Dam-Breach hydrology of the Johnstown flood of 1889–challenging the findings of the 1891 investigation report

Abstract

In 1891 a report was published by an ASCE committee to investigate the cause of the Johnstown flood of 1889. They concluded that changes made to the dam by the South Fork Fishing and Hunting Club did not cause the disaster because the embankment would have been overflowed and breached if the changes were not made. We dispute that conclusion based on hydraulic analyses of the dam as originally built, estimates of the time of concentration and time to peak for the South Fork drainage basin, and reported conditions at the dam and in the watershed.We present a LiDAR-based volume of Lake Conemaugh at the time of dam failure (1.455 × 107 m3) and hydrographs of flood discharge and lake stage decline. Our analytical approach incorporates the complex shape of this dam breach. More than 65 min would have been needed to drain most of the lake, not the 45 min cited by most sources. Peak flood discharges were likely in the range 7200 to 8970 m3 s−1. The original dam design, with a crest ∼0.9 m higher and the added capacity of an auxiliary spillway and five discharge pipes, had a discharge capacity at overtopping more than twice that of the reconstructed dam. A properly rebuilt dam would not have overtopped and would likely have survived the runoff event, thereby saving thousands of lives.We believe the ASCE report represented state-of-the-art for 1891. However, the report contains discrepancies and lapses in key observations, and relied on excessive reservoir inflow estimates. The confidence they expressed that dam failure was inevitable was inconsistent with information available to the committee. Hydrodynamic erosion was a likely culprit in the 1862 dam failure that seriously damaged the embankment. The Club’s substandard repair of this earlier breach sowed the seeds of its eventual destruction.

The time of concentration application in studies around the world: a review

“…Saha et al (2019) conducted a study to address the issues of the ecosystem, sustainable management and the flow regulation impacts of dams in the Halda River in Bangladesh Coleman et al (2016). reanalyzed the 1889 Johnstown flood with a study based on hydraulic analyses of the dam as originally built, estimates of the time of concentration and time to peak for the South Fork drainage basin, and reported conditions at the dam and in the watershed Kim and Sanders (2016).…”

In the shadow of the dam – Hydrology of the Little Conemaugh river and its South Fork, with insights about past and future flooding

“…The 1977 peak discharge was 24,000 cfs. The 1889 discharge into the former Lake Conemaugh behind the South Fork Dam was estimated at around 7,100 cfs just before the dam overtopped ( Coleman et al., 2016 ). The 12-hour depths were the same as those used for the broader basin.…”

“…At South Fork, no discharge thresholds for flooding like those provided for East Conemaugh are provided because no gage data exist prior to 2017, although an estimate of peak discharge was made for the 1977 flood (see section 4.4 ). An estimate for inflow to the 1889 South Fork Dam prior following the severe storm is also available ( Coleman et al., 2016 ), but that value was likely obtained sometime after the peak inflow to Lake Conemaugh. It has been reported that the precipitation of May 30–31, 1889 likely had a return period of at least 100 years ( Coleman et al., 2016 ).…”

“…An estimate for inflow to the 1889 South Fork Dam prior following the severe storm is also available ( Coleman et al., 2016 ), but that value was likely obtained sometime after the peak inflow to Lake Conemaugh. It has been reported that the precipitation of May 30–31, 1889 likely had a return period of at least 100 years ( Coleman et al., 2016 ). The 100-year design storm yielded a modeled peak discharge of 9,161 cfs (259 m 3 /s).…”

“…Previous work by Coleman et al. (2016) used LiDAR and high-resolution GPS data to analyze the former spillways and basin of Lake Conemaugh, including development of a storage-elevation curve.…”