Super interesting article. I'd just add that transporting water from desalinization is a big problem. Desalinization stations are at sea level, so water needs to be pumped all the way to final consumers. This would certainly increase the water cost, which consumers usually are not very willing to pay.
Turns out most of the world population at a low enough altitude above sea level to make it affordable to transport the water, as 94% of the world's population is below 1,600 meter altitude. The transport costs equals desalintion costs at 2000 meter altitude or 1,,600 km horizontal distance at least according to a paper from 2005, and its probably not changed significantly since then.
In detail:
This may help, an old article from 2005. It found that transport costs are less than desalination costs if the water is transported less than 1600 km horizontally or 2000 meters vertically. It doesn't look like desalination costs have changed much since 2005. So I expect this calculation still applies. today.
QUOTE STARTS
Transport costs are assumed to be 6 cents per 100 km horizontal transport plus 5 cents per 100 m vertical transport.
The costs of desalination, here assumed to equal 100 cents/m3, are typically larger than the costs of transport. Indeed, one needs to lift the water by 2000 m, or transport it over more than 1600 km to get transport costs equal to the desalination costs.
So I expect most of the world's population could be supplied with water at at most double the cost of desalination.
I expect the desalination costs would get lower once we have lots of excess energy from renewables to use. That would just reduce the costs for the calculation.
More renewables will increase the price of energy, not decrease it. What good is a wind turbine powered water plant if there's no wind, or a solar powered plant at night? Renewables require subsidies and either backup generation or huge amounts of storage, both of which drive up the cost hugely.
Really interesting. I always thought that desalination was expensive too. I will make the same excersice for Spain, because some regions are water stressed (and it will get worse in the future)
I found a different process (not mature yet) to desalinate water with much less energy. It's been developed in Australia. It is only suitable for agricultural purposes, but maybe a good adaptation measure for the future
I think the key question is the capital cost of the desalination equipment.
If the capital cost is low, then desalination plants can be run off very cheap solar power, 8 hours per day. As a flexible load the solar power cost will be close to nothing - especially given that most desalination is needed in arid sunny regions.
If the capital cost is high, or the operation inflexible, then the equipment will have to run at night, using expensive electricity from gas, or coal, or a battery.
It is the same issue with electrolysis to make hydrogen. If something can run only when energy is cheap, it has a major commercial advantage.
Wind, including offshore wind with much bigger capacity factor, & largely complementary peaks compared to solar over a day & a year, is being built many places. Wires go even more, & batteries can be put anywhere. In most of the world, all 3 are now cheaper together than any other energy source, & still getting cheaper fast.
Geothermal makes up a significant part of the grid in Kenya (51%), Nicaragua, Costa Rica, Philippines, New Zealand, & there’s some in Italy, Mexico, Turkey & elsewhere. The US has the most absolute amount installed though it’s only 0.3% of its grid, & a number of those countries, especially NZ & Nicaragua, also use it as primary energy. Lots of those places can desalinate usefully & more will in the future.
Casey Handmer blogged on this a a couple of years back. His key assumption was that desalinization would use cheap solar electricity. His calculation was 32 sq m of panels would be enough to meet all the water needs of a person (including their share of agriculture use etc.)
Yeah this seems critical. Nobody is plugging their desalination plants into residential electricity outlets! I think $50/MWh is a pretty reasonable price, not $90-130/MWh.
I know you do not identify waste brine in your calculation but it would be a significant disposal problem as it would be so concentrated and should not be sent back to the aquifer or ocean in the volumes noted on a global basis. My experience with skid mounted RO units in factories showed 15-25% waste for potable water produced. I have read the ratio in desalination plants is 2:1 (brine waste to water produced). This would pose a significant source and disposal volume problem for world scale plants.
Anybody dealing with this problem on a sustainable basis?
In theory, by pulling from ocean water sources there would be less water required from fresh water sources, which would increase fresh water outflows to the sea, so bulk effects from removing water would be unlikely.
