Can Solar Powered Desalination Save the World?
In a previous post I presented numbers which show that the only chance humanity has of avoiding a Malthusian Collapse and surviving the next century as fossil fuels run out is if some other energy source can step up to the plate. We require this to continue performing the essential functions on which our societies – and our lives – are critically dependent: fertilizer manufacture, food production, provision of fresh water, a functioning food distribution network, essential transportation, basic electrical needs for everyone so we can maintain social order, and a somewhat intact manufacturing sector enabling these basic technological functions to be maintained indefinitely.
Fresh water and food are of course the most basic of those needs, and they need to be top priority. The numbers highlighted so far indicate an intriguing possibility, in that the amount of solar energy striking the Earth is far greater than we could ever possibly need. The technology for capturing that energy exists today, is dropping in price rapidly, is not complex, can be scaled anywhere from very small to very large, and because of all this, everyone can have access to it in their backyard.
Along with this, most of the planet is covered with water, and sea water can be turned into fresh water using energy. And conveniently, a large portion of the planet’s land area is barren desert or scrub which is not very productive for providing food despite warm temperatures and ample sunshine. The one missing ingredient hindering that land’s productivity is water. What if we were somehow able to bring those three ingredients together – limitless energy from the sun, desalinated sea water, and huge areas of barren deserts – to create new agricultural land that could provide the food that might be able to save humanity?
I have previously pointed out studies showing that the total Net Primary Production (NPP) of the planet has gone down by 10% despite all of our modern agricultural advances, largely because our agriculture simply displaces natural ecosystems with domesticated ones; one kind of NPP is merely replaced with another, subsidized using fossil fuels. But if we were to green up the Sahara, the Middle East, southern Africa, coastal Chile/Peru, Outback Australia, and other currently unproductive expanses then this would be new NPP we’d be creating. We may even be able to maintain global NPP at current levels as fossil fuels run out, and possibly even increase it!
Fortunately, regions that would have the greatest potential agricultural production with the assistance of external irrigation (i.e. hot deserts), also have the greatest potential for solar energy capture. And what’s also fortunate is that solar and wind energy’s inherent variability isn’t so much of a problem for agricultural uses since cloudy spells spanning days or even weeks could be relatively easily buffered with fresh water storage (much easier than energy storage), and irrigation could be curtailed for a few days if needed without too much difficulty. Furthermore, when it’s cloudy out we won’t have such high transpirational demand for water from crops.
It seems at first cut to be a match made in heaven. Here I go through a numerical analysis to investigate if this is a reasonable possibility or just fantasy. I suspect beforehand, as with electric transportation, that the answer to this from a technical and ecological perspective will be, “yes”, but may be tempered with some strong “buts” given how we humans tend to collectively behave, and how we have a disturbing history of taking the opposite actions of what prudent planning would suggest is in our own best interest.
How Does Desalination Work?
There are a variety of processes that can produce fresh water from saline water, and they basically boil down to (pun intended) these:
1) Multi-Stage Flash (MSF). Heat up and boil the water through a series of pressure vessels, then condense the steam. The heat from condensing the steam is used to pre-heat the incoming water. This is a useful process when used in conjunction with thermal electricity power plants. The second law of thermodynamics says that when you produce electricity using heat flow from hot to cold (with the heat coming from burning fossil fuels or splitting atoms, and the cold being a lake or ocean located next door), only a portion of that flow of heat energy can be turned into electricity. Practically, only 40% (for coal) to 60% (for natural gas) can be converted into electrical energy because the magnitude of the difference in temperature dictates the efficiency of the process and for practical engineering reasons (you don’t want to melt your pipes) there is an upper limit to the hot side. Using this waste heat flow to do something else useful besides produce electricity (like heat New York City through underground pipes) is an example of “cogeneration”, which takes advantage of the inevitable waste heat that would otherwise be dumped into the ocean, and instead uses it to pre-heat and boil water. Because MSF is intimately connected to thermal power plants, it is most often used in places like Saudi Arabia that have abundant fossil fuels to power the electricity generation regardless, and which are located right next to the sea for thermal dumping. Its advantage is that it is not critically dependent on the quality of incoming water.
2) Multi-Effect Distillation (MED). Use similar thermal processes as above to play with pressures and force the water to boil at lower temperatures, without the external heat input. The boiling temperature of water drops as the pressure decreases so if it boils at 100oC at atmospheric pressure, you can get it to boil at room temperature by dropping the pressure to something like 1/50th of atmospheric pressure. It is more efficient than MSF.
