Faultory

Case file 38 · History & technology

The dam that could not spill

Johnstown Flood, 1889: why the South Fork Dam failed, and what the record can say about responsibility

Subject
Johnstown Flood
Reading
10 min
Evidence
5 exhibits
Reasoning
3 steps
Level
Intermediate
Black-and-white photograph of the broken South Fork embankment: a wide gap in the earth bank with trees behind, and a flat, dark, sandy floor of the drained lake in the foreground.
1889 The breached South Fork Dam after the Johnstown Flood. Photo: Library of Congress · Public domain · Wikimedia Commons.
Illustrative signature. A lake level that climbs with the rain, then overtops and falls away once the outlet clogs; a stylised water level, not a gauge reading. Drawn to explain the case, not plotted from source data.

01 · The briefing

Begin with the question.

The South Fork Dam was an earth embankment begun in the early 1840s to supply water for a canal in the hills above Johnstown, Pennsylvania. It failed in 1847, and again after heavy rain in 1862, and it was patched rather than fully rebuilt. Later it passed to the South Fork Fishing and Hunting Club, a private group whose members included wealthy industrialists. The club lowered the top of the dam, never replaced the removed discharge pipes, and put fish screens over the spillway. On 31 May 1889, heavy rain filled the lake behind the dam. The screens clogged, so the water could leave only by overtopping the embankment, and the dam breached. The flood reached Johnstown about 14 miles downstream. Three explanations are commonly offered. The first is that the rain was simply too heavy for any dam. The second is that the club's modifications reduced the dam's capacity to pass water. The third is that the structure was already weak from its construction and earlier repairs. Your investigation separates the physical record from the conclusions drawn by the 1891 engineering committee and by later analysis. A finding about cause is not the same as a finding about responsibility, and the record supports only some of the claims made about either.

Colour photograph of the dam site: a green meadow with a wooden fence and walkway, a grassy earth ridge with trees, and a road beyond the ridge.
Where it happenedPhotographed 2013 The South Fork Dam site today, with walkways at the ends of the earth embankment. Photo: David Brossard · CC BY-SA 2.0 · Wikimedia Commons.

The timeline

  1. Construction of the South Fork Dam begins as a feeder reservoir for a canal.

  2. The half-completed dam fails in 1847, and fails again after heavy rain in 1862.

  3. Congressman John Reilly buys the dam for $2,500; the National Park Service says its five sluice pipes were removed while he owned it.

  4. Heavy rain fills the lake; the fish screens over the spillway clog, and water overtops and breaches the embankment.

  5. The flood reaches Johnstown, about 14 miles downstream; about 2,209 people die.

  6. An ASCE committee report concludes that the club's changes did not cause the disaster.

  7. Coleman, Kaktins and Wojno publish a hydraulic reanalysis that disputes the 1891 conclusion.

A researched reconstruction for learning. These reasoning steps simplify a complex record; no score can capture the full history.

02 · The evidence

What does the record show?

Black-and-white 1889 photograph of a shallow stream and grassy bank in the foreground, with a wooden barrel and a fallen branch, and an earth embankment with trees beyond.
On the evidence table1889 'Johnstown Flood No. 28: South Fork Dam & Club House', copyright 1889 by Langill & Darling, as printed on the photograph. Photo: Library of Congress · Public domain · Wikimedia Commons.

Read each exhibit before settling on an explanation. The source links below every exhibit lead to the record behind it.

Exhibit 01

Pipes removed, crest lowered, spillway screened

The National Park Service's history of the dam says that the sluice pipes removed in 1875 were never replaced, that the club lowered the top of the dam, and that it placed fish screens over the spillway. The screens clogged on 31 May 1889, so overtopping became the only way for the water to leave. The page records these changes, but it does not by itself measure how much each one mattered to the outcome.

Source record s1nps.gov

Exhibit 02

Ownership and scale

The Johnstown Flood Museum records 2,209 deaths and says the dam was owned by the South Fork Fishing and Hunting Club, an exclusive club whose members included Andrew Carnegie and Henry Clay Frick. It gives the distance from the dam to Johnstown as 14 miles, and says the reservoir held about 20 million tons of water. Membership describes who owned the dam, not how it behaved in flood.

Source record s2jaha.org

Exhibit 03

The 1891 finding

The 2016 paper by Coleman, Kaktins and Wojno summarises the 1891 report of an ASCE committee, which concluded that the club's changes did not cause the disaster, because the embankment would have overflowed and breached without them. The authors say the 1891 committee's work represented the state of the art for its time. They then challenge several of its observations and assumptions.

Source record s3doi.org

Exhibit 04

The original design

Using a LiDAR-based estimate of the lake volume and reconstructed flood hydrographs, the authors argue that the original dam, with a crest about 0.9 metres higher, an auxiliary spillway and five discharge pipes, had more than twice the discharge capacity at overtopping of the reconstructed dam. They conclude that a properly rebuilt dam would not have overtopped and would likely have survived the runoff. That is a counterfactual reconstruction built on modelled flows, not a direct measurement.

