to the method and satellite data after the year 2015, so analysis of the recent past is less likely to include temporal inconsistencies.</li><li>Use 3-year moving averages to smooth any inconsistencies between years caused by the availability of Landsat images and better account for delays in detection of late-year fires and other losses.</li><li>When comparing time periods such as before or after 2011, or before and after 2015, be conscious that you’re comparing across model changes.</li><li>Apply caution when assessing the trend over time in areas more likely to be impacted by the changes (e.g. Central Africa), and note any changes in the trend aligning with the timing of the method changes (e.g. a large uptick starting in 2013 that may align with the incorporation of Landsat 8).</li><li>Contact the GFW or UMD team if you have questions about model performance change for a particular region.</li></ul>



<h3 class="wp-block-heading">Do the changes impact other uses of the data?</h3>



<p>For those using the tree cover loss map to understand where forest loss is happening, there is no impact. The model changes have impacted the sensitivity of the data to detect tree cover loss, so we should be capturing change even more comprehensively than before.</p>



<p>For those calculating area statistics for a given year but not comparing over time, there is not a big impact. The data can be used to give a relative indication of the extent of tree cover loss across different geographies.</p>



<h3 class="wp-block-heading">What changes will impact tree cover loss data going forward?</h3>



<p>The University of Maryland and GFW are working on an updated version of the tree cover loss algorithm to apply from 2001 onward, thus providing a more consistent record of loss over the entire time series. This “version 2.0” of the data will not only improve detection of loss in earlier years but will also allow us to detect multiple losses in the same pixel over time (e.g., in places where cyclical agriculture or forestry is happening) and distinguish areas of complete canopy removal from partial disturbances (e.g., from selective logging or understory fires).</p>



<p>The “version 2.0” will resolve the inconsistencies in the data related to the algorithm – however, improved satellite data and image availability will always be a factor that causes variation in the detail of detection. Landsat 8 has delivered constant quality and quantity since its launch in 2013, something that will hopefully be continued with the launch of Landsat 9, scheduled for late 2021.</p>



<p>However, as satellite data and methods continue to improve, it’s likely there will always be trade-offs between detecting current change as well as possible and ensuring consistency over time.</p>



<p>The tree cover loss data remains one of the best sources of information on forests out there due to its coverage, spatial resolution, and regular updates. Being aware of the caveats and changes to the data over time helps ensure that we’re properly taking advantage of this valuable resource.</p>



<style type="text/css"></style>

Assessing Trends in Tree Cover Loss Over 20 Years of Data

<p>The tree cover loss data from Hansen et al. 2013, featured on Global Forest Watch, represents the best available spatial data on how forests are changing around the world. It is unparalleled in its global coverage, multidecadal time span, annual update frequency and 30-meter pixel resolution. As such, it is a powerful tool for mapping recent tree cover loss  and has been widely used by governments, academia, private sector and civil society alike to identify and address deforestation.</p>



<p>The data remains so widely used in part because it continues to improve as relevant methods and technologies evolve. Here we describe the changes to the data over time, explore the implications for how this data should be interpreted and used, and offer a glimpse of our plans to improve the consistency of this data in the future.</p>



<h3>How and why has the tree cover loss data changed over time?</h3>



<p>There are a few factors that cause inconsistencies in the data over its 20-year span (read more on the <a href="https://storage.googleapis.com/earthenginepartners-hansen/GFC-2020-v1.8/download.html" target="_blank" rel="noreferrer noopener">UMD download page</a>):</p>



<ul type="1"><li><strong>Algorithm adjustments</strong>: The original algorithm used to map tree cover loss from satellite images for 2001-2012 has been improved upon in subsequent updates. These improvements allowed for annual updates starting in 2013 (coupled with <a href="https://blog.globalforestwatch.org/data-and-research/a-fresh-look-at-forests-2011-2013/?utm_campaign=BLOG:+Trend+Analysis+Rebuttle&amp;utm_medium=bitly&amp;utm_source=GFWBlog" target="_blank" rel="noreferrer noopener">small additions</a> to the 2011 and 2012 data) and have resulted in enhanced detection of loss— particularly from 2015 onwards.</li><li><strong>Improved satellite data: </strong>The algorithm uses every available Landsat image (from the Landsat 5, 7 and, since 2013, Landsat 8 satellites) to detect tree cover loss. The Landsat 8 satellite has the same resolution as previous Landsat missions, but has an improved sensor that can better resolve features on the ground. Incorporation of Landsat 8 data into the loss algorithm resulted in better detections of tree cover loss smaller than an individual pixel (e.g. selective logging) starting in 2013.</li><li><strong>Variations in satellite image availability: </strong>The number of Landsat images available per year has generally increased over time, allowing us more opportunities to “see” changes on the ground. The number of images in 2012 was particularly low given the gap between the decommissioning of Landsat 5 in 2011 and the launch of Landsat 8 in 2013. Clouds and smoke can also delay the detection of tree cover loss if there are no clear images available between when the tree cover loss event happened and the end of the year. This is particularly relevant in the case of <a href="https://blog.globalforestwatch.org/data-and-research/technical-blog-caveats-to-the-2016-tree-cover-loss-data/?utm_campaign=BLOG:+Trend+Analysis+Rebuttle&amp;utm_medium=bitly&amp;utm_source=GFWBlog" target="_blank" rel="noreferrer noopener">late-year fires</a>, which are often not detected until the following calendar year.</li></ul>



<p>In general, the first two factors have resulted in increased sensitivity in detecting tree cover loss, giving us a better understanding of changes that are typically difficult to map, including fires, selective logging and shifting agriculture. Certain areas benefitted more than others from these improvements— the Brazilian Amazon and Indonesia are dominated by large-scale clearing and so loss was already well captured by the initial algorithm, while West and Central Africa, Southeast Asia, and other areas with smaller-scale forest clearing were more affected by the improvements. In European and North American temperate forests, the improved forest loss data after the year 2015 is more sensitive to selective logging. Surface fires that partially degrade the tree canopy are also detected better in the recent data compared to the 2000-2012 original product.</p>



<p>The tradeoff for improved loss detection is the introduction of inconsistencies into the time series. To some extent, this is inevitable – science and data improve over time, and it’s not always possible or desirable to maintain 100% consistency. As an example, consider COVID-19 case numbers: testing is more available and well understood than it was in early 2020, but that doesn’t mean that case counts from early in the pandemic are useless.</p>



<h3>How can the data be used to analyze trends?</h3>



<p>The most robust way to analyze loss trends for a specific area and time period of interest is to perform a sampling exercise using <a href="https://www.sciencedirect.com/science/article/pii/S0034425714000704" target="_blank" rel="noreferrer noopener">appropriate statistical models</a>. The tree cover loss map can be used to identify where change likely happened each year, and a sample of points within and outside such areas can be inspected using Landsat time-series data (e.g., analysis ready data provided by the <a href="https://www.glad.umd.edu/ard" target="_blank" rel="noreferrer noopener">UMD GLAD lab</a>) and high resolution imagery. The reference data collected through such an exercise can be used to develop an unbiased estimate of the area of tree cover loss that occurred, as well as a measure of uncertainty around the estimate. These sample-based estimates are independent of the method changes in the mapped data.</p>



<p>However, that approach is not feasible to implement for every project or analysis when considering the needs for timeliness, the number of areas to assess, or budget. Here are some tips to keep in mind when using the tree cover loss data directly to assess trends:</p>



