A wildfire sparks an average of 60,000 times a year in the United States. Fortunately, it is not just anybody’s guess where it goes from there—whose house it might burn or which town it might decimate.
LANDFIRE Updates Valuable Firefighting Data
LANDFIRE 2020 released this summer for the United States, Puerto Rico, U.S. Virgin Islands
A wildfire demands rapid decision-making to protect the people and property in its path. Scientific data on topography, vegetative conditions, and weather help guide analysts and firefighting teams. This information indicates which way a fire is likely to head and helps prioritize where resources will be needed most.
LANDFIRE, short for Landscape Fire and Resource Management Planning Tools, is a key national data source for the management of wildfires, management of the plant materials that fuel fires, and planning for prescribed fires. The U.S. Department of the Interior (DOI) and the U.S. Department of Agriculture Forest Service (USFS) provide resource support and business leadership for the 20-year-old LANDFIRE program, with data products generated at the U.S. Geological Survey Earth Resources Observation and Science (EROS) Center, in partnership with The Nature Conservancy.
LANDFIRE datasets, which involve a team of dozens working to produce the most accurate and detailed products possible, are ever-evolving. Change, or disturbance, happens continually across the landscape. Trees are harvested, barren or burned areas grow new vegetation, and cities spread. The latest version of the LANDFIRE product suite, called the LANDFIRE 2020 Update, was released this summer for the conterminous United States (CONUS), Alaska, Hawaii, Puerto Rico, and the Virgin Islands.
Certainly the DOI and USFS have plenty of public land to monitor for fire potential, including 193 million acres for the USFS and 450 million acres in the DOI’s national parks, wildlife refuges, and other lands. But fires don’t always start or stop within public borders. The comprehensive maps produced by LANDFIRE include information about all public and private land. Fire managers anywhere can access LANDFIRE products directly, or through a number of other tools and programs that rely on LANDFIRE’s layers of information.
“It’s impressive that you can build these maps that are going to help regular everyday people use it for looking at their backyard, but we have agencies that also use it. The amount of people that are using it is phenomenal,” said LANDFIRE Task Manager Marcine Hyser, a contractor for the USGS.
The western United States has suffered from drought since 2000 and is experiencing longer fire seasons, more and larger fires, and increasing populations living in the Wildland Urban Interface, the transitional area where wildland and human development meet.
In 2022, as of August 22, more than 43,000 fires have begun with nearly 6 million acres burned, both higher than the 10-year average, according to the National Interagency Fire Center (NIFC), and likely to grow significantly during the remaining fall fire season.
Firefighters need all the help they can get.
Landsat Helps Complete the Picture of Change
Before the 2020 update, comprehensive base maps were redone for the LANDFIRE 2016 Remap from original 2001-era information, although updates had been released in between. A 2019 Limited update included changes reported for 2017, 2018, and 2019 for CONUS, but it did not include any unreported changes detected only by remote sensing. The new 2020 effort offers information about both reported and unreported disturbances in 2017, 2018, 2019, and 2020, along with the expected 2022 fuel conditions.
Reported disturbances tend to be those events that happened on public land and changed the landscape, such as fire or blown-down trees. LANDFIRE receives those reports from the federal or state agencies managing the land, or other organizations.
LANDFIRE finds out about a lot of other changes to the landscape through remote sensing, usually from Landsat imagery. The two orbiting Landsat satellites together provide a view of every spot on the Earth’s land surface every eight days, and the archive of imagery stretches back 50 years. This imagery, with its 30-meter pixels, can reveal changes in vegetation, burn scars, and urban development beyond the borders of public land. Timber harvesting on private land in the northwestern and southeastern United States is an activity for which LANDFIRE leans heavily on satellite imagery to be able to detect.
LANDFIRE’s meticulous analysis and verification by team members help boost the accuracy of the product. Then when things like high haze or rain-soaked ground get flagged as change on imagery, human eyes discern whether it’s a true change or a momentary condition.
The process to produce LANDFIRE’s insights may be intensive, but it makes the impossible—characterizing each pixel of land throughout the United States—possible, if not in person, then through satellite imagery.
The reported events are compared to remote sensing disturbances, and then they are all mapped. Additional datasets related to burned areas are also folded into LANDFIRE data, such as Monitoring Trends in Burn Severity (MTBS), which is based at USGS EROS. MTBS maps the fire perimeter and levels of severity throughout burned areas of at least 1,000 acres in the west and 500 acres in the east.
LANDFIRE senior scientist Brian Tolk, a contractor for the USGS, leads the effort to map disturbances, which includes assigning values that essentially indicate the level of confidence the team has in their label of the type of disturbance. A disturbance that’s both reported and detected by satellite imagery has higher confidence than one detected only by imagery, for example. The additional datasets like MTBS also help, Tolk said.
“Not only is it beneficial to us because we’re more confident in their severities because they do it fire by fire, but it’s also nice to have that data for our disturbance products because people who use them then have the MTBS data already built into the product so they wouldn’t have to download it themselves,” he said.
Determining Vegetation and Fuels
The next steps in the LANDFIRE mapping process involve vegetation and fuels, which help determine what a wildfire will do next. Landsat provides information about the type and timing of disturbance on the ground and their effect on lifeforms such as trees, shrubs, and grasses. Some vegetation will burn easily; other vegetation will not. Irrigated cropland is not likely to burn, while a dry pine forest is. Barren ground, such as a mine or gravel pit, will stop a fire, but fire will surge through a thick stand of dead annual grass.