Realistically though, water consumption would increase if it was accessible and low cost in the regions where desalination occurred, so that’s not necessarily a safe assumption.
If necessary, I think it would be possible to divert fresh water from major rivers, mix it with brine, and reinfect it in already brackish areas, or just allow it to flow out to the ocean. That would add cost, but wouldn’t be unimaginable
Using seawater from once through cooling for NPPs or big industrial users, and doing deep ocean return flows is the ideal source and destination for this.
This would also be the type of stream to do direct ocean capture of carbon from, and this is actually where thermal desalination comes in handy - from waste heat, not from dedicated sources.
There’s desalination brine mining, extracting valuable minerals that are already concentrated in the process, which potentially makes the brine less toxic
One of the more cost-effective ways of desalination is using waste heat recovery. The heat rejection from a 1MW diesel generator, for example, can generate 5 cubic meters of water per day.
It will be important to consider thermal desalination if we can use heat from nuclear. Both waste heat and primary heat when electricity demand is low could change the economics on thermal desal.
When water will finally be expensive people will stop wasting it. The fact that a household in the US consume 3 times more than in the UK is just a demonstration of where the problem really is.
Btw, there are a lot of techniques to recycle water and the level of cleanliness for drinking vs showering vs flushing is not identical (https://en.wikipedia.org/wiki/Greywater). The worst things for water are fecal contamination and salt (not talking about chemical pollution, etc).
For example, in the Philippines, they distribute buckets with a filter system to population in the mountains where water is not cleaned properly. These filters can remove 99.999999% of bacteria and viruses and their life span is around 1M gallons. And they cost $35: https://www.wavesforwater.org/
There are also many techniques to collect water from roofs, filter it and reuse it at a local level. In Mexico they have these type of collection systems: https://islaurbana.org/en/
So the technologies already exist, for saving, collecting, recycling/filtering. As much as someone can think about installing solar panel on their roof they can also think about installing a water collection system. More expensive water will make these systems more meaningful, just like solar panels. It's will be interesting to see how this is going to.
Useful and interesting. However, the US is (unsurprisingly) very inefficient in its agricultural water usage. To get a better sense of the realistic economic constraints for agricultural desalination, it might be better to use data from Israel.
> The osmotic pressure of ocean water is approximately 27 atm.
In SI units, that is approximately 2.7 × 10^6 Pa. Therefore, under ideal conditions, the work needed to force one cubic metre of ocean water through the membrane is 2.7 × 10^6 Pa m^3 = 2.7 × 10^6 J = (2.7 × 10^6 / (3.6 × 10^6)) kW h = 0.75 kW h.
Trying to get water desalinated with an energy cost lower than that would be like trying to lift a stone with less work than you get by letting the stone go back to its starting position. In fact, it would enable you to build a perpetual-motion machine, by letting the fresh water fall about 270 metres, going through a turbine, which would extract the 0.75 kW h per cubic metre, and then letting osmotic pressure suck it through the same kind of membrane up to ocean level.
I had no idea the osmotic pressure of ocean water was so large. Since I’d always seen osmosis illustrated with what looked like small tubes one could find in a laboratory, I expected it to lift the salt water a few decimetres at most above the fresh water, rather than an amazing 270 metres. Now I understand why desalinating water with hydroelectric power will probably never be practical for seaflooding (<https://unchartedterritories.tomaspueyo.com/p/seaflooding?utm_source=publication-search>).
Glad that you added the section on agriculture. Few people realize that household usage of water is tiny compared to agriculture. Even within metro areas watering landscaping uses much more water than household use.
If there were a way to have separate distribution systems, then this would not matter, but I don’t see how we will ever get to that.
Desalination costs still have a long way to go, but at least they are going in the right direction.
