Introduction
Beginning in FY2006, the US Geological Survey, National Center for Earth Resources Observation and Science (EROS) and the USDA Forest Service, Remote Sensing Applications Center (RSAC) will map burn severity for fires that burn within the United States. This project will map burn severity using the Differenced Normalized Burn Ratio (DNBR) approach at 30m spatial resolution for all fires in the United States meeting a minimum size requirement. The minimum size will be 500 acres or greater for all eastern fires (east of the -97o longitude) and 1,000 acres or greater for the rest of the US.
The DNBR approach will be used to map burn severity and analyze fire trends for fires going back uninterrupted to 1984. This will be accomplished by using archived and current Landsat imagery. Future fires will be mapped and analyzed using the same techniques to allow continuous monitoring of burn severity trends across the nation.
This work will be an extension of the existing cooperation between these two National Centers that currently provide rapid response burn severity mapping products to Forest Service and DOI Burn Area Emergency Response (BAER) Teams. Satellite imagery and products developed for BAER support will be leveraged for this larger assessment which covers all fires over a longer time period.
This brief paper outlines the objectives of the MTBS project, as well as objectives and needs of BAER projects. There are areas where these two efforts overlap in function and purpose, including image acquisition. However, there are also specific needs for each project that do not coincide with each other.
Purpose and Objectives of MTBS Project
The primary objective of this project is to provide for a national analysis of trends in burn severity. MTBS mapping efforts are meant to inform policy decisions driven by the National Fire Plan and Healthy Forests Restoration Act as well as provide a broad scale means to assess and monitor trends in fire affected landscapes. Because of severe droughts and greater than normal numbers of catastrophic fires in recent years (since 2000), it is essential for the trend analysis to account for potential climate variability and base the assessment on a longer period of time (going back to 1984) than just the last several years.
This project will serve three primary user groups with one set of data and information:
- National policies and policy makers such as the National Fire Plan and Wildland Fire Leadership Council (WFLC), which require information about long-term trends in burn severity and recent burn severity by vegetation types, fuel models, condition classes (data from LANDFIRE), and results and accomplishments (data from NFPORS)
- Field management units that benefit from GIS-ready maps and data for pre- and post-fire management decisions and monitoring
- Existing databases such as Fire Regime and Condition Class (FRCC) and LANDFIRE can integrate burn data produced at compatible spatial scale and resolution for validation and updating geospatial data sets.
The national burn severity-mapping project will be conducted in two time phases. The fires occurring in 2004-2010 will be mapped and reported annually. All fires between 1984 and 2003 will be mapped, analyzed and reported by designated mapping zones through the duration of the project. Burn severity project mapping zones were generated by aggregating MRLC map zones into significant geographical units encompassing broadly similar ecological and climatic patterns. Mapping and reporting of project zones has been prioritized and scheduled based on a number of factors including, fire frequency, density of fire occurrences, ecological significance, and importance to prevailing land management policies (see Appendix A). Fires will be mapped according to the following process and specifications:
- Identify fire locations. For older fires (1984-2003), fire locations are listed and available. For fires from 2004 - 2010, we will work with the National Interagency Fire Center (NIFC) staff and State contacts to obtain locations of fires to map.
- Analysts from EROS and RSAC will meet to select scenes based on the following criteria:
- Fires that are 500 acres or greater for the eastern states and 1,000 acres or greater for the rest of the US
- Geographic locations: whether multiple fires existing on one path/row or a single fire covering multiple scenes
- Seasonal effects: selecting the best timing in a season for pre- and post-fire Landsat scenes based on experiences and lessons learned
- Extent of cloud cover
- Selected Landsat scenes will be submitted for acquisition and processing to the highest standards for geometric and radiometric rectifications. The Normalized Burn Ratio (NBR) will be produced for each scene.
- Analysts will develop Differenced Normalized Burn Ratio (DNBR) images for each fire, and produce DNBR classes. Fire perimeters will be mapped using a semi-automatic approach.
- Mapping will be conducted in a nationally consistent Albers equal-area map projection. Files will be produced in GIS-ready format.
- Analysts will intersect DNBR classes with available vegetation cover types cross-walked to the nationally consistent Ecological Systems classification, and produce tables of acres burned by DNBR classes and vegetation cover types.
- Analysts will overlay DNBR classes with NFPORS fuel treatment project centroids (already available) or polygons (when available) to produce acres burned by severity classes and status of hazardous fuel reduction efforts.
Purpose and Objectives of BAER Projects
BAER teams are dispatched to fires to make an emergency assessment of the immediate post-fire condition of a landscape. Even though a fire may not directly affect adjacent natural and urban resources, the post-fire effects caused by weather events can have significant impact. BAER teams are charged with mapping areas within the burn scar that were most severely damaged by the fire (in terms of soil burn severity), as well as mapping a multitude of elements that are affected by the fire. These elements could include the habitat of threatened and endangered species, transportation needs, risk to human life or property, and water quality due to increased runoff.
For this assessment to be useful, it needs to be completed very quickly. Current BAER regulations require that the emergency assessment be completed within seven days of fire containment. This timeframe can be altered based on the local needs of the resource managers. The BAER assessment is the basis of the request for federal funds to perform treatments. These treatments are meant to help stabilize the landscape in an effort to mitigate some of the inherent dangers following fires. Some treatments may include culvert widening, installation of debris fences, seeding to quickly reestablish a ground cover, and, on a limited scale, reseeding of natural forest tree species.
