An update on geo-engineering and solar power prices.

(note to self)

For many years now, it has been clear to the insiders that there is no hope in achieving serious reductions to greenhouse gas emission by means of international co-operation: the incentives to free ride on the efforts of others is too great and none of the big players is willing to subjugate themselves to a world police that would enforce a deal. So whilst we have all been happily increasing our consumption of fossil fuels year-on-year, the smart money was always on finding some technological fix to global warming that did not require near-unanimous international agreement, whilst simply adapting to the problem in the meantime. That fix could be geo-engineering or a renewable energy source becoming economically competitive with fossil fuels.

So, where are we currently when it comes to geo-engineering and renewables? In terms of geo-engineering the likes of Bill Gates, Richard Branson, the UK Royal Society, and a whole set of EU-US based institutions have been pouring money and time into looking at what can be done. In terms of renewables, the big movers have been Chinese companies and a glut of new ideas that are leading to much cheaper forms of solar power.

To start with solar power first, according to the Bloomberg New Energy Finance’ Solar Value Chain Index the costs per Kilowatt-hour of solar has reduced around 50% in the last 3 years alone, with various new technologies that have the potential of going down much further. They are talking about printing off solar cells, using iron guns to produce them, making solar panels out of a spray-on paint, and various others ideas. One needs to be an expert at this to judge whether it will actually work, which I am not, but the clear reduction in costs that was achieved recently is there for all to see.

So how close is solar to being competitive to fossil fuels in terms of producing for the electricity grid? As a rule-of-thumb, the life-time costs of the cheapest fossil fuels are currently around US 65 dollars per Megawatt-hour whilst solar was still estimated to minimally cost over 200 dollars per Megawatt-hour in 2010. The cheapest fossil fuels are natural shale gas, natural gas turbines, and some forms of coal. Allowing for a halving of the fixed-cost of solar infrastructure in the next year, solar would still cost above 100 dollars per Megawatt -hour. This is competitive with many currently used forms of electricity generation of fossil fuels, including conventional combustion engines with a cost above 120 dollar per Mwh.

What is important to note is that the current trend of solar prices doesn’t have to continue for long for solar to be the stand-out cheapest form of mass-electricity generation for countries with a lot of sunshine. At the moment solar is hence looking like a real potential long-run replacement for many countries. Even at today’s costs, solar would only be marginally more expensive than fossil fuels.

But what are the inherent disadvantages of solar? Well, for one, you need a lot of solar panels to get a decent electricity flow, so cars or planes with solar cells are nowhere near a realistic prospect in terms of mass-transportation. Hence fossil fuels remain the front-running source of energy for our cars and planes, which on their own are enough to guarantee sufficient demand to keep increasing atmospheric CO2 levels.

Also, it needs to be sunny in order to get electricity, which is a problem for a lot of our economy which is reliant on guaranteed energy flows at any time of the day and where people don’t want to have to reboot their computer after every cloud. Since most of our energy needs are connected to industry or in activities that could run on batteries charged up when the sun shines (like charging up the car and the i-pad), there is some mileage for weening ourselves off this ever-ready energy pattern, but it would seem fair to say that it would be hard for us to adjust to only engaging in major economic activities when the sun shines. Hence a remaining technological hurdle is how to store solar energy easily and in sufficiently huge quantities as to allow for a lack of sun for a few weeks. Given the huge amount of constant electricity demand, this big-battery problem still prevents us from adopting solar as our steady supplier of electricity. If we would have to keep relying on fossil fuel generated electricity when the sun doesn’t shine, which would be the current reality if we’d adopt big solar farms for our electricity base-load, then once again we are guaranteed continued increases in the amount of CO2.


One might object to this by saying that in a large electricity grid one can connect the grids of different countries and thus effectively share sunlight with other regions, but electricity transportation over large distances has a remarkably high loss-rate. As this old Global Energy Network Institute report estimated, every 1,000 kilometers of extra distance increases the costs by around 5-10%, or equivalently that there is about a 5-10% loss in electricity when having to transport it another 1000 kilometers. If one then reflects on the fact that the distance between Sydney and Perth is already close to 4000 kilometers and thus involves a 20-40% loss of electricity, it is clear that it probably is not even cost-effective for Australian states to ‘share’ their sunshine, let alone to share sunlight with other countries.