I've grown weary of certain "energy abundance" advocates (Matt Yglesias, Ezra Klein, Eli Dourado) who glibly bring up desalination as a reason why the world needs to increase its total energy use tenfold or more. None of them bother to do the math, because they don't actually care about desalination; they're really just promoting increased energy consumption for its own sake. As this article shows, desalination is not being held back by energy scarcity and will never add more than a few percent onto total energy use, even in desert locations where it's an important source of fresh water.
Out of curiosity, what level of fresh water catastrophe would see us desalinating enough to counteract the dilution occurring, due to glacier melt (assuming that the excess salt is put back in the sea)? At what level would desalination affect ocean acidification, if at all?
Which is to say, I suppose, are there other upsides to desalination beyond smoothing out potable drinking water supplies?
Great article. My two questions would be on the trajectory of energy $ costs and carbon cost of the water. How falling renewables costs might effect the ultimate cost per litre and carbon cost per litre in the future. Both improved dramatically in recent years and expected to continue to do so. Desal is increasingly powered by renewables.
Would these declines be enough to dramatically change the game in terms of adoption and if so, when?
So, 50 litres per day would require 175 Watts of energy, more than achievable with a solar panel. Reverse osmosis is useful for polluted water also, not just sea water, filtering out bacteria and dissolved solids. https://www.researchgate.net/publication/284804889
Super interesting article. I'd just add that transporting water from desalinization is a big problem. Desalinization stations are at sea level, so water needs to be pumped all the way to final consumers. This would certainly increase the water cost, which consumers usually are not very willing to pay.
Turns out most of the world population at a low enough altitude above sea level to make it affordable to transport the water, as 94% of the world's population is below 1,600 meter altitude. The transport costs equals desalintion costs at 2000 meter altitude or 1,,600 km horizontal distance at least according to a paper from 2005, and its probably not changed significantly since then.
In detail:
This may help, an old article from 2005. It found that transport costs are less than desalination costs if the water is transported less than 1600 km horizontally or 2000 meters vertically. It doesn't look like desalination costs have changed much since 2005. So I expect this calculation still applies. today.
QUOTE STARTS
Transport costs are assumed to be 6 cents per 100 km horizontal transport plus 5 cents per 100 m vertical transport.
The costs of desalination, here assumed to equal 100 cents/m3, are typically larger than the costs of transport. Indeed, one needs to lift the water by 2000 m, or transport it over more than 1600 km to get transport costs equal to the desalination costs.
https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2004WR003749
94% of the world's population lives below one mile, or a bit over 1,600 meters above sea level acording to the cartographer Willaim Rankin:
https://urbandemographics.blogspot.com/2017/06/world-population-distribution-by.html
Not many places would be 1,600 km from the sea.
So I expect most of the world's population could be supplied with water at at most double the cost of desalination.
I expect the desalination costs would get lower once we have lots of excess energy from renewables to use. That would just reduce the costs for the calculation.
More renewables will increase the price of energy, not decrease it. What good is a wind turbine powered water plant if there's no wind, or a solar powered plant at night? Renewables require subsidies and either backup generation or huge amounts of storage, both of which drive up the cost hugely.
Really interesting. I always thought that desalination was expensive too. I will make the same excersice for Spain, because some regions are water stressed (and it will get worse in the future)
I found a different process (not mature yet) to desalinate water with much less energy. It's been developed in Australia. It is only suitable for agricultural purposes, but maybe a good adaptation measure for the future
https://www.anthropocenemagazine.org/2024/05/simple-desalination-tech-needs-just-a-dash-of-heat
Using solar power, or using less solar power with heat pumps?
I think the key question is the capital cost of the desalination equipment.
If the capital cost is low, then desalination plants can be run off very cheap solar power, 8 hours per day. As a flexible load the solar power cost will be close to nothing - especially given that most desalination is needed in arid sunny regions.
If the capital cost is high, or the operation inflexible, then the equipment will have to run at night, using expensive electricity from gas, or coal, or a battery.
