Wednesday, April 9, 2014


Catastrophes have a way of catching our attention. A single nearby disaster can lead us to believe that this is the only important threat to us. A case in point: the 1980 Mount St Helens eruption in Washington State killed 57 people, and led to a dramatic increase in volcano research and infrastructure over the ensuing years. Wildfires and floods were on the back burner for awhile in the Pacific Northwest, and people bought a lot of masks that were never used. However, as the technology resulting from the research spreads worldwide, it is becoming increasingly unlikely that a volcanic crisis will ever again evolve into a volcanic disaster.

Sometimes we do not want to learn the lessons. Just two hurricanes in the United States (Katrina and Sandy) killed between them around 1,100 people in 2005 and 2012. Yet people are rebuilding homes on exposed New Jersey coasts and below sea level in New Orleans as if these events never occurred.

More recently, the OSO/SR 530 landslide killed at least 35 people, with 11 still unaccounted for. How are we as a society going to react to this? We are riveted when a woman is devoured by a shark off an Australian coast (New South Wales, March 2014). However, the United States in 2012 had 34,080 traffic fatalities. This contrasts with more than 51,000 deaths in 1980, so it’s clear that if society focuses on a threat long enough, many deaths can be prevented. But do we expend our resources in mandating seatbelts, airbags, and speeding-and-texting enforcement, or do we construct hundreds of kilometers of shark fence? What is a proportional response to a rare, unforeseen disaster?

Q: I live in Australia, but heard of the recent tragedy in Washington State where many people were killed in a landslide. I have some family who live on a steep hillside in the Pacific northwest and am wondering if they are in danger and if it is possible to predict when a landslide will occur. Thanks
- David I

A: No matter where one lives, there is always what I call “locality risk”. If you live in the woods, there are opportunistic and hungry bears and cougars – but far more commonly there are rocks to slip on. If you live in a city, there are people driving over-sized SUVs while texting. I had a very close call last year with a lady combing her hair with one hand while using the other to talk on a phone. On a curve. Locality risk is obvious to people who live in eastern Australia (truly apocalyptic firestorms), the southeastern US (continent-scale hurricanes), the central US (Force 5 tornadoes cutting swaths more than a kilometer wide across entire states), and California (earthquakes to magnitude 7.2 are not uncommon). Every once in a while our attention is caught by a “new” surprise, such as the 1980 eruption of Mount St Helens in Washington State. Volcanoes? We have volcanoes in this country? In December 2004 relatively few people on the planet had ever heard the word “tsunami” – until 250,000 people died around the margins of the Indian Ocean from a single event.

Bottom line: There. Is. No. Safe. Place. 
In flood-prone areas, or in hurricane-risk areas, in earthquake zones, etc., one can buy event-specific insurance to garner at least some protection. However, these policy riders are always expensive, and usually have large deductibles.

Your query probably has to do with the “Oso Landslide” (technically, the “SR 503 Debris Avalanche and Debris Flow”) in Washington State, on March 22, 2014. I listened to a senior scientist in our office who worked there describe what happened, and his speculation as to why, and learned a number of new things about landslides in general, and the Stillaguamish Valley in particular. I learned that typically the landslide height-to-runout ratio - the height where the cut in the hillside began vs the distance from that cut to the toe where the debris flow eventually stops  is commonly greater than 0.3. However, the Oso debris flow moved nearly three times as far as it should have, based on a database of previous landslide events worldwide. It may have reached speeds of 100 kilometers per hour. It removed and displaced a large section of the Stillaguamish River from its bed, creating a blockage that built a temporary lake. I learned that this particular area had experienced small to medium landslides in the past. I learned that the region had experienced unusually heavy rainfall for months preceding the event. Most importantly, I learned that the surrounding hills were not Cascades Range volcanic rocks like most everywhere else in the region, but were instead a large glacial outflow terrace. In other words, a big pile of (wet) dirt and rock.

What appears to be new in this case – and perhaps the reason for the unusually long and destructive runout – is that these glacial terrace sediments apparently were perched on a layer of clay-rich ancient lake bed material. Under the shock of the initial collapse, this may have (along with the overlying water-saturated glacial material) been liquefied by increasing the water pore-pressure in the sediments. Clearly everything was water-saturated, because even after the event, investigators began calling one scarp face “the weeping wall.” This scientist who led our discussion directs a research group that uses 4D mathematical modeling, laboratory-bench-scale physical modeling, and a 90-meter flume to experiment with debris flows. Their research concentrates on how debris flows behave differently with different composite materials and water saturations – and how they start. With all their years of experience, these scientists are only just starting to get a "feel" for when a debris flow in their flume will begin... but it's still impossible to predict. The leader of this group may be the most experienced landslide/debris-flow expert in the world, and he told us that he had never seen or heard of an event before like Oso. 

