No, Lake Powell is not inexorably headed toward “dead pool”

Here is some data about Lake Powell, the big Colorado River reservoir straddling the Arizona-Utah border.

  • Since 2005, the average estimated “natural flow” of the Colorado River at Lee’s Ferry has been 13.5 million acre feet per year, well below the 14.8maf average since 1906.
    • So yes, it’s been dry.
  • Since 2005, releases from Lake Powell have exceeded the Upper Basin’s Colorado River Compact and Mexican treaty obligations by 9.4maf.
    • So despite the fact that it’s been dry, we Upper Basin residents have had enough extra to deliver a bunch of “bonus water” to downstream water users.
  • Since 2005, the total volume of storage in Lake Powell has risen, from 9.169 million acre feet at the end of 2004 to 11.028 million acre feet at the end of 2018.
    • So despite being dry, and delivering extra water downstream, Lake Powell’s elevation has been relatively stable.

This is straightforward empirical stuff.

Brian Maffly has a great piece in the Salt Lake Tribune about the challenges of Colorado River management, with a focus on Lake Powell.

But Maffly weakens an otherwise excellent survey of the river’s issues with the alarmist assertion that “without a change in how the Colorado River is managed, Lake Powell is headed toward becoming a ‘dead pool.'”

It is important when doing journalism about phenomena over time (and any other sort of analysis) to choose periods of record that accurately capture the thing you’re trying to understand and explain. Maffly, and the sources on whom he relied for the story, have cherrypicked a time window that supports the “OMG POWELL DOOMED” argument, but that is misleading.

In support of his argument, Maffly notes that Powell “has shed an average of 155 billion gallons a year over the past two decades.” That is correct, but it’s a troubling choice of time frame. Since 2000, Maffly writes, Lake Powell “has been steadily dropping.” This is false.

The entire loss Maffly describes happened during the first years of the 21st century, the driest five-year stretch on record. Yowza, Lake Powell dropped a lot during the drought!

From 2005 to the present, Lake Powell’s elevation has been relatively stable, not steadily dropping.

Lake Powell storage since the turn of the century

It certainly would be correct, and is important to note, that a repeat of the drought of the early ’00s would be disastrous. Absent changes in management, it would lead us to dead pool. It is important to have policies in place that are ready for that. Increased withdrawals from the basin increase the risk should that happen. But the years that followed the drought of the early ’00s, in which Lake Powell’s levels have been stable despite a river flow depleted by climate change and the crazy policies that continue to ensure excess downstream deliveries, also are important if we’re to assess the impact of current Colorado River operating policies on the risk of Lake Powell reaching dead pool.

The reservoir that seems to be headed far more inexorably toward disaster is Mead, not Powell. That’s a lot closer to what I imagine when I hear “steadily dropping”. Remember that Mead is declining despite deliveries since 2000 of more than 9 million acre feet above the Compact-Mexico obligation of 8.23maf per year.

Lake Mead storage since the turn of the century

To be clear, I believe the Colorado River faces serious challenges. I agree with Doug Kenney, who told Maffly, “At some point, you can’t ignore reality anymore, and the reality is we need to use a lot less water in the Colorado Basin.” Yup. Doug’s right. This includes the Upper Basin. The Lake Powell pipeline and other efforts to take more water from the system seem ill-advised, guaranteed to increase the risk to existing users in both the Upper and Lower basins.

Much of the challenge right now, though, is in the Lower Basin. The ability to maintain Lake Powell’s elevations over the last ten-plus years despite drought and “bonus water” releases from Powell suggests the Upper Basin is in a far better position.

Much of Maffly’s discussion of the overall challenges is excellent. It’s important, especially for Utah water users to hear. But selling that message with “OMG LAKE POWELL DEAD POOL” is at odds with the facts, and undercuts the otherwise important stuff the story has on offer.

A note on data sources:

On Sunday, I crashed Lake Mead.

Crashing Lake Mead

Sunday afternoon, I crashed Lake Mead.

This was not difficult.

Each spring, UNM Water Resources Program students do a case study of a river basin as they’re learning dynamic simulation modeling, linking hydrology, economics, and rules. This year, we’re doing the Lower Colorado River Basin.

In the WRP curriculum, we’re big on understanding the rules by which we manage water as a resource – where they come from, and how they work.

The model results above are a perfect example of George Box’s famous dictum that “all models are wrong but some are useful.” The Goldsim model, which I wrote really fast because I needed something for tomorrow’s lecture, is a super-simple simulation of the water allocation hydrology and rules in the classic Bureau of Reclamation “structural deficit” slide, combined with a Monte Carlo simulation of some reasonable assumptions by me about 21st century inflow hydrology from the Upper Colorado River Basin.

Within a decade in my model, Lake Mead is kinda unusable, doing little more than passing through whatever water comes down from upstream. Under some scenarios, my simulation quickly throws an error, informing me that you can’t have negative water in a reservoir.

The rules here, which govern now much water gets released from the reservoir, are encoded in the Colorado River Compact and the Boulder Canyon Project Act, as interpreted by the U.S. Supreme Court in the 1963 Arizona v. California decision.

Obviously this is the “all models are wrong” part of Box’s dictum. This is a realistic view of the system if we don’t do anything. It’s not a realistic view of the system because it assumes people won’t say “OMG Lake Mead is almost empty we’ve gotta do something!” Which is what they’re now saying, which is why new and better rules are being written.

This spring’s assignment to our water resources students is to experiment with ideas about what those new rules might look like, how they might work.

I’m very much looking forward to this. I think it will be useful.

