Two Decades of Fire-Driven Loss in Unprecedented Detail

Aug 17, 2022||8 minutes
Languages
Tree cover loss due to fires visualized on the GFW map, 2001-2021

Tree cover loss due to fires visualized on the GFW map, 2001-2021

Languages
Category
  • Data
Topics
  • fires

New data from the University of Maryland (UMD), now available on Global Forest Watch (GFW), provides a more detailed view of how fires are driving tree cover loss across the globe. The new tree cover loss from fires data offers a higher-resolution look than ever before at trends in fire-driven loss over the last two decades. Read on to learn what the new data measures, how we can use it and how it differs from existing fire-related data on GFW.  

New Insights Into Trends of Fire-Driven Tree Cover Loss

The new tree cover loss from fires data builds on the existing annual tree cover loss data from UMD by providing additional context on where fires are driving loss. Each 30-m pixel in the annual tree cover loss data is analyzed using Landsat satellite images to determine whether the loss was likely due to stand-replacing fires (which kill all or most living trees). This data allows fire-driven loss to be separated from all other drivers of tree cover loss, like agriculture and forestry. Using this data, detailed patterns of tree cover loss due to fires can be assessed over a 21-year period using a globally consistent definition of fire-driven forest loss for the first time. 

For this data, loss due to fires is defined as natural or human-ignited fires that result in direct loss of tree canopy cover. This could include wildfires, and intentionally set fires (including escaped fires started by humans for purposes related to agriculture, hunting, recreation or arson). Instances where trees are felled and later burned are not included since the initial driver of loss in that case is mechanical removal. Low-intensity and understory forest fires that do not result in substantial tree canopy loss at the scale of a 30-m pixel are also excluded from this definition. 

Going forward, this data will be updated on an annual basis and will allow us to better understand global, regional and local trends in tree cover loss as they evolve over time.  

Tree cover loss due to fires, 2001-2021

Tree cover loss due to fires 2001-2021

How Can We Use the New Tree Cover Loss from Fires Data?

Identifying Areas with the Most Tree Cover Loss Due to Fires

Users can use this data to visualize and analyze the spatial distribution of loss due to fires over time on Global Forest Watch. For example, the global dashboard can help users identify countries with the most tree cover loss due to fires in the previous year.  

Looking at the list, Russia, Canada and the United States clearly had the most tree cover loss due to fires in 2021. After analyzing the downloaded data, we can see that these three countries combined lost 7.8 million hectares or 84% of all fire-related loss in 2021. 

We can then view how fire activity has changed over time for these areas. For example, in 2021, Russia experienced the largest area of tree cover loss the country has seen in 21 years, 82% of which was due to fires. 

Users can also use the GFW map to visualize when and where these fires occurred and select specific years to dig deeper into the data. For instance, looking at the map, it is clear that most of Russia’s fires in 2021 occurred in the central and eastern part of the country. 

Separating Trends in Fire-Related Tree Cover Loss from Other Drivers

In some cases, this data set can also provide insights into trends of forest loss that is not due to fires. Since loss from fires tends to vary significantly year to year depending on climate and other conditions, it can obscure the longer-term trend in tree cover loss from other drivers. 

For example, the annual tree cover loss data on primary forest loss in Brazil shows a large spike in 2016 and 2017 that is not captured by the official deforestation monitoring system for the Brazilian Amazon, PRODES. Using this new data, we can see that much of that spike was due to tree cover loss from fires, which are typically not captured by the PRODES system. Understanding where fire drives forest loss and where it does not helps provide more nuance when interpreting the data and allows more direct comparisons with other data sources.  

What are the Limitations for Using this Data?

In a previous blog, we describe changes to the tree cover loss data over the past 21 years and recommend caution for comparing trends before and after 2015. This new data on fires builds on that tree cover loss data set, so the same cautions apply. While the changes to the tree cover loss data have resulted in improved detection of forest loss in recent years, including better detection of fires in the boreal forests and other regions, it has also introduced inconsistencies over time.

Despite these inconsistencies, the new data features high levels of spatial and temporal accuracy at the global level. Globally, 90% of areas classified as tree cover loss due to fires were classified correctly; the other 10% were false positives. 

Spatial accuracy varied by region but remained relatively high for most regions. Europe and North America had the highest levels of accuracy with false positive rates of 7% and 4%, respectively. Africa had the lowest accuracy and the highest number of false positives, with 39% of the area labeled as tree cover loss due to fires incorrectly classified. The high false positive rate in Africa is likely because forest fires there are relatively rare. 

RegionFalse Positive Rate
North America4%
Europe7%
Latin America26%
Southeast Asia27%
Africa39%
Global10%

There are also cases where there are delays in detecting tree cover loss due to fires, where the loss occurs close to the end of the year but there aren’t enough clear satellite images to detect it until the following year. This is important to note because sometimes spikes we see in loss due to fires can reflect trends from the previous year. Overall, 96% of all fire-related loss events occurred within one year of the mapped loss date. 

What Other Fire Data is on GFW, and What’s Different About the New Data?

The fire data available on GFW falls into two categories: fire activity data and contextual data. Fire activity data allows users to view ongoing fires and analyze the extent of fires in areas of interest, such as a country or user-defined polygon. Contextual data can be used to supplement the fire activity data and help users understand what areas are at risk of fire and the impacts of fire on air quality.  

