Welcome to our new Geotechnical History Blog! This inaugural post, by lead blogger and G-I member Michael Bennett, is presented in five parts. All notes and image credits are in the fifth part.
The South Fork Dam Failure and Johnstown Flood of 1889:
A Civil and Geotechnical Engineering Perspective
by Michael D. Bennett, E.I.T., A.M.ASCE
Dedicated to the memory of the victims of the Johnstown Flood of 1889.
Introduction: Overview of the Tragedy
On May 31st, 1889, the deadliest dam disaster in US history played out in western Pennsylvania. A torrential storm the previous night had drenched the area around the massive Lake Conemaugh and the dilapidated South Fork Dam holding it back. The rain-swollen lake rose ominously throughout the 31st until it eventually overtopped and breached the dam. The two-phase breach sent 16 million tons of water and an avalanche of debris thundering down the valley of the Little Conemaugh River. The wall of water and debris wiped out miles of railroad tracks, small hamlets, and factories. Eventually, it slammed into the prosperous steel hub of Johnstown and killed thousands, injured thousands more, and obliterated the town. The disaster was soon named the Johnstown Flood. Outraged survivors demanded that justice be served to the Pittsburgh tycoons who had owned the lake and the dam. However, the moguls and their powerful allies unduly influenced how politicians, courts, and even the American Society of Civil Engineers handled the aftermath of the disaster.
Traversing the Appalachian Mountains
The events which would culminate in the destruction of Johnstown in 1889 had begun roughly a century earlier. Following US independence, the newly-formed federal government tragically evicted the Native American tribes within the Appalachian Mountains of Pennsylvania. European-American settlers then began pushing westward toward the frontier. Some stopped in the eastern or central Appalachians. Others traveled over the highest and westernmost portion of the range within the Keystone State, the Allegheny Mountains. In the 1790s, settler Joseph Schantz and his family founded a village on the western slopes of the Alleghenies where the Little Conemaugh and Stony Creek (or Stonycreek) Rivers converged into the Conemaugh River. The site consisted, then as now, of a small, flat plain surrounded by near-vertical hills several hundred feet high. Schantz left about a decade later, but the village thereafter bore the anglicized form of his name – Johnstown (Hanna 2021).
The US frontier kept advancing westward during the early 19th century. The increasing personal and commercial traffic through the Allegheny Mountains eventually overloaded the established wagon roads within the range.
Engineers, entrepreneurs, and politicians soon began exploring alternative solutions for transportation through or around the Alleghenies. New York State completed the Erie Canal in 1825 and connected the Hudson River in Albany to the Great Lakes in Buffalo.
Simultaneously, construction began on the Chesapeake and Ohio Canal along the Potomac River in Maryland. The new transportation networks rapidly began raking in toll revenue for their respective states, and the Pennsylvania General Assembly wanted its share. In 1828, the Assembly authorized the Main Line of Public Works, a network of canals and the novel technology of railroads which would connect Philadelphia to Pittsburgh. By 1834, the Commonwealth had finished the Main Line (Burgess and Kennedy 1949, Hanna 2021).
The Main Line of Public Works
Travelers embarking on the Main Line in Philadelphia first took a conventional railroad to Columbia on the Susquehanna River, where they transferred to canal boats. From Columbia, the passengers took canals along the Susquehanna and Juniata Rivers and through the eastern Appalachian Mountains to Hollidaysburg on the eastern slopes of the Allegheny Mountains. Within Pennsylvania, the Alleghenies form the Eastern Continental Divide, separating rivers flowing toward the Eastern Seaboard from those flowing toward the Gulf of Mexico. Therefore, in Hollidaysburg, the travelers transferred again, this time to the Allegheny Portage Railroad (Burgess and Kennedy 1949).
The Portage Railroad climbed the Allegheny Mountains using a combination of flatter segments of conventional railroads joined by steeper inclined planes. The planes were surmounted using stationary steam engines, which drove gargantuan pulleys connected to cables. These cables hauled railcars carrying passengers, freight, and even whole canal boats up and down the 7 to 10 percent grades of the planes. The inclined planes incorporated safety features such as emergency stoppers and cogged tracks. Finally, upon reaching Johnstown, travelers transferred back to canals along the Conemaugh, Kiskiminetas, and Allegheny Rivers for the final journey into Pittsburgh. Overall, those taking the Main Line covered the nearly 400-mile trip in about 4 days (Burgess and Kennedy 1949).
