A landslide has come down across Highway 20 at about 4pm, July 29 (2009). The landslide follows days of high heat and thunderstorms. Here is the rain total for the last few days.

July precipitation totals with landslide location

July precipitation totals with landslide location

I am switching the format around a bit to include Google Maps into the posts for a more interactive view.

According to the WSDOT website: “They have about 300 yards of mud and debris in a 10-foot-deep swath across the highway to remove, and need to ensure the slope is stable before they will reopen the road.” That is a good sized debris flow.

The debris flow appears to have come down somewhere near (or maybe in?) Swamp Creek, a tributary to Granite Creek.

Update

Here is some more information from the fine folks at WSDOT. The amount of material removed for the landslide was approximately 650 (or more) cubic yards of material. A dump truck carries, on average, 10 cubic yards, so we are looking at 65 dump truck loads of dirt. The work was completed by WSDOT road crews (sometimes this get contracted out for landslides) and the cost was probably a bit more than $3,000. That is dependent on a few factors, wages and equipment, but also dirt transportation and deposit (which becomes more costly the further out you have to go).

Here are some images from the WSDOT Flickr Site

SR 20 Debris Flow, looking up slope, of July 29, 2009 (WSDOT Photo)

SR 20 Debris Flow, looking up slope, of July 29, 2009 (WSDOT Photo)

SR20 Debris Flow over roadway, of July 29, 2009 (photo from WSDOT)

SR20 Debris Flow over roadway, of July 29, 2009 (photo from WSDOT)

Advertisements

Each week we will feature a new landslide in Washington State. Washington State is covered with dynamic and sometimes quirky landslides.

Aldercrest Banyon Landslide, Cowlitz County

The Aldercrest Banyon Landslide is one of Washington’s famous landslides. It was the second worst landslide disaster (in cost) in the United States, following the Portuguese Bend Landslide on Palos Verdes Hills in Southern California, 1956, where 130 out of 160 homes on a ancient landslide were damaged or destroyed destroyed when the landslide reactivated.

Location map for the Aldercrest Banyon Landslide

Location map for the Aldercrest Banyon Landslide

Aldercrest Banyon Landslide Map

Aldercrest Banyon Landslide Map

The Aldercrest-Banyon Landslide started moving in February of 1998, many years after a housing development was established on the landslide mass. A description of the events by Dr. J David Rogers of the University of Missouri-Rolla Department of Engineering Geology follows:

“The Aldercrest-Banyon neighborhood in eastern Kelso, Washington began experiencing gross ground movements in February 1998, following 3-1/2 years of above-average rainfall. The initial signs of distress were the breakage of underground utilities. In March 1998 some framing distress was noted on a few homes. On April 10, 1998 a noticeable crack, 2.5 to 6 feet high, developed above the natural crest of slope west of Banyon Drive and north of Cedar Glen Court. Two homes on Cedar Glen Court were evacuated. The City made a valiant effort to patch streets, fill cracks and provide above-temporary above-ground utilities to the affected neighborhoods so that people could remain in their homes as long as possible.”

In October of 1998, a federal disaster declaration was issued by President Clinton for 138 homes affected by the landslide (Wegmann, 2006). The destruction exceeded $70 million, but the buyout for the houses was 30 cents on the dollar and totaled around $30-$40 million.

Map of houses affected by the Aldercrest Banyon Landslide

Map of houses affected by the Aldercrest Banyon Landslide, Map from Dr. J David Rogers

Photo of landslide damage

Photo of landslide damage, Photo from Dr. J David Rogers

Triggers

Deep-seated landslides are difficult to determine the cause of movement. This landslide was a relict landslide that could have been many thousand years old. The original triggers for movement are gone, perhaps an earthquake, a series of storms or prolonged (for years) rain, or maybe perhaps some sort of removal of lateral strength, by a stream or river. In modern day, the causes can be just as difficult to determine and each scientist has their theory, sometimes in agreement with one another, more often, not. For the Aldercrest Banyon landslide, the only theory most people seem to agree on is years of higher than average above rain probably played a big role.