<ul><li>There have been minimal changes to the method and satellite data after the year 2015, so analysis of the recent past is less likely to include temporal inconsistencies.</li><li>Use 3-year moving averages to smooth any inconsistencies between years caused by the availability of Landsat images and better account for delays in detection of late-year fires and other losses.</li><li>When comparing time periods such as before or after 2011, or before and after 2015, be conscious that you’re comparing across model changes.</li><li>Apply caution when assessing the trend over time in areas more likely to be impacted by the changes (e.g. Central Africa), and note any changes in the trend aligning with the timing of the method changes (e.g. a large uptick starting in 2013 that may align with the incorporation of Landsat 8).</li><li>Contact the GFW or UMD team if you have questions about model performance change for a particular region.</li></ul>



<h3>Do the changes impact other uses of the data?</h3>



<p>For those using the tree cover loss map to understand where forest loss is happening, there is no impact. The model changes have impacted the sensitivity of the data to detect tree cover loss, so we should be capturing change even more comprehensively than before.</p>



<p>For those calculating area statistics for a given year but not comparing over time, there is not a big impact. The data can be used to give a relative indication of the extent of tree cover loss across different geographies.</p>



<h3>What changes will impact tree cover loss data going forward?</h3>



<p>The University of Maryland and GFW are working on an updated version of the tree cover loss algorithm to apply from 2001 onward, thus providing a more consistent record of loss over the entire time series. This “version 2.0” of the data will not only improve detection of loss in earlier years but will also allow us to detect multiple losses in the same pixel over time (e.g., in places where cyclical agriculture or forestry is happening) and distinguish areas of complete canopy removal from partial disturbances (e.g., from selective logging or understory fires).</p>



<p>The “version 2.0” will resolve the inconsistencies in the data related to the algorithm – however, improved satellite data and image availability will always be a factor that causes variation in the detail of detection. Landsat 8 has delivered constant quality and quantity since its launch in 2013, something that will hopefully be continued with the launch of Landsat 9, scheduled for late 2021.</p>



<p>However, as satellite data and methods continue to improve, it’s likely there will always be trade-offs between detecting current change as well as possible and ensuring consistency over time.</p>



<p>The tree cover loss data remains one of the best sources of information on forests out there due to its coverage, spatial resolution, and regular updates. Being aware of the caveats and changes to the data over time helps ensure that we’re properly taking advantage of this valuable resource.</p>

<p>Data hilangnya tutupan pohon yang dikembangkan oleh Hansen et al. (2013), yang ditampilkan di Global Forest Watch, merupakan data spasial terbaik mengenai perubahan hutan di seluruh dunia. Data ini unggul dalam cakupan global, rentang waktu puluhan tahun, frekuensi pembaruan tahunan, dan resolusi piksel 30 meternya. Dengan demikian, data ini menjadi alat yang ampuh untuk memetakan hilangnya tutupan pohon baru-baru ini di seluruh dunia dan telah digunakan secara luas oleh pemerintah, akademisi, sektor swasta, dan masyarakat sipil untuk mengidentifikasi dan mengatasi deforestasi.<br></p>



<p>Data ini masih digunakan secara luas karena terus ditingkatkan kualitasnya, seiring dengan berkembangnya metode dan teknologi terkait. Di sini kami menjelaskan perubahan melalui data dari waktu ke waktu, memaparkan bagaimana data ini ditafsirkan dan digunakan, serta menjelaskan secara singkat rencana kami untuk meningkatkan konsistensi data ini di masa mendatang.</p>



<h3>Bagaimana dan mengapa data hilangnya tutupan pohon berubah dari waktu ke waktu?</h3>



<p>Ada beberapa faktor yang menyebabkan inkonsistensiinkonsistensi data selama kurun waktu 20 tahun (baca selengkapnya di <a href="https://storage.googleapis.com/earthenginepartners-hansen/GFC-2020-v1.8/download.html">halaman unduh UMD</a>):</p>



<ul><li><strong>Penyesuaian algoritma:</strong> Algoritma pertama yang digunakan untuk memetakan hilangnya tutupan pohon dari citra satelit untuk 2001-2012 telah diperbaiki dalam pembaruan berikutnya. Peningkatan ini memungkinkan pembaruan tahunan mulai tahun 2013 (dilengkapi dengan <a href="https://blog.globalforestwatch.org/data-and-research/a-fresh-look-at-forests-2011-2013/">sedikit tambahan</a> pada data tahun 2011 dan 2012) dan telah menghasilkan pendeteksian lebih baik akan hilangnya tutupan pohon—terutama dari tahun 2015 dan seterusnya.</li><li><strong>Data satelit yang dimutakhirkan</strong>: Algoritma yang dikembangkan menggunakan semua citra Landsat yang tersedia (dari satelit Landsat 5, 7 dan sejak 2013, satelit Landsat 8) untuk mendeteksi hilangnya tutupan pohon. Satelit Landsat 8 memiliki resolusi yang sama dengan misi-misi Landsat sebelumnya, tetapi memiliki sensor yang lebih baik dan dapat melakukan fitur di darat dengan lebih baik. Penggabungan data Landsat 8 ke dalam algoritma menghasilkan deteksi hilangnya tutupan pohon yang lebih baik daripada piksel individual (misalnya, tebang pilih) sejak tahun 2013.</li><li><strong>Variasi dalam ketersediaan citra satelit</strong>: Jumlah citra Landsat yang tersedia per tahun umumnya makin meningkat seiring waktu dan memberikan lebih banyak peluang untuk “melihat” perubahan di lapangan. Jumlah citra pada tahun 2012 sangat sedikit akibat jeda waktu antara penonaktifan Landsat 5 pada tahun 2011 dan peluncuran Landsat 8 pada tahun 2013. Awan dan asap juga dapat menghambat deteksi hilangnya tutupan pohon jika tidak tersedia citra yang jelas antara saat kejadian hilangnya tutupan pohon hingga akhir tahun. Hal ini sangat relevan khususnya dengan kejadian <a href="https://blog.globalforestwatch.org/data-and-research/technical-blog-caveats-to-the-2016-tree-cover-loss-data/">kebakaran akhir tahun</a>, yang seringkali tidak terdeteksi hingga tahun berikutnya.</li></ul>



<p>Secara umum, dua faktor pertama telah menyebabkan peningkatan sensitivitas dalam mendeteksi hilangnya tutupan pohon, dan memberikan pemahaman lebih baik tentang perubahan yang biasanya sulit dipetakan, termasuk kebakaran, tebang pilih, dan pertanian berpindah. Daerah-daerah tertentu lebih terbantu daripada daerah lainnya akibat perbaikan ini— daerah Amazon di Brasil dan Indonesia didominasi oleh pembukaan lahan berskala besar sehingga hilangnya tutupan pohon sudah dapat dideteksi dengan baik oleh algoritma awal, sementara Afrika Barat dan Tengah, Asia Tenggara, dan daerah lain dengan skala pembukaan hutan yang lebih kecil, terbantu lebih baik oleh perbaikan ini. Di hutan beriklim Eropa dan Amerika Utara, peningkatan data kehilangan tutupan hutan sejak 2015 memberikan hasil yang lebih sensitif terhadap pendeteksian tebang pilih pohon. Kebakaran permukaan hutan yang dapat merusak sebagian kanopi pohon juga terdeteksi lebih baik dalam data terbaru tersebut dibandingkan dengan produk keluaran 2000-2012.</p>