It is the same issue with electrolysis to make hydrogen. If something can run only when energy is cheap, it has a major commercial advantage.
Wind power. Geothermal. CSP. Hydro. PHS.
Ultimately it matters not. As long as the price of electricity is below X, water can be desalinated.
But in those dry sunny regions, solar PV is the technology that will drive prices lower.
Hydro and wind and geothermal do drive the prices down in Iceland, Scandinavia and Canada, but these places don't need desalination.
Wind, including offshore wind with much bigger capacity factor, & largely complementary peaks compared to solar over a day & a year, is being built many places. Wires go even more, & batteries can be put anywhere. In most of the world, all 3 are now cheaper together than any other energy source, & still getting cheaper fast.
Geothermal makes up a significant part of the grid in Kenya (51%), Nicaragua, Costa Rica, Philippines, New Zealand, & there’s some in Italy, Mexico, Turkey & elsewhere. The US has the most absolute amount installed though it’s only 0.3% of its grid, & a number of those countries, especially NZ & Nicaragua, also use it as primary energy. Lots of those places can desalinate usefully & more will in the future.
Casey Handmer blogged on this a a couple of years back. His key assumption was that desalinization would use cheap solar electricity. His calculation was 32 sq m of panels would be enough to meet all the water needs of a person (including their share of agriculture use etc.)
https://caseyhandmer.wordpress.com/2022/11/20/we-need-more-water-than-rain-can-provide-refilling-rivers-with-desalination/
Yeah this seems critical. Nobody is plugging their desalination plants into residential electricity outlets! I think $50/MWh is a pretty reasonable price, not $90-130/MWh.
I know you do not identify waste brine in your calculation but it would be a significant disposal problem as it would be so concentrated and should not be sent back to the aquifer or ocean in the volumes noted on a global basis. My experience with skid mounted RO units in factories showed 15-25% waste for potable water produced. I have read the ratio in desalination plants is 2:1 (brine waste to water produced). This would pose a significant source and disposal volume problem for world scale plants.
Anybody dealing with this problem on a sustainable basis?
A very salineinent question
In theory, by pulling from ocean water sources there would be less water required from fresh water sources, which would increase fresh water outflows to the sea, so bulk effects from removing water would be unlikely.
Realistically though, water consumption would increase if it was accessible and low cost in the regions where desalination occurred, so that’s not necessarily a safe assumption.
If necessary, I think it would be possible to divert fresh water from major rivers, mix it with brine, and reinfect it in already brackish areas, or just allow it to flow out to the ocean. That would add cost, but wouldn’t be unimaginable
Using seawater from once through cooling for NPPs or big industrial users, and doing deep ocean return flows is the ideal source and destination for this.
This would also be the type of stream to do direct ocean capture of carbon from, and this is actually where thermal desalination comes in handy - from waste heat, not from dedicated sources.
There’s desalination brine mining, extracting valuable minerals that are already concentrated in the process, which potentially makes the brine less toxic
Good analysis.
One of the more cost-effective ways of desalination is using waste heat recovery. The heat rejection from a 1MW diesel generator, for example, can generate 5 cubic meters of water per day.
It will be important to consider thermal desalination if we can use heat from nuclear. Both waste heat and primary heat when electricity demand is low could change the economics on thermal desal.
From big industrial facilites as well...
When water will finally be expensive people will stop wasting it. The fact that a household in the US consume 3 times more than in the UK is just a demonstration of where the problem really is.
Btw, there are a lot of techniques to recycle water and the level of cleanliness for drinking vs showering vs flushing is not identical (https://en.wikipedia.org/wiki/Greywater). The worst things for water are fecal contamination and salt (not talking about chemical pollution, etc).
