The Man Who Knew the Wall Wasn't High Enough

Roy Freeboard is not a real person. He is a composite character drawn from the documented record of Army Corps engineers who designed, maintained, and studied the Johnstown Local Flood Protection Project between 1943 and the present. That record includes internal reports, feasibility studies, and post-disaster assessments that tell a story the public version never quite caught up to. The name is a small joke for the hydraulically inclined: freeboard is the distance between the waterline and the top of a channel wall, the margin you hope you never need. We met him at a booth in a diner on Main Street in Johnstown, Pennsylvania. The Conemaugh River runs through a concrete channel visible from the parking lot.

You spent decades working on a flood control system that the community called "flood free." When did you first realize that phrase was going to be a problem?

Roy: It was a problem before I got there. Colonel Wilkes said what he said in '43, that the flood problem was effectively solved, and the city ran with it.¹ By the time I'm reading the design documents in the late sixties, "flood free" is just in the air. Tourism materials. Casual conversation. The way people introduce their town to outsiders.

And you're sitting there with the 1939 channel improvement report open on your desk, and it says, right there in plain language, that the system was designed to handle a flood equal to the March 1936 event. Equal to. Not greater than. Not "anything nature throws at you." Equal to one specific flood that already happened.²

So you know. From day one, you know. The question is what you do with knowing.

What did you do with it?

Roy: My job.

Can you say more about that?

Roy: Look, there's a version of this where I'm the lone voice in the wilderness, banging on doors while the bureaucrats cover their ears. That makes a better movie. The reality is duller and worse. The Corps knew. We published a report in '74, "The Potential for Future Flooding in the Johnstown Area," that showed what water levels would look like under a standard project flood.³ We put it out there. It was not a secret document.

But we couldn't tell people when it would happen, and it turns out that's the only part anyone cares about.

You say "this could happen," and they hear "but it probably won't." You say "the channels are designed for a 1936-type event," and they hear "the channels protect us." The word designed does a lot of quiet work in that sentence. Nobody hears it.

The 1977 AAA Tour Book was still telling tourists Johnstown had "complete flood protection." Three years after your report came out.

Roy: [long pause] Yeah. Someone showed me that after. "Complete flood protection." I think I laughed. Not because it was funny.

Here's the thing that kept me up. The channels did work in '54 against Hazel. They worked in '72 against Agnes.⁴ Every time they held, the myth got stronger. And every time the myth got stronger, the odds of anyone reading our report got worse.

Success was the worst thing that could have happened to us, communication-wise. Thirty-four years of the system doing exactly what it was supposed to do, and all anyone learned was that floods were over.

After the '77 flood, the Corps said water levels would have been eleven feet higher without the channels.⁵ That's technically a defense of the system. How did that feel to say out loud?

Roy: It's true. It is technically, hydraulically, completely true. And it is the kind of thing you say when seventy-six people are dead and you need to explain why the thing you built didn't stop it.⁶

What I'd want your readers to sit with is this: the flood didn't come through our system. Not primarily. It came through Solomon's Run. Sam's Run. Peggy's Run. Little streams off the hillsides that were never part of the project.⁷ The Corps had looked at extending protection to those tributaries back in '72 and concluded it wasn't economically feasible. You'd need projects on every small stream in the area, and the numbers didn't pencil.⁸ So we knew those streams were unprotected. We knew it, we documented it, and we moved on, because the authorization didn't cover it and the money didn't exist.

Then twelve thunderstorm cells parked over that valley for nine hours, and every one of those little streams became a river.⁹

The channels handled the main river flows. They overtopped, yes. The Conemaugh hit 120,000 cubic feet per second against a design capacity of 81,500. But the catastrophic damage, the deaths, that was the small watersheds.¹⁰

Do you think the public can hold that distinction?

Roy: [stirs coffee] I think the public holds a simpler model, and I don't blame them for it. You build a wall, the wall keeps water out. That's the mental model. The idea that a flood control system has a scope, that it addresses river flooding but not flash flooding from a different watershed entirely, that's an engineering concept. It's not intuitive.

