Where Land Slides, Trying to Learn Why
Where Land Slides, Trying to Learn Why
19 October 2009
published by www.nytimes.com
USA —
Water bottle in hand, Dennis Staley crouched over a patch of ground, cleared away surface soil with a finger and poured a couple of drops on the exposed layer.
Dr. Staley and a colleague, Jason Kean, researchers with theUnited States Geological Survey, were taking a break from their work one recent morning in Mission Canyon just above Santa Barbara to offer a show and tell on the interaction of soil and water.
It will soak in on the surface, Dr. Staley said. But just below the surface its going to bead up. Sure enough, the water drops were still visible, as if they had been spilled on a tabletop. So that enhances the amount of runoff that naturally comes off these slopes, he said.
Which, he added, looking at the largely ash-dusted, barren terrain, is why we do all this.
For the two scientists, all this had included hiking down trail-less slopes of loose rock and soil past sharp, charred stumps of manzanita and oak, hauling gauges and sensors, solar-powered communications equipment, a surveying device with a tripod, and assorted mounting poles, clamps, cables, drills and batteries, as well as cement and other supplies. As part of long-term geological survey research on landslide hazards, they were setting up a remote monitoring station to study how and when the runoff from coming rains might pick up soil and rock and become a destructive torrent of mud.
Given that it was denuded by a wildfire just four months before, Mission Canyon is a likely place for a landslide in this case, more properly called a mudslide, or, even more properly, a debris flow. In Southern California and other parts of the West, where the wildfires of spring and summer are followed by the rains of fall and winter, slides are the almost inevitable result.
But mudslides and landslides occur all over the world, often in areas with plenty of vegetation. Just a half-hour south on the Pacific Coast Highway in the small town of La Conchita, plain wooden crosses mark where 10 people were killed when the unburned hillside above the town gave way in 2005. In the past several months alone, typhoons in the Philippines and Taiwan have caused mudslides that buried whole villages, killing hundreds of people and bringing widespread economic disruption.
Last year, according to data compiled by Dalia B. Kirschbaum, a researcher at theNASA Goddard Space Flight Center in Greenbelt, Md., at least 540 landslides set off by rain were reported in published accounts around the world. In all, more than 2,100 people were killed.
Given the toll in lives and money, scientists like Dr. Staley and Dr. Kean are studying how and when slides happen what characteristics of soil and underlying rock, what degree of slope and intensity of rainfall will cause a hillside to fail or a flood of mud to roar down a canyon, demolishing houses or whatever else is in its path. And Dr. Kirschbaum and others are looking at landslide hazards on larger scales, using data from satellites.
The goal is to be able to assess risk and forecast when disaster looms locally, regionally, even globally.
Yet landslides do not get as much attention as other disasters, from the public or from scientists. They are usually thought of as secondary hazards, byproducts of much larger catastrophes like earthquakes, fires or storms. They are highly localized events, often occurring in remote, underdeveloped areas. And they can be baffling: one hillside may collapse in a storm while an adjacent one remains intact.
We have a fundamental problem, really, saidDavid Petley, a professor of geography at Durham University in England and director of a small research group there, theInternational Landslide Center. We understand the process through which these landslides occur incredibly poorly.
Dr. Petley has maintained a database of fatal landslides since 2002, in part, he said, because knowledge of when and where the slides occur is a step toward understanding them. But another reason, he said, is simply to demonstrate that landslides have a big impact.
Landslide risk assessment and forecasting may be an inexact science, but the work by geological survey researchers is about as precise as it gets.
Mission Canyon burned in May in the Jesusita fire, which destroyed more than 8,700 acres, much of it in theLos Padres National Forest. The canyon was a western edge of the fire, and a populated part. Dr. Staley and Dr. Kean had passed the remains of some of the 80 homes that burned usually only the chimney and foundation are left after a wildfire during the drive up to their staging area on a power-line maintenance road.
Thesite for their monitoring station, about 300 vertical feet below at the bottom of a small basin, had been selected through a modeling program developed by Susan Cannon, who directs the geological surveys post-wildfire debris flow studies program in Golden, Colo. The software takes into account factors like slope, soil characteristics and the severity of a burn, which can be measured by before-and-after satellite images of the vegetation, to determine possible sites for debris flows.
