That plan was created in the 1950s and ultimately deemed unfeasible.
You wouldn’t make it all pipe, though. Like the CA Aqueduct, it would be a mix of open canal (which can cheaply have a large cross-section and correspondingly low losses). It transports ~1200 ft^3/sec but has a cross-section of something like 1500 ft^2 in the canal parts, so the flow is <1 ft/s.
Sure, there would be some pipe sections where the canal isn’t feasible. It’s hard to sure how much that would account for without planning out the whole route, but ideally you’d limit it to mountainous parts where you’re doing pumping anyway.
I certainly agree that almond growing in CA depends on absurdly cheap, subsidized water, and that costs would be far higher if they actually paid for the value of their water properly.
California residential water use is about 48 gal/capita-day. A penny per gallon would mean a water bill of $58/mo for a family of 4. Not trivial, but not going to break the bank, either.
Why not? Couldn’t the California legislature just pass a bill that metaphorically tells those farmers to “go suck on a bitter almond?” In other words sure, they keep their land, they just don’t get any more huge discount on water that should benefit everyone instead of a few farmers.
I played around with that a little, and while I was not surprised that a larger pipe decreases the losses, I was surprised at by how significant the effect was.
I was able to repeat the 10638 psi number for 5M feet, 48", 125 cfs, steel. Increase the diameter to 96", and the losses drop to 365 psi. That’s a 29x reduction! You can add a plastic liner to get down to 300 psi.
Despite being twice the surface area, that would probably require less steel, because the internal pressure would be almost negligible. And unlike oil or some other substances, you could get away with a less robust structure, given that there’s not much of an environmental problem with a leak.
I wasn’t aware that the Edmonston (?) Pumping Plant was a transcontinental project.
The difference in scales involved between city/regional systems that significantly take advantage of gravity and pumping water uphill across the Great Plains and Rockies with multiple ranges is non-trivial to say the least.
Note that the CA system loses a lot of water due to leaks and evaporation. Covering the giant canals is not going to be practical. Never mind drilling tunnels which is stupendously ridiculous at this scale.
The system is 700 miles. Not quite transcontinental, but it’s in the ballpark. It’s about half the 1500 mi figure that’s been thrown around.
It’s not trivial. But it’s on the same basic scale as things that we’ve been doing for decades. If it were three orders of magnitude as what’s been done, we could rightly ignore it as science fiction. But it’s not, it’s more like a single order of magnitude at most. From something built in the 60s.
Why not? It’s in the same ballpark of surface area as an interstate highway, except that the covering can be far lighter than a road surface. It’s only about a billion square feet. There’s no reason a thin cover can’t be under $1/ft^2.
In CA, there are proposals to cover the canals with solar panels. I’m not yet 100% convinced of the economics there, but they aren’t totally whack-a-doodle. And it solves both the energy and evaporation problem.
Once again, no one is factoring in the staggering amounts of water we’re talking about here.
When the Colorado River Compact was drawn up, it was based on a water flow of 15 million acre-feet per year (that figure was already overly optimistic.)
One acre foot of water = just under 326,000 gallons of water.
I am not an engineer, and I have a poor grasp of big numbers, so feel free to check my figures. But I’m pretty sure just one million acre feet of water is 326 billion gallons. Just to replace one-fifth of the water (3 acre feet) diverted from the Colorado would require close to a trillion gallons.
Can someone tell me how to move a trillion gallons of water per year? Especially when the lowest point across the Rocky Mountains is 6,800 feet higher than the elevation of Lake Superior.
Consider that the Colorado River doesn’t enter California to any significant extent. Since it doesn’t even approach LA, there is already an aqueduct system that moves about 400 billion gallons per year about 240 miles.
Now, that’s an easier problem than moving a trillion gallons across the Rockies. But it isn’t 1000x harder. Existing systems that have been around for a long time already transport hundreds of billions of gallons of water per year, over hundreds of miles, and crossing multi-thousand-foot mountains. It doesn’t require sci-fi technology.
To be clear, I’m not actually advocating building such a system. Just that despite the impressive numbers, it’s not anything particularly far-out. The arguments against it would be that it’s better to pursue efficiency, use desalination, etc.
We aren’t just crossing multi-thousand foot mountains, we’re pushing hundreds of millions of gallons water uphill for more than 800 miles. If we build a canal, it will intersect several major rivers. If we build pools at those intersections, we’ll need an elaborate system of dams at each point to regulate the flow of water. And we aren’t even considering the water losses due to evaporation, seepage, etc.
Dr. Strangelove and I aren’t actually debating the possibility of the giant project. But I think at some point the cost-benefit analysis will show that it’s cheaper just to close the Southwest United States and have everyone move back to the Rust Belt.
I reckon a couple of us are.
Yep. 326 billion US gallons or 271 billion imperial gallons … same volume of water.
Gravity and evaporation.
As I noted in my subsequent post, the water must go uphill.
Pumping horizontally for 800 miles and then a few thousand feet up isn’t much different from pumping up a gradual incline across 800 miles. It gets worse if there are lots of ups and downs, since you might need pumps and you can probably only recover about 80%, but some of that could use tunnels instead (the Colorado River Aqueduct has 92 miles of tunnels).
