Step 1: Identify the Landslide.
Step 2: Avoid the Landslide.
Step 3: Live!
Step 1: Identify the Landslide.
Step 2: Avoid the Landslide.
Step 3: Live!
Your odds of death by landslide are one in a million per year. If you get caught in one, you can't do much to increase your odds of survival. Luckily, landslides frequently give warning before collapsing. These are a few precursors to a landslide, telling you to run before it's too late.
Reactivated La Chichita Landslide in Ventura, California. Photography credit: Mark Reid/U.S. Geological Survey
Those one-in-a-million odds of death by landslide come with some caveats. Landslides are gravity-driven mass movements. If you're in the flatlands with no hills in sight, you can consider yourself pretty safe from landslides. Depending on the underlaying geology, you might still have problems with subsidence or sinkholes, but you're unlikely to be buried in tons of rock. If you live on a steep slope with landslides all around, your odds might be higher.
The odds of death by landslide are engineered odds: one-in-a-million per year was selected as an appropriate number for acceptable risk by policy makers. In places where the risk is naturally higher, mitigation measures are taken until the odds drop to acceptable levels. This isn't quite universal around the world — Hong Kong spends more money on its landslide-prevention program, changing their odds to 1 in 10 million per year, while the Philippines spends less money and has consequently higher odds at 1 in 100,000 per year.
Trying to survive a landslide has more to do with luck and quick reflexes than anything you can carefully plan. If your car gets thunked by a falling rock, slam on the brakes, hop out, and run. If your entire home starts to collapse around you, ball up and try to protect a pocket of air, like you would in a snow avalanche. If you're outside and see one coming, get uphill, preferably on the inside bend of a curve where the landslide will spill up the outside edge like a race car banking around a corner. It's not much to work with, but if you're lucky, you might survive.
Far better advice is to try to identify landslides in advance, and either fix them or avoid them before they happen. Geologic Survey landslide hazard programs are typically under-funded, so don't blindly rely on the government to proactively tell you somewhere is dangerous. Here's some of the signs of an impending landslide to put you on your guard:
Just because a landslide has already happened does not mean that an area is safe from landslides. In fact, observing old landslides is a good indication that the area has unstable geology, and that more landslides are likely in the future. Multiple landslide events in the same place can be retrogressive, piecemeal, or reactivated.
The Steelhead landslide in Oso, Washington and the La Conchita landslide in Ventura, California are both reactivated landslides. A reactivated landslide is where something changed at an old, semi-stable landslide, triggering a new failure at the same location. If you look around an area and see a whole lot of old landslide scarps and deposits, it's a pretty good indication that even if a particular landslide doesn't reactivate, the underlaying geology makes the whole region unstable and susceptible to landslides.
Rescuers searched for survivors after the first landslide in Badakhshan, Afghanistan. They were likely buried in the second landslide. Photography credit: AP/Ahmad Zubair
Earlier today in Afghanistan, a landslide buried a village, then a few hours later, another landslide at the same location buried people searching for survivors. Although details at this point are very scarce, this is probably a piecemeal failure, where the landslide failed in several distinct pieces in a relatively short period of time. Due in part to the strange physics of extremely large landslides, these smaller landslides are physically less catastrophic than it would be if everything failed at once, but they can be far more devastating if rescuers are on-site and get caught in the secondary events.
Finally, a retrogressive landslide is when the top scarp marking the beginning of the failure starts creeping backwards. That doesn't sound too scary, until you realize that means any nearby bystanders think they're safe until they suddenly aren't.
This is a video of a retrogressive failure in leda clay under a railway track near Stackpool, Ontario, Canada. In it, the landslide starts progressively eroding farther and farther and farther back. I'm very glad this did not happen at the Baltimore event, or the neighbours could have lost more than their cars. Leda clay, or quick clay, is particularly prone to retrogressive failure as the internal structure collapses. The most stunning example of this is a 1978 failure in Rissa, Norway, where over 300,000 square meters of farmland eroded into the river within hours.