Colorado River data

Storage in lakes Mead and Powell

In the years I’ve been working on Colorado River stuff, I’ve collected a lot of data. Most of this is stuff the river managers at the Bureau of Reclamation have generously shared with me. Some I’ve assembled myself by transcribing old paper/pdf records of historic documents. Some of the Bureau’s most important data is already easily publicly available (most notably the Natural Flow Database, Lake Mead elevation, Upper Basin reservoir elevation and storage histories, and the 24-month studies). But over the years, the accumulation on my hard drive(s) has turned into a useful personal resource for my research and writing if I wanna look up, say, what the trends are in water use in Yuma or some such.

I’ve been meaning for a while to find a way to make what I’ve got more accessible. So I’ve started uploading data to Github. First up is the data for the graph above, for which I get a lot of requests. It’s total end-of-water-year storage in Lake Mead and Lake Powell, stitched together in a more easily usable form from a couple of different USBR datasets. I update my personal copy sorta monthly with the Bureau’s forecasts and make scary graphs to post on Twitter. (Scary Lake Mead graphs make great clickbait.) The stuff on Github is just actual year end numbers.

Two motivations: I’m teaching a Colorado River-related course this spring at the University of New Mexico Water Resources Program – we’ll need data! And as Eric Kuhn and I finish up the new book, we’re starting to think about how to make the key data and documents used in the book easily available. Not sure this is the platform, but this’ll give me an experimental toy to think about how to do it.

MWD increases its Lake Mead withdrawals

For background:

Meanwhile the Metropolitan Water District of Southern California has increased the pressure with an ultimatum of its own. Absent a DCP, Met is at risk under the current rules of having its banked water stranded in Lake Mead. DCP would relax those restrictions, but without a DCP, Met has made clear it has no choice but begin taking that water out beginning in January. (Daniel Rothberg explains all this here.) This means Lake Mead drops faster, increasing everyone’s risk, but especially Arizona’s. (Tapping foot, glaring, arms crossed, at Arizona.)

Happy New Year, Lake Mead dropped another foot in 2018

By the numbers, Lake Mead began 2019 at elevation 1,081.47 feet above sea level, down a foot in 2018. This comes in spite of yet another Upper Basin release of 9 million acre feet from Glen Canyon Dam last year, an extra 770,000 acre feet above the 8.23 million acre feet the Lower Basin can rightly expect under the Law of the River. Since 2000, the Upper Basin has delivered an excess of 9.8 million acre feet from Glen Canyon Dam above and beyond the 8.23 million acre feet expected every year under the current interpretation of the Law of the River. In that time, the level of Lake Mead has declined 115 feet.

Worth repeating.

In the 21st century, water users in the Lower Basin have received 9.8 million acre feet of bonus water, above and beyond their legal expectation under the Law of the River. This is not about climate change or drought. This is about using too much water.

Eric Kuhn and I are in the finishing stages of a new book looking at the history of our understanding of the hydrology of the Colorado River – what we knew about how much water the river really has, and how that knowledge shaped river management institutions. I don’t want to give away too much here (Buy our book! As soon as it’s published!), but our reading of the history makes clear that this situation was essentially known or knowable a century ago.

the now-superfluous spillways of Hoover Dam

Over the winter break, in clarifying some of the book’s final bits, I’ve revisited John Wesley Powell’s Report on the Lands of the Arid Region. As the historian Sara Porterfield pointed out at last year’s Colorado Mesa University Upper Basin Forum (a bit of her argument here), Powell hagiography has become an intellectual crutch, but he’s nevertheless a good guide. In the 1878 introduction to the first edition of Arid Lands, Powell talked about the crucial interplay between three things: descriptive science (How much water is there?), engineering (What kind of stuff can we build to use the water?), and institutions (What kind of rules do we need to govern the cooperative enterprise needed to pull this off?). He and others of the day I’ve been reading use a funky old word that I like to describe what is needed – “prevision”. By this they mean a sort of foresight, a prognostication. “[A] wise prevision, embodied in carefully considered legislation, is necessary,” Powell wrote.

This is the bit that isn’t working so well.

If you look at the math, it was abundantly clear by 2000 that, as Terry Fulp noted in a passage in my last book, “Lake Mead will go down.” (You can go ahead and buy it now, or wait until March it’s coming out in paperback.) Our “prevision” hasn’t been so great.



“What happens if we turn here?”

Alley, Albuquerque’s Old Town

I turned down this alley in Albuquerque’s Old Town on a bike ride this morning, amazed that I’d never noticed it before.

I spend a lot of time on the bike. I’m long past my youthful zoomy lifestyle, but in my slow and plodding way I have come to deeply value my time wandering Albuquerque. My friend Maria Lane (an actual professor of geography) calls it “geography by bike”. It’s based on a “what happens if we turn left here” approach to my city, and this year I logged 4,400 miles of it, nearly all here in Albuquerque. After a few years of working too hard and riding too little, it’s rejuvenated my view of my city. And me. Remember that feeling of freedom when you were a kid and grabbed a sack lunch and rode off to wherever?

And yes, I know the number of miles. Yes, I am a geek. Yes, I GPS every ride.

John’s geeky 2018 Albuquerque bike ride map

And I try to hit all the alleys.

DCP by the numbers

I’m forever needing to look up these numbers, sticking them here so I can find them easily, and y’all can as well. It’s the cuts by Lake Mead elevation levels for the various parties under the Colorado River Drought Contingency Plan and the related Minute 323 between the U.S. and Mexico:

DCP cuts