Comparison of GFW Fire Data

Fire Activity

 

Data Set

Resolution

Data Source

Update Frequency

Date Range

Best Use

Tree Cover Loss from Fires

30m

UMD/GLAD

Annual

2001-2021

Calculating extent of stand-replacing forest fires

MODIS Burned Areas*

500m

NASA

Monthly (Two-month delay)

2000-Present

Calculating burned area extent for all land covers

VIIRS Fire Alerts

375m

NASA

Twice daily

2012-Present

Detecting fire locations in near-real-time (day/night)

MODIS Fire Alerts*

500m

NASA

Twice daily

2000-Present

Detecting fire locations in near-real-time (day)

Contextual

Global Fire Weather Index

10km

NASA

GFWED

Daily

Present

Current fire risk

Global Air Quality Index

n/a

World Air Quality Index

US EPA

Every 15 minutes

Present

Fire-related effects on air pollution

 

 

*The Terra satellite that carries the MODIS sensor will be decommissioned next year, along with associated products including the MODIS near-real-time alerts and burned areas data set. VIIRS will continue to be the main source of near-real-time fire alerts.

Fire Activity Data

Tree Cover Loss from Fires Data

The tree cover loss from fires data differs from other fire activity data on the GFW platform because it focuses solely on stand-replacing forest fires, while other fire data sets do not make a distinction between stand-replacing fires and other fires. This is important because forests act as carbon sinks and contain most of the land-based carbon on the planet, which is released when they burn. 

The tree cover loss from fire data also tracks fires at a much finer resolution than the other fire activity data; however, it is also updated less frequently, making it less useful for monitoring ongoing or very recent fire activity. 

Other Fire Activity Data

  • Near-real-time alerts (MODIS and VIIRS): MODIS* and VIIRS active fire alerts provide users with near-real-time data that is updated daily and can be used to monitor and respond to ongoing fires. The alerts can detect varying fire sizes depending on the fire temperature, time of day and presence of cloud cover or heavy haze.  
  • Fire extent over time (MODIS burned areas): Although updated less frequently than near-real-time alerts (monthly instead of daily), the MODIS burned area data* is better suited for assessing trends in fire extent over time than fire alerts because it identifies the area of land that has burned.  

Users can analyze fire activity data on the GFW map and dashboard, and filter the data to only view fires that occur in primary forests or other areas of conservation importance. Signing up for a myGFW account also gives users the option to create and analyze custom areas of interest. Users can also stay up to date on fire activity in their custom areas by subscribing to notifications through their myGFW account.

Contextual Data

  • Fire risk (Global Fire Weather Index): The Global Fire Weather Index can be used to identify areas where climate conditions make landscapes more vulnerable to fires, allowing users to understand the current risk of burning. 
  • Health impacts (Global Air Quality Index): The Global Air Quality Index measures current air quality of particulate matter which is strongly impacted by smoke and haze from fires. With over 15,000 stations worldwide, this helps users measure the risk of health effects related to air pollution.  

The new tree cover loss from fires data is a valuable addition to the existing suite of fire-related data on GFW and will enable all users to better track the impact of fires on forests worldwide to develop more effective management responses. 


To explore fire data on GFW, visit the map and dashboard.

View a webinar in English, Spanish, Portuguese and Bahasa Indonesia to learn more about tree cover loss from fires and other fire-related data on Global Forest Watch.

Category
  • Data
Topics
  • fires

Explore More Articles

Ripe Cocoa pods from a cocoa farm in Ghana.
Feb 14, 2024|Data|10 minutes

Ending Deforestation from Cocoa in West Africa with New Data-Driven Resources

Two new data-driven resources provide a shared view of priority areas in West Africa and can help realize a a deforestation-free cocoa sector.

Aerial view of the Amazon Rainforest
Jan 18, 2024|Data|8 minutes

Comparing Forest Extent in 2020 from Global Forest Watch and the Forest Resources Assessment

This blog compares the forest extent in 2020 for data from UMD on GFW and the FAO Forest Resources Assessment and explains the differences.

Side by side comparison of UMD tree cover loss and JRC Tropical Moist Forest data
Jan 08, 2024|Data|10 minutes

Differences Between Global Forest Watch’s Tree Cover Loss Data and JRC’s Tropical Moist Forest Data Explained

We explain key differences between two data sets that track forest change in the tropics: UMD tree cover loss and JRC Tropical Moist Forest.

Explore More Articles
Ripe Cocoa pods from a cocoa farm in Ghana.
Feb 14, 2024|Data|10 minutes

Ending Deforestation from Cocoa in West Africa with New Data-Driven Resources

Aerial view of the Amazon Rainforest
Jan 18, 2024|Data|8 minutes

Comparing Forest Extent in 2020 from Global Forest Watch and the Forest Resources Assessment

Side by side comparison of UMD tree cover loss and JRC Tropical Moist Forest data
Jan 08, 2024|Data|10 minutes

Differences Between Global Forest Watch’s Tree Cover Loss Data and JRC’s Tropical Moist Forest Data Explained

fetching comments...