The war being waged against wildfires from Southern California to Greece and Australia is almost as complex as the infernos themselves. Innovative computer mapping tools advance, as do airborne imaging techniques that can look straight through black smoke for views of emerging dangers no firefighter ever sees. However, some crews battle blazes on bulldozers older than they are, and funding is tight all around. Still, the breakthroughs keep coming. This is the first part of a five-part series looking at firefighting:
Part I: THE FIRES DOWN BELOW: ‘LOOK-DOWN' TECHNOLOGY
Part II: UNDERSTANDING WILDFIRE BEHAVIOR AND PREDICTING ITS SPREAD
Part III: WHAT'S REALLY HAPPENING ON U.S. FIRELINES
Part IV: CATCHING WILDFIRE ARSONISTS RED-HANDED
Part V: SMART SOLUTIONS GOING FORWARD
The Viz Lab is a large, dimly lit, war room dominated by huge, computer-generated maps projected onto dark walls. Its tool kit includes an array of links to information and imaging feeds gathered by satellites, airplanes, unmanned aerial vehicles (or UAVs) and helicopters from sources like NASA and Google Maps. The lab is bent on delivering real-time (or pretty darned close) computer mapping and imaging to a wildfire's first responders so they'll know just what the blaze is doing, where and when.
Data fusion is the name of the game at the San Diego State University's Immersive Visualization Center — layering sophisticated weather, atmospheric, smoke and fire data and images onto, say, a topographical Google Earth map. It provides an illuminating picture for emergency operations chiefs who urgently need to pinpoint trouble spots and interpret fast-changing developments.
Once, fire perimeters were indicated by simple black lines on old-fashioned land maps — best guesses made from the field without benefit even of GPS. Now, satellites or aircraft use "look down" technology to create 3-D topographical images of what lies below dark, billowing smoke. Tools distinguish live from burned vegetation and show in various colors rapidly updated information on a blaze's "hot spots" and accelerating or subsiding dangers.
"It's absolutely dramatically more useful," explained Eric Frost, co-director of the Viz Lab.
The Viz Lab normally focuses on geographic information systems research for homeland security and disaster relief. But it also proactively tracks everything from brush fires on its doorstep to natural disasters worldwide. Last February, for example, it helped map wildfires in Australia that killed 173 people. "It takes less than half a second to go from here to Australia on fiber optics," Frost noted.
The lab's before-and-after imaging of 2005's Indonesian tsunami benefited those rushing to maximize disaster response efforts. And after Hurricane Katrina, Frost's team observed that an oceanfront casino had relocated to a new spot — atop a vital four-lane highway. Priceless information.
The Viz Lab can do all kinds of clever stuff with wildfires such as gathering data on flame heights and temperatures, and wind direction fluctuations, allowing tweaks in suppression tactics.
In October 2007, San Diego County's numerous raging wildfires killed seven people and forced 500,000 to evacuate. Plagued by the notorious, fast-moving Santa Ana "devil winds," the blazes burnt more than 325,000 acres.
With next-generation radar from the National Weather Service, smoke movement can be viewed less than 10 minutes out of real time or even faster in a less data-rich mode. "During these fires," Frost said, "images were about three minutes from real time." Because it elongates in the direction of the major wind, smoke tells a really important story for those planning suppression efforts.
Layering images of the fire spread onto maps showing the positions of houses, roads and infrastructures, then layering on images showing where brush and wildland had been burned or spared, was enormously informative.
NASA's MODIS (for "moderate resolution imaging spectroradiometer") satellite imagery proved very effective. It can be processed to show smoke very well. NASA's Earth Observing-1 imagery also was used to see through the smoke to the fires below.
During major emergencies, data and satellite imagery pours in to the Viz Lab from dozens of sources, notably NASA and the military. But for the 2007 San Diego fires, new ground was broken when NASA staffers, along with Google programmers and engineers, physically descended on the Viz Lab to work side by side with its team and wrestle with problems like overwhelmed computers and getting near real-time images — free of charge — to all who needed them. When the Viz Lab's own computer systems crashed, a switch to the Web servers of the nearby California Institute for Telecommunications and Information Technology saved the day.
