Lake Powell is one foot from switching off its own power

Climate & Energy 4 min 80 sources

A federal forecast puts the Colorado River's second-biggest reservoir a single foot above the level where its dam stops making electricity — while the US clean-energy retreat picks up speed.

Key takeaways

A federal forecast puts Lake Powell one foot above the level where its dam stops making electricity — proof that a slow decline can end in a sudden cliff, not a gentle slope.
The US is stepping back from clean energy — offshore wind halted, carmakers retreating to petrol, Tesla's profit down — even as new solar and wind stay cheaper to build than fossil plants.
The grid is the real pressure point: AI data centres are driving demand up, a central-states operator warned of possible blackouts, and household electricity bills are rising faster than inflation.

One foot from the cliff

The number to hold onto this week is one foot. That is the gap between where Lake Powell’s water is forecast to sit next spring and the level where the Glen Canyon Dam can no longer make electricity.

A federal analysis from the Bureau of Reclamation, released Wednesday, projects the reservoir will fall to 3,491 feet by March [3]. The dam’s “minimum power pool” — the level below which water no longer flows through the turbines — sits at 3,490 feet [3]. Lake Powell is the Colorado River’s second-largest reservoir, and it is already down to about a quarter of its capacity [35].

Here is the mechanism, in plain terms. Glen Canyon Dam makes power by letting reservoir water fall through spinning turbines. There is a fixed height below which the water intakes sit dry — no flow, no spin, no electricity. That can happen while there is still a great deal of water in the lake. Drop a little further, to what engineers grimly call “dead pool,” and water can’t pass the dam at all. The dam supplies hydropower to millions of people across seven western states, so this is not an abstract line on a chart.

The immediate cause is a drought made worse by a warming climate, which has already pushed the combined storage in the river’s biggest reservoirs to an all-time low [3]. Scientists at the University of Colorado Boulder, in a report bluntly titled Dancing with Deadpool, warn that the river’s flow keeps coming in below what managers plan for [35]. Meanwhile the seven states that share the river are deadlocked over how to divide the shortfall before current rules expire [3]. On the water itself, marina operators at Lake Powell are relocating docks and boat ramps to chase the retreating shoreline [5].

A dry summer, spread wide

The Colorado is the sharp edge of a broader drought. In Puerto Rico, the government has declared emergency water rationing as its own drought worsens [61]. Across the Atlantic, a scientific analysis found that record heat is “sucking Europe dry.” The heat is evaporating water from rivers, soils and plants at a rate made 80 times more likely by fossil-fuel warming in the west [50]. Much of southern England is under a hosepipe ban, and roughly nine million tonnes of grain are at risk across the continent [50].

The heat has a human toll that is only now being counted. France recorded about 5,700 more deaths than usual during a historic June heatwave [54]. That is the kind of number that arrives weeks after the weather does.

The US steps back from clean energy

While the water falls, the American energy transition is losing official support. The Trump administration this week moved to halt offshore wind development, citing national security [4]. It follows last year’s repeal of the Inflation Reduction Act, the Biden-era law that had steered hundreds of billions toward clean power. Industry analysts argue the economics of solar and batteries may carry the transition forward with or without Washington [7].

Carmakers are reading the same signals. General Motors told investors it has essentially no new Chevrolet or Cadillac electric models planned for the next few years, pivoting back toward petrol engines [24]. Honda is dropping the only fully electric vehicle it sells in the US [49]. And Tesla reported a sharp fall in quarterly profit, squeezed by the discounts it is offering to move cars [62]. The pull-back is a business decision as much as a political one — for now, the buyers aren’t there in the numbers the companies planned for.

None of this changes the underlying cost picture. Fresh analyses this week still found new solar and wind cheaper to build than new fossil plants in most markets, even as their costs tick up [19][17].

The grid is the pressure point

The strain is landing hardest on the electricity grid. A grid operator across the US central states warned this week of possible rolling blackouts during peak demand [8]. The fastest-growing new load is artificial-intelligence data centres, whose appetite for power is forcing utilities to rethink how much generation they need to build [15]. For households, the bill is already moving: retail electricity rates are rising faster than inflation, with more increases set to come [12].

The quiet story: a fading sunshade

The under-covered piece this week is about clouds. Since 1979, low-lying clouds over the oceans have been expanding, reflecting more sunlight back to space and quietly shading the planet. That has masked some of the warming greenhouse gases would otherwise have caused, according to research reported this week [69]. Scientists now expect that trend to reverse. Clouds have long been the biggest uncertainty in climate models [69]. If a sunshade we didn’t design is starting to lift, the warming already baked in could arrive faster than the averages suggest.

02 · Lesson · why it matters

Some declines are a slope. Some are a floor with a trapdoor.

A reservoir can fall for twenty-five years and lose nothing that matters — until the single foot that switches off the power for millions.

The fall isn’t the danger. The line is.

