The 1977 Toccoa Flood
Report of Failure of Kelly Barnes Dam Flood and Findings
by Federal Investigative Board
December 21, 1977
While it is not possible to conclude the specific cause(s) of the
dam failure, the investigations and available evidence support the
development of certain probable failure hypotheses and at the same
time throw out consideration of some possible modes of failure.
Information received and records available indicate that the dam
and/or spillways were modified, repaired, and/or extended over the
years. Some witnesses referred to previous spillway and/or dam
failure. These activities and the causes leading to them could not
be specifically documented. However, documentation shows there was
construction about 1939-1940. Additional construction or repair
occurred about 1948-1949. The reservoir was essentially empty in
January 1955 for some unknown reason. Reports were received that
sandbags were used in the past to prevent overtopping of the crest. The
crest elevation at that time is not known. A responsible source
reported that the main earth spillway on the left abutment was constructed in
1965-1967 and sandbagging was not needed thereafter. This may have
been an enlargement of the spillway since the 1955 aerial photos
show a left abutment spillway. The 1949 and 1955 photos and the
1954 survey plot show two upstream inlet structures. Information
from fishermen and visitors with recent knowledge of the dam and
lake refer to only one structure. The conditions of these
structures at the time of failure is undetermined.
The Board addressed its attention to any possibilities of failure
and circumstances that could have contributed to the failure. These
possibilities are addressed individually below.
Tornado. Severe storms occurred in the area including a small
tornado the night of the failure, but the possibility of a twister
in the area of the dam which could have caused trees to fall and
open the fill to failure appears very remote. Inspections of trees
in the vicinity of the dam indicated no tornado effects.
Animal burrows. No evidence was found in the remaining sections of
the dam to indicate that animal burrows contributed to the failure.
Lightning. Although severe electrical activity apparently occurred
prior to the time of failure, there is no evidence to indicate that
lightning caused any structural damage.
Earthquake. A check with earthquake monitoring sources at the
Georgia Institute of Technology, the University of South Carolina,
and the National Earthquake Center in Denver indicates that the
only disturbance in the Toccoa area was limited to thunderstorms.
Sabotage. There was no evidence of sabotage.
inspections and measurements of the high-water marks in the
reservoir and the crest of the dam indicate that overtopping did
not cause the initial dam failure.
Laboratory tests on foundation and embankment soils show that
these materials are not dispersive clays, and did not cause the failure.
Laboratory tests indicate that the materials, while having a low
density, have sufficient cohesive strength and plasticity so that liquefaction
was not a cause of failure.
The following mechanisms, conditions, failure modes, other than
those above, were examined more extensively because of their more
likely being factors contributing to the failure.
Piping. Piping through the embankment, and the subsequent removal
of material from the embankment in the vicinity of the low-level
spillway pipe, is a possible cause. If not the actual cause, it may
have been a contributory factor. In addition, seepage in the area
of the old root mat may have caused piping that contributed to the
Structural collapse of Intake. Structural collapse of the
low-level spillway intake structure, flashboards, and/or other parts
of the structure would allow water to flow under full head into the
partially collapsed and blocked pipe. This could have caused an
opening of the pipe at a joint or weak spot resulting in further
slope destruction on the downstream face.
Slope failure. This
appears to be a distinct possibility, particularly on the
downstream slope when the previous slope failure is considered
along with the possibility of the development of tension cracks upslope of the
previous failure together with a computed factor of safety that is
marginal. The long period of rain would have saturated tension
cracks, if they existed, and the entire downstream slope would have
become essentially saturated and even more susceptible to failure.
A local downstream slope failure similar to that observed in 1973
could have caused limited breaching allowing localized overtopping.
This concept would corroborate the hydraulic computations.
Seepage. A local slide
would have foreshortened the seepage paths and could have resulted
in rapid erosion, and/or piping,
Local Failure/Pipe Rupture. Since the outlet pipe was
observed in the slide area in the 1973 photo, an additional similar
slide could have further damaged the pipe sufficiently to cause
internal piping and/or erosion
that subsequently resulted in a collapse of the embankment. This
concept would also corroborate the hydraulic computations.
Dam Condition. The dam was in generally poor condition. It
can be concluded that in the vicinity of the breach, the dam was heavily
vegetated and root systems extended deep into the embankment. The embankment
slopes were steep. A pipe from a low-level spillway penetrated the
dam in the downstream direction. In the vicinity of the pipe the
embankment had experienced a local slope failure and seepage for some period of time.
This pipe was reported to have collapsed during construction and to
have been repaired at that time. The welded steel penstock pipe, exposed by
excavation after the dam failure, had previously collapsed, ripped,
been repaired, and apparently partially collapsed again. Other
contributory factors that cannot be disregarded was a design, or
lack of same, that included a rock crib dam within the earth
embankment, together with several feet of lake sediment and old roots, open
ended spiral riveted pipe in the earth embankment, a concentration
of three pipes in a localized area, pipes through the dam that were
not permanently and/or only partially plugged, and a lack of repair
of local slope failures.
The Board could not determine a sole cause of failure. It does
conclude that a combination of factors caused the failure. The
most probable causes are a local slide on the steep downstream
slope, probably associated with piping and/or an attendant
localized breach in the crest followed by progressive erosion,
saturation of the downstream embankment, and subsequently a total
collapse of the structure.
/s/ROBERT L. CRISP, JR., Chairman, U.S. Army Corps of Engineers, Department of the Army
/s/WILLIAM E. FOX, National Weather Service, Department of Commerce
/s/ROBERT C. ROBISON, Soil Conservation Service, Department of Agriculture
/s/VERNON B. SAUER, U.S. Geological Survey, Department of the Interior