Are your friends and family at risk?

What can they look for? Is there a lot of open ground up-slope from their house that could be exposed to heavy rain? Do their foundations anchor in any sort of bedrock - or just thick soil? If there are old trees in the area, do they have bases that appear to bend into the hillside? This latter is a sure sign of ground creep. If the slope above their home is mostly other houses, paved streets and sidewalks, and the trees above and below them are straight, they probably have little to fear.  

As discussed in an earlier chapter, we cannot predict earthquakes. We can generally predict tornadoes by a few minutes to hours, and hurricanes with perhaps a few days warning. We can forecast these if we have enough data on previous events, especially in the case of large regions like the southeastern US, southern California, or the San Francisco Bay Area. By forecast, I mean to provide a percentage likelihood that an event of a certain magnitude will take place within a fixed span of time (usually 30 years). Forecasting is different from predicting, however. Predicting implies foreknowledge of the where and the when of an event. It implies that a warning can be given (like a siren for an impending flood) and people can be evacuated beforehand. Ideally, a disaster can thus be mitigated to be “only” an economic crisis. Forecasting, on the other hand, is largely suited to inform building codes, emergency preparedness, and to calculate actuarial data for insurance rate purposes. It may help you make a better-informed decision about accepting a job somewhere.

One of the few destructive events that scientists CAN consistently predict in the medium to long term are volcanic eruptions – if the volcano is adequately instrumented. However, even this is imperfect – we can often predict an approximate time of an eruption, especially as the magma approaches the surface, but we do rather poorly when it comes to predicting size and duration of a volcanic eruption.

Can we predict landslides? 

No – no more than we can predict earthquakes. Can we forecast landslides? Not really – they are localized events, and not regional events where we can gather meaningful statistics. Each landslide is like a human or a bear – it has its own unique characteristics, or “personality.”

If you are living in a flat area, however, it’s probably safe to say you need not fear a landslide. With sufficient geological mapping, we can get a sense of whether a landslide is possible in a given area: Are there steep slopes nearby? Are the steep slopes hard rock like granite, or are they hydrothermally altered or mixed rock types like we commonly find in volcanic terrains? A more dangerous end member is something like the unconsolidated glacial terrace deposits surrounding part of the Stillaguamish Valley. It is even more dangerous if there is geologic evidence of previous slides in the area. It gets more dangerous still if the area is prone to earthquakes or heavy rains, such as in Los Angeles. And it could get even worse: if there has been a huge fire or clear-cutting, followed the next year by heavy rains (such as Vernonia, Oregon, in 2007), then you lose even the limited protection of vegetation anchoring the soil of a slope.

In retrospect, the Oso area had several of these risk factors: heavy rains, unconsolidated sediments piled 180 meters high, evidence of previous landslides. However, there had not been any recent clear-cutting, nor had there been a fire in the area. There had not been any seismic activity, nor any human activity that could have triggered the mass movement. It just happened.

Perhaps we can say that landslides/debris flows are a risk one assumes when building in a place with a nice view. A son and a cousin who live near mountains in different parts of the Los Angeles area each separately experienced a large wildfire nearby, followed the next year by large mudslides. Neither regarded the mudslides as worth much thought – but they were not living in expensive hillside homes, either. It was the smoke and flames earlier that caught their attention and distressed them the most. For both, the fire was the more immediate and palpable threat, even though both fire and landslides were probably equally as dangerous to human life.

Oso is apparently just one of those rare, remarkable anomalies that could not have reasonably been predicted. It just happened - in one tiny fraction of the all the landslide-prone areas in Washington State. Initial mapping suggests that it’s an isolated situation - the glacial outwash terrace deposits to not extend very far up or downriver from Oso. The area is being monitored now with helicopter-dropped USGS “Spider” instrument packages and time-lapse cameras, but these don’t help the 46 dead or missing. It may help protect the survivors - however it’s hard to imagine people rebuilding in this area. 

Your friends and family are probably as safe as you are – or anyone else.  

No comments:

Post a Comment