3) Reverse Osmosis (RO, or SWRO for “Sea Water Reverse Osmosis”). This uses high pressure pumps to force water through a semi-permeable membrane which allows water to pass through but not salts. It’s called “reverse osmosis” because the two sides of the membrane want to equilibriate in salinity but we provide external energy to force them to diverge. Sea water has an osmotic pressure, which is the pressure required to separate water from its dissolved salts. This is the force required to push water through the membrane. It is 30 bars (or 30 atmospheres of 14.7 psi each), which is 440 psi total. This will begin the process of reverse osmosis. In practise, however, about 1,000 psi is required to get reasonable flow. This leads into one of the disadvantages of RO, which is that the water must be crystal clear to avoid fouling of the membrane. The pre-treatment to achieve this can be expensive.
And because the water on the salty side gets saltier as water passes through the membrane, the osmotic pressure would tend to rise and then require greater pressure, and therefore greater energy use. To prevent this from occurring and to maintain the salinity optimally low, a constant supply of new sea water must flow by the membrane. This necessitates a brine discharge flow from the pressure side of the membrane back out to the ocean. This amounts to about 58% brine discharge back to the sea and 42% fresh water through the membrane, which means you need to suck up over twice as much sea water as gets desalinated. This saline discharge back out to sea is required for all the other processes as well, not just RO.
The saline outfall must be located a distance from the intake pipe since you don’t want to be sucking up your saline discharge water again! You also might be wondering: Won’t the ocean get continually saltier as a result? That depends. If the intake is in the open ocean then, no, it won’t, because that fresh water that got desalinated will become part of the hydrological cycle and be recycled back to the ocean via rain or river runoff. Furthermore, the ocean has chemical processes that prevent it from getting too salty – it precipitates out salt.
On the other hand, in places with limited circulation with the open ocean and tons of evaporation from desert climates, like the Persian Gulf, this can indeed be a problem. In fact, Saudi Arabia and the surrounding small countries have desalinated so much water from the Persian Gulf that it now has a salinity up to 60 parts per thousand, or almost twice that of standard sea water.
RO and MSF are the most common types of water desalination processes. RO is most suitable for locations that do not have immediate access to large amounts of fossil fuels as sources of heat. It only requires electricity (and sea water) to run. It can be expected to increase in prominence as we run out of fossil fuels since we won’t have that easy access to heat sources we currently enjoy from burning fossil fuels. That heat could be provided by solar thermal plants but this would require that the solar farm be located adjacent to the desalination plant, which would be prohibitive for most locations. By contrast, electricity can be transported long distances from inland deserts to coastal desalination plants. Of course, nuclear power plants could be located next to the ocean but we all know how oceanside nuclear power plants fare in earthquake / tsunami events…
The above desalination techniques are not independent of each other and often a facility will use a hybrid to take maximum advantage of the various forms of waste energy available from the fossil-fuel-to-electricity process. For example, RO and MSF are often paired up.
4) Solar Stills. These are the simplest of desalination setups and can vary in scale from tiny to large. They simply flow sea water through some sealed chambers under glass in the sunlight, and the greenhouse effect evaporates some water. This is condensed somewhere else and in doing so produces fresh water. They have the advantage that they are relatively cheap and low tech, meaning that local people with limited technological capabilities can maintain them. Their disadvantage is their large area footprint per volume of fresh water produced. This paper goes through the many different direct solar desalination techniques available.
5) Electro-dialysis. An electrical current passing through the sea water causes ions to move through an electrodialysis stack consisting of alternating layers of cationic and anionic ion exchange membranes, leaving purer water behind.
How Much Land Would Be Needed to Double Food Production?
Let’s say that in order to save the world from a Malthusian Collapse we will have to produce the same amount of food as we do now, but over again. How do I arrive at that figure? Well, it sounds reasonable and it makes for a good base for calculations. It should account for the difference needed to make up for decreased productivity of current farmland when fossil fuels run out, plus additional food demand from more people. I previously estimated our food production to be 1 cubic mile of oil equivalent energy per year. I will not account for the rest of the plant matter that must be produced along with food like leaves, stems, and roots, because the productivity figure I’ll use assumes it’s an Iowa corn field or equivalent, which produces 10 tonnes of corn product per hectare per year (it’s about the same for rice). Since this new desert agriculture will presumably be happening in the tropics and subtropics, we will therefore enjoy year-round productivity, so I’ll double that to 20 t/ha/yr.
If we multiply 20 t/ha/yr by the ratio of 17/45, or the energy ratio between carbohydrates and oil, we get 60 barrels of oil equivalent energy (in the form of plant tissues) produced per hectare per year.
Since the world consumes 1 cubic mile of food per year, at 26 billion barrels per cubic mile we’d need 500 million hectares of land to do this. This works out to an area 1,000 km by 5,000 km. That seems reasonable if you consider the vast expanses in Africa, Arabia, Australia and coastal Chile / Peru. It’s actually about a third the size of the Sahara. Not too tall an order! Additionally, this desalinated water could be used on existing marginal agricultural areas to boost their productivity, which wouldn’t be new agricultural land. So we aren’t going to have to colonize Mars to find that much barren desert, after all.
