Hydrologic Investigations Atlas HA-613
The Kelly Barnes Dam on Toccoa Creek near Toccoa, Ga., failed at approximately 1:30 a.m., November 6, 1977, after a period of intensive rain. Thirty-nine people were killed and damages were estimated at $2.8 million by the Federal Disaster Assistance Administration and the U.S. Department of Housing and Urban Development (Roland Serabia and Carl Badger, oral commun., 1978).
Immediately after the flood, hydrologists of the U.S. Geological Survey obtained hydrologic information to document the disaster and to test the Survey's dam-break model procedures.
President Jimmy Carter, at the request of Governor George Busbee of Georgia, authorized the U.S. Army Corps of Engineers to make a technical assessment of the Kelly Barnes Dam failure with the assistance of other Federal agencies having appropriate expertise. The Corps of Engineers, the U.S. Geological Survey, the Soil Conservation Service, and the National Weather Service formed a technical Federal Investigative Board that conducted the technical assessment and published its evaluation as, Report of Failure of Kelly Barnes Dam, Toccoa, Georgia, in December 1977. The history of the dam and meteorological and hydrologic findings of the Board are repeated herein to complement the additional detailed hydrologic and hydraulic data presented in this atlas.
Toccoa Creek Basin: Toccoa Creek lies immediately north of Toccoa, Ga., and flows in an easterly direction to Lake Hartwell (fig. 1). The drainage basin above the damsite has an area of 4.6 square miles and it is heavily wooded with pines and deciduous trees. The Toccoa Creek basin is in the Piedmont physiographic province and soils are of the Pacolet-Wedowee-Chandler association, clay loam to loam subsoil. Stream channels are fairly steep and have slopes that generally exceed 100 feet per mile.
Between the dam and Toccoa Falls, the average low-water channel width of Toccoa Creek is 40 feet. Below the falls the low-water channel is about 50 feet wide and the flood plain ranges from 100 to 400 feet in width. The Toccoa Falls College trailer village was located on the flood plain where the width is approximately 350 feet. The flood plain expands downstream from the trailer village, but becomes restricted to a width of 200 feet upstream from Georgia Highway 17.
Kelly Barnes Dam and Lake: Kelly Barnes Dam was about 400 feet long, 20 feet wide at the crest, and 40 feet high at the maximum section. The dam was concave upstream. Figure 2 shows the lakebed and the broken dam.
The history of the dam as determined by the Federal Investigative Board (1977) is summarized as follows: The dam went through various stages of development, first as a rock crib dam and then with subsequent stages as an earth dam. The rock crib dam was completed about 1899 to impound water for a small hydroelectric plant located near the foot of the falls. About 1937, the Toccoa Falls Bible Institute, which later became Toccoa Falls College, was interested in developing a more dependable power source and decided to build an earth dam over the rock crib dam with equipment provided by a local manufacturer. Figure 3 shows part of the rock crib dam and the penstock pipe to the powerplant that was incorporated in the new dam.
After World War II, the earthfill was raised to a point where an earth spillway on the left side of the valley (facing downstream) could be utilized, and a low point on the rim on the right side away from the dam would serve as a secondary spillway during high flows. This installation served as a power source for the Toccoa Falls Bible Institute until 1957. At that time, power generation was stopped but the lake continued to be used for recreation.
The normal pool elevation for Kelly Barnes Lake was about 1,137 feet above NGVD (National Geodetic Vertical Datum of 1929). The lake had an impoundment of about 18 million cubic feet (410 acre-feet) and a surface area of about 42 acres. The maximum lake level reached before the dam failed was 1,141.6 feet, with an impoundment of about 27 million cubic feet (630 acre-feet). The elevation of the low point on the crest of the dam was approximately 1,147 feet, based on surveys of the remaining embankment after the failure.
The surface-area elevation and volume-elevation curves for Kelly Barnes Lake are shown in figure 4. Bench marks used for topographic surveys, high-water profiles, and cross sections are listed in table 1 and shown in figures 9 and 10.
Topographic maps of the lakebed and the dam were prepared immediately after the break by the U.S. Geological Survey (figs. 5 and 6). The earth spillway at the left end of the dam (facing downstream) apparently carried all normal lake overflow. The crest of the spillway was 1,136.7 feet above NGVD (596.2 feet above datum). A low area, minimum elevation 1,139.8 feet (599.3 feet above datum), in the approach road about 1,100 feet upstream on the right rim, served as a secondary spillway.