If you watch the video of the retaining wall collapse in Baltimore, you can hear people complaining they had been trying unsuccessfully to get this road repaired for three years. Dave's Landslide Blog points out that you can spot the impending failure on Google Street View by the tension cracks. It gets even more obvious if you check archival images: In the image from August 2009, the road is in disrepair, but has no clear cracking. In the image from September 2011, the road has long, cracks parallel to each other and to the slope, and even the sidewalk is starting to split from the stress. The same features are not present on an equally poorly-maintained road a few blocks away but not above a retaining wall.
Comparison of the Google street-view images in 2009 and 2011. Note the development of linear cracks along what will be a landslide scarp in 2014.
Those are tension cracks created by the stress of geological material pulling apart. They often mark the eventual landslide scarp, the top edge of the failure zone. You can identify them as road segments that constantly need long, thin patches, or out in the wild as actual cracks in the ground. Tension cracks above an existing landslide can hint at a future reactivation.
Tension cracks at Burma Lines Camp, Fanling. Photography credit: Geotechnical Engineering Office of Hong Kong
If you find tension cracks on flat ground, they're still being created by extensional stress, but are more likely related to a fault than a landslide.
This can be extremely obvious. In the Baltimore event,you can watch the cars slowly tilt and tip during the initial slow portion of the landslide before the whole thing collapses. More dramatically, check out this video fromHastings old town, East Sussex. A few rocks fall from the cliff shortly before massive towers of rock topple into the ocean.
(For the landslide pedantic, this is a rock topple, not a rock slide, as the motion is rotational and not transverse.)
Movement can also be more subtle for an earthflow or a creep, where the ground is moving at maybe centimeters per year. Deformed fences, damaged structures, or subsidence are all indicators of stress, but the one I see most commonly is trees who have bent up in a J-curve as the ground slips out from under them.
Trees on a creeping hillside. Photography credit: Mika McKinnon
A single tree doing something gnarly could be just a mis-angled seedling twisting towards the light, or an entire forest might be from snow-loading of young saplings. But a patch of angled forest on a slope or J-curved trees somewhere that never gets snow, can be a good indicator that the ground is less solid than it seems. While so far the motion has been slow, the slope could catastrophically collapse as a rapid landslide, especially if you find tension cracks farther uphill.
Inside structures, signs of slow shifting can be floors tilting, doors no longer closing property, or broken utilities. Creaking and cracking can also be warning signs. Alas, these are also indicators of things as benign as, "An increase in humidity has swelled the wood" or "An old house settling on its foundations," so use caution when self-diagnosing an impending failure.
Changes in water flow are Not Good. The sudden arrival of water in a spring, seep, or wet ground somewhere that is usually dry indicates that something changed, and change is not to be trusted. Worse, all that water is saturating the pore space, lubricating any failure surfaces and increasing pressure that can drive a failure.
The sudden disappearance of water is just as ominous. Water levels in a creek suddenly dropping can be due to an upstream obstruction like a landslide-dam that will break and send a debris flow rushing downstream. A debris flow is a very wet, very mobile landslide, where water is loaded with trees, mud, rock, and everything else caught in the torrent. In this video, notice the low water level preceding the arrival of the debris flow surge:
While I can't confirm it with publicly-available documents, rumour among geotechnical nerds is that the railway-landslide in Ontario was triggered by a change in water flow. Apparently, a drainage culvert was clogged with ice, which suddenly failed and suddenly drained the ditch. The rapid down-draw was enough to trigger the quick clay collapse, leading to the retrogressive failure.
Changes in water are downright creepy when working underground in a mine. The sudden absence of a constant drip-drip-drip can be terrifying enough to trigger an evacuation just in case.
This list doesn't cover every possible sign of an impending landslide, nor does everything on this list immediately lead to a landslide. It's a start, warning signs to help you recognize when it's time to hired an expert to assess a slope's stability, or to make you more aware that something is amiss. To learn more about the disasters that impact your area most, I recommend checking out municipal hazard information, or sign up for an introductory geology course.
If you witness a landslide, please report it to the USGS Did You See It? database. Do you feel more comfortable with some types of disasters than others?