FAQs

The Forest Service RAVG program includes fires from 2007 and later that enclose at least 1000 acres of forested National Forest System (NFS) lands. For NFS Regions 8 and 9 (the eastern U.S.), the threshold has been 500 acres since 2016. Data for most fires that qualify should be available on this website within 45 days after fire containment, although some will be postponed until the following summer ("extended assessments") and some with very mild burn effects may not be included at all. To see the mapping status of various qualified fires, see the mapping status" page on this website. A larger set of RAVG data, including data for select fires on Department of the Interior (DOI) lands are available from the Burn Severity Portal's Interactive Viewer (https://burnseverity.cr.usgs.gov/viewer/?product=RAVG).

Use the data query tool on the Data Access page to search for RAVG data by fire name. If multiple fires merged into a single footprint, they may be published as a single dataset under the name of one of the original fires. Likewise, if the fire was part of a named complex, it may be published under the name of the complex. To locate these fires, use the "Advanced Query" to find the set of fires that would include the fire of interest (e.g., all fires from a given Forest Service region and year) and scan the "Related Fires" column for the desired fire.

The advanced query on the Data Access page allows the user to filter based on year, state or NFS region, and fire size. All fires that meet the filter parameters will be displayed in the resulting table. The user can view or download data for individual fires, or view summary graphics within the query results. Another option is to use the Burn Severity Portal's interactive viewer (https://burnseverity.cr.usgs.gov/viewer/?product=RAVG), which has additional query options and allows the user to combined data bundles for multiple fires into a single download.

For links to annual national mosaics of raster products and combined perimeter layers, see the National Datasets section of the Data Access page.

RAVG data are typically published within 45 days after fire containment. In cases where there is suitable imagery available sooner, the data may be published sooner. In cases of long-duration fires where the fire is not officially contained but fire activity is minimal, the data may be generated and published before formal containment. Finally, data for fires that burn in the fall or early spring may not be published until the following summer. To check on the status of fires slated for mapping, see the RAVG mapping status page (https://burnseverity.cr.usgs.gov/ravg/mapping-status).

There are several possible reasons for a fire being absent from this website:

1. The fire doesn't meet program criteria. Generally, only wildland fires that meet the standard RAVG criteria have data published on the RAVG website. The primary criterion is that the fire includes at least 1000 acres of forested National Forest System land (500 acres for Regions 8 and 9). Data for 2007 and later are included.

2. Suitable imagery is not yet available. Sufficiently clear pre- and post-fire imagery is required for each RAVG assessment. In order to improve the assessment quality, the analyst may choose to wait for near-ideal post-fire imagery rather than using an earlier but less-than-ideal scene. Most fires are assessed using imagery acquired no later than 45 days after fire containment; however, for fires that burn into the Fall or that occur during winter or early Spring, the presence of snow and/or excessive terrain shadows may dictate waiting until late Spring or Summer for imagery. The Mapping Status page shows the status of fires that are expected to meet RAVG criteria but that have not yet been assessed.

3. The fire is included with another assessment. If two or more large fires merge and are assessed as a single burned area, they will appear in the database under the name of only one of the original fires. Similarly, fires in a named complex may appear under the name of the complex or under individual fire names. To locate these fires, use the "Advanced Query" to find the set of fires that would include the fire of interest (e.g., all fires from a given Forest Service region and year) and scan the "Related Fires" column for the desired fire.

4. The fire pre-dates RAVG. The RAVG program was implemented on a national scale in 2007. Fires from years before 2007 are not included in the RAVG database.

5. The fire severity was extremely low. Occasionally the signature of a low-severity fire will be undetectable in satellite imagery. This is more likely for a surface fire under closed canopy cover. In this situation, the RAVG analyst may choose not to conduct the assessment.

For additional information about missing fires, contact the RAVG program (see the Contact Us page).

Sometimes multiple named fires will be mapped together and listed under the name of one fire (usually the largest) or a fire complex. Similarly, a fire complex may be listed under the name of one of its constituent fires. To find these fires, the following is recommended: 1. Use the Advanced Query on the Data Access page to find a collection of fires that would include the fire of interest. For example, query fires that occurred in Oregon in 2017. 2. Scan the "Related fires" column of the results table for the fire of interest. 3. If desired, download the data bundle and extract a spatial subset using a polygon that covers your individual fire.

Forest Service units can request non-standard assessments by contacting the RAVG program (see the Contact Us page). Requests will be considered on a case-by-case basis.

This website includes only fires that meet the Forest Service RAVG program criteria. RAVG data for some other fires (e.g., fires on Department of the Interior lands) can be found on the Burn Severity Portal (https://burnseverity.cr.usgs.gov/viewer/?product=RAVG). In addition, Forest Service units can request non-standard assessments by contacting the RAVG program (see the Contact Us page). Requests will be considered on a case-by-case basis.

 

For a list of standard RAVG data products, see the Products section on the Background, Products and Applications page. The metadata file associated with each RAVG assessment also includes product descriptions.

The three RAVG burn metrics (percent basal area (BA) loss, percent canopy cover (CC) mortality, and Composite Burn Index (CBI)) are determined from models that relate field-derived burn data to satellite-imagery-based indices (dNBR, RdNBR, and RBR).

The colors represent distinct ranges of the several burn metrics (e.g., 75-90% basal area loss). "Warm" colors (i.e., reds and yellows) indicate greater change and "cool" colors indicate lesser change. The table below shows the classes for each of the classified RAVG products, along with the associated value and color in both original ("classic") and current color schemes. The hexadecimal color code is also provided for reference.