Rainfall Rates for the Cowlitz County Area, from DGER RI35

Rainfall Rates for the Cowlitz County Area, from DGER RI35

Karl Wegmann (who was a landslide geologist here at the Washington Geological Survey) states in his publication that 6 years of increased rainfall correlates to a period of increasing landslide activity (Wegmann, 2006). Others reduce this somewhere between 2-4 years of increased precipitation. As a friend of mine would ask about now, but there are many areas of above normal rainfall on the graph, why this time, why this event? We might want to look at when the housing district first was established in this area. The housing development was in full swing in the area in 1975-76 with what I suspect is a septic system, although it could be sewer. With an increase in houses comes an increase in house related activities, watering the lawn, roofs pouring out water in the downspouts, concentrating water, etc. All things bad for landslide stability. Increased in rainfall would rates could have contributed to the reactivation of the landslide as well.

Geology

The subsurface in any landslide is an important characteristic to study. The area surrounding the landslide was mapped by Walsh et al., 1987. The deposits that the landslide sits on is known as the Troutdale Formation, which is appoximately 2 to 14 million year old Columbia River deposited gravels, sands, silts, and clays. Under the Troutdale is the Cowlitz Formation(roughly 38 million years old), depositing silt, sand, and mud in a near-shore marine deposition environment.

Geologic formations associated with the Aldercrest Banyon Landslide, image from Dr. J David Rogers

Geologic formations associated with the Aldercrest Banyon Landslide, image from Dr. J David Rogers

The formation creates areas of weaknesses at the contact between the Troutdale and Cowlitz Formation (and probably somewhat between the upper and lower Troutdale Formation).The increase in water over time probably contributed to increasing pore water pressure between the contact. This combination is present in many areas around the Kelso area and is probably responsible for much of the instability in the area.

The Aldercrest Banyon Landslide got the attention of many people. Its destruction caught the attention of the legislature, who initiated a landslide mapping project within the Washington Geological Survey. The landslide also propelled counties to look closer at landslide hazards to prevent another Aldercrest Banyon Landslide.

Reference

Walsh, T.J., Korosec, M.A., Phillips, W.M., Logan, R.L., and Schasse, H.W., 1987, Geologic Map of Washington Southwest Quadrant. Washington Division of Geology and Earth Resources Geologic Map GM-34.

Wegmann, Karl W., 2006, Digital landslide inventory for the Cowlitz County urban corridor, Washington; version 1.0: Washington Division of Geology and Earth Resources Report of Investigations 35, 24 p. text, 14 maps, scale 1:24,000. [accessed Mar. 6, 2008 at http://www.dnr.wa.gov/ResearchScience/Topics/GeologyPublicationsLibrary/Pages/pub_ri35.aspx]

There is a new blog out for Washington’s Geologic Hazards.
http://blog.seattlepi.com/washingtongeohazards/
The blog will be covering all geologic hazards to Washington State, not just landslides.

Each week we will feature a new landslide in Washington State. Washington State is covered with dynamic and sometimes quirky landslides.

Ribbon Cliff Landslide, Chelan County

The Ribbon Cliff landslide moved on December 14th, 1872, during (or shortly following) the 1872 earthquake. The landslide moved into the Columbia River and blocked it for several hours.
The earthquake location has been sort of a mystery until recently, as paleoseismic scientists have tracked down an approximate location for the landslide. The landslide was centered somewhere near Lake Chelan and caused some fairly interesting stories. Excerpted from the USGS website:

“Most of the ground fissures occurred at the east end of Lake Chelan in the area of the Indian camp; in the Chelan Landing-Chelan Falls area; on a mountain about 19 kilimeters west of the Indian camp area; on the east side of the Columbia River (where three springs formed); and near the top of a ridge on a hogback on the east side of the Columbia River. These fissures formed in several localities of differing physiographic environments. Slope failure or settlements or slumping in water-saturated unconsolidated sediments may have produced the fissures in areas on steep slopes or near bodies of water. Sulfurous water was emitted from the large fissures that formed in the Indian camp area. At Chelan Falls, “a great hole opened in the earth” from which water spouted as much as 9 meters in the air. The geyser activity continued for several days, and, after diminishing, left permanent springs.”

It would take some time, but it would be interesting to try and find the locations of slope failures from this landslide.