<p>Sisi buruk dari peningkatan hasil deteksi kehilangan tutupan pohon ini adalah terjadinya inkonsistensi dalam serial waktu. Sampai batas tertentu, hal ini tidak dapat dihindari – sains dan data berkembang dari waktu ke waktu, dan tidak memungkinkan untuk berharap atau mempertahankan konsistensi 100%. Sebagai contoh, mari kita lihat jumlah kasus COVID-19: pengujian covid saat ini jauh lebih tersedia dan lebih mudah dipahami dibandingkan awal 2020, tetapi tidak berarti bahwa jumlah kasus di awal pandemi tidak ada gunanya.</p>



<h3>Bagaimana data dapat digunakan untuk menganalisis tren?</h3>



<p>Cara terbaik untuk menganalisis tren hilangnya tutupan pohon untuk area dan periode waktu tertentu yang dipilih adalah mengambil sampel dengan <a href="https://www.sciencedirect.com/science/article/pii/S0034425714000704" target="_blank" rel="noreferrer noopener">model statistik yang sesuai</a>. Peta hilangnya tutupan pohon dapat digunakan untuk mengidentifikasi lokasi kemungkinan terjadinya perubahan lahan setiap tahun, dan sampel titik di dalam dan di luar area tersebut dapat diteliti menggunakan data serial waktu Landsat (misalnya, data siap analisis yang disediakan oleh <a href="https://www.glad.umd.edu/ard" target="_blank" rel="noreferrer noopener">lab UMD GLAD</a>) dan citra satelit beresolusi tinggi. Data referensi yang dikumpulkan melalui kegiatan tersebut dapat digunakan untuk mengembangkan estimasi yang tidak bias dari area hilangnya tutupan pohon yang terjadi, serta menjadi ukuran ketidakpastian pada estimasinya. Estimasi berbasis sampel ini tidak bergantung pada perubahan metode dalam data yang dipetakan.</p>



<p>Namun, cara ini tidak dapat diterapkan pada setiap proyek atau analisis ketika mempertimbangkan ketepatan waktu, jumlah area yang akan dinilai, atau anggaran. Berikut adalah beberapa petunjuk yang perlu diingat saat menggunakan data hilangnya tutupan pohon secara langsung untuk menilai tren:</p>



<ul><li>Hanya ada sedikit perubahan pada metode dan data satelit setelah tahun 2015, sehingga analisis masa lalu cenderung tidak mencakup inkonsistensi temporal.</li><li>Gunakan pergerakan rata-rata 3 tahun untuk meminimalisir data yang tidak konsisten antar tahun yang diakibatkan ketersediaan citra satelit Landsat dan lebih baik memperhitungkan keterlambatan deteksi kebakaran akhir tahun dan kehilangan tutupan lahan lainnya.</li><li>Saat membandingkan periode waktu seperti sebelum atau setelah 2011, atau sebelum dan setelah 2015, ketahuilah bahwa Anda membandingkan seluruh perubahan model.</li><li>Berhati-hatilah saat menilai tren dari waktu ke waktu di area yang lebih mungkin terkena dampak perubahan (seperti Afrika Tengah), dan catat setiap perubahan tren yang terjadi seiring waktu perubahan metode (misalnya kenaikan besar mulai tahun 2013 yang mungkin terjadi di waktu yang sama dengan penggunaan Landsat 8).</li><li>Hubungi tim GFW atau UMD jika Anda ingin bertanya tentang perubahan kinerja model untuk wilayah tertentu.</li></ul>



<h3>Apakah perubahan tersebut mempengaruhi penggunaan data lainnya?</h3>



<p>Tidak ada dampak tertentu bagi pihak yang menggunakan peta kehilangan tutupan pohon untuk menganalisis terjadinya perubahan tutupan hutan. Perubahan model telah mempengaruhi sensitivitas data untuk mendeteksi hilangnya tutupan pohon, sehingga kita dapat mendeteksi perubahan secara lebih menyeluruh dari sebelumnya.</p>



<p>Demikian juga tidak ada pengaruh yang besar saat menghitung statistik secara meluas untuk tahun tertentu tetapi tidak membandingkan dari waktu ke waktu. Data tersebut dapat digunakan untuk memberikan indikasi relatif tentang tingkat kehilangan tutupan pohon di berbagai bentang lahan.</p>



<h3>Perubahan apa yang mempengaruhi data kehilangan tutupan pohon di masa mendatang?</h3>



<p>University of Maryland dan GFW sedang mengerjakan versi terbaru dari algoritma kehilangan tutupan pohon untuk diterapkan mulai tahun 2001, agar menghasilkan catatan kehilangan yang lebih konsisten di seluruh rangkaian waktu. Data "versi 2.0" ini tidak hanya mampu meningkatkan deteksi kehilangan di tahun-tahun sebelumnya, tetapi juga dapat mendeteksi sejumlah kehilangan di piksel yang sama seiring waktu (misalnya, di tempat-tempat di mana siklus pertanian atau kehutanan sedang berlangsung) dan membedakan daerah-daerah yang lahannya ditebangi seluruhnya dari daerah yang mengalami gangguan parsial (misalnya, dari tebang pilih atau kebakaran di bawah permukaan tanah).</p>



<p>“Versi 2.0” ini akan menuntaskan inkonsistensi dalam data yang terkait dengan algoritma, tetapi peningkatan ketersediaan data dan citra satelit akan selalu menjadi faktor penyebab variasi dalam detail pendeteksian. Landsat 8 telah memberikan kualitas dan kuantitas yang konstan sejak diluncurkan pada tahun 2013 dandiharapkan akan berlanjut dengan peluncuran Landsat 9, yang dijadwalkan pada akhir tahun 2021.</p>



<p>Namun, seiring dengan data citra satelit dan metode yang terus diperbarui, kemungkinan akan selalu ada kompromi antara mendeteksi sebaik mungkin perubahan saat ini dan memastikan konsistensi dari waktu ke waktu. Data kehilangan tutupan pohon masih menjadi salah satu sumber informasi terbaik tentang hutan karena cakupan, resolusi spasial, dan pembaruan rutinnya. Dengan menyadari kekurangan dan perubahan pada data dari waktu ke waktu, kita bisa memanfaatkan sumber daya yang berharga ini dengan sebaik-baiknya.</p>

Menilai tren hilangnya tutupan pohon dalam kurun waktu data 20 tahun

Peter Potapov

<p>Los datos sobre la pérdida de cobertura arbórea de Hansen et al. (2013) presentados en Global Forest Watch (GFW) constituyen la mejor información geoespacial disponible para analizar cómo están cambiando los bosques en todo el mundo. Estos datos son de una calidad inigualable, pues presentan una amplia cobertura mundial, abarcan varias décadas, se actualizan anualmente y cuentan con una resolución de píxeles de 30 metros. Así pues, constituyen una herramienta de gran valor para mapear la pérdida reciente de cobertura arbórea a escala mundial y su uso se ha extendido entre gobiernos, el mundo académico, el sector privado y la sociedad civil a fin de identificar y abordar la deforestación.</p>



<p>Uno de los motivos por los que estos datos siguen siendo muy populares consiste en que su calidad mejora a medida que evolucionan los métodos y las tecnologías pertinentes. En este documento, describimos los cambios a lo largo del tiempo, exploramos las posibles repercusiones sobre cómo debería interpretarse y utilizarse esta información, y presentamos una perspectiva de nuestro plan para mejorar la consistencia de los datos en el futuro.</p>



