On paper, the Cumberland River loses a fifth of its hundred-year flood right where the Stones River pours in. Lester doesn't trust the river. He trusts the drainage areas — and they say two labels in Nashville's federal flood study are sitting one station off. Eight days to a yes. Eighty-seven to a reprint. And five editions of the book, laid side by side.
He'd told Atlanta he was continuing upstream into the Cumberland, and he'd meant it. Every county on this stretch of the river has a Flood Insurance Study, and every one has a Summary of Discharges — the table that says how much water the hundred-year flood carries past a named point. This morning it was Davidson's turn. The Cumberland through Nashville is Region 3's river; hydrology for every region runs out of headquarters, which made it his.
He read the 1% column the way he reads everything — slowly, one number at a time, because that's the only way the numbers stay where they're printed. One hundred ninety-eight thousand. One hundred fifty-five thousand. Downstream of the Stones, the hundred-year flood got smaller.
He always writes that second line. The feeling of wrongness arrives before the cause of it, and if you don't write the feeling down you'll forget you had it twenty minutes into the wrong hypothesis. He pulled the neighbors' tables. Nobody else on that stretch of river printed 198,000. The county downstream printed 208,000 at its own dam and 197,000 at its far line — nearly twenty-eight hundred square miles further down.
The first hypothesis was Old Hickory. Everything on this reach lives below a Corps dam, and regulated flow doesn't behave like an open river — attenuation can shave a peak. But twenty-two percent, in exactly the reach where a major tributary comes in? He drew a box around it: true, and beside the point.
The second was the convention. "At the confluence" can mean the river just above a tributary, or just below it with the tributary counted in, and a reader who assumes the wrong one can manufacture a discrepancy out of nothing. So he stopped reading FEMA's numbers and went and got his own — USGS StreamStats, one physical place at a time. Old Hickory Dam at 8:25 AM: 11,674 square miles. The Stones confluence, Stones counted in: 12,683. The Harpeth confluence: 14,125 with the Harpeth, 13,257 without.
Every receipt came back stamped REGULATED, because every point sat below the dam. That was fine. He wasn't asking for a flow. He was asking for the shape of the land, and the shape of the land doesn't care who runs the dam. Both ways, at both confluences, the labels still didn't fit.
He rolled his chair back and looked at the parking lot for about ninety seconds. Each label sat one station downstream of the number printed beside it. The row that said Stones River carried Old Hickory Dam, to a tenth of a square mile. The row that said Harpeth River carried the Stones. And the actual Harpeth confluence wasn't in Davidson's table at all. Next to it, the neighbors read clean. It looked like only Davidson.
Slide the labels home and the 198,000 sits at the dam, the 155,000 just below the Stones. The labels moved. The drop didn't. He circled that line, and he didn't throw it away.
Sarah Morrison had been at her desk since seven, which meant something was already wrong somewhere else in the building. He walked her through it — the column, the dam, the convention, the slide. "Davidson only?" "Davidson only." "If you put this on your website, somebody's going to send you a letter by Thursday." "I'm not putting it on the website. I'm calling Henry Holiday."
Henry Holiday was the state's flood insurance program coordinator — the man who'd forwarded the Bedford package to Atlanta. "After Bedford you could probably call Atlanta direct," Sarah said. "Bedford went through the state. This goes through the state," Lester said. "I'm a hydrologist at a state DOT. I don't get to skip the people whose job this is."
"Call him. Send it. Copy me." Then, as he stood up: "This is going to matter." "Nobody's flood elevation moves because a label's wrong," he said. Sarah said: "Yet."
The call took eleven minutes. The write-up took the rest of the morning — not because the finding was complicated, but because how you write a finding decides whether anyone opens the attachments. Draft one sounded like a man guessing. Draft two sounded like a committee. Draft three opened the way a phone call ends, gave the short version, then the table, then the thing that first caught his eye: a 198,000 that no other station in four counties printed, higher even than the 197,000 far downstream.
