Failure of the Teton dam: geotechnical aspects4 July 2002
On 5 June 1976, the Teton dam, located on the Teton river in Idaho, US, failed, releasing a wall of water that was reported to be 22.9m high. Fourteen people died as a result of the flood and large areas of cropland were destroyed, along with livestock and buildings downstream. The possible reasons behind the failure of the 93m high dam are discussed in the first of a series of two articles on the disaster
During the middle of the last century, the US Bureau of Reclamation(USBR) built dams to help control flooding, improve irrigation and provide power and recreational areas. The Teton dam was just another on a long list of dams to be constructed. US Congress authorised its construction in 1964 but this was delayed until 1972. Original plans called for an earth dam that would be approximately 93m high creating a 27.4km long reservoir with a 333M m3 capacity. Overall, the dam would require about 7.6M m3 of earth and 14,051m3 of grout.
The dam was located in Fremont County, Idaho, near the Idaho-Wyoming border in the Teton Mountain range. The area had a basalt rock foundation that would prove to be problematic to the construction and integrity of the dam. On 3 June 1976, during initial filling, small springs were noticed downstream of the dam. Taking this as an indication of water seepage, the face of the dam was examined.
On 5 June 1976, seepage was discovered on the dam face and steps were taken to ensure the safety of the dam and power plant. Workers tried to fill newly created holes with rubble but the voids continued to grow. Three dozers fell into the holes before hope was abandoned for the survival of the dam. A vortex was observed behind the dam one hour before failure. Within five hours of the first observed seepage from the dam face, it failed. At its peak release, the flow was estimated to be 28,300m3/sec. A wall of water rushed down the valley that was reported to be 22.9m high.
After the floodwaters had receded, US Congress passed the Teton Disaster Relief Bill providing US$400M for the flood recovery program. Five people were drowned with another nine deaths associated with the flood. Large areas of cropland were destroyed along with livestock and buildings downstream. Overall, construction costs for the dam were US$55M. A pale number to the nearly US$1B in damages reported.
Soil piping is believed to be the most likely mode of failure. An explanation of the mechanics of soil piping is included below, along with a description of pre-existing conditions, construction, and other possible modes of failure.
Teton dam was composed of five distinct zones. The core of the dam, zone one, was composed of locally borrowed wind blown silt. Being impermeable, this zone is the most critical zone of earthen dams. The existing zones can all be thought of as filters for the zone inside it, essentially protecting the inner zone from erosion. Each zone is comprised of coarser material than the previous zone. The primary focus will be on zone one.
During the construction of zone one, several deficiencies were noted by construction inspection personnel for USBR. These include:
• Inadequate preparation of foundation rock prior to placement of zone one fill.
• Improper mixing of zone one fill.
• Severely inadequate testing of zone one fill.
• Difficulty obtaining optimum water content.
During the preparation of foundation rock, several issues regarding the general topography of the key trench walls were also not addressed properly. Loose rocks, overhangs, abrupt changes in slope and open jointing created a ragged condition. There were also many open joints left ungrouted in the foundation rock, some exceeding six inches in width. This did not comply with the USBR's specifications, which stated that 'loose rock shall be removed from foundation contacts and rock cliffs, ledges, overhangs, and sharp irregularities shall be reduced to provide satisfactory foundation contours.'
Foundation rock in this condition does not permit proper compaction and bonding of zone one material with the key trench walls. Ungrouted cracks create a conduit for water through the grout curtain, allowing it to enter zone one material from upstream.
Zone one material was placed in early spring. Daily snow removal was a common activity. For a two-week period in March 1975, snow removal was the only task. This snowmelt increased the water content of the material well past optimum. To remedy the situation, the contractor placed a thin layer of dry soil on top and attempted to mix it with the wetter soil below. Adequate mixing was not achieved, resulting in distinct seams with differing water contents.
USBR specifications recommended that earthwork control tests are performed every 1529.2m3 of zone one fill placed. Due to inadequate staffing, the first 5352.2m3 had no control tests performed. Testing was performed on the remaining fill but much of it was severely inadequate. Without control testing it is impossible to determine the properties of the fill material.
Moisture control for zone one placement was unusually difficult due to the nature of the material as well as improper irrigation at the borrow source. For the first three weeks of placement, the borrow area was not irrigated. This resulted in fill material that was considerably dry of optimum. There were six days during this period in which water content was 4% or more dry of optimum. Water contents in this range have an unacceptably increased permeability. This also contributes to improper compaction bonding between lifts.
These four construction deficiencies contributed to the failure of the dam. However, construction deficiencies are expected in any project. These must be protected against by an overly defensive design. Unfortunately, this design was not defensive enough.
Possible modes of failure
The failure most likely began developing well before it was visible. Once visible, failure had progressed too far for any possibility of repair. There are several possible modes and contributing factors to the failure. These include:
• Piping due to cracking of zone one material. Post failure test data from left embankment soil indicated that differential settlement could have occurred. This could have been caused by any or all of the previously mentioned construction deficiencies. Hydraulic fracture due to arching could have occurred in the deep key trenches due to low lateral stress. However, the US Department of the Interior investigation concluded that neither of these modes of failure was likely.
• Flow through grout curtain. Due to the fractured nature of the foundation rock, it is impractical to expect that the grout curtain was tight. This could have allowed water flow into zone one core, again, resulting in internal erosion.
• Piping. Due to the fracture rock base, water could have entered zone one through ungrouted fissures. Due to a lower hydraulic pressure on the downstream side of the dam, water would naturally move in that direction. This could lead to piping. Due to construction deficiencies above, all three of these piping scenarios were possible.
While these modes of failure are possible, all physical evidence was washed away on 5 June 1976. The exact mode of failure will never be known. However, inadequate protection of the zone one core from internal erosion remains the undeniable cause of failure.
Today, very little is left of the US$55M project that was the Teton dam. Because of the Teton dam failure, major overhauls were made to USBR procedures for design and construction of dams. Dams were designed with a higher factor of safety.
More rigorous soil testing was performed during both design and construction. Inspection was also significantly increased.
Possibly even more significant than changes that were made to the Bureau's dam policy, were emergency response changes. US President Carter earmarked more money in the 1977 federal budget for the National Dam Inspection Act of 1972. This allocation provided for complete inspections of high hazard dams across the country.
In 1979, Carter formed the Federal Emergency Management Agency (FEMA) by merging several disaster assistance agencies.
Related ArticlesA war against nature