I agree with snowthx’s proposal that you just need to pump to the Colorado River basin, and let the existing system do the rest. There is already plenty of capacity along the way, given that the Colorado River mostly flows along a giant canyon. Some of the pumping losses can be recovered by existing dams. And the way things are looking, the whole system is under-capacity anyway, so this would just act to supplant it. Maybe some aqueducts would need to be beefed up here and there, but that’s easier than building an entire new system.
The Southwest has tremendous solar capacity and will provide clean energy for much of the country in a few decades. It seems like water would be a decent trade for that.
Yep, which is why you need evaporation to move the trillion gallons of water per year up hill.
The only truly mountainous part is from Fort Collins, CO to the headwaters of the Colorado River, about a 42-mile stretch. Other than that, I can’t think of much territory between Chicago and the Rockies where you’d be able to convince water to move very far west without a pump. Poking around on Google Earth, it looks like there’s about a 70-mile stretch in western IL where you run downhill as you move west toward the Mississippi River valley, and a 60-mile stretch in western IA where you run downhill as you move west toward the Missouri River valley, but that’s about it. Going west across NE and eastern CO, it’s pretty consistently uphill.
I’m a mechanical engineer, so I can understand the relationships between pipe size, flow velocity, and head loss. A civil engineer, especially one who does municipal water supply design for a living, would be better positioned to eyeball what’s considered a good compromise between installation costs (size/thickness of pipe), operating costs (maintenance, pump power), reliability, and carrying capacity (CFM). A flow velocity of 10 ft/s was my first guess, and it looks like that’s kind of in the right range for large-diameter pipes. This table shows recommended max flow velocity for pipe diameters up to 24 inches, at which the recommended max is about 14 ft/s. The recommended max for a 48-inch pipe is probably considerably greater, and suggests that 10 ft/s (and the 2.5 ft/s your 96" pipe would have) would mean the pipe is being underutilized.
That is very impressive. It demonstrates that water favors large scale.
No such thing as underutilized in an absolute sense. The propellant pipes on a rocket engine flow at hundreds of feet/sec. 10 ft/s is almost stationary in comparison. Electrical engineering is the same way. There are cases where sending multiple amps through a wire only as thick as a hair is the right answer. But you’d want 100x that for wiring your house.
At any rate, I agree that there is going to be some compromise based on the costs involved. Just eyeballing things, I suspect the right compromise is going to be between a 10,000 psi head pressure and 300 psi. Which end it’s closer to depends on things like the cost of electricity and steel.
The Colorado River Aqueduct has a 1500 ft^3/s capacity, and the main line has 16 ft tunnels. That’s 7.5 ft/s velocity, but the tradeoff might be toward lower speeds as the aqueduct gets longer.
It’s quite the drop. I expected more like a 16x reduction, since it’s 1/4 the velocity and these kinds of things tend to scale with velocity squared. But it’s better than even that. Maybe something to do with reduced shear forces as well.
Oh, I agree. Things like the 300-400 subsidiary projects and the six nuclear power stations just to run the pumps and « oh, by the way, we’ll need to move the city of Prince Rupert, BC » kind of put a damper on the plan. But it shows the scale of the problem.
Exactly this. The problem is that you are consistently going uphill, unless you are going double the distance or more where feasible to take advantage of contour lines… (I’ve seen irrigation canals that do that). With a consistent uphill, you need many regularly spaced pump stations - it’s not a matter of one big pump station to send it all the way, otherwise you end up with the absurdly high pressure requirements and a pipe consistently ready to leak, and canal coverings that need to be strong enough to contain pressure, not merely evaporation shields.
Then if you pump it down an existing riverbed to complete the journey, you have to consider what the additional flow would do to the overall environment. If it’s a significant amount, do you have to raise bridges? Move shoreline structures? Do you adjust for seasonal variation in the river itself? How much is lost to evaporation and ground seepage?
This is a really good example of people just not seeing the realities of the situation.
The weather in the zone from the Great Lakes to and thru the Rockies is very different from most of CA.
A “thin cover” won’t last either a winter of blizzards or a spring of thunderstorms. Never mind tornados.
And what is the water in the canal going to be like across hundreds of miles of -10 degree weather? How are pumps going to pump water uphill if the water is slushy or worse?
The old family farms of my youth were fed by irrigation canals from a reservoir a good ways off. And temps could get well under -20. But they drain them for the winter.
Turning off the water to the SW during the winter is a big problem since that’s the prime growing season for winter-time fresh produce. Okay, so you can pump during the good weather and store the water for later use in winter, right? Well, how does that affect your numbers if these canals are unused for a 1/3 of the year? How much bigger capacity will you need to send the same amount of water per year?
As to putting solar panels over canals. Not Well Thought Out. Solar panels are hot. They absorb a lot of sunlight. That’s going to add heat to the water. And since you can’t seal the whole thing up, this will increase evaporation.
Just put them on dry land. There’s a lot of that.