<h3>¿Cómo y por qué han cambiado los datos sobre la pérdida de cobertura arbórea a lo largo del tiempo?</h3>



<p>Existen varios factores que causan inconsistencias en los datos durante sus 20 años de uso (puede obtener más información en la <a href="https://storage.googleapis.com/earthenginepartners-hansen/GFC-2020-v1.8/download.html" target="_blank" rel="noreferrer noopener">página de descargas de la Universidad de Maryland</a>):</p>



<ul><li><strong>Ajustes del algoritmo: </strong>El algoritmo original utilizado para mapear la pérdida de cobertura arbórea entre 2001-2012 a partir de imágenes satelitales se ha mejorado con sucesivas actualizaciones. Gracias a estas mejoras, se pudieron hacer análisis anuales a partir de 2013 —con <a href="https://blog.globalforestwatch.org/data-and-research/a-fresh-look-at-forests-2011-2013/" target="_blank" rel="noreferrer noopener">pequeñas incorporaciones</a> a los datos de 2011 y 2012— y han permitido mejorar la detección de las pérdidas, especialmente a partir de 2015.</li><li><strong>Mejores datos satelitales:</strong> El algoritmo utiliza todas las imágenes obtenidas mediante los satélites Landsat disponibles para detectar la pérdida de cobertura arbórea (de los satélites Landsat 5, 7 y, desde 2013, del Landsat 8). La resolución del Landsat 8 es la misma que la de  misiones Landsat anteriores, pero cuenta con un sensor mejorado que ofrece imágenes del terreno de mayor calidad. Desde 2013, la incorporación de los datos del Landsat 8 al algoritmo de pérdidas ha mejorado la detección de las pérdidas de cobertura arbórea de un tamaño menor al de un píxel individual (como la tala selectiva).</li><li><strong>Variaciones en la disponibilidad de imágenes satelitales:</strong> En general, el número de imágenes Landsat disponibles ha aumentado con el tiempo, lo que nos ha permitido visualizar mejor los cambios sobre el terreno. En 2012, el número de imágenes disponible fue bajo debido a la brecha entre el retiro del Landsat 5 en 2011 y el lanzamiento del Landsat 8 en 2013. Asimismo, las nubes y el humo pueden retrasar la detección de las pérdidas de cobertura arbórea si no se dispone de imágenes nítidas recabadas entre el momento en el que se produjo dicha pérdida y el final del año en cuestión. Esto es especialmente relevante en el caso de <a href="https://blog.globalforestwatch.org/data-and-research/technical-blog-caveats-to-the-2016-tree-cover-loss-data/" target="_blank" rel="noreferrer noopener">los incendios que se producen a finales de año</a>, puesto que a menudo no se detectan hasta el comienzo del siguiente año calendario.</li></ul>



<p>Por lo general, los dos primeros factores han aumentado la sensibilidad a la hora de detectar la pérdida de cobertura arbórea, lo que nos permite un mejor entendimiento de los cambios que suelen ser difíciles de mapear, como los incendios, la tala selectiva y la agricultura itinerante. En ciertas zonas, estas mejoras han tenido una mayor repercusión: la deforestación a gran escala está muy extendida en la Amazonía Brasileña e Indonesia, por lo que el algoritmo original ya había documentado gran parte de estas pérdidas; por el contrario, las mejoras tuvieron más impacto en África Occidental y Central y el sudeste de Asia, así como en otras zonas en las que la deforestación se produce a menor escala. En los bosques templados de Europa y América del Norte, los datos de mayor calidad sobre la pérdida de cobertura arbórea registrados a partir de 2015 son más sensibles a la tala selectiva. Además, los datos más recientes detectan mejor los incendios superficiales que degradan parcialmente el dosel forestal, en comparación con los recabados para el período 2000 y 2012.</p>



<p>La desventaja resultante de detectar mejor las pérdidas es la aparición de inconsistencias en las series cronológicas. Esto es inevitable hasta cierto punto, ya que la ciencia y los datos mejoran con el tiempo y no siempre es posible o deseable mantener una consistencia total. Un ejemplo son las cifras de casos de COVID-19: ahora hay una mayor disponibilidad de pruebas y entendemos mejor cómo usarlas que a principios de 2020, pero ello no implica que los recuentos de casos efectuados al comienzo de la pandemia carezcan de utilidad.</p>



<h3>¿Cómo podemos usar los datos para analizar las tendencias?</h3>



<p>El método más fiable para analizar las tendencias de las pérdidas en una zona concreta y un periodo de tiempo específico es llevar a cabo un ejercicio de muestreo mediante los <a href="https://www.sciencedirect.com/science/article/pii/S0034425714000704" target="_blank" rel="noreferrer noopener">modelos estadísticos adecuados</a>. El mapa de la pérdida de cobertura arbórea resultante puede emplearse para identificar los puntos en los que es probable que se produjeran cambios cada año, y se puede estudiar una muestra de puntos tanto dentro como fuera de dichas zonas por medio de las series cronológicas de datos Landsat —por ejemplo, datos listos para su análisis y proporcionados por el <a href="https://www.glad.umd.edu/ard" target="_blank" rel="noreferrer noopener">laboratorio Global Land Analysis and Discovery de la Universidad de Maryland</a>— e imágenes de alta resolución. Los datos de referencia recopilados en este tipo de muestreo pueden servir para desarrollar una estimación no sesgada de la zona de pérdida de cobertura arbórea, así como una medida de la incertidumbre de dicha estimación. Los cálculos basados en muestras son independientes de los cambios metodológicos aplicados a los datos resultantes del mapeo.</p>



<p>Sin embargo, ese enfoque no puede aplicarse a cada proyecto o análisis si tenemos en cuenta los plazos, el número de zonas que han de evaluarse o el presupuesto. A continuación, se enumeran algunos consejos para utilizar los datos sobre la pérdida de cobertura arbórea directamente en la evaluación de las tendencias:</p>



<ul><li>A partir del 2015 ha habido modificaciones mínimas a los métodos o los datos satelitales, por lo que es menos probable que haya inconsistencias si se hace un análisis de un periodo más reciente.</li><li>Utilice un promedio móvil de tres años para compensar las inconsistencias interanuales causadas por la disponibilidad de imágenes Landsat y para integrar mejor los retrasos en la detección de los incendios de finales de año y otras pérdidas.</li><li>Cuando compare periodos anteriores o posteriores a 2011 ó 2015, tenga en cuenta que los periodos comparados utilizan versiones distintas del modelo.</li><li>Tenga cautela al evaluar las tendencias a lo largo del tiempo en zonas con una mayor probabilidad de verse afectadas por los cambios (por ejemplo, África Central) y preste atención especial a los cambios que se produjeron en las tendencias cuando se introdujeron modificaciones metodológicas (como un repunte pronunciado que empiece en 2013 y que tal vez se deba a la incorporación del Landsat 8).</li><li>Póngase en contacto con el equipo de GFW o de la Universidad de Maryland si tiene preguntas sobre cambios en el rendimiento del modelo en una región específica.</li></ul>



<h3>¿Influyen los cambios en otros usos de los datos?</h3>



<p>No hay impactos para quienes utilicen el mapa de pérdida de cobertura arbórea para entender dónde se está produciendo la pérdida de bosques. Los cambios en el modelo han repercutido en la sensibilidad de los datos para detectar la pérdida de cobertura arbórea, por lo que deberíamos captar los cambios de forma aún más exhaustiva que antes.</p>



