A River Wearing a Suit Three Sizes Too Big — One Hydrologist's Reckoning With the Colorado's Foundational Lie

In 1976, two researchers at the University of Arizona published a 70-page bulletin that should have detonated the legal foundation of water management in the American West. Charles Stockton and Gordon Jacoby had used tree-ring data to reconstruct 400 years of Colorado River streamflow, and their conclusion was blunt: the river's long-term average flow at Lees Ferry was approximately 13.5 million acre-feet per year, roughly 10 percent less than the observed record and far below the 16.5 million acre-feet already committed to states and Mexico under the 1922 Colorado River Compact.¹ The compact had been negotiated during the wettest quarter-century in at least four centuries. Every dam, every delivery contract, every city water system built since 1922 was premised on a river that was larger than it actually was.

The message, according to the TreeFlow archive maintained by the University of Arizona, "did not resonate for very long."²

What happened inside the Bureau of Reclamation when this paper landed? The institutional record is curiously quiet. No Bureau hydrologist is documented as the first internal recipient of the Stockton-Jacoby findings. No memo surfaces in the accessible archive saying we have a problem. The gap is itself information: the finding existed; the institutional channel for receiving it did not.

Darlene "Dar" Sandoval, who may or may not have worked as a Bureau of Reclamation hydrologist in the Lower Colorado Region office in Boulder City, Nevada, during the late 1970s, and whose name does not appear in any personnel record we've been able to locate, agreed to speak with us about what it was like to read the Stockton-Jacoby paper and understand its implications from inside the institution built on the old number.

You were handed the Stockton-Jacoby paper in late 1976. What was the assignment?

Dar: My supervisor wanted a two-page technical summary. "Flag anything operationally relevant." Two pages. I remember sitting there thinking, okay, the operationally relevant part is that the entire Colorado River Compact is premised on a wet anomaly, so do I put that on page one or page two?

Had you suspected the number was high before you read the paper?

Dar: Suspected is generous. I'd watched it get smaller. When I started at the Bureau, the long-term average at Lees Ferry was cited as about 16 million acre-feet. By the mid-seventies, with more years of data, it had already dropped to around 15.³ You don't need tree rings to notice a trend line.

But there's a world of difference between "the average is drifting down" and "the average was never what we said it was." The tree rings weren't showing a trend. They were showing that the baseline was wrong. A trend you can manage. You adjust projections, build in a buffer, write a memo that says "conditions may be changing." A wrong baseline means the legal architecture of the Colorado River, the compact, the Mexican Water Treaty, the delivery contracts, Arizona v. California, all of it, was built on a measurement taken during a four-hundred-year outlier.⁴ You can't memo your way out of that.

So what did you write?

Dar: Three drafts. The first one was honest. The second one was diplomatic. The third one was two pages.

What was in the first draft?

Dar: The math. Stockton and Jacoby's reconstructed average was 13.5 million acre-feet. The compact allocates 15 million between the basins, plus 1.5 to Mexico. That's 16.5 committed against 13.5 available. I wrote, and I remember this because I stared at the sentence for a long time: "The tree-ring reconstruction suggests a structural deficit of approximately 3 million acre-feet per year between committed allocations and probable long-term supply."⁵

Three million acre-feet. That's roughly the entire annual allocation to the state of Arizona. That's the Central Arizona Project, which was at that moment being built, concrete being poured, tunnels being bored, to deliver 1.5 million acre-feet a year to Phoenix and Tucson.⁶

What happened to that sentence?

Dar: It got revised to something like "the reconstruction merits further study in the context of long-range planning assumptions." Which is the bureaucratic equivalent of screaming into a pillow.

Were you told to soften it?

Dar: Nobody had to tell me. I understood the institution I worked for. The Bureau of Reclamation doesn't set water allocations. Those come from the compact, from Congress, from the Supreme Court. We operate infrastructure within that framework. So when I write a memo saying the framework's foundational number is wrong, I'm not raising a technical question. I'm raising a legal and political question that is, in the most literal sense, above my pay grade. My supervisor's pay grade. His supervisor's pay grade. You'd have to go all the way up to God, and God would refer you to the seven basin states.

You mentioned La Rue earlier, off the record. Can you say more?

