Identification Information Citation Originator: USDA Forest Service, Geospatial Technology and Applications Center, BAER Imagery Support Program Publication Date: 2017-09-26 Title: Burned Area Reflectance Classification (BARC) Data Bundle for the MR SEPARATION Fire occurring on the Willamette National Forest - 2017 Publication Information Publication Place: Salt Lake City, Utah, USA Publisher: USDA Forest Service Geospatial Data Presentation Form: raster digital data Online Linkage: http://activefiremaps.fs.fed.us/baer/download.php?year=2017 Description Abstract: These data products are preliminary burn severity assessments derived from Landsat 8 OLI data. The pre-fire and post-fire subsets included were used to create a differenced Normalized Burn Ratio (dNBR) image. The dNBR image attempts to portray the variation of burn severity within a fire. The severity ratings are influenced by the effects to the canopy. The severity rating is based upon a composite of the severity to the understory (grass, shrub layers), midstory trees and overstory trees. Because there is often a strong correlation between canopy consumption and soil effects, this algorithm works in many cases for BAER teams whose objective is a soil burn severity assessment. It is not, however, appropriate in all ecosystems or fires. It is expected that BAER teams will adjust the thresholds to match field observations to produce a soil burn severity. Purpose: These data were created by the USDA Forest Service Geospatial Technology and Applications Center (GTAC) to support Burned Area Emergency Response (BAER) teams. Supplemental Information Fire Name: MR SEPARATION Agency: USDA Forest Service Land Management Unit: Willamette National Forest Type of assessment: Emergency Fire size (as mapped): 17977 acres Pre-Fire Imagery Date and Path/Row: 2016-08-21; 045/029 Post-Fire Imagery Date and Path/Row: 2017-09-09; 045/029 Output Dataset Projection: NAD_1983_UTM_Zone_10N Datum: D_North_American_1983 Spheroid Name: GRS_1980 Mapping Comments: None Product List or4410112190920170810_20160821_l8_refl_utm.tif = Pre-Fire Landsat 8 OLI reflectance at sensor data subset, scaled by 400 and converted to integer (band order (1-8): B, G, R, NIR, SWIR-1, SWIR-2, Coastal Blue, Cirrus). For best viewing set RGB to bands 6, 4, 3. or4410112190920170810_20170909_l8_refl_utm.tif = Post-Fire Landsat 8 OLI reflectance at sensor data subset, scaled by 400 and converted to integer (band order (1-8): B, G, R, NIR, SWIR-1, SWIR-2, Coastal Blue, Cirrus). For best viewing set RGB to bands 6, 4, 3. or4410112190920170810_20160821_20170909_dNBR.tif = Continuous dNBR product scaled by 1000; theoretical range of values is -2000 to 2000 or4410112190920170810_20160821_20170909_dNBR_barc256_utm.tif = BARC256, 256-class (0-255); continuous image representing preliminary estimate of burn severity. This dataset can be adjusted by the user, if needed, to refine the thematic BARC4 product and/or define a new BARC4. It is anticipated that users will adjust the breakpoints between classes, then recode to their desired 3 or 4 classes. or4410112190920170810_20160821_20170909_dNBR_barc4_utm.tif = BARC4, Four category preliminary estimate severity classification. Subset values: 1 = unburned / undetectable (Dark Green) 2 = low severity (Cyan) 3 = moderate severity (Yellow) 4 = high severity (Red) Time Period of Content Multiple Dates/Times Single Date/Time Calendar Date: 2016-08-21 (pre-fire image) Single Date/Time Calendar Date: 2017-09-09 (post-fire image) Currentness Reference: Severity data correspond to the date of the post-fire image. Status Progress: Evaluation of methods in process Maintenance and Update Frequency: As needed Spatial Domain Bounding Coordinates West Bounding Coordinate: 578027.845371 East Bounding Coordinate: 599597.845371 North Bounding Coordinate: 4900791.90388 South Bounding Coordinate: 4877121.90388 Keywords Theme Theme Keyword Thesaurus: none Theme Keyword: Wildland Fire Theme Keyword: Wildfire Theme Keyword: Burned Area Emergency Response Theme Keyword: BAER Theme Keyword: differenced Normalized Burn Ratio Theme Keyword: dNBR Theme Keyword: Fire Severity Theme Keyword: Burn Severity Theme Keyword: USDA Forest Service Theme Keyword: Landsat 8 OLI Theme Keyword: Landsat 8 OLI Place Place Keyword Thesaurus: none Place Keyword: Willamette National Forest Place Keyword: MR SEPARATION Fire Access Constraints: Thresholded, preliminary severity estimates (BARC4 and BARC256) are only delivered to BAER teams. Further distribution is at the discretion of the BAER team leader. Pre-imagery, post-imagery, and unthresholded data are publicly available on http://activefiremaps.fs.fed.us/baer/download.php. Access to pre-imagery and post-imagery may be limited when sensors other than Landsat or other government remote sensing assets are used due to product licensing restrictions. Use Constraints: There are no restrictions on use, except for reasonable and proper acknowledgement of information sources and limitations as preliminary/draft data. Data Set Credit: USDA Forest Service Native Data Set Environment: ERDAS Imagine, ESRI ArcGIS Data Quality Information Positional Accuracy Horizontal Positional Accuracy Horizontal Positional Accuracy Report: Landsat data are terrain corrected using a USGS digital elevation model with less than 1/2 pixel RMS error. Accuracy may vary for other sensors. Lineage Process Step Process Description: These data products are derived from Landsat 8 OLI data. Pre-fire and post-fire scenes are analyzed to create a differenced Normalized Burn Ratio (dNBR) image. The dNBR image portrays the variation of burn severity within the fire. The pre- and post-fire images are terrain corrected and further processed to convert top of