Big ol’ Alaska. The state that can swallow Texas, Montana, and California with room for a New England-size dessert. The Last Frontier is vast, yet threadbare of roads. Settlements are motes on the terrain. There’s so much space, so much of it inaccessible, that the US Forest Service—charged by Congress with keeping track of the nation’s timber—readily admits that more than one-quarter of the state’s forest has never been inventoried. But that’s about to change.
The Forest Service has always counted trees, mostly so timber companies would know how many two-by-fours could be culled from a given forest. Climate change, though, has changed the agency’s focus. Trees trap carbon, pulling it out of the atmosphere, and climatologists need to know exactly how much. With that data, they can build better models, and policymakers can prepare better tomorrows. So this summer, the Forest Service joined some earth science geeks from NASA to find a way to map more of Alaska’s woods—and the carbon stored within.
The mission revolves around a sensor called G-LiHT, or Goddard’s Lidar, Hyperspectral & Thermal Imager. Built by a small team of earth scientists, biologists, and technicians at NASA’s Goddard Space Flight Center, it’s a Frankenstein: three sensors in one, measuring carbon density, tree health and soil conditions. Bruce Cook, G-LiHT’s lead tinkerer, says similar machines have flown before in the lower 48 and other places you’d expect scientists to care about carbon storage, like the Amazon.
But nobody’s ever flown a stateside mission over an area so large and remote.
The Tanana Valley, located near the state’s center, is an Iowa-size chunk of boreal forest nestled between two mountain ranges. To the south, it’s framed by the towering Alaska Range—home to Denali—and to the north, by the smaller but still impressive Tanana-Yukon Uplands. Northwest of the valley is Fairbanks: home base for Cook and his tree-wrangling NASA cohort.
For five weeks this summer, Cook, two technicians, and a pilot stayed in an apartment near the University of Alaska. On a typical morning, weather permitting, the pilot and one technician would drive to the local airstrip, climb into a two-prop Piper Cherokee and fly to the valley. (The other two would stay behind at their so-called “frathouse” and process data.)
In the Cherokee’s cockpit, G-LiHT sat to the right of pilot Thaddeus Fickel, acting as his armrest, snack tray, and drink holder. Fickel would fly south, G-LiHT’s sensors strafing the landscape, until the slopes of the Alaska Range turned him around, about 240 miles out. Then he’d continue north, low and straight, towards the Uplands. “Flying between those two mountain ranges feels like you’re going back and forth in a half-pipe,” says Cook.
But structure is only part of the carbon picture. Trees lose some of their storage capacity if they’re stressed or diseased or dead, and some species are better at storage than others. That’s where sensors two and three come in. A hyper-spectral camera picks up super-subtle color changes, which can reveal a tree’s age, type, and even its health condition. And G-LiHT’s thermal camera uses heat signatures to determine whether soil is frozen, melted, or completely dried out—important for predicting how the forest will fare in the future.
Fickel would fly low at 1,100 feet, passing over trees, tundra, glaciers, and U-shaped valleys—left behind by glaciers now melted. And in the back, Cook or one of his fellow technicians was plugged into a laptop, watching fat packets of data stream in from the sensor in the co-pilot’s chair. Every second G-LiHT fired 150,000 laser beams, and snapped 75 frames from each of its cameras. Five weeks of flying netted them over 25 terabytes of data.
With that much data, little things start to go wrong. Each flash of the laser, every click of the camera, is annotated by a nanosecond-accurate clock, a nanometer-precise GPS, and a cadre of gyroscopes and accelerometers that correct for the plane’s constant jostling by the wind. All that pinging and clicking and tracking racks up minuscule errors. So, the data needs to be calibrated with what’s actually growing out of the ground.
For the Tanana Valley project, that meant sending 30 Forest Service employees and University of Alaska grad students into the woods on foot. “We chose the Tanana Valley because it was the most accessible of the uninventoried forests in Alaska,” said Hans-Erik Andersen of the Forest Service, who planned the ground-truthing effort.
In Alaska, though, “accessible” means trekking several miles from a trailhead, rambling through the bush in ATVs, or taking a boat 40 miles up a river. Some plots were so remote that a helicopter had to drop the team off—and then they still had to tromp through miles of mud before reaching one of their survey sites, located about every 130 miles along the plane’s path. Once there, they would size up the trees and soil, armed with measuring tape, shovels, and waterproof notepads. (Plus shotguns and bear spray, because it’s Alaska).
That might sound like the hard work. But now that the survey is done—they finished up in August—the real challenge falls to Andrew Finley, a Michigan State University statistician specializing in forestry. “It’s very easy for us to make maps that look really pretty,” says Cook. “But the people who will use them want to make sure you are certain about the quality of the data inside.” Though Finley went on no Alaskan adventures, and his role is inarguably unglamorous, the Tanana Valley inventory hinges on his work.
Once the lidar data has been processed, it is used to make point clouds showing the forest structure. 
Bruce Cook/NASAFinley’s the one who has to make sure every laser strike and camera shot is properly annotated by time, location, and orientation. He’ll assign every point and pixel with a rating that tells how accurately it reflects the truth on the ground—essential for the scientific analysis to come. And he’ll extrapolate from the mapped swaths to fill in the gaps where the Cherokee didn’t fly. By the end, the sensor flew over 13,000 miles—a trip from the north pole to the south pole would have been shorter—but the valley is so vast that the plane still had to candy stripe it, with six miles separating each flight line.
Every step of the Tanana Valley mission, from gathering the data by air and land to computationally chewing through it, was gargantuan. But there’s still a lot of work to do. The Tanana Valley counts for barely a fifth of Alaska’s unmapped public forestland. The team is hopeful they’ll get funded to do the rest and are also looking for opportunities to take G-LiHT abroad. Like to Siberia, which has larch forests that are crucial to carbon cycling but less charted than even Alaska’s remote woods. “There are a lot of stories going on in this landscape,” says Cook. “And we’re just starting to untangle what the effects of a warming climate might be.” Peppering the canopy with lasers might not seem like the most obvious way to understand climate change, but it’s a start.
How a Flying Laser Built a 3-D Map of a Massive Alaskan Forest - Wired
No comments:
Post a Comment