How Much Water Would this Require?
From this source it appears that we would need 1.5 liters per second per hectare for an arid climate. This works out to 5 meters of “rain” per year, which interestingly is well beyond what would sustain a rainforest. The dreary Pacific Northwest typically only gets between 1 and 3 m per year, and the wettest place in North America, Henderson Lake on Vancouver Island, gets 6 m a year. The wettest place in the world, in the Choco province of Colombia’s Pacific coast, gets up to 10 m per year. Vancouver gets 1.5 m and Victoria 0.6 m (Victoria is in the rain shadow of the Olympic Mountains, like Seattle). Chicago gets 1 m, New York a little over a meter, and Lima, Peru gets none (interesting how the wettest and driest places on Earth are separated by only a few thousand kilometres of South American coastline).
This shows how dependent our agricultural systems are on artificial irrigation. 5 m per year is 50,000 m3/ha which works out to 25 trillion cubic meters of water a year for all of this agricultural land.
How Much Energy Would this Require?
It takes about 3 kWh to produce a cubic meter of desalinated water (for perspective, this is how much energy one of those little portable plug-in space heaters would go through in 2 hours).
So to desalinate enough water for a hectare we’d need 500,000 megajoules (MJ) per year. To bring it into oil equivalent energy (ignoring thermal efficiencies; this is just for accounting purposes), this gives 90 barrels of oil per hectare per year of required energy to desalinate the water. For 500 million hectares, we’d therefore need 40 billion barrels of oil equivalent energy a year.
This is only the energy needed to desalinate the water, but we’d also need to pump it hundreds of kilometres inland and up in elevation to get it to where it’s needed. So I will double this and assume we need 80 billion barrels of oil. This is just 3 cubic miles of oil equivalent energy per year. That’s not too far into outer space since we currently use 3 cubic miles of “technical” energy per year right now in the forms of coal, oil and gas, with a little bit of hydro, wind, solar and nuclear thrown in there too.
We also need to add in all the other energy required to produce food, like to manufacture fertilizers and to drive machinery around. This is 18 GJ/ha for an Iowa corn field, which is fairly minor compared with our desalination requirements of 500 GJ/ha. I will consider it a rounding error, as well as all the other energy needed to process, store, and distribute that food (which is popularly accounted for in the, “It takes 7 Calories of fossil fuel inputs to provide 1 Calorie of food on your plate” statistic), since we’d have most of that infrastructure already in place anyways; we’d just be getting food from a different source.
How Much of an Increase in Desalination Capacity are We Talking About?
The world currently desalinates over 60 million m3 per day which is 22 billion m3 per year. This is about the same as the flow in New York’s Hudson River, or ¼ the flow of the Nile, 1/25th that of the Mekong, and 1/300th the flow of the Amazon.
So to produce 25 trillion m3 a year we’d need to increase desalination capacity by 1,000 times. Not a small feat, but not out of this world either. Here is the recent trend in global desalination capacity:
It’s been growing by a consistent 16% annual rate over the last couple decades. If that trend continues, we’ll be where we need by 2060! No problem! Right when we’ll presumably be running out of fossil fuels and be needing that desalination! Of course, we’ve all been warned of the dangers of ruthless extrapolation, and this exponential growth in desalination capacity could be somewhat constrained at some point by energy scarcity. That’s where (hopefully) solar power comes in! (And nuclear and wind too, I guess…)
A typical large desalination pant today produces about 4,000 cubic meters per hour. For perspective, 4,000 m3 per hour is what you’d get out of a gushing 3 foot diameter pipe. So we’d need about 700,000 more of these typical desalination plants! That would be one desalination plant for every 10,000 people.
How Much Desert Area Would Be Needed to Power this Desalination?
For an example solar power plant, the Ivanpah facility in California has a maximum nameplate capacity of 400 MW but this is not what you’d get averaged out. Yearly it is expected to produce 1,000 GW-hrs of electricity. This works out to 100 MW on a continuous basis. And the site is 1,400 ha in size, or 14 square kilometres, or a square 4 km by 4 km.
100 MW equals 511,000 barrels per year. Since we’d need 80 billion barrels total to do all this desalination, this equals 150,000 Ivanpah’s. This adds up to an area of 2,000,000 square kilometres, or a square 1,500 km by 1,500 km. Not too huge. It’s half the area of the agricultural land this water would be supplying.
Will We Really Need to Desalinate This Much Water?