atmosphere reflectance. The Normalized Burn Ratio (NBR) is computed for the pre- and post-fire images using the following formula: (NIR Band - SWIR Band) / (NIR Band + SWIR Band) = NBR The Differenced NBR is computed to determine severity by subtracting the post-fire NBR from the pre-fire NBR: (PreNBR - PostNBR) = dNBR The BARC products are a generalization of the raw, continuous dNBR dataset. Both BARC products have been resampled to unsigned 8-bit GeoTIFF and are easily viewed and edited within ArcGIS. The classes represented on the BARC products are created with thresholds chosen by an analyst at GTAC. These thresholds can be roughly related back to original dNBR values by multiplying by 5 and then subtracting 275 (for example, a BARC256 value of 100 would relate to a dNBR value of 225). The BARC thresholds used on this particular fire are: Unburned / Undetectable: All values less than or equal to 64. Low: All values greater than 64 and less than or equal to 97. Moderate: All values greater than 97 and less than or equal to 193. High: All values greater than 193. General descriptions of the severity classes are below: Unburned / Undetectable: This means the area after the fire was indistinguishable from pre-fire conditions. This does not always indicate the area did not burn (i.e. canopy may be occluding the burn signal). Low: This severity class represents areas of surface fire with little detected change in cover and little detected mortality of the dominant vegetation. Moderate: This severity class is between low and high and means there is a mixture of detected effects on the dominant vegetation. High: This severity class represents areas where the canopy has high to complete consumption. Spatial Data Organization Information Direct Spatial Reference Method: Raster Raster Object Information Vertical Count: 1 Spatial Reference Information Horizontal Coordinate System Definition Planar Grid Coordinate System Grid Coordinate System Name: NAD_1983_UTM_Zone_10N Universal Transverse Mercator UTM Zone Number: 10N Transverse Mercator Scale Factor at Central Meridian: 0.9996 Longitude of Central Meridian: -123.0 Latitude of Projection Origin: 0.0 False Easting: 500000.0 False Northing: 0.0 Planar Coordinate Information Planar Coordinate Encoding Method: row and column Coordinate Representation Abscissa Resolution: 30.0 Ordinate Resolution: 30.0 Planar Distance Units: Meter Geodetic Model Horizontal Datum Name: D_North_American_1983 Ellipsoid Name: GRS_1980 Semi-major Axis: 6378137.0 Denominator of Flattening Ratio: 298.257222101 Distribution Information Resource Description: Downloadable Data Metadata Reference Information Metadata Date: 2017-09-26 Metadata Contact Contact Information Contact Organization Primary Contact Organization: USDA Forest Service, Geospatial Technology and Applications Center (GTAC) Contact Person: Justin Epting Contact Address Address Type: mailing address Address: 2222 W. 2300 S. City: Salt Lake City State or Province: UT Postal Code: 84119 Contact Voice Telephone: 801 975 3755 Contact Electronic Mail Address: baerimagery@fs.fed.us ******************************************************************************** The following section should be completed after the BARC has been field validated by the BAER team and this entire metadata file should be returned to GTAC with the Soil Burn Severity data. BAER Assessment Completion Date: 10/10/2017 BAER Team GIS Analyst: Jeremy Hobson BAER Team Leader: Frederick Levitan Soil Burn Severity Analyst(s): Sarah Brame Original BARC256 thresholds: Unburned / Undetectable: <=76 Low: >76 and <=110 Moderate: >110 and <=198 High: >198 Soil Burn Severity thresholds: Unburned / Undetectable: <= 50 Low: > 50 and < = 98 Moderate: >98 and < = 200 High: > 200 Sequence of steps used to create Soil Burn Severity data: 1. Some areas within the Sepration fire were masked due to cloud cover, etc or were missing soil burn severity calculations. In order to fill in these no data gaps, the initial BARC map was converted to a polygon feature class. Polygons were built for the masked areas and assigned to the designated soil burn severity class. This includes some spot fires outside the primary fire perimeter which did not receive initial soil burn serverity calculations. 2. Vector data containing data updates was converted back to a raster dataset, classified within the four soil burn severity classes. 3. The existing BARC data was reclassified using soil burn severity thresholds noted above. 4. The converted raster dataset representing edited voids was mosaicked with the reclassified BARC data to produce the final SBS dataset for the entire fire area. Additional Comments: One spot fire to the north was missing BARC data; the polygon feature class was edited to reflect observed soil burn severity, as were some additional spots surrounding the fire. Note, for anlaysis, the final SBS grid was converted back to a polygon feature class. One additional edit occurred to this dataset. For the final SBS raster dataset, this edited version polygon feature class was converted to a raster dataset. Hence the edited area was rasterized and won't match the final acreage calculated per the polygon feature class. The Separation BARC map overlapped with the Nash BARC map. For anlaysis purposes, once the data was converted to a polygon feature class, the Separation fire perimeter was used to erase data extending within and across the Nash fire. The Nash fire perimeter was similarly used to erase data from the Nash grid that was extendings within and across the shared boundary. Consequently, each respectives fire data was cut to the shared boundary, for analysis purposes. The final SBS grid was exported with a similar footprint as the original BARC map.