The Science of Fighting Wildfires Gets a Satellite Boost

The Science of Fighting Wildfires Gets a Satellite Boost

11 September 2017

published by www.wired.com

USA – ASH IS RAINING down on the city of Portland. Thousand-degree fireballs are forcing evacuations near Salt Lake City, Seattle, and parts of Northern California. In Missoula, Montana, the sun burns a bloody red even even at high noon.

These days, the American West looks less like an Ansel Adams postcard and more like the kingdom of Mordor. Across nine states, nearly 1.5 million acres are on fire. And according to President Trump’s natural resources czar, you can blame it all on the hippies.

Allow me to explain: Last month, Interior Secretary Ryan Zinke traveled to his home state of Montana for an on-site briefing on the Lolo Peak fire, burning just south of Missoula and turning the valley’s air into toxic nubulae. Joined by Agricultural Secretary Sonny Perdue and members of Montana’s Republican delegation, Zinke pointed the finger not at drought or climate change, but on mismanagement resulting from lawsuits by “environmental extremists.” It’s not a new argument from the pro-industry camp, but Zinke’s comments are reigniting some old debates about the best way for people to manage forests that have been turned into tinderboxes by decades of overgrazing, fire suppression, and extended droughts.

Climate change only makes the cycle more vicious: Less water leads to more fuel leads to more fire means more money to fight fires means less money to manage forests, and now we’re back to too much fuel. Since 1985, global warming has nearly doubled the annual number of acres burning in the western US.

So what is a scrappy, resource-strapped agency like the Forest Service to do? That’s where science comes in. An emerging consensus suggests that officials should spend less time thinning out forests where a fire might hit, and more time figuring out what the specific conditions are when a fire actually does. But to do that, they’ll need some help from outer space.

Challenging the Tree-Hugger Theory

Lincoln, Montana, located 75 miles east of Missoula on Highway 200, is surrounded on all sides by the thick stands of spruce, fir, and pine that make up the western sections of the Helena National Forest. More than a decade ago, the area was dealt a double whammy—a series of summer droughts and a plague of pine bark beetles—that left the forest littered with big, dead, beetle-hole-ridden trees that were barely still standing. Called snags, this brittle timber is far more likely to fall in a fire, making for much more dangerous wildfire fighting conditions. And overlapping downed trees can make it virtually impossible to cut fire lines. So a few years ago the Forest Service began developing a plan to thin trees, selectively log, and do controlled burns on about 4,800 acres four miles north of Lincoln. The idea was to build a more resilient forest to mitigate the intensity of any fire that might come across the landscape.

The project was to start this spring. But in February, two conservation groups filed a lawsuit against the Forest Service, saying the agency failed to comply with federal laws requiring proper environmental impact assessments. And they alleged the project would disrupt essential habitat for endangered species like grizzly bears and lynx that live in the area. In May, a US district judge issued an injunction to stop all management actions from proceeding until the case could be settled.

Then came summer, with its dry heat and powerful thunderstorms. In July, lightning sparked a fire on a steep, isolated slope and a month later another storm started a second blaze. For firefighting purposes, they are now considered to be a single incident, which to date has burned 17,722 acres, mostly within the project area.

Here’s the straightforward logic of Zinke’s scapegoating: Environmentalists block the Forest Service from lowering the fuel load on the land, land catches on fire, and now it’s harder to put out. Thanks, tree-huggers.

But fire scientists say it’s more complicated than that. Many question the ecological (and economic) value of thinning forests out, for three big reasons. One, the evidence for its efficacy is both scant and at times contradictory. Two, probabilistic risk assessments show that the thinning doesn’t really help much because the likelihood of a fire starting close enough to interact with thinned areas is negligibly small. And three, in the worst weather conditions—dry, hot, and most importantly, windy—no amount of thinning or selective logging is going to make much difference.

A case in point: that Park Creek fire burning outside of Lincoln. It started on a remote slope that wasn’t slated for any prescribed burns or dead tree removals. But such treatments wouldn’t have made much difference anyway, according to Carl Seielstad, a fire ecologist at the National Center for Landscape Fire Analysis at the University of Montana, because the closest road is more than mile away, at the bottom of a slope. If you know anything about fire behavior, you know it moves much faster uphill. And in this case there wasn’t much in that direction, except more trees. “Without any roads in this area there was nothing for firefighters to anchor to,” says Seielstad, pointing at a 3D rendering of the fire’s path he’s pulled up on his computer. “It’s fair to say that regardless of treatment, this area would probably have been impossible to contain.”

Seielstad has been fighting, researching, and teaching about wildfires for 17 seasons. He says on the first day of his classes every semester he tells his forestry department freshman the founding principle of fire management: weather, topography, and fuels determine a fire’s behavior. But fuels are the only one you can do anything about. Even so, he says, the impact of thinning forests is mostly speculation. Its “effectiveness is hard to study because you can’t control any other variables out in the wild,” Seielstad says. “Sometimes it reduces the speed and intensity of a fire, sometimes it does the opposite.”