For both disturbed and undisturbed areas, the vegetation layers describe the existing vegetation type, among more than 800; the percent cover of tree canopy; and the average height of the dominant plant. On-the-ground data points of vegetation are used in relation to the satellite image surface reflectance to extrapolate across the entire country. The layers also include information about the vegetation that likely existed in each area before Euro-American settlement, known as biophysical settings.
Fuel assignments follow disturbance and vegetative determinations. In addition to live trees, shrubs, and grasses, fire also finds fuel in litter on the ground, such as dead leaves, branches, pine needles, wood chips, and downed trees.
Although the most recent input for fire behavior fuel models is from 2020, the fuels team can adjust the determination to reflect 2022 conditions by considering how the fuels would predictably have progressed in the length of time since a disturbance like a fire, with a given level of severity.
Sites vary according to factors like climate and seed sources, so the help of a couple dozen regional experts is key to the process. Damage from insects and disease can create different fuel characteristics for a site as well, varying in severity and timing.
“If a fire modeler is going to be pulling these data in, they need to know where the disturbance has happened and how that affects the fuels. So, if they’re trying to understand where a new fire is going to go, when they model the fire behavior, it’s going to account for previous disturbances, and the fuels will be different. The rate of spread and the flame length will all change in the fire behavior model,” said fire scientist Inga La Puma, a contractor for the USGS serving as the LANDFIRE technical lead.
Among other layers in the LANDFIRE 2020 Update are topography, which was updated with the latest available 3D Elevation Program high-quality data and provides key factors of fire behavior, such as elevation and slope; and the new operational roads feature, which lets firefighters in the area figure out how to access a fire.
Former Hotshots Put Their Experience to Work
When Charley Martin and Tobin Smail work on their portion of LANDFIRE—fuel assignments—the contractors approach it with more than knowledge of datasets and computer modeling. They both served with interagency hotshot crews, including 20 years for Martin, who served the last five as superintendent of the Forest Service’s Sante Fe hotshots. “I think I’ve been to every state except for Massachusetts on fires, and several countries as well,” Martin said.
This real-world experience matters.
“I have a lot of actual fuels background, especially in Oregon,” Martin said. “Going out and doing prescribed fire and learning about fire and then analyzing what it does on the ground, rather than just sitting in the office and determining what it does, I think that’s really important.”
Martin and Smail both were tapped to help in government roles with fire management, fire fuels, and fire behavior modeling in the western United States and eventually ended up drawn to LANDFIRE. “Having that experience of being on the ground really helps with assigning fuel models and knowing what canopy fuel looks like in different parts of the country and understanding how fire propagates up through it in different areas, different locations, and different vegetation types,” Smail said.
LANDFIRE’s production team is based at USGS EROS in Sioux Falls, South Dakota, and some team members live there, but the program benefits from team members who can relate to geographic regions all over the country, too. Martin lives in Oregon, Smail lives in Washington, and La Puma lives in New Jersey; other states represented include Montana, Minnesota, Ohio, and Florida.
Fire Support Systems, Smoke Modelers, and Researchers Turn to LANDFIRE
Two major users of LANDFIRE data are the interagency Wildland Fire Decision Support System (WFDSS), which is intended to assist agencies with firefighting decisions and documentation, and the Interagency Fuel Treatment Decision Support System (IFTDSS), an application that guides any land manager with fuels treatment, analysis, and planning. Smoke modelers also rely on LANDFIRE fuels information.
“We’re providing data and tools, and if users feel like there’s adjustments that can be made, they can do that for fuel models and make them fit how they need to for their area,” Tolk said.
But fire agencies aren’t the only users of the valuable data. “LANDFIRE is used for a lot of different studies outside of operational fire,” La Puma said. Businesses or organizations interested in determining the risk of fire for an area use it. The nonprofit group First Street Foundation created a fire risk database for the whole country based on LANDFIRE datasets. Habitat researchers also find LANDFIRE’s vegetation data useful in studies and restoration efforts.
The LANDFIRE 2020 Update, with four years of vegetation and disturbance data, and LANDFIRE 2019 Limited, with three years of limited data, represented a tremendous effort by team members. But they are now looking toward a faster timetable: to be able to supply LANDFIRE’s users with data on an annual basis.
Smail said LANDFIRE is up to the challenge of producing an annual product. “It’s a lot of work. But as we do it, we’re getting better, and we’re developing things faster.”
The public—the fire managers and the people they’re trying to protect—will benefit most. “If they can have the most current data as possible, that’s what they’re after,” Martin said. “If they could get it near real time, that’s the best for them.”
‘We Transcend Boundaries’
Tolk has been part of LANDFIRE since its prototype days—well before Landsat imagery became free in 2008. At the time, the LANDFIRE team had to limit the number of scenes they bought and search through all of them to choose only high-quality scenes. After the archive opened up, “we could go through and download stacks of data instead of just one,” Tolk said.
Since then, automation has helped develop scene composites to reduce gaps in data, which can be more likely in cloudier areas such as the Northeast.
“The best part about the LANDFIRE data is that we transcend boundaries,” Tolk said. “Fires and disturbances, they don’t care where (jurisdiction) boundaries are, and they’ll pass through. To have consistent data across the U.S. of same pixel size—everything is matched up so they don’t have to get all these different pieces from 20 different places—it’s a simple system that allows them to, with as little effort as possible, pull that data and information in.”