Source record s3doi.org

Exhibit 05

Competing flow estimates

The same paper argues that more than 65 minutes would have been needed to drain most of the lake, rather than the 45 minutes cited by most sources, and estimates peak flood discharges of roughly 7,200 to 8,970 cubic metres per second. Those figures depend on the lake volume, the breach geometry and time estimates, so they are model results that other analysts can contest.

Source record s3doi.org

03 · Your reasoning

Follow the strongest explanation.

Choose an interpretation, then compare it with the evidence. You can revise your choice. The goal is understanding, not a perfect score.

Reasoning step 01 of 03

Which observation most directly bears on whether the club's modifications changed the dam's capacity to pass water?

Evidence-based explanation

Overtopping depends on how much water a structure can pass. The removed pipes and the screened spillway are central to that question, and both the National Park Service account and the 2016 analysis emphasise them. The members' identities and the flood's reach do not answer it.

Weigh it against

Exhibit 01Pipes removed, crest lowered, spillway screenedExhibit 04The original design

Reasoning step 02 of 03

Why did the water rise over the crest on 31 May 1889?

Evidence-based explanation

The National Park Service record places the immediate failure at the spillway and crest: once the screens clogged, the lake had no outlet but over the top. Heavy rain was the trigger, but the 2016 analysis argues that the rain alone does not show a rebuilt dam would have failed.

Weigh it against

Exhibit 01Pipes removed, crest lowered, spillway screenedExhibit 04The original design

Reasoning step 03 of 03

Which statement about responsibility for the disaster does the record support?

Evidence-based explanation

The 1891 and 2016 analyses share the physical facts but differ on assumptions about inflow, timing and the original design. The record therefore supports a careful statement about contribution to the failure, but it cannot assign legal or moral blame.

Weigh it against

Exhibit 03The 1891 findingExhibit 04The original design

04 · The findings

The chain behind the outcome.

The physical sequence is well documented. The South Fork Dam had been built as a feeder reservoir, had failed before, and had been patched rather than fully rebuilt. Its sluice pipes were removed, in 1875 according to the National Park Service, although the 2016 analysis says the date is unclear and notes that the 1891 committee attributed the removal to the club. The club lowered the top of the dam, never replaced the pipes, and placed fish screens over the spillway. When heavy rain filled the lake on 31 May 1889, the screens clogged, leaving the only outflow over the top of the embankment. The dam was overtopped and breached, and the released water reached Johnstown about 14 miles downstream. The Johnstown Flood Museum records 2,209 deaths. What is disputed is how much the modifications mattered. The 1891 ASCE committee concluded that the club's changes did not cause the disaster, because the embankment would have overflowed and breached regardless. Coleman, Kaktins and Wojno (2016) argue the opposite. Using a LiDAR-based lake volume, hydrographs and a reconstruction of the original design, they report that the original dam, with a crest about 0.9 metres higher and five discharge pipes, had more than twice the discharge capacity at overtopping of the reconstructed dam, and that a properly rebuilt dam would likely have survived. Their peak flow estimates of roughly 7,200 to 8,970 cubic metres per second, and their drainage times, depend on modelling assumptions that other analysts could contest. The careful conclusion is that the physical changes reduced the dam's capacity to pass floodwater, that the failure ran through overtopping at a blocked outflow, and that engineers still disagree about the size of the club's contribution. Legal and moral responsibility is a different question, and this case does not answer it.

Black-and-white photograph of a wrecked street: a row of tents at left with a few people standing, piles of timber and debris, a horse and cart, and damaged brick buildings behind.
How it ended1889 Tents and wreckage in Johnstown after the flood, identified with the arrival of the American Red Cross. Photo: Unknown author · Public domain · Wikimedia Commons.

Contributing causes and conditions

  1. Earlier construction stoppage and the 1862 breach left the embankment damaged, according to the National Park Service account.
  2. The discharge pipes were removed (when, and by whom, is disputed) and never replaced, reducing the outlets available for high water.
  3. The crest was lowered, which reduced the margin before water could pass over the top.
  4. Fish screens over the spillway clogged, so the lake could drain only by overtopping.
  5. Heavy rain filled the lake on 31 May 1889, and the embankment was overtopped and breached.

OutcomeJohnstown Flood

Take the method with you

  1. A structure's capacity depends on its outlets as well as its height, so a change that looks minor can alter how the structure fails.
  2. When two careful analyses share the same facts but differ on assumptions, identify which assumptions carry the conclusion.
  3. A finding about cause is not the same as a finding about responsibility, and a case file should keep the two separate.

You can read findings at any time. Exploring every reasoning step marks this case as complete in your notebook.

Open another investigation

The open record

Sources & further reading

Read the original record and its scope before relying on a conclusion. Source references support the exhibit summaries; Faultory’s reasoning prompts are editorial reconstructions.

  1. s1
  2. s2
  3. s3

Written and curated for Faultory by Yunus Emre Vurgun, with AI-assisted research and drafting. Sources and interpretations can need correction. Read our editorial standards or report an issue.

From the journal

Field notes that use this case

Keep investigating