The Federal Investigative Board (1977) concluded that separate inlets for a low-water spillway and a welded steel penstock to the powerplant existed at the time of the dam failure (fig. 7). These facilities were described by the Board as follows: "The low-level spillway had a rectangular shaped masonry inlet which could be shut off with stop logs or flashboards. The inlet for the welded steel penstock was also masonry and controlled with a slide gate at the upstream toe of the dam. . . . Some historical witnesses remember the penstock slide gate but do not recall the masonry structure. After the failure, this gate was retrieved from below the damsite and found to be closed. The conduits for the spillway outlet and new penstock were both welded plate pipes approximately 30 inches in diameter. A historical witness stated that during placement of fill over the spillway conduit, the pipe began to cave in and had to be reinforced with metal struts. . . . At this time, or during previous construction, the old spiral riveted penstock pipe had been abandoned, but not removed. Conflicting information suggests the possibility of a series of construction events during the 1940's. A 1954 Survey of Buildings, Roads, Streams, and Lakes of the Toccoa Falls Institute locates the overflow pipe (low-level spillway), 24-inch conduit (welded steel penstock), and the intake structures.
"Information available indicates that the fill was constructed to its final height in the late forties. However, a study of stereopairs of aerial photographs taken on January 17, 1955, shows the reservoir to be essentially empty at that time. The inlet structure, two earth spillways:, and the pipe leading to the powerhouse are clearly visible. In later years a heavy growth of vegetation became established on both the upstream and downstream faces and apparently obliterated the masonry intakes from view. Again, information received from historical witnesses conflicts as to whether or not two structures existed at the time of failure. Judging from the debris found downstream and the remains of the welded steel penstock, it is the Board's opinion that both existed at the time of failure."
The Federal Investigative Board (1977) reported:
"A number of observers informed the Board of seeing almost continual seepage on the downstream slope of the dam near the point of exit of the spillway pipe. Photographs obtained from one source, taken in 1973, show that a large embankment slide had occurred on the downstream face of the dam. . . . This 12-foot high, 30-foot wide slide of unknown depth, was observed on the lower one-third of the downstream slope in the area of the current failure section, which was the highest point of the dam. The slide left a two-foot, vertical, scarp face and partially exposed the end of a pipe. The slide, at the time of the photograph, was apparently not recent because of the existence of established vegetation on the slide area. The picture shows the pipe to be essentially clogged with silt and trash, but a very slight seepage was observed coming from the pipe. The area near the pipe contained water that was discolored by iron oxide, sulfur, or some other matter indicating little or no flow. The area adjacent to the recent dam failure was wet and spongy a week after the failure and seepage water was still coming from the toe near the right abutment (looking downstream) at that time."
The Board also stated,
"The rock type at the site is a biotite gneiss of the Carolina gneiss series. . . . A layer of soft alluvium, from two to five feet thick [overlying solid bedrock] was observed in both faces of the breach of the dam. This material extends from the proximity of the old crib dam upstream - an undetermined distance. The layer appears to be quite extensive in the foundation of the earth dam. The lower six inches of the layer of the alluvium contains an extensive root mat. The alluvium appears to be composed principally of soft, plastic clay and (or) silt with some layers of fine silty sand."
Meteorological conditions. - The NWS (National Weather Service) described meteorological conditions in the report by the Federal Investigative Board(1977) as follows: "Before rain began the ground was already wet from heavy rainfall of 3 1/2 to 4 1/2 inches, which fell on October 25 - 26. The rainfall began on Wednesday morning, November 2, and ended by midnight on Saturday, November 5. A strong high pressure area, centered over New England, was bringing Atlantic moisture into Georgia on Wednesday and this produced the rainfall at the beginning of the storm period. Meanwhile, a more important development was taking place in Texas where a strong low pressure area was developing at upper levels in the atmosphere. This intense upper level low was located near Shreveport, Louisiana, on Thursday morning, November 3; New Orleans, Louisiana, on Friday morning, November 4, Mobile, Alabama, on Saturday morning, November 5; Centreville, Alabama, on Sunday, November 6. As this slow-moving upper level storm moved closer to Georgia, more and more moisture was brought in on southerly winds from the Gulf of Mexico. Lifting of the air by the higher terrain of north Georgia accentuated the lifting processes provided by the atmospheric storm patterns. A plentiful supply of moisture, lifting of the air, and slow-moving storm system are the ingredients of a long, heavy rain event.