And we never made it intuitive. We let "flood free" do our communication for us. "Flood free" doesn't come with an asterisk.

Sixty flood insurance policies in the whole city before '77.¹¹

Roy: Sixty. Out of the entire housing stock. Eighty percent of the flood victims GAO interviewed said they believed Johnstown was floodproof. Forty-one of sixty-seven specifically cited the Corps project as their reason.¹²

That's not ignorance. That's trust. They trusted us because we had built something that worked for thirty-four years against every flood that showed up. We just hadn't earned that trust against the flood that hadn't shown up yet. There's a difference, and it's a difference that only matters once.

You're describing a pattern that exists right now in hundreds of communities behind levees and seawalls designed for historical conditions.

Roy: I know I am. [half-smile] That's why I'm sitting here, isn't it?

Is there something you'd want to say to an engineer maintaining one of those systems today?

Roy: [long pause]

Your files know things your community doesn't. That's not a comfortable position. It's also not a temporary one. You will not fix it with a report. You will not fix it with a press release. You might not fix it at all.

But you have to try, because the alternative is sitting in a diner forty-nine years later explaining why sixty insurance policies wasn't enough.

Roy Freeboard finishes his coffee and glances out the window toward the channel. The concrete lining is visible from here, eighty-three years old, recently cleared of sediment by a Corps crew funded through the Bipartisan Infrastructure Law.¹³ The freeboard looks adequate. It always does.

Footnotes

1. Kozlovac, Joseph P. "Adventures in Flood Control: The Johnstown, Pennsylvania Story." Graduate paper, NOAA, April 1995. Citing Army Corps of Engineers, Pittsburgh District, Johnstown Channel Improvement, 1939. https://www.nwrfc.noaa.gov/floods/papers/Johnstown_flood.pdf ↩

2. Army Corps of Engineers, Pittsburgh District. Johnstown Channel Improvement. Pittsburgh, PA, 1939, pp. 5-7. Accessed via Kozlovac 1995. ↩

3. Kozlovac 1995, citing Corps internal report, "The Potential for Future Flooding in the Johnstown Area," 1974. ↩

4. Explore PA History / Pennsylvania Historical and Museum Commission. "Johnstown Local Flood Protection Project Historical Marker." https://explorepahistory.com/hmarker.php?markerId=1-A-348 ↩

5. Heritage Johnstown, "1936 & 1977 Floods," citing Army Corps of Engineers estimate. Confirmed in GAO 1978 and Kozlovac 1995. https://www.heritagejohnstown.org/attractions/johnstown-flood-museum/flood-history/1936-1977-floods/ ↩

6. U.S. General Accounting Office. The Johnstown Area Flood of 1977: A Case Study for the Future. CED-78-114. May 1978. Death toll of 76 confirmed, 10 missing. https://www.govinfo.gov/content/pkg/CZIC-hg9983-u55-1978/html/CZIC-hg9983-u55-1978.htm ↩

7. Tribune-Democrat, July 2017, quoting Richard Burkert, Johnstown Area Heritage Association. Via GovTech: https://govtech.com/dc/articles/New-3-D-Technology-Gives-Closer-Look-at-River-Walls-in-Johnstown-Pa.html ↩

8. GAO 1978, p. 3. ↩

9. Kozlovac 1995, sourcing USGS Professional Paper 1211, 1977. ↩

10. GAO 1978. Design capacity and actual 1977 discharge figures. ↩

11. Kozlovac 1995, citing GAO 1978. ↩

12. GAO 1978, Appendix IV. Survey of 67 flood victims. ↩

13. U.S. Army Corps of Engineers, Pittsburgh District. "Scoop and restore: Army Corps removes sediment from river channels designed to safeguard Johnstown from catastrophic floods." Army.mil, October 30, 2024. https://www.army.mil/article/280965/ ↩