UPDATE - March 25, 10:43 p.m. ET: Local search and rescue officials recovered two more bodies from the mudslide on Tuesday, bringing the official death toll to 16. Another eight bodies were discovered on Tuesday, but officials were unable to extract them. The unofficial death count is 24, with 176 people still missing. Read the latest updates, here.

When the massive debris flow of trees, rocks and mud swiftly buried the Steelhead Drive Development in Oso, Wash., on Saturday morning, killing at least 14, it did so without any warning. With at least 176 people still listed as missing since the landslide hit, this could be the second-deadliest landslide in U.S. history.

According to the U.S. Geological Survey (USGS), landslides in the U.S. typically cause between $1 billion and $2 billion worth of damage, and kill more than 25 people each year. They happen in all 50 U.S. states plus its territories, and can take many forms.

Unlike deadly hazards such as tornadoes and flash floods, for which the U.S. spends billions each year in order to issue advanced watches and warnings, the federal government spends a little under $3 million per year on a national landslide program, according to Jonathan Stock, director of the USGS Innovation Center for Earth Sciences in California.

Image: Ted S. Warren/Associated Press

“We are a very small national program,” Stock says. “We don’t have the capabilities to do widespread monitoring.”

Instead, the USGS, working with the National Weather Service (NWS) and state and local agencies, has put together a “patchwork quilt” of monitoring and experimental warning programs, based upon rainfall and soil moisture and pressure measurements. One such program has been in place near Puget Sound, but did not cover the area where the March 22 landslide occurred.

As the Seattle Times reported, the hillside above the Steelhead Drive development has a long history of producing large, sudden landslides, the most recent of which occurred in 2006. This raises questions about why there was such a large housing development at the bottom of the steep hillside — and why proper landslide monitoring was not already taking place.

The Washington landslide may lead to a reevaluation of the haphazard way in which the U.S. currently monitors for and warns of landslide risks. “I think there’s going to be a debate about why we don’t have a more robust monitoring system,” Stock said.

Experts told Mashable that there are several reasons why there is no landslide warning network comparable to, for example, the more than 100 Doppler weather radars that scan the skies nationwide for signs of impending tornadoes and other hazardous weather.

Stock points to three factors that make establishing such an early warning network so difficult, even on a local level, in addition to the harsh fiscal realities on Capitol Hill. First, he said, the phenomenon is intermittent. Second, landslides occur across a broad geography, and it is difficult to narrow down the most vulnerable locations.

And finally, there is the challenge that the most important data is found underground, and can’t be seen by the naked eye staring at a hillside, until it is too late.

The challenge of predicting landslides is quite different from that of predicting severe weather events. For one thing, landslides in any particular location are extremely rare. There are often decades between slides, during which time officials may question why money is being spent to monitor such a low risk (albeit high impact) event.

Tornadoes, meanwhile, happen hundreds of times per year — though their frequency in any community outside of "Tornado Alley" is far lower.

When deciding where to put their monitoring equipment, scientists have to anticipate which areas are at greatest risk of landslides, Stock says. But that risk assessment is often not available until after they do the monitoring. This leads to a chicken-and-egg problem.

“We have a tremendously difficult time running that experiment," Stock recalls. "Where do you put the instruments if we don’t know where an event is going to happen?”

A host of factors go into determining when a vulnerable hillside gives way — including the type of soils in a particular area, how much water infiltrates such soils, and whether human activity has destabilized the hillside.

Even earthquake forecasting (which no one does on a local or national level) might be easier, Stock says, since geologists can detect where strain is building up along known fault lines.

Image: Mark J. Terrill/Associated Press

“Even if we wanted to put it up, I’m not sure how we would make that national [landscape warning] network,” Stock said.

In Washington State, heavy rainfall — roughly 200% of typical precipitation during the past three months, plus snowmelt — has saturated hillsides, elevating landslide risks.

According to Lynn Highland, a geographer at USGS’ National Landslide Information Center in Golden, Colo., the landslide danger may stay high even after the rain stops.

“Sometimes it takes days or weeks after heavy rain for certain slopes to become saturated, and much of this is unseen, due to the fact that water moves underground, as well as seeping in from the surface. Earthquakes can make the situation even worse,” Highland says.

“As far as warnings, we know that certain thresholds of rainfall can initiate landslides," she added. "However, the geology and soils are so different, even in small areas, that we can only get rainfall thresholds on a limited basis.”

Rainfall thresholds, above which landslide risks are higher, exist for the Puget Sound area of Washington, southern California, and parts of the Appalachians, she said. “There is not enough data for other parts of the country, although there are landslide-initiation rainfall rates for certain wildfire burned areas — but these are only good for about 3 years, until vegetation grows back."

Rex Baum, a research geologist with the USGS in Golden, Colo., told Mashable that monitoring when rain levels go above a particular threshold is the most promising basis for warning of elevated landslide risks. But this would still mean other landslides, which are not triggered by heavy rains, would often go unwarned.

“The potential for making reliable early warnings with landslides is also somewhat limited to rainfall induced landslides,” Baum says. “Sometimes landslides occur due to other causes, and making any kind of prediction for when those slides occur is very difficult.”

The USGS is trying to improve scientists’ understanding of landslide dynamics. But its program is dwarfed by the size and scope of the hazard faced by communities, particularly given the rapid population growth in and around vulnerable areas of the West.

In the Bay Area, Stock oversees a network of four monitoring installations, each of which costs about $10,000, not including personnel costs incurred while maintaining them. He says the program in San Francisco is geared toward finding out exactly what soil moisture and pressure conditions trigger landslides in that area, which could have broad implications for other areas prone to such geological phenomena.

Landslides typically occur there once every six or seven years, Stock said.

The deadly Washington State landslide was similar to the 2005 La Conchita landslide in Santa Barbara, California in which 10 people died. Both events happened in known regions of historical landslide activity, Stock says — but in both cases, local residents did not foresee the possibility that the next landslide could be far worse.

In La Conchita, as in Snohomish County, there had been other slides in recent history, including one in 1995. But local residents had not expected a sudden, rapid failure to the extent of the 2005 landslide that soon after.

“The memory of a particular landslide fades very quickly,” Stock said.

Highland said USGS researchers are stymied in part by a lack of a reliable landslide inventory for the U.S., which is a prerequisite for conducting susceptibility studies.

“As far as warnings ahead of time, [there is] not really a good way of telling when a slope will fail," Highland said. "Sometimes there are warning signals like ground cracking, springs suddenly appearing on a hillslope, or houses slowly moving out of kilter. Landslide instrumental monitoring has come a long way, but still hasn't been perfected.”

Stock says more resources need to go into landslide risk evaluation if future tragedies are to be avoided, noting that the $3 million annual budget for the USGS landslide program is less than the cost of a high-end home in the Bay Area.

“We haven’t invested the kinds of resources in going after this that we have in other challenging scientific areas,” he says.

Have something to add to this story? Share it in the comments.