Burn severity is important, Dr. Cannon said, because heating the soil really changes how rainfall infiltrates into it. In places like Santa Barbara, she said, naturally hydrophobic chemicals in the chaparral become volatilized when the vegetation burns, creating a water-repellent layer in the soil.
Dr. Kean said ash from the fire might also play a role, plugging up pores in the soil structure. Either way, he said, as the show and tell demonstrated, its almost like a parking lot.
Among the instruments he and Dr. Staley set up at the site was a simple rain gauge. Rainfall is the key parameter for all this, so were always measuring rain, he said. They also installed a device to measure the moisture content of the soil the more saturated, the more runoff there will be.
The main piece of equipment, installed on a boom over a channel at the base of the canyon, is an acoustic sensor to gauge the runoff during a rainstorm. It measures the height of the flow, and when the data is coupled with measurements of the channel profile made using the surveying equipment, a rough volume can be calculated. Pressure data, from a transducer installed in a groove cut into the base of the channel, can determine whether the runoff is only water not necessarily a good thing, since flash floods can be destructive, too or whether the conditions were such that soil and rock became mixed in.
The data is constantly sent by cellular link to Dr. Keans computer in Golden. Even if no debris flow occurs, the information will be useful, adding to the body of knowledge about slide hazards and helping to refine Dr. Cannons model. The data is also sent to theNational Weather Service, which uses it and information from a network of other rain gauges to decide whether to issue local alerts about potential mudslides.
If theyre anticipating or measuring a certain intensity of rainfall over a given duration of time, theyll pull the trigger on a watch or a warning, Dr. Staley said. Its a program thats doing pretty well, but wed like to improve it and reduce their false alarm ratio.
Dr. Staley and Dr. Kean are part of a relatively small team, with about 15 scientists studying debris flows and other landslides. But by global standards the geological survey effort is large.
While there are other relatively small-scale risk analysis and warning programs around the world Hong Kong, for example, has a system of sensors to help protect residential areas that back up against steep hills many slides, particularly in less-developed countries, happen with little or no warning and no prior scientific assessment that the site was at risk.
AtNASA Goddard, researchers are working to provide risk assessment on a much larger scale. Were in the very early stages of trying to use space observations to look on a global basis where these kinds of hazards are occurring, said Robert Adler, an atmospheric scientist and rainfall expert.
The system uses near-real-time rainfall data from a NASA satellite applied to a global map of terrain ranked according to a susceptibility index. If rainfall exceeds a threshold in an area, Dr. Adlers algorithm will rate that area as likely to have landslides.
Were not in the warning business, for sure, Dr. Adler said. For one thing, the resolution of the data is low; they can forecast that landslides are more likely within a square area about 15 miles on each side. That is not useful for warning people to evacuate, he said, but it may help relief agencies, to give them a signal for prepositioning supplies.
Dr. Kirschbaum is working to improve the system, by creating higher-resolution maps of areas that are prone to landslides. She has used data from Hurricane Mitch, which caused many slides in Central America in 1998, to improve the susceptibility index for that region. Improving the resolution of rainfall data will have to wait, though, for new satellites.
The very apparent problem with this global issue is that theres uncertainty with every aspect of this, Dr. Kirschbaum said. Rainfall data has its flaws. Surface data has trouble in its resolution and accuracy.
For the geological survey scientists, resolution is not a problem. Sometimes they put out more rain gauges, because they find that rainfall can vary even across a small basin. The more data points on rain, Dr. Kean said, the better we can understand the variability.
They keep instruments at a site for only a year or two; after that the vegetation will have grown back enough to help stabilize the soils. Already at Mission Canyon, greenery is sprouting from the bases of burned-out trees.
While they were working in Santa Barbara they already had their eye on the Station Fire, a huge blaze near Los Angeles that destroyed more than 100,000 acres in September. Perhaps there, or at another future site, they would use Lidar, a device that scans the terrain with laser pulses to create an extremely precise map. With Lidar, we can create a virtual basin, Dr. Kean said. Every nook and cranny should be represented.
By repeating the measurements throughout the rainy season, we can see how sediment is moving through the watershed, Dr. Staley said.
And if there is a debris flow, he continued, well know exactly how much material came out of the basin.
Of course, that would involve hauling even bigger equipment down some slippery, burned-over slope. Its a gear-intensive effort, Dr. Staley said. But the information we can get out of it is pretty incredible.