Hence the problem of energy storage is a serious one for solar and we are still waiting for improvements in battery-efficiency to consider solar as an alternative to the fossil fuel electricity generators which can deliver power whenever we need it. This problem of course also besets wind-energy, for which the prospect of large future reductions in costs is much less rosy.

Furthermore, solar panels need setting up and they have to be kept clean, things that become much cheaper to do if one is setting up many of them in a single spot. Hence solar is unlikely to replace combustion engines as a means of delivering local energy or energy to residential homes: too much hassle and only viable with subsidies or in places where it is hard to get constant supplies of fossil fuels. Some major structures like large boats and sky-scrapers are a different matter though, so one should expect more medium scale uses of solar, although there too the problem of energy storage is a major one.

In short, the price reductions for solar is exceptionally good news for our way of life: given the big price reductions that bring solar close to parity with existing fuels, and given the near inexhaustibly huge supply of solar (the sun sends about 6000 times more solar energy to us than we humans generate from all sources), the future of our industrial modern societies based on cheap energy looks very bright. We might have to adapt to the intermittent nature of the energy flow if we can’t crack the energy storage problem, but at least we now have a liveable alternative. The substitute source of energy to fossil fuels is hence in sight even though it may take a decade or two before its better than what we currently use.

Then the topic of Geo-engineering. Since the landmark 2009 report by the Royal Society (which I extensively reviewed previously), engineers have been dreaming up a lot of new stuff, with particularly hopeful possibilities in the area of Solar Radiation Management (SRM). Front-runners are the ideas of spray-gunning the atmosphere in order to create more clouds, and sending dust particles up into the air.

The spray-gun idea is of a charming simplicity: clouds are white and reflect a lot of sunlight. Hence if you can create yourself more clouds, you cool the earth. How do you create more clouds? Well, clouds are made of water vapour. That vapour arises naturally from the sun heating water, but you can also try to do it yourself by putting water into a plane and delivering the vapour into the atmosphere where you want it (Neukermans A, Cooper G, Foster J, Galbraith L, Ormond B, Johnston D, Wang Qin (2011). Supercritical saltwater spray for marine cloud brightening. Geophysical Research Abstracts, 13, EGU2011-9655-1).

The technology hence has many potential advantages: because one would be in the business of creating thousands of clouds every day, one gets a very sensitive instrument for geo-engineering. You get to decide where you want to cool, just at what temperature you are going to stop cooling, and one can easily experiment with small regions without seriously upsetting the balance of the planet. After all, Nature experiments with clouds all the time and a few more or less wont unbalance the earth. So the technology can be safely tested and experimented with and has great advantages in terms of timing and delivery.

What are the problems? Well, for one, we don’t quite yet seem to be able to produce a fine enough mist quickly enough. After all, you want to be able to do this quickly and thus convert thousands of liters of sea-water into a cloud in a matter of minutes. Yet, sea-water is salty and thus corrosive, and there are all kinds of things in sea water that would clog up any tiny holes. Nature solves this by simply heating the water and thus having salt-free water molecules rising up in the air, but that solution is not open to us because the sun warming the sea water was precisely the problem we are trying to address, not add to. Hence we still need to sort out the problem of quickly filtering sea water and misting it.

A secondary problem is sheer coordination: if we end up with thousands of planes misting the atmosphere then one would be looking at a whole network of airports and cooperating countries. The countries most suited, i.e. close to the North Pole, might actually discover they don’t want to halt warming and thus fail to cooperate.

The bigger problem is that it might not work: aeroplane delivery of water-vapor is an intriguing idea but there are only computer simulations that suggest it is do-able at reasonably low costs. The computer models can easily be off by a magnitude of 10 or more in terms of how much cloud needs to be created to get enough cooling. Just think about it: if we would have to create 10% more clouds in the world, we would be talking about an artificial vapour with the size of America. That’s too much vapour and planes to realistically be able to muster, so one has to hope that it would require no more than a hundredth of this area. I am personally skeptical on this point and would thus not be surprised if some new computer model in the coming years would say it’s a hopeless plan, but we will see.

