By Chris Winters Herald Writer
OSO — The March 22 mudslide that wiped out a neighborhood and killed 43 people was largely triggered by a previous landslide in the same area in 2006.
This “remobilization” of the 2006 slide was far more dramatic and devastating than previous slide activity, and as a new report released Tuesday makes clear, there is no clear cause of what made that old slide come back to life, aside from the presence of water in the soil.
The report, released Tuesday morning, is published by the GEER Association. GEER, which stands for Geotechnical Extreme Events Reconnaissance, mobilizes teams of geologists and engineers to the sites of natural disasters that damage infrastructure.
While the county is facing multiple lawsuits from people saying they should have been warned about the landslide risk, the report also notes that there are no national or state guidelines concerning levels of risk due to natural landslides that warrant government action.
The goal of GEER is to gather as much data as possible as quickly as possible, before natural geological forces begin to obscure evidence of the disaster, and then release its findings publicly to spur on further research.
The report said that logging on the Whitman Bench above and behind the slope that collapsed might have led to increased groundwater in the slope.
But that was just one possible factor among many others contributing to the slide, and it was not the most significant factor even to the amount of groundwater in the slope, the experts said.
More significant to triggering the catastrophe was the large amount of rainfall in the weeks leading up to the slide, and many other factors that created conditions ripe for the slope to fail.
“It was really on the edge, and it doesn’t take a lot to set it off,” said Joseph Wartman, a professor of civil and environmental engineering at the University of Washington who co-led the team with Jeffrey Keaton, principal engineering geologist with AMEC Americas in Anaheim, Calif.
The team found that the March 22 slide took place in two stages.
The first stage started lower down the slope on Hazel hill and was the more devastating of the two. Stage 1 was the reactivation of the 2006 slide, sending soil and debris downhill, across the North Fork Stillaguamish River and through the Steelhead Haven neighborhood to wash up the slopes on the south side of the valley.
The second stage came a few minutes later, triggered most likely by the loss of support provided by the slope lower down, and involved the collapse of the main scarp of the hillside, causing it to drop 350 feet down, and pushing more debris across the newly exposed slope to hit the trailing edge of the first stage of the slide.
What triggered the initial collapse of the first stage, however, is not clear.
“The fact that we identified so many factors instead of just one just points to how many threats there were to that slope,” Wartman said.
Those contributing factors all point to a complex system in the Stillaguamish River valley, he said. Some of the identified factors might have contributed to the initiation of the slide, others to how the slope collapsed afterward, and others still to how the debris ran out across the valley, he said.
The team identified 10 possible triggers for the initial stage of the slide:
Three weeks of intense rain preceding the slide.
Groundwater entering the slide zone from the Whitman Bench, the top of the hill, which may have been affected by past logging activity.
More groundwater coming in from the Headache Creek basin.
The gradient of the hillside after the 2006 slide was flatter, and surface drainage less effective, allowing more groundwater into the area.
The 2006 deposits could have been loose enough to be subject to liquefaction, in which the soils are compacted so much they lose their surface coherence and can flow like a liquid.
Subsurface soil deposits consolidating and leaving voids that allow more water to seep in (a process called dilation).
Loss of soil strength on the shear surface of the 2006 slide.
Soil piping, in which denser soil deposits tend to push ground water through looser deposits. During periods of rapid infiltration (such as heavy rains), that can destabilize the slope.
The 2006 slide deposits below the surface had still not achieved equilibrium and were therefore more easily destabilized.
The toe of the initial slide incorporated river water and saturated river valley soil, giving the slide more fluidity.
Some of the possible contributing factors were supported by discoveries of interesting phenomena in the field.
The team found sand boils on the site, for example. Sand boils are an occurrence when underground water pressure pushes sandy soil up to the surface in a way that looks like a small volcano. They are evidence of liquefaction happening below the surface.
They found evidence of 15 other large slides in the valley over the past 6,000 years, and through carbon dating of tree bark in deposits and other means, calculated their frequency at anywhere from 400 to 1,500 years apart.
Though the scope of the report did not extend to analyzing how logging activity over the years might have contributed to the slide, it did find that its role would be primarily limited to contributing to increased groundwater in the slide area. Even then, its role was secondary.
“The place we saw the strongest effect of groundwater contribution was actually groundwater seeps coming in from the Headache Creek basin,” said David Montgomery, a professor of Earth and Space Sciences at the University of Washington who was on the team.
The team was comfortable in ruling out other possible effects of logging, such weakened roots of vegetation left on the slope, Montgomery said.
Further analysis of the effects of logging, and indeed of any larger scale studies to pinpoint causes, would take much more than the four days of field work and budget of less than $10,000 that GEER provided, he added.
The team also considered — but deemed unlikely given the evidence they found — that the entire slope collapsed in one massive event, or that the slide started at the top of the hill with a failure of the Whitman Bench.
With the publication of the report, it may be that other organizations will pick up where GEER left off, Wartman said.
One gap in the team’s knowledge is what was going on below the surface, because GEER was not able to do any drilling on or in the vicinity of the slide.
The study also listed conclusions that have more policy ramifications, such as considering past slide zones when mapping areas for zoning, considering both cumulative amounts and short-duration intensities of rainfall in assessing the likelihood of a slide, and revisiting the methods by which potential runout zones are calculated.
When standing on the site of the Oso landslide, it’s easy to believe that the slide’s runout across the entire valley could not possibly have been predicted, Wartman said.
But when comparing it to the record buried in the ground, the March 22 slide was within the scale of previous slides, he said.
“When you actually look at the numbers, you see that it actually does lay out,” Wartman said.
One final lesson for policy makers is an emerging distinction between landslide hazards and landslide risk. Landslide hazards would determine how likely a slide is to occur in a specific area. Landslide risk would determine possible impacts to human life or property.
The Stillaguamish River valley is full of landslide hazards, Wartman said, but Oso also was a high-risk location because people lived there. About 25 homes below the slide were occupied full-time; another 10 were vacation spots. All were in compliance with county codes requiring setbacks from landslide zones, the report noted.
“Studies conducted in the decades preceding the Oso Landslide clearly indicated a high landslide hazard at the site,” the report said. “However, these studies were primarily focused on the impacts of landslides to the river versus the impacts to people or property.
“In addition, it does not appear that there was any publicly communicated understanding that the debris from a landslide could run-out as much as one kilometer, as it did in the 2014 event.”
Implementing more risk-based landslide policies may lead to a clearer understanding of natural hazard danger, even years or decades into the future after a major event has faded from recent memory.
“Landslides won’t always behave in the future the way they’ve behaved in the past,” Montgomery said.
“That plays into public perception of both likelihood and risk. That’s the key lesson,” he said.
Chris Winters: 425-374-4165; firstname.lastname@example.org. Twitter: @Chris_At_Herald