Moving information around sounds simple. It's not. The true head-scratcher remains how to convert gigantic maps that only supercomputers can run into micro images and tiny files that can be transmitted to fire chiefs and ultimately to firefighter's individual GPS units. The challenge lies in extracting and creating bite-sized files for tiny networks out of the big picture, "rather than this massive thing that is going to fracture the system," Frost said.
Droning On and On
Military and National Guard air support arrived too late in San Diego in 2007 for the liking of some. Frost, a true believer that communication, networking and relationship-building pays off, saw it differently. One triumph: "NASA very graciously agreed to bring the Ikhana Predator B here and to pay the bill, which was way over $1 million."
The Ikhana UAV (or drone, as remotely piloted aircraft, with their military connotations, are often called), is usually used on long-term earth science observation missions and as an airborne platform for developing and evaluating new UAV technologies, but it has been used to fight California fires since 2007. It collects real-time thermal-infrared data and can fly for up to 30 hours at high altitude carrying science and technical research sensors as well as atmospheric and remote-sensing instrumentation.
Hyperspectral imaging instrumentation, for example, collects, measures and analyzes reflected light in the form of a spectral fingerprint and by doing so, can see what we cannot. It can spot a vehicle hidden by vegetation and can reveal what lies below clouds and smoke so precisely that it can even distinguish a maple tree from an oak. Watching a fire perimeter's fluctuations shows where a fire is bursting out and where efforts to battle it are working; and with state-of-the-art imaging, good quality information can be on the ground to emergency operations centers or fire battalion chiefs in 10 minutes.
In 2007, while National Guard, Air Force and Navy aircraft provided images, and the U.S. Forest Service delivered more environmental information, back in the Viz Lab, Google brainiacs helped build software on the spot to create topographical maps embodying all the available data and imagery.
The armory of information-gathering methods grows. Synthetic aperture radar (or SAR), for instance, can differentiate between burned and new trees, and can also be used at night or to see through fog or clouds. It's invaluable to have precise, up-to-date locations for all remaining fuels — everything flammable from trees, brush, plants and timber, to houses and other structures - so is infrared imaging that can read moisture levels in shrubs and grasses and monitor when they get dangerously low and combustibility is rising.
Data fusion maps also enable close monitoring of hazardous materials in relation to smoke movement and can give advance warning when a fire is headed for a danger spot like a large propane tank or chemicals. That, in turn, allows officials to advise people living downwind to get away. It also helps rescue and relief agencies set up their facilities out of harm's way.
Precisely how most wildland firefighters toil in the dark was chillingly revealed by the Viz Lab's 2007 high-resolution images. Staffers saw hidden perils like fire creeping up a hillside behind firefighters, a blaze splitting off and hot spots out in front. Firefighters know when embers are flying over their heads but, especially if they're flying over a hill, they don't know what they are doing behind them.
The revelations drove home the courage of firefighters: "The fire's burning behind them and they're caught in the squeeze, and yet they're still standing there," said Frost, his awe obvious.
The Ikhana, which is 36 feet long with a 66-foot wingspan, can't fly close to large fires or in intense winds, so it's not the optimum aircraft for imaging wildfires. (Just as well, since few wildfires have an Ikhana on scene.) A solution now gaining traction is to use small, high-speed UAVs — some less than 10 percent the size of an Ikhana for perhaps as little as 1 percent of the cost — to carry in pods the high-tech instruments needed to image wildfires.
SDSU recently acquired three new UAVs, made by RP Flight Systems in Texas, with a wingspan of just 4 ft. Yet the SDSU drones can give better imagery — typically, they fly 400-500 ft. above ground — for far less money than the Ikhana.