Then the dust particles, also known as dimming, or ‘aerosols’. I have written about this before and the advantage of this one is that it’s a proven technology. Volcanoes proved it for us. The Mt Pinatubo eruption in 1991 caused a global cooling by belching huge volumes of dust particles into the atmosphere, proving that dust can cool the earth.

Despite some people saying we don’t know how to dust the atmosphere, we humans have also done it. Until about 20 years ago we put a lot of dirty particles into the air by having dirty coal power stations, unfiltered car exhausts, and various other unfiltered industrial emissions. This lead to large-scale dimming to the extent that the amount of sunlight hitting the earth was reducing by up to 4% per decade from 1960-1990.

What happened to dimming? We started to clean up because of concerns over the local environment. People don’t like smog and haziness in their own cities, so governments have mandated industries and electricity generators to clean up and no longer send particles in the air. This has reversed the global dimming trends, such that our days are once more full of gloriously clear sunshine. And quite probably also means global warming is resuming on a faster upward trajectory.

It is not too hard to guess what can be done: by re-adopting our dirtier ways we can resume the dimming process. Furthermore, there are scientists trying to perfect what we stumbled upon by accident. We can make the dust we send up more reflective, more buoyant (so it stays up for longer), less degradable, and less annoying to people.

The big problem with this solution is again one of sheer scale: when we were dimming we were belching up an awful lot of stuff into the atmosphere. We were effectively blocking out an area the size of Australia and we were still warming up the planet! It is even worse: since then we have added a lot of extra Greenhouse gasses, meaning that we’d have to take dimming to an extra level to be of potential help. To make an impact, we’d have to re-designate large areas that we don’t care about, such as, say, the Pacific Ocean, as dimming territories and continuously belch up huge volumes of dust. That in turn requires a massive industrial exercise since it would involve sending huge volumes of resources to tiny island in the middle of distant oceans in order to send it up. It is not at all clear yet that this is affordable in terms of what we are willing to pay to stop global warming (which is not much).

Summarising, geo-engineering is probably do-able though we don’t yet know the true costs. We can safely assume it would be minimally in the order of hundreds of billions of dollars every year. And the uncertainties are such that we are easily 20 years off knowing enough to be able to implement it. During that 20 years, we can safely say that we will keep going through the cheapest energy sources – fossil fuels, whilst solar energy is promising to be the go-to source once the cheapest forms of fossil fuels have run out. If improvements in solar and battery technology are spectacular, it may even muscle out fossil fuels within this decade as the major provider of base-load energy.

In short, there have been very hopeful developments for the sustainability of our current way of life on this planet in the last 2 years.

Author: paulfrijters

Professor of Wellbeing and Economics at the London School of Economics, Centre for Economic Performance

18 thoughts on “An update on geo-engineering and solar power prices.”

  1. I have heard that salt batteries are almost economic with natural gas turbines. The idea is that you use solar power to heat up salt and the molten salt to power steam turbines. The salt gives you a battery life of about 14 – 20 hours. So it should soon be possible to have a network of solar farms providing base load power for 99% of the time.


  2. Solar power has the benefit of generating maximum power during times of maximum demand (hot, sunny days) when spot prices are high. And it can be installed on-site in many cases,where the retail price of electricity is more relevant than the wholesale price. So it’s getting close to being competitive.
    The cheapest way to store power for the night is reverse hydro, but you need a dam nearby. Personally I like the idea of keeping electric cars plugged in to the grid when not in use and using their batteries to buffer between supply and demand.


  3. James, Matt,

    reverse hydro would entail severe capital investments (which was the whole original problem with solar): not just do you need to have a dam nearby, but you need to build another one downstream to store enough water to pump up again. Plus big pumps. Add to that the 20% energy loss you get from pumping it up and then letting it go down again, and you can see how we are once again in uneconomical territory unless solar is much cheaper than fossil fuels.

    Having electric cars be the battery for the grid sounds like a neat idea, but it is simply not enough batteries for a couple of weeks without the sun. A battery car runs only for a few hours without recharging, not nearly enough for itself for a few weeks, let alone enough for itself and the rest of the economy. One thus would need a lot more batteries to store enough energy. For the same reason, salt batteries lasting us a day are just not enough. Just think about it: we are not yet willing to have trains run only if its been sunny somewhere in the last 3 days! Reverse hydro at least has the potential of storing enough energy, but the capital costs are stupendous and you need the right geography close by.


