Laserfiche WebLink
<br />Earthfi 11 in the dam was suffi ci ent ly fi ne-grai ned to be fro5t- <br />suscept i b 1 e. Soi l-property changes induced by numerous freeze-thaw cyc 1 es <br />could have been a possible cause of failure. However, no ice lenses nor <br />frozen ground were found during investigations at the damsite in July 1983; <br />any changed soil characteristics would be reflected in the geotechnical <br />investigations of the fill material. <br /> <br />Results of an undrained embankment-stability analysis by the Office of <br />the State Engineer (1983) indicated a safety factor of only 0.6 existed in the <br />dam for the assumed phreatic conditions. Assumption of an undrained condi- <br />t ion, based on the results of permeabil ity tests, is unreal i st i c. Some type <br />of steady-state drainage must have existed in the dam which kept pore pres- <br />sures below critical values and prevented the dam from failing. <br /> <br />During investigations at the dam5ite, numerous pieces of lead caul king <br />that was used to form the seal between the outlet pipe and the gate valve <br />(fig. 6) were found. This lead caulking was corroded and deteriorated to the <br />point that it would be unlikely to form a water-tight seal. The use of lead <br />caul ki ng to secure the pi pe and valve was not in comp 1 i ance wi th ori gi na 1 <br />plans and specifications approved by the Colorado State Engineer in 1902, <br />which indicated that the valve would be encased in concrete. <br /> <br />Cause of Failure <br /> <br />Data gathered by the Office of the State Engineer (1983) indicated that <br />the most 1 i ke ly cause of the Lawn Lake dam failure was deteri orat i on of the <br />lead caulking used for the connection between the outlet pipe and the gate <br />valve. The resulting leak eroded a pipe-shaped cavity in the earthfill <br />(called piping), creating a void along the outlet pipe. The leak continued to <br />remove easily eroded earthfill and may have reduced the strength and stability <br />of the dam embankment by affecting the pore-pressure distribution. Such a <br />scenario of leakage, piping, and resulting embankment failure required the <br />prerequisite conditions of a closed or nearly closed outlet gate and a near- <br />full reservoir, to supply the head necessary for accelerated leakage and <br />subsequent progressive piping through the years. Just prior to failure, the <br />outlet gate was nearly closed and the reservoir was full. <br /> <br />The void created along the outlet pipe would have become enlarged by the <br />internal erosion of embankment material. Campers in the vicinity of Lawn Lake <br />the night before the failure reported hearing a noise that sounded like strong <br />winds. This sound indicated that the reservoir may have been discharging <br />through the dam for at least 3 to 4 hours before the hole enlarged <br />sufficiently--or embankment failure occurred--above the outlet pipe, causing a <br />total breach of the dam where the outlet works were located (figs. 7A,B). <br /> <br />Surveyed breach dimensions are: Depth of 28 ft, top width of 97 ft, and <br />bottom width of 55 ft. Following the complete failure of the dam embankment <br />at about 0530 MOT, the resulting outflow of water reached its peak discharge <br />very quickly (estimated to be within 10 min). The resulting flood peak of <br />18,000 ft3/S (see "dam-Break Modeling") then proceeded down the Roaring River <br />toward Horseshoe Park and the Cascade Lake dam. <br /> <br />12 <br />