<p>Los cambios tampoco afectarán de manera significativa a quienes usen los datos para calcular estadísticas de superficie de un año determinado sin analizar su evolución. Los datos pueden utilizarse para formular una indicación relativa de la magnitud de la pérdida de cobertura arbórea en diferentes zonas geográficas.</p>



<h3>¿Qué cambios influirán en los datos de pérdida de cobertura arbórea en el futuro?</h3>



<p>La Universidad de Maryland y GFW están desarrollando una versión actualizada del algoritmo de pérdida de cobertura arbórea para aplicarlo a partir de 2001, lo que daría lugar a un registro más consistente de las pérdidas a lo largo de toda la serie cronológica. Esta "versión 2.0" de los datos no solo mejorará la detección de las pérdidas de años anteriores, sino que también nos permitirá detectar múltiples pérdidas en el mismo píxel a lo largo del tiempo (por ejemplo, en lugares en los que la agricultura o la silvicultura son cíclicas) y diferenciar las zonas en las que el dosel se ha eliminado por completo de las zonas que presentan perturbaciones parciales (como las originadas por la tala selectiva o los incendios del sotobosque).</p>



<p>La "versión 2.0" solucionará las inconsistencias de los datos relacionadas con el algoritmo; no obstante, la mejora de los datos satelitales y la disponibilidad de las imágenes siempre será un factor que propicie variaciones en la minuciosidad de la detección. Desde el lanzamiento del Landsat 8 en 2013, la cantidad de imágenes generadas ha sido constante, al igual que su calidad; se espera que la situación continúe así con la presentación del Landsat 9, prevista para finales de 2021.</p>



<p>Sin embargo, los datos satelitales y los métodos seguirán mejorando, por lo que es probable que siempre existan conflictos entre detectar, de forma óptima, los cambios actuales y garantizar la consistencia a lo largo del tiempo. &nbsp;</p>



<p>Los datos relativos a la pérdida de cobertura arbórea siguen constituyendo una de las mejores fuentes de información sobre los bosques que existen debido a su cobertura, resolución espacial y actualizaciones periódicas. Ser conscientes de las limitaciones y los cambios en los datos a lo largo del tiempo ayuda a asegurar que este recurso tan valioso sea aprovechado debidamente.</p>

Evaluación de las tendencias relativas a la pérdida de cobertura arbórea a partir de datos recopilados durante 20 años

As satellite data has improved over time, tree cover loss detection has improved with it. Here's what that means for analyzing forest trends

How Tree Cover Loss Data Has Changed Over Time | GFW Blog

https://secure.gravatar.com/avatar/3dc298e411f65b199fc3734919d6934e?s=96&d=mm&r=g

Sarah Ruiz

Fires

Learn about the complex relationship between forests and fires.

Fires | Global Forest Watch Blog

agriculture

agriculture | Global Forest Watch Blog

Cerrado

Cerrado | Global Forest Watch Blog

Fire Ban

Fire Ban | Global Forest Watch Blog

Luiz Amaral

Marcelo Matsumoto

Thailynn Munroe

<p>Fires across Brazil caught the world’s attention this August, with fears that burning — and subsequent damage to the Amazon rainforest — would only intensify when peak fire season hit in September. However, Global Forest Watch (GFW) analysis of NASA data shows that fire alerts in the Amazon decreased by 34% between August and September.</p>



<h3 class="wp-block-heading">Amazon Fires Are Down. Can We Thank Brazil’s Fire Ban?</h3>



<p>Fire season in Brazil typically peaks in September, so this reduction in fires in the Amazon is unexpected. Other forested regions in Brazil showed an increase in fire alerts in September, including the Cerrado — a tropical savanna where fire is a more natural feature of the landscape. In the wetter, Amazon region, fires are rarely natural and almost all burning is related to human activity.</p>



<div class="imagecenter"><img decoding="async" class=" wp-image-3569 alignright" style="padding-bottom: 10px;" alt="fire counts, August to september" src="https://dev-global-forest-watch-blog.pantheonsite.io/wp-content/uploads/2019/10/Fire-Alerts-03.png" width="100%">Amazônia—the Amazon biome— was the only region of Brazil that saw a drop in fires in September, which deviates from the normal upward trend in fires across the country.</div>



<p>So what explains the 34% decrease in fire alerts in the Amazon? It likely wasn’t the weather, as September was slightly drier than normal. The other potential explanation is the fire ban enacted at the end of August. During the peak of the crisis, when both national and international media attention was focused on the Amazon, President Jair Bolsonaro issued a TV and radio announcement declaring “zero tolerance” for environmental crimes. Law enforcement was dispatched to fight the fires and the government enacted a two-month ban on the use of fire to clear land. This ban was in effect across Brazil, but stricter rules applied within the Amazon, where we have seen the greatest reduction in fire alerts.</p>



<h3 class="wp-block-heading">Forests on Privately Owned Land Disproportionately Affected</h3>



<p>Even though private properties make up only 48% of the land in the Amazon, they accounted for 62% of the fire alerts detected in the Amazon region in August and September. These fires are usually set to prepare agricultural land for production – an authorized practice before the ban.</p>



<div class="imagecenter"><img decoding="async" class=" wp-image-3569 alignright" style="padding-bottom: 10px;" alt="land use categories in Brazil" src="https://dev-global-forest-watch-blog.pantheonsite.io/wp-content/uploads/2019/10/Fire-Concentration-02.png" width="100%"></div>



<p>While the fires typically start on farmland, they often affect nearby forests. Farmers are legally required to set aside up to 80% of their land for conservation purposes and much of this land is forested. Over one-third of the fire alerts on private lands in August and September were detected in primary forests.</p>



<p>Roughly 14% of fires over the same time period occurred on conservation units and indigenous lands in the Amazon, which comprise 24% of the region and tend to be even more forested than private lands. Fires in conservation units are never legal, and these areas may also be affected by runaway fires from nearby land-clearing activities. Fires on indigenous lands could be part of the land management practices used by the community, since the fire ban did not apply to indigenous lands. Alternatively, these fires could also represent illegal encroachment from outside actors.</p>



<div class="imagecenter"><img decoding="async" class=" wp-image-3569 alignright" style="padding-bottom: 10px;" alt="Distribution of fires alert" src="https://dev-global-forest-watch-blog.pantheonsite.io/wp-content/uploads/2019/10/PieChart.png" width="100%"></div>



<h3 class="wp-block-heading">Dry Forests of the Cerrado Are Still Burning</h3>



<p>Although fires have gone down significantly within the Amazon, they are still burning in the Cerrado. The dry forest of the Cerrado is a unique ecosystem that plays a critical role in storing carbon and maintaining watersheds. Fires can occur naturally here, and fire alerts detected from satellite cannot differentiate natural fires from fires set by humans. So far this year, fire counts for the Cerrado are in line with historical averages; however, the drying effects of climate change and increased agricultural activity in the region could result in increased fire threat to these forests in the future.</p>



<div class="imagecenter"><img decoding="async" class=" wp-image-3569 alignright" style="padding-bottom: 10px;" alt="fire concentration August to september" src="https://dev-global-forest-watch-blog.pantheonsite.io/wp-content/uploads/2019/10/Fire-Concentration-01-1.png" width="100%">The highest concentration of fires moved from the Amazon biome—where the majority of Brazil’s primary forest is located— to the Cerrado and other parts of central Brazil in September.</div>



<h3 class="wp-block-heading">Future Fires Seasons</h3>



<p>The fire ban is scheduled to lift today, October 28, although Brazil’s fire season can continue into November. The reduction in Amazon fires seen in September suggests that recent attention given to this region – including the fire ban, public discourse and environmental enforcement policies – may have had a positive impact on curbing fires. Other critical forest biomes in Brazil have not received the same attention nor experienced the same reduction in fires.</p>



<p>Continued public and government attention to this issue will be critical for the remainder of this year’s fire season and beyond. The immediate aftermath of this year’s Amazon fire crisis demonstrates just how important this spotlight is.</p>



<hr> 

<div class="attribute" style="”padding-bottom:" 10px;”=""> 

BANNER PHOTO: Satellite image of Brazil fires, August 22, 2019. Photo by NASA Goddard Space Flight Center.