  4. kme,

    costs of additional length in electricity transport are not quite linear, and the 5-10% figure was crude to begin with, which is why I gave a 20-40% range which includes your 35% at the upper end.
    As to sharing, with severe losses, many countries are already sharing. The French share their nuclear electricity at night with other countries.


  5. When are the renewable proponants going to acually admit that renewables don’t exist when full life cycles are considered? The standing answer is “never.” It is not convenient to admit the planet would have to be heavily mined to find all the rare earth elements required by current day PV solar, modern magnets required for efficient wind turbine, and batteries. Let’s dig up planet earth to stop producing carbon which is a building block of all life on earth. Let’s ignore the actual earth data that shows all warming models and prediction of warming have failed. And we can replace our food crops for crops that make fuels. I am all for spending big for research in true renewables, and that includes small pilot plants to work out technolgoies. But not for subsidizing renewables that are not renewable when full lifecycle is considered. Also, society needs to focus on waste reduction and efficiency, not doing without.


  6. Scott,

    that’s far too negative a view for me. Sure, renewables are not economically viable at present and it is unhelpful to pretend otherwise. Yet, by the same token, we will eventually run out of fossil fuels anyway and technology has time and again proven to improve and adapt to problems. So in any scenario we are going to transition to other power sources and, given that nuclear is quite expensive on a life-cycle basis, renewables should not be written off so easily. I see this very much in horse-race terms and am noting that solar is rapidly looking more like a winner.
    As to the point about rare minerals as inputs being the hard constraint that will foil solar: if one of the inputs into PVs turns into a scarce, super-expensive deal-breaker, we will probably find substitutes, just as we found substitutes for the oil lamp and the wax candle.
    As to ‘all warming models have failed’, lets not get side-tracked into that discussion here, apart from saying that I by and large go with the IPCC as reflective of the scientific consensus on this. What I am skeptical about is the pretend-solutions, not the existence of the problem.


  7. How does nuclear power fit in? Is it as simple as fossil fuels vs renewables? Haven’t there been similarly large advances in nuclear power technology and corresponding decreases in life-cycle costs?


  8. bob,
    nuclear is simply too expensive at the moment. With the current designs, it is certainly not competitive against fossil fuels. I am indeed not aware of major cost reductions in nuclear, at least not in terms of plants that governments are prepared to live with. With every nuclear accident the cost of the next nuclear plant just goes up and up. Having said this, the reliability of nuclear would put it ahead of renewables at the moment, so it is only the promise of further cost reductions in solar that put it on the map.


  9. Paul – more nitpicking.

    Reverse hydro does not usually require extra pumping equipment – you just run everything in reverse, using the Frances turbine as a pump and the dynamos as motors. That’s how it is generally done worldwide today. If you have a really high head you’ll have Peltier turbines and would need separate high-pressure pumps, but that’s moderately rare.

    And your lower-level storage dam can be relatively cheap because its a big shallow reservoir at the bottom of the hills and hence can often use a simple earth wall (it doesn’t need to generate electricity or be a flood mitigator – they’re jobs for your hideously expensive concrete arc dam higher up).

    Your points about the efficency losses and there being few large hydro opportunities left to exploit, though, remains.


  10. More broadly, you need to distinguish between energy storage that allows you to match evening peak demand with daytime solar production – a problem, but one that is beatable even with current technology – and catering for prolonged bad weather lasting days or even weeks – a much tougher problem.

    One thing helping the latter problem is that storms block out the sun but they also generate wind – wind and solar power are to some extent complements rather than rivals.

    Another thing is the widespread availabilty of cheap natural gas with fracking technology. If we’re happy to minimise, rather than eliminate, our carbon emissions then filling in a bad week or two a year with natural gas power might work (in the way some places have a desalination plant on standby only for use in prolonged droughts. Maybe we could even just sit gas turbines atop hydroelectric or solar thermal turbines, with a gas pipe to the turbine hall, and so minimise the additional capital outlay.