Ribbon Cliffs Landslide Location Map

Ribbon Cliffs Landslide Location Map

Ribbon Cliff Landslide

Ribbon Cliff Landslide

The landslide toe is a bit difficult to determine. The Rocky Reach Dam, built in 1962, formed Lake Entiat, which inundated the toe of the Ribbon Cliff Landslide. Older topographic maps indicate a probable toe (which is the loosely drawn toe on the map above). The landslide was drawn to the area mapped out in Madole et al, 1995.

The landslide location is along the Columbia River and was probably subjected to undercutting and oversteepening from erosion. Evidence of landslides (one just north of the Ribbon Cliff Landslide) is evident when looking at aerial photos of DEMs. The landslide itself was probably not a rock fall/topple event, but more of a translational landslide, that carried the mass relatively intact.

One of the more interesting summaries of the landslide can be found on this website. It gives a general summary of the landslide (plus some amazing oblique pictures of the landslide).

Controversy
It turns out (like most things in science) that not everyone agrees the Ribbon Cliffs Landslide is from the 1872 earthquake. The 1872 date, beyond the account of the 15-year old youth living in a cabin 3 km up stream, is also based on the presence of Mt. St. Helens Volcanic ash (set W) near the top of the undisturbed talus deposit (absolute maximum date). However, a dendrochronology study by Kienle et al (1978) reported that two of the oldest trees examined and three younger trees examined pointed to a failure date prior to 1872. No evidence was found within the tree rings of disturbance. This could be explained by the mass of the landslide did considerably disturb the trees (if the landslide mass moved as a whole).

Reference

Kienle, Clive F., Jr.; Farooqui, Saleem M.; Strazer, Robert J.; Hamill, Molly L., 1978, Investigation of the Ribbon Cliff landslide, Entiat, Washington: Shannon & Wilson, Inc., 26 p., 23 figs., 2 plates.

Madole, Richard F.; Schuster, Robert L.; Sarna-Wojcicki, Andrei M., 1995, Ribbon Cliff landslide, Washington, and the earthquake of 14 December 1872: Seismological Society of America Bulletin, v. 85, no. 4, p. 986-1002.

One of the things I have been pondering for awhile is if lahars should be entered into a landslide database. Technically, they are landslides, they act very much like hyperconcentrated flows or debris flows. So, have your say:

Each week we will feature a new landslide in Washington State. Washington State is covered with dynamic and sometimes quirky landslides.

Koontzville Landslide, Okanogan County
This landslide is part of the 1961 USGS publication Landslides along the Columbia River valley, northeastern Washington.

Location Map for the Koontzville Landslide

Location Map for the Koontzville Landslide

Koontzville Landslide Map

Koontzville Landslide Map

An excerpt from USGS Professional Paper 367 on the Koontzville Landslide:

“The Koontzville landslide involved the entire village of about 35 houses, one store, and a section of State Highway lOA [replaced by State Highway 155]. The village was built in 1934 and 1935. … Old landslide materials extend from river level almost to the top of the terrace, or to an altitude of about 1,600 feet. Little or no landslide activity was noticed before the 1948 flood. There may have been some slight highway settlements or minor movements owing to irrigation and river-bank erosion below the highway, but no property damage from landslides was reported. In the fall of 1948 (about the time of the Seatons landslide movement) one resident of the area had trouble with water pipes parting and resorted to flexible hose connections to keep his water system operating. So far as is known, this marked the beginning of reactivation of the ancient slide. The slide has moved many times since. Movements are recorded on the following dates:

December 23, 1951;
November 10 or 11, 1952;
November 27, 1952;
and January 10, 1953.