Further complicating matters is the fact that current models used to predict fire behavior aren’t particularly useful for forests that have been attacked by bark beetles, because scientists don’t yet have a lot of good data for how bug-butchered timber actually behaves. And, as Steven Running, a climate scientist who studies forest carbon (and shares a Nobel Peace Prize with former Vice President Al Gore for his work on the IPCC’s first global warming report) points out, the vast majority of forests around the world don’t come with detailed plot records. Meaning that scientists are always making assumptions about how old and how dense a forest is, what kinds of species make up a given hillside, and how fuel loads are distributed on the landscape. “When it comes to fire danger we don’t have much of an idea about how much dead material is lying around on the ground,” he says. “So when lightning strikes we don’t really know what’s out there, not in any detailed way.”

Until about a month ago, Running directed the intimidatingly named Numerical Terradynamic Simulation Group at the University of Montana. (He was one of a number of faculty who took a buyout package offered by the University to staunch recent budget woes.) He used massive, high-powered NASA satellites to measure the Earth’s daily rate of photosynthesis. He thinks forest managers should be using something similar (though much, much smaller) to better track, model, and plan for the West’s future fires.

For more than a decade, satellites have beamed down thermal data captured from high above the earth’s atmosphere to inform fire management systems of the location and severity of hot spots. But it’s only in the last few years—thanks to miniaturization and the democratization of remote sensing technologies—that such tools have started to become affordable. “When I started, satellites had the resolving power of 1 square kilometer,” Running says. “Now they’re good enough to get down to the level of a single tree. And they’re cheap.”

He envisions a world where flocks of nanosatellites constantly circle the planet, snapping pictures, waiting for a job to come in from a forest manager who has spotted a lightning strike or a rogue campfire ember. With the coordinates uploaded, a small sat could fly over to the area of interest, collect a bunch of images before the smoke gets too dense, and then send it to a computer for processing and analysis. With the right algorithms, in just a few hours forest service officials on the ground could have a real-time map of where the fuels were and thus where the fire was most likely to go. Combine that with already well-used weather and water stress models and you could have a much better plan much earlier, Redding says. “That capability is pretty much ready to go right now,” he says. “It’s just a matter of using it.”

The Canadian Model

You know who is using it? Canada. This year, Natural Resources Canada—the ministry responsible for managing the country’s minerals, metals, and forests—enlisted the help of 178 tiny satellites, each one weighing barely 11 pounds, to give managers a head start on any fires burning in the nation’s 857 million acres of forest. The satellites, nicknamed Doves, are owned by Planet, a San Francisco-based company started by former NASA engineers. Together with 5 RapidEye sats, the company is able to snap photos of 70 million square miles every day—the entirety of Earth’s landmass, and then some. And every day they upload about 7 terabytes of data to a fire detection company in Vancouver called Tanka. Using machine learning to parse smoke and flame from fire-free forests across Canada’s vast wilderness areas, Tanka then sends NRC the coordinates of any fires that have started in the last 24 hours, along with information about how big they are, how fast they’re growing, and what kind of fuel is in the area.

Tanka’s CEO, Nikola Obrknezev, says once the company has a full fire season under its belt he plans to begin talks with government agencies in the US, Australia, and Chile about offering a similar image-based surveillance service. The US, however, already has plans for its own early warning networking devoted entirely to wildfires. First conceptualized in 2011 by engineers at NASA’s Jet Propulsion Laboratory, FireSat aspires to be a constellation of 200 orbiting thermal sensors, capable of detecting fires as small as 35 feet across, within 15 minutes of the time they begin. The project is now being developed by San Francisco-based startup Quadra PI R2E, which plans to launch its first batch of 20 sensors sometime next year.

The goals of FireSat and Tanka are the same—detect fires as they happen, so that forest managers can deploy resources to contain them while they’re small, saving money, timber, property, and yes, clean air to breath. These systems will also help the Forest Service and the NRC make decisions about whether or not to contain a fire in the first place. In a natural cycle, the lodgepole and ponderosa pine forests of the Northern Rockies would burn every 8 – 30 years. And while all-out fire suppression has been the dominant policy in the Forest Service for the last few decades, that’s starting to shift.

Looking at Seielstad’s computer screen, which now shows all the fires burning in Montana, a pattern starts to jump out. What at first look like random splotches and cutouts on the landscape all have something in common. The places they’ve stopped growing all butt up against the borders of old fire scars—land that burned in the recent past, either from wildfires or intentional burns. “In my opinion, the best fuel treatment is fire itself,” Seielstad says. The fires burning right now are a form of future insurance for nearby communities. “Right now if you went out in the streets of Missoula and said these fires and this smoke were actually a good thing, you’d get tarred and feathered,” he says. “But five years from now, I think people will be really grateful.”

No one in Montana is excited about rebranding as the Big Smoke State. But fires, like floods and hurricanes, are inevitable. And attempting to fireproof the nation’s forests when no one knows where the next lightning strike will come is a Sisyphean task, with or without environmentalists nipping at your heels. It’s all going to burn at some point. The question is, will forest managers be able to make informed decisions about how and when and whether to let them go when that time comes? Having a few hundred heat-seeking satellites at their disposal certainly won’t hurt.