I arbitrarily based my calculations on doubling current global food production using desalinated irrigation. But we would unlikely need that much water in order to sustain the existing population. Let’s use a more realistic figure — let’s say that we’ll merely need to desalinate enough water to replace current groundwater use, since most groundwater is in decline. Even this may be an overestimate of desalination requirements for agriculture as groundwater also gets replenished; it’s just that we remove more than is recharged.
From this paper we see that 545 billion cubic meters of groundwater are used for irrigation per year. This is only 2% of the 25 trillion figure I used above. So let’s just factor everything down by 50 times. We’d need 0.06 cubic miles of oil equivalent energy to desalinate this much water. This is 2% of the world’s current technical energy demand of 3 cubic miles. Not too tall of an order.
It would require a desalination capacity of only 25 times what we have now. That’s almost a non-issue. We’ll probably be there in a few years regardless.
What are the Economics of this Solar Powered Agriculture?
What kind of ballpark are we talking about? The average new large RO plant that produces 100,000 m3 water per day has a capital cost of around $200 million. Let’s say the agricultural waterworks required to distribute that water costs about $50, so $250 million. It would consume 110 GW-hrs of electricity a year which is 1/10th of the Ivanpah’s output. Ivanpah will cost $2 billion so the share going to our sample RO plant would cost $200 million. Add that in and we get a total of $450 million capital expenditure for that RO plant, the solar power plant to run it, and the waterworks to distribute the water.
The water it produces could support 730 hectares of land, and with 20 t/ha/yr of corn production at $320 per tonne, this would total $5 million of revenue per year. So $5 revenue on $450 investment is about a 1% return. Not very good, but very close to being something worth looking at. If food scarcity causes corn prices to rise then it could easily move into profitable territory. Also not considered is yearly operating and maintenance costs for everything, which will reduce the return on investment.
For a reality check, we see from this page that prime Iowa farmland currently goes for about $7,000 an acre or $17,000 a hectare. Our sample 730 hectares costs $450 million or $616,000 per hectare. But our land will be producing food for the whole year, not just half, so let’s double the Iowa farmland’s price to $34,000 per hectare to compare apples to apples. Thus my analysis pegs our artificial farmland 18 times more expensive than real Iowa farmland on a per-tonne-of-corn-produced basis. But I did not include the cost to irrigate the Iowa corn field (which increases yields by 63%)
Based on my crude analysis I don’t think that’s too bad. I’m sure it would be possible to go into the economics of this in more detail but for my purposes this is a good enough order of magnitude approximation which shows that this idea isn’t too far out to lunch.
Is This All Realistic?
Would we really be able to provide this much desalinated water in time to save the world? Don’t forget that all this increased desalination capacity will have to compete with other additional needs for fresh water – municipalities and mining. How much desalinated water are we really going to need? Probably somewhere between my 25 trillion cubic meters per year and 545 billion cubic meters per year. While it’s unlikely that we’ll come anywhere close to desalinating enough water to be able to double world food production, it is definitely within the realm of reality to expect that we’ll desalinate enough to offset declines from groundwater. The question is whether we will use renewable energy to power this, or continue with our reliance on fossil fuels (i.e., coal) and thus increase our depths of ecological overshoot.
But the prospect I find interesting is that if we are able to pull off an energy transformation that would be in the order of powering this level of desalination, then by default we should be able to continue maintaining the energy inputs to existing agriculture fairly easily, to manufacture fertilizers via the Haber process, to synthesize hydrocarbons using the Sabatier process, and hopefully to use all of these resources more judiciously due to their increased scarcity (no more disposable plastic bags – what an incredible waste of a valuable resource those things are, plus all the ecological damage they cause when they inevitably end up in the oceans). Furthermore, the industrial production of nitrogen fertilizer is not a super energy intensive activity (using about 1% of world energy supply today), so it would not be difficult to synthesize the nitrogen fertilizers needed to sustain our existing agriculture even with a fairly rudimentary renewable energy infrastructure. With a combination of all these things, maybe we won’t see agricultural production drop off precipitously when fossil fuels run out. Phosphorus supply may present a greater problem, however, as there is no source other than mining it.
One of the challenges with desalination is that it is a fairly technologically complex operation. It requires a technologically competent society to run and manufacture desal plants. If society falls off a cliff and enters chaos as we experience energy decline then it may not be feasible to expect countries to focus their efforts on desalination. And in the event of war, desalination plants would be sitting ducks as targets to take out so as to cripple nations. They could be built underground to help alleviate this vulnerability.
It’s a huge challenge, but I think there is still opportunity for hope if we can collectively change our behaviours and our understanding of how technology and the world’s energy systems work. This is what I find to be more unlikely since the mainstream media seems to be doing what it can to keep the reality about humanity’s predicament from being generally understood by the average person.