In contrast to the diminishing rate of movement observed in the Seaton slide since 1948, the Koontzville slide seemed to move more and at more frequent intervals in successive years to and including the spring of 1953. Local residents have noticed that their houses cracked and moved each weekend during low stages of the Columbia River, which corresponded to drops in river level due to power operations at Grand Coulee Dam. Many houses and the store have been severely damaged, the springs have changed their courses, large fissures have crossed the village area, and each year the slide has worked farther back into the hillside. In 1952, a fissure connected the Koontzville slide with the Seaton slide along the silt-granite contact (fig. 20). The displacement in 1955 extended all along the bedrock outcrops between Seatons Grove and Koontzville. Vertical movement along this bedrock scarp ranges from a few inches to 5 feet. Before the 1948 movement there was a light-colored zone on the granite immediately above the contact with the surficial deposits which ranged in width from 0 to 15 feet. Above this zone, all the granite wall is much darker due to weathering and organic growths. This light-colored zone may represent the amount these slides moved down following an earlier Columbia River flood such as the one in 1896. Geologically, Koontzville is in a setting where the sequence of Pleistocene deposits is the most favorable for landsliding. A preglacial channel of Peter Dan Creek underlies Koontzville and because of this geologic setting ground-water conditions are very high.

Figure 20 from USGS Professional Paper 367

Figure 20 from USGS Professional Paper 367

Conditions similar to this have been described in the Reed terrace area, and they can be anticipated, almost without exception, where deposits of silt and clay now occupy the area of confluence of preglacial valleys with the main valley. The causes of the initial reactivation of this ancient landslide seem to parallel those outlined for the Seatons landslide. The causes of the periodic movements, however, are not well understood. In 1953, the Corps of Engineers drilled three test holes in the slide to obtain undisturbed samples of the soil and to install gauges to record pore-water pressures throughout the year.”

The factors pertaining to the Seatons Landslide movement were:

“Many factors influenced the renewed landslide action in this area, of which the following seem the most important:
1. The unusually heavy rainfall during the spring and summer of 1948.
2. The high water in the Columbia River during the flood of May and June, 1948, undoubtedly resulted in a higher water table throughout the entire slide area.
3. The flood eroded and unloaded the toe of the slide, which is on the outside of a bend in the river where erosion would be greatest.
4. Melt water from the heavy snowfall in the winter of 1948 and 1949 kept the slide lubricated and moving after sliding began.
5. Very deep freezing in the winter of 1948 and 1949 may have had some effect in extending old slide cracks and in damming ground water.
6. Seatons Lake was created in a key position at the head of the ancient slide so that it kept much of the lower part of the ground saturated. Springs on the lower slopes of the hill produced more water when the level of Seatons Lake was higher, and the lake surface was raised purposely at times to make the springs at lower altitudes flow at a greater rate for irrigation.
7. The material at the toe of the slide consisted of silt J and clay thinly mantled with sand, gravel, and boulders. Silt and clay could be observed pushing through the gravels at several places along the toe of the slide. Since the construction of Grand Coulee Dam, a replacement supply of sand and gravel to cover and protect the silt and clay from erosion had been largely cut off.
8. The extensive use of this area for homes, gardens, irrigated tracts, and roads had undoubtedly been a factor in encouraging the renewed activity of the slide. Renewed activity might have been postponed if the natural cover of grass and sagebrush had not been removed and if the streams had been kept in their natural channels. The principal spring, which flowed a full stream through a 2%-inch pipe, supplied the entire area with domestic water. The other two springs in the drainage above were about the same size. The small stream, which was seasonally diverted into Seatons Lake, flowed between 0.5 and 0.6 cfs, even in dry years. The stream probably flowed about 1 cfs in the early spring and during unusually wet seasons. The supply of water to the main spring was cut off during the slide of November 1948, but the flow of water was restored to about normal by driving a pipe into a small seep which broke out near the spring. The spring water was milky for several days before it cleared.”

Looking at the 10m DEM, it looks like they missed a rather large earthflow that came down.

Earthflow on the Koontzville Landslide

Earthflow on the Koontzville Landslide

Although this probably didn’t play much of a role in the Koontzville movement in the 1950’s, except for the overall instability in the area. Finding information on this landslide has been difficult. I would only assume it has mostly stabilized out, as houses still dot the landscape.

Reference:
Jones, Fred Oscar; Embody, Daniel R.; Peterson, Warren Lee, 1961, Landslides along the Columbia River valley, northeastern Washington: U.S. Geological Survey Professional Paper 367, 98 p., 6 plates.

Landslides can be destructive, destroying houses, infrastructure, and kill or injure people. However, we don’t usually think about landslides being bad for your health.