Dar: Eugene La Rue. USGS hydrologist. In 1916, seven years before the compact was signed, he calculated a lower flow estimate and warned that the 1905-to-1922 period was abnormally wet. He wrote to Herbert Hoover offering his expertise to the compact negotiators. He was not invited.⁷

So when I read Stockton and Jacoby, I wasn't holding a first warning. I was holding a second warning, sixty years later, saying the same thing with four centuries of evidence behind it. And I thought: if they ignored La Rue when the stakes were a piece of paper, what happens when the stakes are Hoover Dam and Glen Canyon and a half-built aqueduct to Phoenix?

What happened next, in 1977?

Dar: Drought. Severe, one-year drought. And for about five minutes I thought, okay, maybe this is the moment the institution pays attention. The river was confirming what the tree rings said.

Did it?

Dar: No. Because then 1983 happened.

Record inflows. Twenty-five million acre-feet. So much water that the outlet tunnels at Glen Canyon Dam were physically damaged; cavitation tore chunks out of the concrete.⁸ The dam we built because the compact said there was enough water nearly broke because there was too much water. You can't write satire that good.

And the tree-ring findings?

Do you think the message could have been delivered differently?

Dar: I think about this more than I should. The problem wasn't the message. The problem was that there was no institutional process for receiving it.

The Bureau's planning function used the allocation framework as a fixed input. A given. The tree-ring data said the given was wrong. But there was no form for that. No channel. No meeting where someone says, "Let's revisit the foundational assumption of the Colorado River Compact." That meeting didn't exist. It couldn't exist, because the compact wasn't ours to revisit. We were the plumbers. Nobody asks the plumber whether the architect got the blueprints right.

So who was supposed to act on it?

Dar: That's the question I never answered. Congress? The seven basin states? The Supreme Court? Everyone was waiting for someone else to go first. And meanwhile, the Central Arizona Project kept getting built, Las Vegas kept growing, and the river kept being the river, which is to say, a 13.5-million-acre-foot river wearing a 16.5-million-acre-foot suit.

Last question. If you could go back?

Darlene Sandoval's personnel file, if it ever existed, has not been located in the Bureau of Reclamation's archived records. Readers interested in the documented institutional history should consult Eric Kuhn and John Fleck's Science Be Dammed: How Ignoring Inconvenient Science Drained the Colorado River (University of Arizona Press, 2019), particularly Chapters 17 and 18.⁹

Footnotes

1. Stockton, C.W. and Jacoby, G.C. (1976), Long-Term Surface Water Supply and Streamflow Levels in the Upper Colorado River Basin, Lake Powell Research Project Bulletin No. 18, UCLA. https://www.treeflow.info/content/first-reconstruction-streamflow-lees-ferry-1976 ↩

2. TreeFlow, "The First Reconstruction of Streamflow for Lees Ferry (1976)." https://www.treeflow.info/content/first-reconstruction-streamflow-lees-ferry-1976 ↩

3. TreeFlow records that the observed long-term average natural flow at Lees Ferry dropped from 16 MAF/year (as measured with data through 1945) to 15 MAF/year (as measured with data through 1975). https://www.treeflow.info/content/first-reconstruction-streamflow-lees-ferry-1976 ↩

4. National Academies Press, Colorado River Basin Water Management (2007), Chapter 2. https://www.nationalacademies.org/read/11857/chapter/4 ↩

5. The 3 MAF structural deficit figure is derived from the gap between 16.5 MAF in compact commitments and Stockton-Jacoby's reconstructed 13.5 MAF long-term average. See Ge, Shemin et al., "Fixing the Flawed Colorado River Compact," Eos (2023). https://eos.org/features/fixing-the-flawed-colorado-river-compact ↩

6. Congressional Research Service, "Management of the Colorado River" (updated 2026). https://www.congress.gov/crs_external_products/R/PDF/R45546/R45546.48.pdf ↩

7. Kuhn, Eric and Fleck, John, Science Be Dammed (University of Arizona Press, 2019). La Rue's 1916 calculations and his letter to Hoover are documented in the book's archival research. https://www.barnesandnoble.com/w/science-be-dammed-eric-kuhn/1131178701 ↩

8. TreeFlow documents that in 1983, record inflows of approximately 25 MAF caused severe cavitation damage to Glen Canyon Dam's outlet tunnels. https://www.treeflow.info/content/first-reconstruction-streamflow-lees-ferry-1976 ↩

9. Kuhn and Fleck (2019), Chapters 17 ("Tree Rings and Climate Change") and 18 ("Reality Sets In"). https://www.barnesandnoble.com/w/science-be-dammed-eric-kuhn/1131178701 ↩