 <p></p>

Originally appeared on <a href="https://www.wri.org/blog/2019/10/brazil-s-fire-ban-correlates-reduction-amazon-wildfires-ban-lifts-today" target="_blank" rel="noopener noreferrer">WRI Insights</a>. 
<p></p>
Download the raw data for these analyses <a href="http://gfw-files.s3.amazonaws.com/fires/final_fires_database/FINAL%20DATABASE%20-%20FIRES%20-%20BR%20-%20MODIS%20-%20AUG-SEP%20-%202019%20-%20Simplfied.xlsx" target="_blank" rel="noopener noreferrer">here</a>. Data for this analysis was sourced from NASA, Imaflora and ESALQ. 
</div> 

 

<hr>



<style type="text/css"></style>

Brazil’s Fire Ban Correlates a Reduction in Amazon Wildfires. The Ban Lifts Today.

<p>Fires across Brazil caught the world’s attention this August, with fears that burning — and subsequent damage to the Amazon rainforest — would only intensify when peak fire season hit in September. However, Global Forest Watch (GFW) analysis of NASA data shows that fire alerts in the Amazon decreased by 34% between August and September.</p>



<h3>Amazon Fires Are Down. Can We Thank Brazil’s Fire Ban?</h3>



<p>Fire season in Brazil typically peaks in September, so this reduction in fires in the Amazon is unexpected. Other forested regions in Brazil showed an increase in fire alerts in September, including the Cerrado — a tropical savanna where fire is a more natural feature of the landscape. In the wetter, Amazon region, fires are rarely natural and almost all burning is related to human activity.</p>



<div class="imagecenter"><img class=" wp-image-3569 alignright" style="padding-bottom: 10px;" alt="fire counts, August to september" src="https://dev-global-forest-watch-blog.pantheonsite.io/wp-content/uploads/2019/10/Fire-Alerts-03.png" width="100%">Amazônia—the Amazon biome— was the only region of Brazil that saw a drop in fires in September, which deviates from the normal upward trend in fires across the country.</div>



<p>So what explains the 34% decrease in fire alerts in the Amazon? It likely wasn’t the weather, as September was slightly drier than normal. The other potential explanation is the fire ban enacted at the end of August. During the peak of the crisis, when both national and international media attention was focused on the Amazon, President Jair Bolsonaro issued a TV and radio announcement declaring “zero tolerance” for environmental crimes. Law enforcement was dispatched to fight the fires and the government enacted a two-month ban on the use of fire to clear land. This ban was in effect across Brazil, but stricter rules applied within the Amazon, where we have seen the greatest reduction in fire alerts.</p>



<h3>Forests on Privately Owned Land Disproportionately Affected</h3>



<p>Even though private properties make up only 48% of the land in the Amazon, they accounted for 62% of the fire alerts detected in the Amazon region in August and September. These fires are usually set to prepare agricultural land for production – an authorized practice before the ban.</p>



<div class="imagecenter"><img class=" wp-image-3569 alignright" style="padding-bottom: 10px;" alt="land use categories in Brazil" src="https://dev-global-forest-watch-blog.pantheonsite.io/wp-content/uploads/2019/10/Fire-Concentration-02.png" width="100%"></div>



<p>While the fires typically start on farmland, they often affect nearby forests. Farmers are legally required to set aside up to 80% of their land for conservation purposes and much of this land is forested. Over one-third of the fire alerts on private lands in August and September were detected in primary forests.</p>



<p>Roughly 14% of fires over the same time period occurred on conservation units and indigenous lands in the Amazon, which comprise 24% of the region and tend to be even more forested than private lands. Fires in conservation units are never legal, and these areas may also be affected by runaway fires from nearby land-clearing activities. Fires on indigenous lands could be part of the land management practices used by the community, since the fire ban did not apply to indigenous lands. Alternatively, these fires could also represent illegal encroachment from outside actors.</p>



<div class="imagecenter"><img class=" wp-image-3569 alignright" style="padding-bottom: 10px;" alt="Distribution of fires alert" src="https://dev-global-forest-watch-blog.pantheonsite.io/wp-content/uploads/2019/10/PieChart.png" width="100%"></div>



<h3>Dry Forests of the Cerrado Are Still Burning</h3>



<p>Although fires have gone down significantly within the Amazon, they are still burning in the Cerrado. The dry forest of the Cerrado is a unique ecosystem that plays a critical role in storing carbon and maintaining watersheds. Fires can occur naturally here, and fire alerts detected from satellite cannot differentiate natural fires from fires set by humans. So far this year, fire counts for the Cerrado are in line with historical averages; however, the drying effects of climate change and increased agricultural activity in the region could result in increased fire threat to these forests in the future.</p>



<div class="imagecenter"><img class=" wp-image-3569 alignright" style="padding-bottom: 10px;" alt="fire concentration August to september" src="https://dev-global-forest-watch-blog.pantheonsite.io/wp-content/uploads/2019/10/Fire-Concentration-01-1.png" width="100%">The highest concentration of fires moved from the Amazon biome—where the majority of Brazil’s primary forest is located— to the Cerrado and other parts of central Brazil in September.</div>



<h3>Future Fires Seasons</h3>



<p>The fire ban is scheduled to lift today, October 28, although Brazil’s fire season can continue into November. The reduction in Amazon fires seen in September suggests that recent attention given to this region – including the fire ban, public discourse and environmental enforcement policies – may have had a positive impact on curbing fires. Other critical forest biomes in Brazil have not received the same attention nor experienced the same reduction in fires.</p>



<p>Continued public and government attention to this issue will be critical for the remainder of this year’s fire season and beyond. The immediate aftermath of this year’s Amazon fire crisis demonstrates just how important this spotlight is.</p>



<hr> 

<div class="attribute" style="”padding-bottom:" 10px;”=""> 

BANNER PHOTO: Satellite image of Brazil fires, August 22, 2019. Photo by NASA Goddard Space Flight Center.

 <p></p>

Originally appeared on <a href="https://www.wri.org/blog/2019/10/brazil-s-fire-ban-correlates-reduction-amazon-wildfires-ban-lifts-today" target="_blank" rel="noopener noreferrer">WRI Insights</a>. 
<p></p>
Download the raw data for these analyses <a href="http://gfw-files.s3.amazonaws.com/fires/final_fires_database/FINAL%20DATABASE%20-%20FIRES%20-%20BR%20-%20MODIS%20-%20AUG-SEP%20-%202019%20-%20Simplfied.xlsx" target="_blank" rel="noopener noreferrer">here</a>. Data for this analysis was sourced from NASA, Imaflora and ESALQ. 
</div> 

 

<hr>

GFW analysis of NASA data shows that fire alerts in the Amazon decreased by 34% between August and September. Was brazil's temporary fire ban effective?