As of the 2022 fire season, classified (thematic) RAVG products use a color scheme designed to meet the intent of Section 508 of the Rehabilitation Act. Specifically, it is intended to be more accessible to those with common color vision deficiencies ("color blindness"). As before, higher values are represented in "warm" colors (i.e., reds and yellows) and lower values in "cool" colors (i.e., blue and blue-green colors). Color definitions for each value are available in the attribute table for GeoTIFFs and in legends associated with other graphic products.

Note: A minor change to the color scheme was applied in 2024. In 2025, annual national mosaics will be republished with the current color scheme applied. Raster datasets for individual fires may be updated similarly in the future.

RAVG assessments are based on models that relate imagery-derived indices to field-based burn severity measures for forest vegetation. They are not calibrated for other vegetation types and other disturbances. Although RAVG data may indicate the relative severity of other disturbances or in other vegetation types, the burn severity values cannot be directly applied to such cases.

Because RAVG burn measures are pixel-based, raster format (e.g., GeoTIFF) is the most natural way to provide the data. Users who prefer a vector representation can convert the files using a GIS tool, such as the "raster to polygon" tool in ArcGIS Pro, assigning the raster "value" attribute to the polygons.

The original RAVG models are based on data collected from fires that occurred in the Sierra Nevada and northern California. They have traditionally been applied to fires across the U.S. In an effort to improve model accuracy, regional models were developed based on data from fires in other regions. Currently, regional models include the "southwest" model, based on data from Arizona and New Mexico, and the "northwest" model, based on data from Oregon, Washington, northern Idaho, western Montana, and northwestern Wyoming. 
 

When one of the regional models is applied to a RAVG dataset for a given fire, it will be noted in the associated metadata. As of 2024, the southwest model is used by default for fires in the southwest region and the northwest model is used for extended assessments in the northwest region (approximated by a bounding box with latitude between 43 and 49 degrees north and longitude between 109 and 118 degrees west). The national model is used elsewhere and for most older datasets. As older fires are reprocessed, the analyst may choose to apply the regionally appropriate model. 

The vegetation classes included in the RAVG summary tables are derived from a temporally appropriate version of the LANDFIRE Existing Vegetation Type (EVT). The hundreds of EVT classes are grouped into the following eight RAVG classes, based primarily on the EVT sub-class attribute: 

- Open-canopy evergreen 

- Closed-canopy evergreen 

- Open-canopy deciduous 

- Close-canopy deciduous 

- Open-canopy mixed evergreen and deciduous

- Closed-canopy mixed evergreen and deciduous 

- Pinyon-juniper 

- Non-tree (grass, shrub, and non-vegetation) For some purposes, the first seven classes are collapsed into a single "tree" class. 

For more information about EVT, see https://landfire.gov/evt.php. Note: RAVG vegetation classes are used to determine how many "forested" acres a fire includes and are provided for context; however, they do not play a direct role in any of the burn metric calculations.

The RAVG, MTBS and BAER Imagery Support programs are three distinct but related post-fire assessment activities managed by GTAC in partnership with DOI agencies. MTBS is conducted jointly with the DOI's USGS EROS Center. For BAER imagery support, GTAC responds to requests from Forest Service BAER teams and EROS responds to requests from DOI teams. The three programs have many similarities, such as the use of moderate resolution multispectral satellite imagery, but differ in their methods and protocols, as well as in their intended audiences. For more information about the MTBS and BAER imagery support programs, see https://www.mtbs.gov and https://burnseverity.cr.usgs.gov/baer, respectively.

For MTBS data, go to www.mtbs.gov. or https://burnseverity.cr.usgs.gov/.

For BAER-related data, go to https://burnseverity.cr.usgs.gov/baer/baer-imagery-support-data-download or https://burnseverity.cr.usgs.gov/.

The thematic version of the RAVG Composite Burn Index (CBI) is similar to the MTBS and BAER Imagery Support burn severity products, with each having four severity classes: unburned (unchanged or very low severity), low severity, moderate severity, and high severity. In some applications, the three products can serve the same purpose; however, because each is based on different protocols and often on different imagery, differences in the results can be expected. Note that the other two RAVG burn measures are estimates of the percent change in live basal area and canopy cover relative to the pre-fire conditions, and not properly termed "burn severity".

Overviews of remote sensing principles and application are available from many university and government websites, including NASA (see, for example, https://appliedsciences.nasa.gov/get-involved/training/english/arset-fundamentals-remote-sensing or https://appliedsciences.nasa.gov/sites/default/files/2022-11/Fundamentals_of_RS_Edited_SC.pdf).

Data masks are applied to RAVG products to identify "unmappable" areas; that is, to pixels where modeled RAVG metrics would clearly not be valid for any of several reasons. The analyst typically applies masks to areas with open water, snow, dense smoke or clouds, deep shadows caused by smoke or clouds, and active fire. Data gaps in Landsat-7 imagery caused by the failure of the scan-line corrector in 2003 are also masked. To preserve as much data as possible, the analyst may choose to not mask haze, thin clouds, or light smoke that is unlikely to significantly obscure the imagery bands used in RAVG models (near infrared and short-wave infrared). Similarly, light snow cover may be left unmasked. For practical reasons, small areas that would be masked otherwise (e.g., a 1-acre patch of active fire) or long, narrow features (e.g., a one-pixel-wide stretch of river) may be omitted from the mask. The numerical value of masked pixels varies by product but is always outside of the range of valid values. For example, masked pixels in the classified raster datasets have value 9, whereas valid class values are 1 to 4 for the 4-class CBI, 1 to 5 for the 5-class CC loss, and 1 to 7 for the 7-class BA loss raster.