For example, in the Philippines, they distribute buckets with a filter system to population in the mountains where water is not cleaned properly. These filters can remove 99.999999% of bacteria and viruses and their life span is around 1M gallons. And they cost $35: https://www.wavesforwater.org/
There are also many techniques to collect water from roofs, filter it and reuse it at a local level. In Mexico they have these type of collection systems: https://islaurbana.org/en/
So the technologies already exist, for saving, collecting, recycling/filtering. As much as someone can think about installing solar panel on their roof they can also think about installing a water collection system. More expensive water will make these systems more meaningful, just like solar panels. It's will be interesting to see how this is going to.
Useful and interesting. However, the US is (unsurprisingly) very inefficient in its agricultural water usage. To get a better sense of the realistic economic constraints for agricultural desalination, it might be better to use data from Israel.
> If we wanted to get energy use lower than 1 kWh per m3 we’d need to innovate and develop a different process.
As far as I understand it, we’d actually need to break the laws of physics to go significantly below that.
According to Wikipedia (<https://en.m.wikipedia.org/wiki/Osmotic_pressure#Applications>),
> The osmotic pressure of ocean water is approximately 27 atm.
In SI units, that is approximately 2.7 × 10^6 Pa. Therefore, under ideal conditions, the work needed to force one cubic metre of ocean water through the membrane is 2.7 × 10^6 Pa m^3 = 2.7 × 10^6 J = (2.7 × 10^6 / (3.6 × 10^6)) kW h = 0.75 kW h.
Trying to get water desalinated with an energy cost lower than that would be like trying to lift a stone with less work than you get by letting the stone go back to its starting position. In fact, it would enable you to build a perpetual-motion machine, by letting the fresh water fall about 270 metres, going through a turbine, which would extract the 0.75 kW h per cubic metre, and then letting osmotic pressure suck it through the same kind of membrane up to ocean level.
I had no idea the osmotic pressure of ocean water was so large. Since I’d always seen osmosis illustrated with what looked like small tubes one could find in a laboratory, I expected it to lift the salt water a few decimetres at most above the fresh water, rather than an amazing 270 metres. Now I understand why desalinating water with hydroelectric power will probably never be practical for seaflooding (<https://unchartedterritories.tomaspueyo.com/p/seaflooding?utm_source=publication-search>).
Typo nitpicking:
> But when suffer from drought,
Isn’t a subject missing there?
Glad that you added the section on agriculture. Few people realize that household usage of water is tiny compared to agriculture. Even within metro areas watering landscaping uses much more water than household use.
If there were a way to have separate distribution systems, then this would not matter, but I don’t see how we will ever get to that.
Desalination costs still have a long way to go, but at least they are going in the right direction.
Thank you for this.
I've grown weary of certain "energy abundance" advocates (Matt Yglesias, Ezra Klein, Eli Dourado) who glibly bring up desalination as a reason why the world needs to increase its total energy use tenfold or more. None of them bother to do the math, because they don't actually care about desalination; they're really just promoting increased energy consumption for its own sake. As this article shows, desalination is not being held back by energy scarcity and will never add more than a few percent onto total energy use, even in desert locations where it's an important source of fresh water.
Love this thought experiment. Brilliant! 💡
Out of curiosity, what level of fresh water catastrophe would see us desalinating enough to counteract the dilution occurring, due to glacier melt (assuming that the excess salt is put back in the sea)? At what level would desalination affect ocean acidification, if at all?
Which is to say, I suppose, are there other upsides to desalination beyond smoothing out potable drinking water supplies?
Great article. My two questions would be on the trajectory of energy $ costs and carbon cost of the water. How falling renewables costs might effect the ultimate cost per litre and carbon cost per litre in the future. Both improved dramatically in recent years and expected to continue to do so. Desal is increasingly powered by renewables.
Would these declines be enough to dramatically change the game in terms of adoption and if so, when?
Interesting thank you Hannah.
So, 50 litres per day would require 175 Watts of energy, more than achievable with a solar panel. Reverse osmosis is useful for polluted water also, not just sea water, filtering out bacteria and dissolved solids. https://www.researchgate.net/publication/284804889