Brazil’s Fire Ban Correlates a Reduction in Amazon Wildfires. The Ban Lifts Today. | Fires | Global Forest Watch Blog

Climate

Learn about the role forests play in the global climate.

Climate | Global Forest Watch Blog

climate

climate | Global Forest Watch Blog

climate change

climate change | Global Forest Watch Blog

Data

Data | Global Forest Watch Blog

Emissions

Emissions | Global Forest Watch Blog

Mitigation

Mitigation | Global Forest Watch Blog

restoration

restoration | Global Forest Watch Blog

David Gibbs

Nancy Harris

Frances Seymour

<p>Protecting tropical forests is essential for achieving the climate goals of the Paris Agreement. <a href="http://bit.ly/2xPrHn3" target="_blank" rel="noopener noreferrer">Global Forest Watch Climate</a> recently released estimated carbon dioxide emissions associated with the <a href="http://bit.ly/2xQVN9H" target="_blank" rel="noopener noreferrer">2017 tropical tree cover loss data</a>, and the numbers demonstrate more of what we already knew. If tropical tree cover loss continues at the current rate, it will be nearly impossible to keep warming below the pledged two degrees Celsius.</p>
<aside id="textbox">
<div><em>Global Forest Watch’s annual tree cover loss data provides an unbiased proxy for which nationally-reported estimates of tropical deforestation can be assessed. GFW Climate builds on this global dataset to estimate associated carbon emissions from aboveground biomass clearing.</em></div>
</aside>
<p>Below are five numbers that demonstrate just how important tropical forests are in preventing further climate change, and how much more visibility they need on the global climate change mitigation agenda.</p>
<p>If tropical deforestation were a country, it would rank third in carbon dioxide-equivalent emissions, only behind the mega-emitters <a href="http://www.wri.org/blog/2017/04/interactive-chart-explains-worlds-top-10-emitters-and-how-theyve-changed" target="_blank" rel="noopener noreferrer">China and the United States of America</a>.</p>
<p>Annual gross carbon dioxide emissions from tree cover loss in tropical countries averaged 4.8 gigatons per year between 2015 and 2017. Put another way, tropical tree cover loss is now causing more emissions every year than 85 million cars would <a href="http://css.umich.edu/sites/default/files/Carbon_Footprint_Factsheet_CSS09-05_e2018_0.pdf" target="_blank" rel="noopener noreferrer">over their entire lifetime</a>.</p>
<div class="imagecenter*"><img decoding="async" class=" wp-image-3569 alignright" style="padding-bottom: 10px;" src="https://dev-global-forest-watch-blog.pantheonsite.io/wp-content/uploads/2018/10/18_TCL_Emission_By_Numbers-03-1.png" alt="emissions rankings" width="100%" /></div>
<p>It’s getting worse. Average annual emissions in the last three years were 63 percent higher than in the preceding 14 years.</p>
<div class="imagecenter"><img decoding="async" class=" wp-image-3569 alignright" style="padding-bottom: 10px;" src="https://dev-global-forest-watch-blog.pantheonsite.io/wp-content/uploads/2018/10/18_TCL_Emission_By_Numbers-02.png " alt="emissions" width="100%" /></div>
<p>Across all Nationally Determined Contributions (NDCs) under the Paris Agreement, forests represent a <a href="https://www.nature.com/articles/nclimate3227" target="_blank" rel="noopener noreferrer">quarter</a> of all planned emission reductions by 2030. Although the rate of tree cover loss <a href="http://bit.ly/2xQVN9H" target="_blank" rel="noopener noreferrer">declined in some places</a>, the most recent data for 2017 suggests that many countries are moving in the wrong direction to fulfill these goals. Average emissions from 2015 – when the Paris Agreement was signed – through 2017 were 63 percent higher than the average over the 14 years prior (3.0 Gigatons per year versus 4.9 Gigatons per year).</p>
<p>It’s not just carbon in the trees that make forests important for climate. Forests hold critical local cooling power – to the measure of more than 2 air conditioning units per tree.</p>
<p>Beyond their role in the global carbon budget, forests also have an important role in regulating climate at a local level by shading the ground and transpiring water. Every 100 liters of water a tree transpires (just a fraction of what many trees release each day) provides the equivalent of running 2 central air conditioning units for a day (the <a href="https://www.sciencedirect.com/science/article/pii/S0959378017300134" target="_blank" rel="noopener noreferrer">cooling equivalent of 70 kilowatt-hours</a> of electricity).</p>
<p>On the other hand, deforestation can <a href="http://science.sciencemag.org/content/351/6273/600" target="_blank" rel="noopener noreferrer">increase local air temperature</a> in the tropics and temperate zones by 1 degree Celsius and increase daily temperature variation by almost 2 degrees Celsius in the tropics and 2.85 degrees Celsius in the temperate zone.</p>
<p>Tropical forests are <a href="http://www.wri.org/publication/ending-tropical-deforestation-tropical-forests-and-climate-change-latest-science" target="_blank" rel="noopener noreferrer">8 percent of the problem but 23 percent of the solution</a>.</p>
<div class="imagecenter"><img decoding="async" class=" wp-image-3569 alignright" style="padding-bottom: 10px;" src="https://dev-global-forest-watch-blog.pantheonsite.io/wp-content/uploads/2018/10/18_TCL_Emission_By_Numbers-01-1.png" alt="solution vs the problem" width="100%" /></div>
<p>About 8 percent of global emissions currently come from tree cover loss in tropical forests, but these same forests are capable of providing 23 percent of the cost-effective climate mitigation needed before 2030. NDCs still fall far short of the total mitigation needed to keep 2030 emissions in line with a two degrees Celsius scenario, and about 7.1 gigatons of carbon dioxide can be mitigated annually through the management, protection, and restoration of tropical forests, mangroves and peatlands. That’s equivalent to the total carbon dioxide-equivalent emissions of <a href="https://www.climatewatchdata.org/countries/compare?locations=EU28%2CRUS%2CJPN" target="_blank" rel="noopener noreferrer">Russia, the European Union and Japan combined in 2014</a>. This potential comes from the avoided emissions through stopping deforestation and degradation as well as the removal of atmospheric carbon that takes places through forest growth and restoration.</p>
<p>Forests receive just 3 percent of available climate mitigation finance.</p>
<p>Despite this potential, forest-related finance, even for countries with high rates of deforestation, accounts for <a href="https://climatepolicyinitiative.org/wp-content/uploads/2017/10/2017-Global-Landscape-of-Climate-Finance.pdf" target="_blank" rel="noopener noreferrer">less than 3 percent of global</a> climate mitigation-related development funding. To reach global climate goals it’s critical that national and local actors alike double down on the proven strategy of reducing deforestation to mitigate climate change.</p>
<hr />
<div class="attribute">
<p>BANNER PHOTO: Restored forests in Costa Rica. By <a href="https://www.flickr.com/photos/worldresourcesinstitute/34830483474/in/album-72157685675061986/" target="_blank" rel="noopener noreferrer">Luciana Aaron Minnick/WRI</a>.</p>
<p>Originally appeared on <a href="https://www.wri.org/blog/2018/10/numbers-value-tropical-forests-climate-change-equation" target="_blank" rel="noopener noreferrer">WRI Insights</a>.</p>
</div>
<hr />
<style type="text/css"><span data-mce-type="bookmark" style="display: inline-block; width: 0px; overflow: hidden; line-height: 0;" class="mce_SELRES_start">﻿</span><span data-mce-type="bookmark" style="display: inline-block; width: 0px; overflow: hidden; line-height: 0;" class="mce_SELRES_start">﻿</span></style>