"Rainfall reports are received by the National Weather Service from radio stations WLET and WNEG in Toccoa. Since these stations are located some distance from the drainage area above Kelly Barnes Dam, a rainfall survey was made. Three additional storm total values were obtained . . .. [The location and storm totals of hourly and daily rainfall stations from the Federal Investigative Board (1977) and National Oceanic and Atmospheric Administration (1977) for the period November 2 - 6 are shown in fig. 1.] These measurements, interviews with local residents, radar echoes, and hourly values from nearby recording stations [at Burton Dam, Ga., and Long Creek, S.C.] were used in estimating the rainfall distribution for the drainage basin above Kelly Barnes Dam. [See fig. 8.] Although radar echoes indicated that the heaviest rainfall probably occurred between 6:30 pm. and 7:30 pm. on November 5, the estimated maximum 1-hour value is shown between 6:00 pm. and 7:00 pm. since points are plotted on the hour.
"The rainfall was light for the first two days with slightly over 1 inch falling in the Toccoa area by Friday morning, November 4. Rainfall intensity increased somewhat for the next 24 hours, with a storm total of approximately 2 1/2 inches by 8:00 am., Saturday morning, November 5. The storm total increased to approximately 3 1/2 inches by noon on Saturday and then the rainfall was apparently light until about 6:00 p.m., when showers moved into the area. From 6:30 pm. to 7:30 p.m., radar echoes [from Athens, Gal] indicated periods of torrential rainfall with frequent intense lightning and thunder. A small tornado apparently touched down about 5 miles southeast of Kelly Barnes Dam at approximately 8:00 pm. The damage to trees was limited to only 100 yards in width and about one-fourth mile in length, but the tornado was indicative of the severity of the weather. After 8:00 p.m. the showers were less frequent and had practically ceased by midnight. Estimated basin rainfall for the entire period is about 7 inches, with almost 3 1/2 inches occurring between 6:00 pm. and midnight on November 5."
The Dam-Break Flood
Between Kelly Barnes Dam and Highview Road the flooded area was defined by a survey of high-water marks (figs. 9 and 10. Note: Explanation for fig. 9 map is shown on fig. 10.). Seventy-two cross sections were obtained. A detailed analysis of peak discharges, flood frequency, and flood profiles is presented below.
Peak discharges and flood frequency. - The peak discharges summarized in table 2 were computed on basis of cross sections, water-surface profiles, roughness coefficients, and, at some places, bridge geometry. Peak discharge measurement sites A-G are shown in figure 10. Site H is 1.6 mi downstream from site G and not shown in figure 10.
1 Peak inflow to Kelly Barnes Lake estimated from unit-hydrograph computations.
The computed peak discharge of Toccoa Creek at site A near the head of Kelly Barnes Lake, for the flood of November 5 - 6, was 830 ft3/s (cubic feet per second). The peak inflow of 980 ft3/s to the lake was estimated at site B by hydrograph synthesis with a rainfall-runoff model. Based on a regional flood-frequency analysis (Golden and Price, 1976) for small drainage basins, the estimated recurrence interval of this flood is about 10 years. The recurrence interval is the average time interval between actual occurrences of peak flows of a given or greater magnitude. Although the recurrence interval represents the long-term average period between floods of a specific magnitude, such floods could occur at short intervals or within the same year.
At the site of the discontinued U.S. Geological Survey gaging station on Panther Creek (drainage area, 32.5 square miles), 5 1/2 miles north of Toccoa (fig. 1), peak discharge for November 5 - 6 was 6,600 ft3/s. Based on a frequency analysis of 48 years of station record, this flood had a recurrence interval of approximately 15 years. The Panther Creek basin adjoins the Toccoa Creek basin.
The peak outflow of 400 ft3/s from Kelly Barnes Lake before dam failure was estimated by means of slope-area and flow-over-road computations at sites C and C1, respectively. Flow occurred in both the primary earth spillway (site C) near the left end of the dam and the secondary spillway (site C).
Although the dam-break flood undoubtedly moved downstream as a flood wave, it produced a pool- and-riffle type of flow as shown by the water surface profiles in figure 11.