Due to the short timeframe, any intermediate data used to create the final soil burn severity map must be delivered quickly after fire containment. RSAC and EROS provide assistance by creating classified DNBR images (referred to as the BARC – Burned Area Reflectance Classification). Analysts on the BAER teams adjust the BARC product based on field observations and ultimately create a soil burn severity map. In an effort to deliver data to BAER teams quickly, the post-fire imagery acquired for a fire may not be ideal. Often, there will be clouds over a portion of the burn scar. In other cases, due to the satellite revisit time, the best imagery may actually be acquired shortly before fire containment and have smoke and active fire spots visible in the imagery. In certain regions of the United States, the fire season lasts well into the fall when the sun angles are lower, causing deep shadows in mountainous terrain that create difficulties for image processing and interpretation.
Despite all these listed limitations, BAER teams will often make use of the best available data, regardless of the completeness or overall accuracy. In nearly all cases, the BARC product will improve the efficiency and precision of burn severity mapping compared to the singular use of traditional methods (helicopters, hard copy maps, and observations from scenic overlooks).
BAER and MTBS Overlap
Despite the fact that RSAC or EROS map fires with satellite imagery for BAER purposes, BAER fires will again be mapped during the MTBS project. However, the timing of post-fire images will be much different. Instead of acquiring imagery as close to fire containment as possible, imagery will be acquired for the purposes of an extended assessment, often during the peak of green in the next growing season. In most cases, this will be the next spring or summer following the fire. However, in some ecological regions (e.g., Southwestern United States), the next growing season may be in the same calendar year depending on weather and growing conditions.
On the surface there is an apparent duplication of effort between BAER and MTBS. However, the difference in purpose for the information that’s developed in each program clearly justifies separate efforts. More appropriately, BAER and MTBS mapping efforts should be viewed as separate processes that contribute to a suite of information for a given fire. Consistency in data type and method used to generate each product will enable the conjunctive use of those products for purposes beyond the original scope of either program.
Due to the differences in purpose and timing between BAER/BARC and MTBS products, resources dedicated to each program will be separate. During the fire season, RSAC and EROS will continue to have analysts dedicated to mapping active BAER fires. The mapping expertise necessary for each program is similar and will be shared when appropriate and feasible. However, during fire season, resource prioritization will be toward immediate post-fire support. MTBS project planning will use BAER support resources opportunistically but rely on its own dedicated resources to achieve production goals. Post-fire imagery necessary to support BAER and MTBS will be funded separately by the respective programs.
BAER efforts benefit from the MTBS project because all imagery purchased for the project will be available as pre-fire image choices for current BAER projects. Each image purchased by the USDI and USDA for fire is immediately placed in an archive, significantly reducing the cost for repeat purchases. As fires mapped for MTBS are completed, those images are entered into an archive and made available for a reduced price ($80) instead of the normal full price ($580 – for Landsat 5 TM imagery). Not only will this provide less expensive pre-fire imagery for BAER fires, but it will greatly increase the number of scene options available. In cases where burn severity is optimally mapped shortly after a fire, the MTBS project will benefit from BAER data acquisitions and analysis. Such cases are likely in vegetation communities where burn severity characteristics are quickly obscured due to re-growth, ash loss and other factors (e.g., desert shrublands, rangelands, and other ecosystems TBD).
Because the MTBS project will utilize many more Landsat scenes than a typical fire year, EROS systems engineers are making improvements to the imagery ordering and processing protocols. Engineers are refining the image geo-registration process at EROS and the processing speed. BAER incidents will benefit from the increased quality of the Landsat imagery, as well as the increased speed in delivery.
MTBS as an Extension of BAER
While the MTBS project will not replace or interfere with BAER mapping efforts, this project should be viewed as an extension to current BAER projects. As mentioned above, even if a fire was mapped for a BAER assessment, it will be mapped again during the next growing season as part of the MTBS project. Nearly everything about a BAER-mapped fire and an MTBS-mapped fire will be the same, with the major exception of timing. Both efforts use Landsat imagery to create DNBR images, the foundation for the BARC used by BAER teams.
Any mapping effort through the MTBS project can also act as a first step in a long-term monitoring project for local resource managers. Such an effort is currently underway on the Cerro Grande Fire (burned during 2000) in the Santa Fe National Forest, New Mexico. Analysts have secured Landsat imagery between 1999 (pre-fire) and 2005 and created vegetation indices highlighting the amount of photosynthetic activity present both within and without the fire perimeter. The comparison of the burn scar and an adjacent control site help analysts and resource managers draw conclusions about vegetation regeneration in the context of resource management objectives.
Conclusions
BAER mapping efforts are for specific and immediate assessment purposes. BAER teams try to find ways to stabilize the land burned by a fire in order to mitigate potential hazards to various natural and urban resources. MTBS mapping efforts are meant to inform policy decisions driven by the National Fire Plan and Healthy Forests Restoration Act as well as provide a broad scale means to assess and monitor trends in fire affected landscapes. Data created from the MTBS and BAER projects can be used to support and maintain a variety of resource data and modeling processes, including vegetation classifications, habitat and fuels maps, fire models, and invasives/pests/pathogens risks. While the actual product will have virtually the same spatial and thematic properties, BAER and MTBS do not operationally overlap. Despite that, the BAER community will benefit from greater efficiencies, costs, and options resulting from the MTBS project, and the resource community as a whole will benefit from a more comprehensive set of post-fire information.