By the Numbers: The Value of Tropical Forests in the Climate Change Equation

Protecting tropical forests is essential for achieving the climate goals of the Paris Agreement. <a href="http://bit.ly/2xPrHn3" target="_blank" rel="noopener noreferrer">Global Forest Watch Climate</a> recently released estimated carbon dioxide emissions associated with the <a href="http://bit.ly/2xQVN9H" target="_blank" rel="noopener noreferrer">2017 tropical tree cover loss data</a>, and the numbers demonstrate more of what we already knew. If tropical tree cover loss continues at the current rate, it will be nearly impossible to keep warming below the pledged two degrees Celsius.

<aside id="textbox">
<div><em>Global Forest Watch’s annual tree cover loss data provides an unbiased proxy for which nationally-reported estimates of tropical deforestation can be assessed. GFW Climate builds on this global dataset to estimate associated carbon emissions from aboveground biomass clearing.</em></div>
</aside>Below are five numbers that demonstrate just how important tropical forests are in preventing further climate change, and how much more visibility they need on the global climate change mitigation agenda.

If tropical deforestation were a country, it would rank third in carbon dioxide-equivalent emissions, only behind the mega-emitters <a href="http://www.wri.org/blog/2017/04/interactive-chart-explains-worlds-top-10-emitters-and-how-theyve-changed" target="_blank" rel="noopener noreferrer">China and the United States of America</a>.

Annual gross carbon dioxide emissions from tree cover loss in tropical countries averaged 4.8 gigatons per year between 2015 and 2017. Put another way, tropical tree cover loss is now causing more emissions every year than 85 million cars would <a href="http://css.umich.edu/sites/default/files/Carbon_Footprint_Factsheet_CSS09-05_e2018_0.pdf" target="_blank" rel="noopener noreferrer">over their entire lifetime</a>.
<div class="imagecenter*"><img class=" wp-image-3569 alignright" style="padding-bottom: 10px;" src="https://dev-global-forest-watch-blog.pantheonsite.io/wp-content/uploads/2018/10/18_TCL_Emission_By_Numbers-03-1.png" alt="emissions rankings" width="100%" /></div>
It’s getting worse. Average annual emissions in the last three years were 63 percent higher than in the preceding 14 years.
<div class="imagecenter"><img class=" wp-image-3569 alignright" style="padding-bottom: 10px;" src="https://dev-global-forest-watch-blog.pantheonsite.io/wp-content/uploads/2018/10/18_TCL_Emission_By_Numbers-02.png " alt="emissions" width="100%" /></div>
Across all Nationally Determined Contributions (NDCs) under the Paris Agreement, forests represent a <a href="https://www.nature.com/articles/nclimate3227" target="_blank" rel="noopener noreferrer">quarter</a> of all planned emission reductions by 2030. Although the rate of tree cover loss <a href="http://bit.ly/2xQVN9H" target="_blank" rel="noopener noreferrer">declined in some places</a>, the most recent data for 2017 suggests that many countries are moving in the wrong direction to fulfill these goals. Average emissions from 2015 – when the Paris Agreement was signed - through 2017 were 63 percent higher than the average over the 14 years prior (3.0 Gigatons per year versus 4.9 Gigatons per year).

It’s not just carbon in the trees that make forests important for climate. Forests hold critical local cooling power – to the measure of more than 2 air conditioning units per tree.

Beyond their role in the global carbon budget, forests also have an important role in regulating climate at a local level by shading the ground and transpiring water. Every 100 liters of water a tree transpires (just a fraction of what many trees release each day) provides the equivalent of running 2 central air conditioning units for a day (the <a href="https://www.sciencedirect.com/science/article/pii/S0959378017300134" target="_blank" rel="noopener noreferrer">cooling equivalent of 70 kilowatt-hours</a> of electricity).

On the other hand, deforestation can <a href="http://science.sciencemag.org/content/351/6273/600" target="_blank" rel="noopener noreferrer">increase local air temperature</a> in the tropics and temperate zones by 1 degree Celsius and increase daily temperature variation by almost 2 degrees Celsius in the tropics and 2.85 degrees Celsius in the temperate zone.

Tropical forests are <a href="http://www.wri.org/publication/ending-tropical-deforestation-tropical-forests-and-climate-change-latest-science" target="_blank" rel="noopener noreferrer">8 percent of the problem but 23 percent of the solution</a>.
<div class="imagecenter"><img class=" wp-image-3569 alignright" style="padding-bottom: 10px;" src="https://dev-global-forest-watch-blog.pantheonsite.io/wp-content/uploads/2018/10/18_TCL_Emission_By_Numbers-01-1.png" alt="solution vs the problem" width="100%" /></div>
About 8 percent of global emissions currently come from tree cover loss in tropical forests, but these same forests are capable of providing 23 percent of the cost-effective climate mitigation needed before 2030. NDCs still fall far short of the total mitigation needed to keep 2030 emissions in line with a two degrees Celsius scenario, and about 7.1 gigatons of carbon dioxide can be mitigated annually through the management, protection, and restoration of tropical forests, mangroves and peatlands. That’s equivalent to the total carbon dioxide-equivalent emissions of <a href="https://www.climatewatchdata.org/countries/compare?locations=EU28%2CRUS%2CJPN" target="_blank" rel="noopener noreferrer">Russia, the European Union and Japan combined in 2014</a>. This potential comes from the avoided emissions through stopping deforestation and degradation as well as the removal of atmospheric carbon that takes places through forest growth and restoration.

Forests receive just 3 percent of available climate mitigation finance.

Despite this potential, forest-related finance, even for countries with high rates of deforestation, accounts for <a href="https://climatepolicyinitiative.org/wp-content/uploads/2017/10/2017-Global-Landscape-of-Climate-Finance.pdf" target="_blank" rel="noopener noreferrer">less than 3 percent of global</a> climate mitigation-related development funding. To reach global climate goals it’s critical that national and local actors alike double down on the proven strategy of reducing deforestation to mitigate climate change.

<hr />

<div class="attribute">

BANNER PHOTO: Restored forests in Costa Rica. By <a href="https://www.flickr.com/photos/worldresourcesinstitute/34830483474/in/album-72157685675061986/" target="_blank" rel="noopener noreferrer">Luciana Aaron Minnick/WRI</a>.

Originally appeared on <a href="https://www.wri.org/blog/2018/10/numbers-value-tropical-forests-climate-change-equation" target="_blank" rel="noopener noreferrer">WRI Insights</a>.

</div>

<hr />

<style type="text/css"><span data-mce-type="bookmark" style="display: inline-block; width: 0px; overflow: hidden; line-height: 0;" class="mce_SELRES_start">﻿</span><span data-mce-type="bookmark" style="display: inline-block; width: 0px; overflow: hidden; line-height: 0;" class="mce_SELRES_start">﻿</span></style>