Improving the Main Line: The Western Reservoir and Dam
The Pennsylvania General Assembly needed to keep traffic moving along the Main Line to keep profits coming into the state’s coffers. However, the civil engineers of the Commonwealth soon noticed that the Main Line canals, like the adjacent rivers, often froze over during Pennsylvania’s frigid winters and ran low during the state’s hot, dry summers. The latter issue was especially pressing near Hollidaysburg and Johnstown, since the watersheds upstream of both towns were relatively small (Burgess and Kennedy 1949, Hanna 2021).
The civil engineers soon proposed a solution which would both keep the Main Line open and keep toll revenue flowing into the Pennsylvania treasury. They suggested impounding reservoirs on both sides of the Alleghenies to supplement the canals. In the mid-1830s, the state sent civil engineer Sylvester Welch to survey potential sites for both reservoirs. Based on his recommendations, the state located the aptly-named Western Reservoir for the Conemaugh canals in Cambria County along the South Fork of the Little Conemaugh River. The site lay about 14 miles upstream, or 8 miles due east-northeast, of Johnstown. Welch recognized that the drop of about 450 feet from the surface of the planned reservoir to Johnstown would give the waters plenty of elevation head for their fairly long journey to the Conemaugh (Coleman 2018, Francis et al. 1891, Hanna 2021).
Welch also developed preliminary designs for the dam at the Western Reservoir. He bluntly stated that water could not be allowed to overtop an earthen dam if one were built. Clearly, the standard of care, or best practice, for the design of earth dams in the mid-1830s included the prevention of overtopping. Welch elaborated that the leakage of water through an earthen dam could be prevented either by building a masonry core within the dam or by constructing the upstream half of the dam using puddled clay. During the puddling process, wet clay is placed in thin layers, making it nearly watertight (Unrau 1979).
Next, the Commonwealth tasked civil engineer William Morris with finalizing the design of the dam for the Western Reservoir. First, Morris had laborers dig shafts and tunnels at the proposed site of the dam to confirm both the presence of bedrock and the availability of good-quality material for the dam. Such a practice may well be considered a forerunner of the 21st-century geotechnical investigation. Present geological references explain that the site of the dam is underlain by the Casselman Formation, which contains thin, interbedded layers of sedimentary rock such as limestone, sandstone, shale, siltstone, and claystone (PA DCNR 2023, Unrau 1979).
Morris then used his crew’s findings to convert Welch’s preliminary design for the dam at the Western Reservoir into a final, constructible design. Per Welch’s recommendations, Morris designed an earth dam with an upstream half built of puddled clay for the site. The dam would be 72 feet high and about 860 feet long, and would taper from being 225 feet wide at its base to being 10 feet wide at its crest. Workers would excavate a trench to bedrock beneath the future dam and would fill the trench with puddled clay to make the dam watertight. This feature, named a puddle trench, would serve the same function as a cutoff trench below a 21st-century dam (Kaktins et al. 2013, Unrau 1979).
Morris also included a drainage outlet at the base in his design for the dam to allow the operators of the future Western Reservoir to use its full depth to supply the Main Line canals west of Johnstown. The outlet would consist of five cast-iron discharge pipes in the center of the future dam emptying into a brick discharge culvert. Each pipe would be 24 inches in inner diameter and about 80 feet long, and would be operated from a masonry control tower in the reservoir. The outlet would also be used to safely lower the reservoir before storms or whenever the water rose too high. Finally, Morris realized that the floods which sometimes inundated the South Fork of the Little Conemaugh River presented “little danger” to the dam if “proper channels [were] constructed for their discharge”. Accordingly, he included in his design spillways at each end of the dam (Kaktins et al. 2013, Unrau 1979).
Contractors’ work crews supervised by field engineers started constructing the Western Reservoir and its dam in the spring of 1840. The laborers first cleared and grubbed the future dam site and reservoir bed, then dug and backfilled the puddle trench. Next, the crews built the masonry foundation for the discharge pipes and began constructing the brick discharge culvert. As work progressed, though, the US economy entered a recession. Times quickly grew particularly tough in Pennsylvania, since the General Assembly was still paying off the immense cost of building the Main Line. The finances of the Keystone State had gotten so precarious by the spring of 1841 that the Commonwealth paused construction of many infrastructure projects, including the Western Reservoir, to help keep the state solvent (Coleman 2018, Francis et al. 1891, Hanna 2021, Unrau 1979).