The downside? They can carry some key imaging programs, but these very small models can't hold them all. However, their other advantages compensate. The Ikhana has to fly at far higher altitudes to comply with FAA rules. Meanwhile, since these small, civil UAVs have no military use and aren't federal assets, they neatly avoid governmental red tape. (For instance, having to clear classified data slowed down the transmission of the Ikhana's imaging in San Diego in 2007.)
SDSU's small UAVs can deliver infrared imaging of plant life, and with forward-looking infrared (or FLIR), can show where the fire is by reading heat. What they can't view, Frost noted, is, "the composition of the smoke and the great details of the vegetation. And having SAR would provide far higher-resolution images, too, and extremely accurate detections of any changes going on."
That said, these UAVs clearly are a welcome new tool. Improved models are definitely viable and will likely be developed as budgets permit, so it can be expected that the civil use of UAVs in wildfire fighting will expand. With the optimal craft, Frost theorized, "you could fly around a fire, collect imagery, process it, send it down to the ground, and maybe get a fire perimeter every 10 minutes."
On the Ground
There are also entirely different "look down" technologies. In 2003, now-retired California Department of Forestry and Fire Protection Chief Bill Clayton envisioned sensors and digital cameras on tall towers detecting backland fires almost instantly, then relaying data and images back to operations chiefs and on to firefighters who could see imminent threats for themselves on laptop computers. We're not quite there yet, but we are inching forward.
The Wildland Fire Detector System, an optical system currently used in Turkey's governmental agency, the General Directorate of Forestry, is now being evaluated by private clients and governmental agencies in Oregon and California, Australia, Chile and Greece. Smoke is always the first indicator of fire — because tall trees and dense vegetation can hide flames — so having smoke detected in the first five minutes or so could dramatically reduce damage. The WFDS can monitor grasslands, but it shines in forest settings by detecting smoke far sooner than flames might reach the crowns of tall trees, finally making them visible to observers.
"The system consists of remote monitoring locations that scan the area and feed information to the control tower," Dennis Akers, Wildland Detection Systems' CEO, wrote in an e-mail. "At the control center — which can be unattended — the system analyzes the data stream and issues an alarm if a wildfire has started. The system has the capability to identify the GPS coordinates of the fire."
Control centers monitor up to 16 remote sites simultaneously with Windows-based computers and microwave links. Since visual and audible alarms don't require independent operators, noted Akers, "they can be integrated into the existing public safety system, reducing operating cost."
Each remote monitoring point has a digital video camera, power (usually solar-charged battery power), and a communications line (usually point-to-point microwave). A single remote sensor can cover 10,000 acres, or more than 300 square miles. And new technology, Akers was careful to note, has successfully eliminated a high false alarm rate.
The California Department of Forestry and Fire Protection itself has more than 350 remote automated weather stations — towers with computerized sensing equipment that samples weather conditions hourly, then transmits data to a satellite. The U.S.'s second largest fire department, CalFire covers more than 31 million acres of rural California. In 2008, the agency alone responded to 3,593 fires involving 380,310 acres. (California's total firefighting efforts that year, including those of federal and county agencies, involved 6,255 fires and 1,593,690 acres.)
And the U.S. Forest Service, teamed with the U.S. Geological Survey Center for Earth Resources Observation and Science, is using the Burned Area Emergency Response Imagery Support program for fast satellite imagery and data delivery. The Forest Service's Remote Sensing Applications Center also uses the MODIS Active Fire Mapping Program, which can detect fires as small as 100 square meters burning at 1,500 degrees Fahrenheit.
Cloud cover, or the fire's position relative to the MODIS sensor, can hamper detection. But, it offers the public a general overview of fire information.
Still, people whose lives and homes may be in peril crave their updates in real-time, not every four to six hours. And while few systems can cope when 100,000 people descend on a Web site to download a PDF file, the Viz Lab can now handle about 10 million hits a day. It will also gladly host Web sites free for emergency response agencies in the midst of a wildland blaze or other disaster. All they have to do is ask.
* * *
In Part II, we look at efforts being made to understand wildfire behavior and to predict its spread.
Are you on Facebook? Become our fan.
Follow us on Twitter.