Hacker News Reader

Comment by no_wizard | Hacker News Reader

Could we feasibly just run desalination stations from the coasts and pipe the water - like oil - to fill up these reservoirs? Setting cost aside for a second, is there any reason not to do something like this?

View on HN

Anything is possible for enough money but I don't think you understand how much money what you're proposing would cost. In electricity alone desalination costs around 3kwh/m^3. 3.7mwh/acre foot of water. Last year we got 9.2 million acre feet of water in the Colorado River of the allocated 15 million acre feet. So if we say we want to supplement flows with just 2 million acre feet of desalination water, that's 7 Pwh of electricity. Per year. That's approximately 1000x the annual generated power of the largest nuclear plant in the world.

And we haven't even pumped that water 3700 feet up from sea level to the elevation of lake Powell.

This is a great comment. Thank you for doing the math. I thought the word "just" was doing a lot of work in the original comment.

Hm. Your numbers and mine don't match by a few orders of magnitude. I came up with 2.5GW sustained, which is (2.5GW * 24h * 365) 21900 GWh ~= 21TWh/yr. Big, but not the same as your value.

An acre foot is 1233 m^3, and at 3kWh/m^3, that's 3.7 MWh/acre foot. Math checks out so far.

Times 2000000, that's 7400000 MWh / 7400 GWh / 7.4 TWh.

I think you're off by a factor of 1000 somewhere in your math. You're looking at ~the output of a large nuclear plant, not 1000 of them.

Or my math is wrong. That's entirely possible too.

I get your numbers. I think OP skipped a prefix somewhere in the mega-giga-tera-peta hierarchy.

To put it in perspective, Bitcoin mining alone is more than three times that.

> That's approximately 1000x the annual generated power of the largest nuclear plant in the world.

Put another way, it's almost 40x the annual generated power of the State of California.

We don’t need to pump the water to lake Powell. We let Powell refill itself and use the water instead on level ground in California.

I mean, if you don't care in the slightest about cost of the plants, the energy required, the land needed to run the giant tunnels... maybe? It's probably technically possible within the constraints of "I have a full country's resources available and can demand they be used to satisfy my absurd requirements and won't face huge blowback for using resources this way."

But you can pencil it out. Cuts next year are 2-4 million acre feet of water. What would it take to bring that in?

An acre foot is 325,851 gallons, give or take. So, four of those is ~1.3e12 gallons of water for a year. A year has 31,536,000 seconds, so you're looking at a mere 41,000 gallons per second, or about 6500 cu*ft/s. Which is a good sized river's flow.

It's also 155 m^3 of water per second, or about 3.5 billion gallons a day.

I'm seeing [0] a plant in San Diego running 50M gallons a day on 35MW, so you'd need ~70 of those plants (or about 2.5GW) to manage your water supply. California's purring away at 33GW right now, so the energy required, while massive, is feasible.

Trying to figure out how to move 40k gallons a second a distance of 1000 miles exceeds my physics skills, but you could probably figure it out with the right calculators or CFD handwaves. But it's not going to be cheap.

Seems... somewhat easier to reduce demand, though.

[0]: https://www.energy.gov/sites/default/files/2019/09/f66/73355...

40k gallons per second is on the order of the capacity of the Edmonston Pumping Plant of the State Water Project, so it's likely doable with existing technology.

And if you're doing mega projects on that scale, you'd just pipe water over the rockies from the midwest or divert the Columbia down the California coast.

Or cut into the Gulf of Mexico and flood Death Valley.

Generally it would be cheaper to use more efficient farming methods.

The real solution is to renegotiate the Colorado River Compact acknowledging that historical flow is several million acre feet less than what is allocated. Desalination and pumping is prohibitively expensive and would require as much interstate cooperation as reopening the original deal. And at the end you have hundreds of millions of gallons of brine to deal with.

Setting cost aside? No. It costs a lot because it requires lots of energy which probably means lots of fossil fuels burning.

There’s other ways to mitigate that cost. Israel for instance runs many desalination plants from if I recall correctly solar power. We could do the same with solar and wind.

Never the less, yes, because I believe cost can be managed and figured out over time, where as I don’t know what other mitigating factors would be the consequence of such a choice

Do you have a source on the solar power Israel desalination? They use natural gas and coal.

Israel also recycles 90% of it's water. It first goes to urban use, then it's recycled for agriculture. California could do the same. https://www.waterworld.com/wastewater/article/16202781/israe...

If only the cities were before the farms instead of after - reuse of city water for agriculture is a great idea.

It would require a tremendous amount of energy. Lake Powell is more than 1km above sea level, which means each liter of water would require 10kJ just in gravitational potential energy.

The reclamation effort is 1 maf (million acre-feet) of water over a year, which if supplied continuously from sea level with perfectly efficient pumping and transport corresponds to ~10 GW of power. A typical nuclear power plant generates 1GW, so without any transport losses, you'd need 10 new plants running full-time just to move the water up the hill.

Add in the desalination challenges and transport losses, and you're talking what, 20-40 new nuke plants?

> Setting cost aside for a second, is there any reason not to do something like this?

The cost would be so massive that, even if the answer to that question is "no", we shouldn't do it.

We call them “desalination” plants, but the salt doesn’t just disappear. The more fresh water you have them produce, the more salt (brackish water) you need to manage as well.

So even while you may imagine the ocean as a limitless supply of water, there’s only so far you can scale desalination.

To be fair the ocean is still pretty big. Also, water doesn’t disappear from the earth…

Why not pipe the water from the pacific northwest, similar to what was done with the Colorado river.

Lots and lots of mountains. On the east sides of the Cascades are massive dry and drought stricken areas (which are also farmland too). This same question can be asked about simply piping water from one side of the cascades to the other, let alone to the central valley

Cough... https://en.wikipedia.org/wiki/North_American_Water_and_Power... ...cough

This is done in China:

https://en.m.wikipedia.org/wiki/South–North_Water_Transfer_P...

What makes you think there's excess water there?

California already has a version of this. It is called the California Aqueduct. It takes water in the north and sends it to the south.

https://en.m.wikipedia.org/wiki/California_Aqueduct

You can see a portion of it driving on I-580.

The problem is that Northern California hasn't been getting enough water either. The rains have been falling further north. Think Seattle.

Also the aqueduct is fed from the existing river system. The major rivers in Oregon and Washington don't go into California AFAIK.

Looks like you would have to divert the Columbia River. Haha.

You wouldn't have to pump it back up to refill the reservoirs: just slow the release from the reservoirs by the same amount you replace we desalinated water.

I just want to mention, that, upon reading the comments, I now understand why this is a terrible idea

Yes. Extremely high energy usage.

That would take 10-20 years to build out...what will you do next year?

Bottled water. Very big bottles.

Reply on news.ycombinator.com