Washington State has a complex geology. Much of the western Cascades is made up of accreted terrains , composed of both oceanic and continental rocks. Parts of these terrains contain asbestos (which occurs naturally, despite a relatively large number of people believing it artificial). Asbestos is a fairly blanket term for a wide variety of minerals, some harmless, some very dangerous. The most well known example of the dangers of asbestos can be seen in Libby, Montana, where vermiculite mining with occurrences of fibrous tremolite asbestos caused widespread health problems and death for many of the residences and workers.

In Washington State, asbestos outcrops across the state. Most of the outcrops are small, uneconomical to mine or develop and probably pose little danger with limited exposure. However, some larger deposits occur in Snohomish, Skagit, Whatcom, Kittitas, and Klickitat Counties. These deposits can cause weakness within rocks and are sometimes associated with weak, friable material, places where we would expect landslides to occur. The prime landslide that contains asbestos in Washington State is the Swift Creek Landslide in Whatcom County. The landslide material is composed mostly of serpentinite, a friable, weak rock in terms of stability with high amounts of chrysotile. Its origin was probably an uplifted oceanic plate that was probably composed of ultramafic material, such as dunite that was then metamorphosed and transformed into serpentinite. The landslide has produced a significant amount of material which has been transported downhill into the valley below, depositing chrysotile laden sediments. These sediments, especially during flood events, deposit in places where people can come into long-term exposure, which can result in long-term health problems.

Swift Creek might be the most well known landslide to contain asbestos in Washington State, but since asbestos occurs throughout Washington State, many other landslides have the potential to contain asbestos. This map represents deep-seated landslides that have the potential to contain asbestos within them.

Washington State Asbestos-DSLS Map

Washington State Asbestos-DSLS Map

This map is not a perfect representation, as available data is scarce. The map was created by overlaying identified asbestos occurrences found in Bulletin No. 37, Inventory of Washington Minerals (Valentine and Huntting, 1960) with the 100k geologic units (with slight modification on unit selection). The units that were identified with asbestos occurrences were then intersected with deep-seated landslides from DGER Washington’s Statewide Landslide Database (the database is located within the menu). These deep-seated landslides are of all ages, from relict to active. Points were then selected at the centroid of the polygons to create a point file of landslides that potentially contain asbestos materials. It isn’t a perfect method by any means, but it at least gives us an idea that more of these landslides probably exists throughout Washington State. I am in the process of intersecting the landslide layer with ultramafic units known to contain serpentinite, which will help expand and potentially more accurately capture landslides potentially containing asbestos.

Reference
Valentine, Grant M.; Huntting, Marshall T., reviser, 1960, Inventory of Washington minerals; Part I–Nonmetallic minerals; 2nd edition: Washington Division of Mines and Geology Bulletin 37, Part I, 2nd ed., 2 v.

Each week we will feature a new landslide in Washington State. Washington State is covered with dynamic and sometimes quirky landslides.

Pe Ell Landslide, Pe Ell, Lewis County

The Pe Ell Landslide failed during the December 3rd Storm of 2007, closing State Route 6 just west of Pe Ell.

Pe Ell Landslide - Photo by WSDOT

Pe Ell Landslide - Photo by WSDOT

The debris avalanche/slide flowed across the highway and pushed a truck into the living room of the house across the way. Remarkably, most of this was caught on tape by the residences of the house.

Pe Ell landslide impact of house - WSDOT Photo

Pe Ell landslide impact of house - WSDOT Photo

On December 11, Kelsay and I arrived at the landslide. The drive through the Chehalis valley was spooky to me, a lingering stench filled the air and misery could be seen all around. Home after home, farm after farm all showed damage from the floods. By time we arrived, WSDOT had already arranged for an emergency contract with Scarsella (on December 9th) to begin work on clearing State Route 6. Unfortunately, with all of the heavy equipment working on the site, we decided to stay on the periphery of the landslide and investigate the damage to the structures.

The damage was localized to the western lobe of the landslide. It impacted the houses at a low speed, warping and pushing them.