Figure 11. - Profile of flood on Toccoa Creek, Kelly Barnes Dam to Highview Road
Peak discharges were computed at sites A, D, E, and F using the standard slope-area computation procedure described by Dalrymple and Benson (1967). The procedure assumes steady flow conditions that obviously did not exist immediately downstream from the broken dam. The effects of unsteady flow were minimized by using short reaches for slope-area computations. Immediately downstream from sites D, E, and F, high-water profiles indicate that the flow changed from tranquil to rapid state. At these sites, peak discharges computed by the critical-depth method agree closely with those computed by the slope-area method.
Discharges were computed at Highview Road and Georgia Highway 184 (sites G and H) using a standard contracted-opening procedure as described by Matthai (1967).
Conditions for making discharge computations were considered good at sites A and H; fair at sites B, F, and G; and poor at sites C, D, and E.
Graphs showing the relation of discharge to drainage area for the observed flood, together with the 10-, 50-, and 100-year regional flood frequency curves (Golden and Price, 1976), are shown in figure 12. No attenuation of the peak discharge was evident between the dam and Toccoa Falls College (sites D to E). Extreme attenuation occurred between Toccoa Falls College and Highview Road (sites E to G) due to increased overbank storage and rapid changes in stage. Flow at sites F and G was 60 percent and 27 percent, respectively, of the peak discharge at site E.
Because this disaster occurred at night, eyewitness accounts of the timing of the flood wave were not consistent. According to one witness, the creek was out of its bank for about an hour at Forrest Hall Dormitory, at the upstream end of the college campus. Other witnesses indicated that the flood waters were out of bank for only about 40 minutes. An estimated discharge hydrograph at Forrest Hall Dormitory, 0.7 mile downstream from the dam, based on these reports, is shown in figure 13.
Flood profile.- Water-surface profiles, determined from a field survey of high-water marks left by the flood of November 6, are shown in figure 11. Average depths in the main channel above Toccoa Falls were about 17 feet. Depths in the vicinity of the college ranged from about 21 feet at Forrest Hall Dormitory to about 18 feet at the trailer village. Depths downstream from Georgia Highway 17 averaged about 15 feet.
Controls for pool-and-riffle flow are located near stations 517, 486, 480, and 465 and at Georgia Highway 17. These controls are at obvious valley constrictions, but some of the resulting backwater could have been caused by temporarily lodged debris. Extreme scour was evident at these controls.
Water-surface elevations on the left bank differed from those on the right bank by as much as 10 feet in the curves along the main channel because of superelevation resulting from high velocities. However, even in these high-velocity areas, ponded conditions existed in nooks near the mouths of tributaries.
The Flood Without Dam Break
Streamflow estimated by rainfall-runoff model. - Computations were made using a simplified version of the U.S. Geological Survey rainfall-runoff model (Dawdy and others, 1972) to develop a hydrograph of inflow to Kelly Barnes Lake for November 5 - 6 (fig. 14).
The model was calibrated by reproducing the observed discharge at site A of 830 ft3/s from Toccoa Creek into the lake. The model, with streamflow and reservoir routing techniques, was used to estimate discharges (fig. 12) in the reach between Toccoa Falls and Georgia Highway 17 that would have occurred if the dam had not broken, and also if there had been no dam.
Profile estimates without the dam break. - Computed water-surface profiles for assumed conditions without the dam break, both with and without effects of lake storage, are shown in figure 11 above. The two profiles were computed by means of the U.S. Geological Survey step-backwater routing program (Shearman, 1976) using the discharges obtained from the rainfall-runoff model. Cross sections used in the analysis were surveyed after the flood by the U.S. Geological Survey and the U.S. Army Corps of Engineers. The stream channel at these cross sections was generally scoured by the flood. Thus, the computed profiles represent only the minimum water-surface elevations that would have occurred had the dam not broken, and are not representative of conditions prior to the dam break.
A qualitative reconnaissance of sediment sources and sediment deposition was conducted on November 10 and 11. The operation of earthmoving equipment downstream from Toccoa Falls, however, prevented a qualitative evaluation of sedimentation conditions and limited this investigation to the reservoir and the stream channel upstream from Toccoa Falls. The following is a description of observations made during this reconnaissance.