Lake Powell has been dropping for a long time. Foot by foot, drought year after drought year, the shoreline has crept back and the marinas have chased it. Through all of that falling, the Glen Canyon Dam kept doing its job: water fell through the turbines, the turbines spun, the lights stayed on across seven states.

So the news this week isn’t really that the water is low. It has been low for years. The news is that a federal forecast now puts it one foot above a specific line — the height where water stops reaching the turbines at all. Above that line, the dam works. Below it, it doesn’t. Not “works a little less.” Doesn’t.

That gap between how the fall feels and what the line does is the thing worth understanding. It runs through far more than one reservoir.

Linear until it isn’t

Most declines we meet are slopes. A little less rain, a little smaller harvest. A little less sleep, a little more tired. The output moves smoothly with the input, so the size of the change tells you roughly how much trouble you’re in. You can watch a slope coming.

A threshold is different. It is a floor with a trapdoor. For a long stretch, the input drops and the output barely notices. The dam at 3,600 feet and the dam at 3,492 feet are both, from the outside, just a working dam. Then you cross one particular level and the state flips all at once. The last foot does what the previous two hundred didn’t. The system was linear the whole way down, right up until it wasn’t.

The reservoir is a clean example because the physics is blunt: the water intake sits at a fixed height, and water either reaches it or it doesn’t. But the shape is everywhere. Things that change smoothly, until a level is crossed and they change kind.

The gauge reads “working” until it reads “off”

Here is why thresholds fool careful people. From inside the smooth part, the warning signs look identical whether you are a mile from the edge or a foot from it. The dam’s output at 3,600 feet and at 3,491 feet is the same: full power, everything nominal. The gauge that says “still generating” cannot tell you how close the cliff is, because the gauge measures the thing that hasn’t broken yet.

So we reason from the only evidence we have: it’s been fine so far. Twenty-five years of falling water and the power never failed — why would this year be the one? That reasoning works beautifully on a slope and fails completely on a floor. A threshold punishes exactly the confidence that a long, uneventful decline builds up.

The number that matters, then, isn’t the trend. It’s the distance to the edge. And that distance is the one number a smooth decline is worst at showing you.

You are standing on some of these floors

This isn’t a desert problem. The same week’s news carries another one: a grid operator warning of rolling blackouts. A grid absorbs every surge in demand — a hot afternoon, a million air conditioners — until the surge that exceeds its supply. Then it doesn’t dim gently. It sheds whole neighbourhoods at once. A pension fund is solvent, solvent, solvent, until enough people want out at the same time. A bridge carries every truck until the one that snaps it.

You live on these floors too, further from the headline than it looks. A savings buffer covers every surprise bill until the one that empties it. A body tolerates the extra drink, the skipped checkup, the ignored ache, until it doesn’t. A friendship absorbs slights until one too many. None of these decline in a way you can feel approaching, because the whole nature of a floor is that it holds — until it stops holding.

The point isn’t to fear every downhill. Most declines really are slopes, and treating them as cliffs is its own mistake. The skill is telling which is which: is this a thing that gets gradually worse, or a thing that works fine and then fails whole?

The floor was drawn before the drought

There is one more layer, and it is easy to miss. That specific line Lake Powell is falling toward wasn’t set by the drought. It was set long before, by choices that now pose as plain fact.

In 1922, the states divided the Colorado River’s water using flow measured in an unusually wet stretch. They handed out shares adding up to more water than the river carries in an ordinary year. The over-promise sat invisible for decades while the reservoirs were full — a threshold built quietly into the arrangement, waiting. The drought didn’t create the shortage. It revealed a promise that was always over-drawn, and pushed the river toward a floor that the original deal had already installed.

That’s worth sitting with. Many of the cliffs we approach were placed by an old decision made when the buffer was full and no one was thinking about the edge.

What a threshold asks of you

Seeing this shape should make you humbler, not cleverer. From inside a slow decline, no single seat can see how close the edge is. Not the dam engineers, not the seven states still arguing over the water, not the household whose power it is. Everyone is reading the same “still working” gauge, and the gauge is honest and useless at the same time.

Here is the whole. A system falls smoothly toward a line it can’t see. The people watching it have one piece of evidence — it’s been fine so far — and that is precisely the evidence a floor turns against them. Knowing that doesn’t tell you where every edge is. It tells you to stop trusting the trend. The number that matters is the distance to the floor — in the reservoir, and in a life — while there is still more than a foot to spare.

03 · Lab · your turn

The Distance to the Edge

Manage a falling reservoir year by year and feel how the power gauge reads fine right up to the foot where it fails, so the only safe move happens before anything looks wrong.

04 · Hope · carry this

The same forecast that shows the cliff also shows it a year out, measured to the foot — and a danger you can see coming while there is still margin to move is one people have, in the end, always found a way to steer around. Knowing exactly where the edge sits turns out to be the first thing that keeps you off it.