The State of Pennsylvania initially intended the pause in construction at the Western Reservoir to be temporary. Ultimately, though, it lasted a decade. The Commonwealth’s finances stayed weak throughout the 1840s, and the General Assembly used whatever limited money did become available to repair other infrastructure. A cholera epidemic also delayed work for several years. Political and legal wrangling over how best to finance and finish the reservoir only extended the delay in construction. Meanwhile, the Main Line continued to struggle with winter freeze-overs and low summer levels on its canals, which exacerbated the Keystone State’s monetary woes. The incomplete dam at the Western Reservoir site languished in the elements and sustained several partial washouts, which likely compromised its long-term integrity (Coleman 2018, Francis et al. 1891, Hanna 2021, Unrau 1979).
Eventually, though, the Commonwealth regained financial stability. Early in 1851, the General Assembly allotted funding to complete the Western Reservoir. The original contractors resumed construction almost immediately under the supervision of field engineers still directed by William Morris. Morris, assuredly eager to use the new funds wisely, value engineered his design by changing the planned control tower from masonry to wood. The work crews quickly finished both the tower and the discharge culvert. The laborers next built the upstream half of the dam using puddled clay while constructing its downstream half using weathered rock. The crews covered both faces with riprap to further guard against erosion, using larger riprap on the downstream face. The specifications governing their workmanship on the dam were strict. Morris prohibited the use of “light, spongy, alluvial, or vegetable material” within the dam and required that poorly-done earthwork be removed and correctly repaired (Coleman 2018, Francis et al. 1891, Hanna 2021, Unrau 1979).
Over the next year, the dam steadily rose from the future bed of the Western Reservoir as construction progressed. The crew completed the dam by the summer of 1852, at which point impounding of the reservoir began, and finished supporting construction the next year. Initially, the reservoir was filled to a depth of 50 feet, well below its maximum safe depth of 60 feet. This was done to allow the engineers and contractors to address potential problems with the dam. When none arose, the reservoir was finally filled to a depth of 60 feet. The practice of waiting to completely fill a new reservoir remains standard for new dams nearly 175 years later (Coleman 2018, Hanna 2021).
The new dam at the Western Reservoir was among the largest in the US in the early 1850s. The spillways for helping to drain the reservoir during floods were equally imposing. The main spillway at the dam’s northeast abutment was about 10 feet deep. It had been hewn through sandstone bedrock, and its minimum width, which by definition controlled its peak discharge, was roughly 70 feet. The auxiliary spillway at the dam’s southwest abutment was about 3 feet deep and had a consistent width of roughly 70 feet. It had not been excavated to bedrock, perhaps due to budget constraints (Coleman 2018, Kaktins et al. 2013, Unrau 1979).
Researchers from the University of Pittsburgh at Johnstown (UPJ) performed hydraulic analyses of the original Western Reservoir and dam in the 2010s using contemporary software. They began by taking LiDAR scans of, and soil samples from, the former reservoir bed. The team then constructed a storage-elevation curve for the site. This curve, which accounted for soil formation since 1889, allowed the researchers to estimate the capacity of the reservoir. The researchers found that the original reservoir could hold about 4.3 billion gallons of water occupying roughly 574 million cubic feet in volume and weighing nearly 17.9 million tons. The team also estimated the peak discharges in cubic feet per second (CFS) through the main spillway (5,350 CFS), the auxiliary spillway (900 CFS), and the discharge pipes and culvert (700 CFS). The researchers determined that, overall, the reservoir had a discharge capacity of approximately 6,950 CFS when filled to capacity (Coleman 2018).
Morris never marked the auxiliary spillway on his plans for the dam at the Western Reservoir, and some scholars dispute whether he truly supported its construction. Surveyors, however, later confirmed the presence of a depression at the southwestern abutment. This, regardless of Morris’s intentions, would have functioned as an auxiliary spillway during floods. Moreover, the onsite engineers could surely have directed the contractors to fill the depression, and thus increase the capacity of the reservoir, had Morris so desired. Regardless, the auxiliary spillway never negatively impacted the performance of the dam during the reservoir’s service life. Maintenance workers noted only minor leaks, which they readily repaired, in the dam during its early years. The State Engineer inspected the reservoir and dam in 1856 and judged both to be in “excellent condition” and capable of serving the Main Line for decades to come (Coleman 2018, Francis et al. 1891).
The opinions and conclusions expressed in this article are those of the author only and do not necessarily represent the views of ASCE or the Geo-Institute.