The breached section of the dam was a major source of gravel (2.0 - 64 mm), sand (0.062 - 2.0 mm), and finer material. Field measurements indicate that about 13,000 cubic yards of material was removed from the dam. Other sources of sediment in this size range include soil eroded from the valley walls and reservoir deposits that were eroded during stream incision. Most of the sand and finer material was transported over Toccoa Falls. One large deposit of sand, however, occurred on the inside of a broad bend about 400 feet upstream from Toccoa Falls. The areal extent of this deposit was not measured because of debris, however, its thickness is evident in figure 15. The relative amounts of gravel transported through or deposited in the reach cannot be reliably estimated. Deposits of gravel-size material are present throughout the reach, generally intermixed with larger material or in association with debris piles.
Material larger than gravel was derived from bedrock exposed on the valley walls and from material stored on the flood plain. In the relatively narrow valley downstream from the dam, evidence indicates that the water had sufficient power to pluck large pieces of bedrock off the walls of the valley. Well-developed jointing in the bedrock seemed to define the size and shape of these large rectangular blocks. Most of these bedrock blocks were deposited about 500 feet downstream from the dam, where the valley widens. This deposit is the first area of large-scale sediment deposition on the flood plain below the reservoir. The deposit formed a partial dam across the stream channel.
Figure 16 shows the flood's capability for transporting large boulders, which probably had been stored on the streambed. The photograph, taken about 0.3 mile downstream from the breached dam, shows a large boulder that was apparently in suspension when it hit the tree.
Sediment transported by the flood can be divided by particle size, source area, and mode of transport. Sand and finer material derived primarily from the breached dam, dissected reservoir deposits, and eroded soil cover were, for the most part, transported through the reach and over Toccoa Falls. These fine materials fonr.ed most of the wash load of the flood wave. Cobbles (64 - 256 mm) and boulders (larger than 256 mm) derived from the streambed and surrounding bedrock were moved only short distances and thus constitute the bed-material load of the stream. Gravel-size material that was derived from both the breached dam and the streambed forms a transition particle size between wash loads and bed-material load.
Approximately nine houses, 18 house trailers, two college buildings, and many motor vehicles were completely demolished. Four houses and five college buildings were damaged by water. Only two houses downstream from Georgia Highway 17 were damaged.
The embankment at Toccoa Falls Drive, the oxidation pond above Georgia Highway 17, and parts of the main channel were scoured. Two bridges on Toccoa Falls Drive and the culvert at County Farm Road were completely destroyed. The highway embankments at Georgia Highway 17 were washed out at both ends of the bridge, and one of the bridge abutments at Highview Road was destroyed. The water-supply pipe for the city of Toccoa was broken and the city's water supply was contaminated for several days.
Dalrymple, Tate, and Benson, M. A., 1967, Measurement of peak discharge by the slope-area method: U.S. Geological Survey Techniques of Water-Resources Investigations, Book 3, Chapter A2,12 p.
Dawdy, D. R., Lichty, R. W., and Bergmann, J. M., 1972, A rainfall-runoff simulation model for estimation of flood peaks for small drainagebasins: U.S. Geological Survey Professional Paper 506-B, 28 p.
Federal Investigative Board, 1977, Report of failure of Kelly Barnes Dam, Toccoa, Georgia: U.S. Army Corps of Engineers, National Weather Service, Soil Conservation Service, and U.S. Geological Survey, 37 p.
Golden, H. G., and Price, McGlone, 1976, Flood-frequency analysis for small natural streams in Georgia: U.S. Geological Survey Open-File Report 76-511, 75 p.
Matthai, Howard F., 1967, Measurement of peak discharge at width contractions by indirect methods: U.S. Geological Survey Techniques of Water-Resources Investigations, Book 3, Chapter A4,44 p.
National Oceanic and Atmospheric Administration, 1977, Climatological Data, 1977, v. 81, no.11,16 p.
Shearman, J. O., 1976, Computer applications for step-backwater and floodway analysis: U.S. Geological Survey Open-File Report 76-499, 103 p.
Toccoa Falls Institute, 1954, 1954 Survey of buildings, roads, streams, and lakes: Toccoa, Georgia, Toccoa Falls Institute unpublished report.
[U.S.] National Bureau of Standards, 1977 ed., The International System of Units (SI): National Bureau of Standards Special Publication 330, 41 p
First posted May 28, 2010
This publication is in PDF format.