Pe Ell landslide impact to a house - DNR/DGER Photo

Pe Ell landslide impact to a house - DNR/DGER Photo

Pe Ell Landslide impact to second house - DNR/DGER Photo

Pe Ell Landslide impact to second house - DNR/DGER Photo

Meanwhile, WSDOT was working hard on figuring out the landslide. The WSDOT Geotechnical Division has access to many really neat tools to help with their investigations. Here is a 3-D representation of the landslide mass created by their division:

Pe Ell Landslide 3D Model - WSDOT Geotechnical Division

Pe Ell Landslide 3D Model - WSDOT Geotechnical Division

They also compiled a small scale geologic map of the landslide mass (with an amazing aerial photo of the landslide):

Pe Ell Landslide Geologic Map - WSDOT Geotechnical Division

Pe Ell Landslide Geologic Map - WSDOT Geotechnical Division

In the end, WSDOT removed over 47,000 cubic yards of material to stabilize the landslide mass at a cost of around $4 million dollars. The project was completed on March 13th, 2008, over three months after the storm.

The landslide prompted a debate on logging, landslides, and highway safety. The landslide itself was logged weeks before the storm. The interesting part, this landslide wasn’t caused by root strength loss, it was probably too deep anyway to have much impact. The lack of canopy, however, might have played a roll in the landslide initiation. Canopy plays a role in reducing the rate rainfall from reaching the ground (to a certain point) or slow melting of snow on the ground by reducing rain rates and buffering changing temperature. It is difficult to say in an intense storm how much it might have slowed the rainfall, or reduced snow melt (by reducing the warm rain and temperature from reaching the snow), but the lack of trees, even with this intense rainfall, probably did increase the likelihood for its initiation.

Cause aside, the cost of repairing these landslides is expensive. This is just one of probably hundreds of landslides to fall on our highway systems each year. Figuring out why these landslides fail and if we can either mitigate or possibly find better management practices to help reduce landslides would help save millions of dollars and reduce injury and death.

Things have been a bit busy this week with family coming into town and a shift in projects at my work. However, I will be back on track next week and start adding in more landslide information.
Recently, Lee, the librarian here at the Washington Geological Survey, gave me an extra copy of the Geological Survey Professional Paper 367 (Landslides Along the Columbia River Valley, Northeastern Washington). I will be adding in these landslides to our landslide database and probably pluck a few out here and there to add into this blog.
Hope everyone has a safe 4th of July. While watching aerial bursts and swinging around sparklers, remember to think about the chemical and geological sciences that went into creating such magical fires.

From Wikipedia:

Colors in fireworks are usually generated by pyrotechnic stars—usually just called stars—which produce intense light when ignited. Stars contain five basic types of ingredients.

* A fuel which allows the star to burn
* An oxidizer—a compound which produces (usually) oxygen to support the combustion of the fuel
* Color-producing chemicals
* A binder which holds the pellet together.
* A Chlorine Donor which provides chlorine to strengthen the color of the flame. Sometimes the oxidizer can serve this purpose.

Some of the more common color-producing compounds are tabulated here. The color of a compound in a firework will be the same as its color in a flame test (shown at right). Not all compounds that produce a colored flame are appropriate for coloring fireworks, however. Ideal colorants will produce a pure, intense color when present in moderate concentration.
Color Metal Example compounds
Red Strontium (intense red): SrCO3 (strontium carbonate)
Lithium (medium red): Li2CO3 (lithium carbonate)
Orange Calcium: CaCl2 (calcium chloride)
Yellow Sodium: NaNO3 (sodium nitrate)
Green Barium: BaCl2 (barium chloride)
Blue Copper halides:CuCl2 (copper chloride), at low temperature
Purple Potassium or Strontium + Copper: KNO3 (potassium nitrate) or SrCl+ + CuCl+ (Strontium chloride + Copper chloride)
Gold Charcoal, iron, or lampblack
White Titanium, aluminium, or magnesium powders

The brightest stars, often called Mag Stars, are fueled by aluminium. Magnesium is rarely used in the fireworks industry due to its lack of ability to form a protective oxide layer. Often an alloy of both metals called magnalium is used.

Many of the chemicals used in the manufacture of fireworks are non-toxic, while many more have some degree of toxicity, can cause skin sensitivity, or exist in dust form and are thereby inhalation hazards. Still others are poisons if directly ingested or inhaled.