He copied Sarah. He hit send at 1:27. He sat in the truck for twenty minutes, not listening to anything. Then he drove to Thornton's.
Earl Wallace was already there. Earl is always already there — a retired actuary, the Lucky Number man, who still sets odds on tornado grids and Broadway beats, and who has the specific gift of being unable to skim. He read the legal pad for thirty seconds. "So the names are wrong." "The drainage areas are real. They're just sitting next to the wrong names." "And everybody who's used it since reads the label and takes the number." "Because the label and the number read as a unit. You don't audit the label." "Unless you're a tool." "Unless you're a tool."
"Same thing as the fountain machine," Earl said. "Nobody bled the line, because nobody was responsible for it. Then somebody put a laminated sheet on the wall. The sheet didn't fix the air. It made the air visible."
"That's going to be a strange email for somebody to open," Earl said. "You know what's stranger? That it took one guy on a laptop on a Tuesday."
The reply was short, and it was from Henry Holiday. The review was done. The drainage areas and the flows would stay exactly as printed. The location descriptions were wrong, and they should go back to the way the earlier edition had them. FEMA would issue a Notice to User, so the correction reached everyone who used the book.
The 198,000 had been in his email — the thing that first caught his eye. The reply kept every flow as printed. So here is what he wrote back, and here is what the yellow legal pad said.
Then he went and got the earlier editions. Every one he could find.
Not forty years. Not a scar from the 1980s. One rewrite in February 2022, carried into June 2024 — long enough for anyone who opened Table 9 to read past it. And the edition the state pointed him back to was only half right: sixteen hundred feet upstream of the Stones, the Stones isn't in the river yet. That number already had it.
Earl was already there. Lester slid one sheet across the counter, printed from FEMA's Map Service Center.
The reprinted volume came with it, and it didn't go back to 2017. Beside 11,673.9 it said Dam at Old Hickory Reservoir. Beside 12,691.4 it said At the confluence of Stones River. Exactly where the drainage areas had put them on April 21. The map, the database and the flood layer were unaffected. Only the words in the table changed, plus a reprint note in the front of the book.
"And the drop?" Earl said. "Still there." "And how are you feeling about it." "Fine. The labels got fixed. That's what I sent. The rest is the next letter." "Lester." "Yeah." "It is, a little, your job past this point."
He read it down the column anyway, one number at a time. With the words fixed, there was nothing left to blame on the words.
The river picks up a thousand square miles at the Stones, and every flood but the ten-year gets smaller for it — fourteen percent at the 25-year, nineteen at the 50, twenty-two at the 100. The rarer the flood, the more of it goes missing. The labels fit the land now. The water still doesn't.
He tore the sheet off, folded it once, and put it in his shirt pocket. Not sent. Not thrown away.
Then he did what he tells everybody else to do, and believed the water. Two gages bracket that reach: one in the tailwater at Old Hickory, one downtown at Nashville, below the Stones. He pulled both. May 2010, the biggest flood either gage has recorded since 1950: 212,000 at the foot of the dam on the second, 188,000 at Nashville on the third. March 1975, the next biggest, did it too, smaller. In the two biggest floods since 1950, the river picked up the Stones and still came out lighter downstream.
One more thing had been in his own email since April, in a row he’d read straight past. The county on the other side of the dam prints two Cumberland rows where it meets Davidson. At river mile 216.2 at Old Hickory Dam: 11,694 square miles, 158,000 at the 100-year. Davidson County boundary: 13,589 square miles. The dam sits on the Davidson line, and StreamStats reads 11,674 there. The river doesn’t get to 13,589 until the Harpeth pours in, past Nashville. The flows on that boundary row are 115,000, 128,000, 140,000, 155,000 and 190,000: Davidson’s row below the Stones, digit for digit.
Every one of these starts the same way: a number that doesn't sit right, and a man who writes it down before he decides what it means. The surge tower, the SUPERSHEET, the FEMA letter sent from a parking lot, two labels one station off, and a blank field that read as a passing check. One page holds all of it.
Where this one goes next