  11. DD,

    we’re now discussing here at core instead of at troppo, hej? Ok.
    About the pumps: sure at the moment there is a simple two-way system, but nowhere yet do we have the situation of storing the kinds of loads we would be talking about (potentially Terra watts). The lower level dam can indeed be a simple thing, though not too simple: you dont want it to leak to much and if we are talking the storage of Twatts then we are talking about sizeable lakes for which you dont just have a crude earthen wall.
    As to peak shaving, it is indeed clearly the case that tiding over a couple of hours of peak demand can be done with current battery technology, but winters and a long and large cloud cover are a different kettle of fish.
    Then to wind/solar/gas combos: as soon as one is talking about a combination of technologies each individually capable of supplying the energy demand, one is talking about multiples of the cost of any of them: you have to bear all the fixed costs of every option in your combination, and these fixed costs are a major part of these lifetime costs (even for fossil). And costs are crucial. So a combo including solar is only economically viable if solar is much, much, cheaper than fossil fuels. And, as I keep saying, individual countries might experiment and tinker with more expensive options as a feel-good policy but the world as a whole is going to jump from cheapest to the next-cheapest.


  12. Sheesh! the cheapest Fossil fuels $65? Loy Yang does it for about $20 and if we built new Brown Coal it would be a lot cheaper. One or two new Large BC plants in Victoria replacing one or two old ones would gain efficiencies equivalent to the contribution of all the renewables in Australia.

    It is often forgotten that while there are modest cost improvements in renewables. New fossil fuel generation makes massive gains. Not to mention if the cost of a roof full of panels making 1kW reduces by 50% in cost. Those same clever manufacturers are making airconditioners 80% cheaper that consume 10kW!

    All the solar power put into Victoria has cost Billions and generates about 1/4 of 1% of our power needs.

    Solar is great in remote locations but even if you got the panels for free it is just decoration. You still need all the coal and gas base and peak capacity to cover for when it is not shining which is 80% of the time.

    Pumped Hydro is infeasible – all the good locations have been taken around the world. There is no more.

    We have been making batteries for a century and they are not getting significantly better. Learning curves require an increase in production. Its easy to forget every conventional car has one and there are millions of them in telecommunications and all other industries. Globally there have been incentives, especially militarily, for decades to make a cheap lightweight battery.

    I think Scott is on the right track… this is not negative – these are facts.

    We shouldn’t waste any more money.


  13. Rick,

    The 65 dollar is lifetime costs, which ithink includes a lot of the infrastructure.
    The objections you raise are precisely why people are talking about a world power grid which would nullify the battery problem. Unfortunately the grid is super expensive.


  14. I can’t let this one go. Look you can’t compare total infrastructure cost for brown coal with Solar PV. The total infrastructure cost of Solar PV is infinite at 7pm when you want the electricity.

    You could make a stab and say that it is equal to the cost of total infrastructure cost of solar pv plus the total infrastructure cost of battery storage.

    If you do these sums even with the outlandishly optimistic learning curves you get 25X the cost of brown coal and that is very conservative.

    Imagine if you pulled up at the gas station and the attendant announced your fuel tank cost $2500 today sir!

    To the other point about transmission – look if you are sending wind and solar across transmission lines you had better beiieve they are costing 3X to 5X the cost of TX of coal. So add that onto your NPV.

    Its about time MBS started getting real about economics and looking at the bleeding obvious – we are generating from brown coal because it is incredibly profitable and therefore we should generate more!!!


    1. Rick,

      agreed that on current technology solar is still window-dressing, but one should not be so pessimistic about the future. What solar needs in the long run is a world grid so that it doesnt matter whether it is night, winter, or cloudy when you need the energy. I had a go in the clubtroppo thread at calculating the cost of this (

      Suppose we would indeed put a world-grid in place based on huge solar farms in the major desserts of this world, connecting North with South, and East with West. That kind of grid would mean it is no longer relevant that it is winter in one place, summer in another, dark in one place, or rainy somewhere: by combining the regions in a super-grid, one is basically down to adding enough solar panels to the desert operations to provide for base-line world energy demand.

      What would such a grid cost? Well, first consider the length of the grid: the grid has to go close to all the major population centers in order to deliver massive electricity supplies, and it needs to connect up places like Australia with China. Thinking of the length of the earliest railway lines, we would thus easily be looking at a minimal grid of, say, 200,000 km.

      Then consider the currents on it. At present consumption levels, the grid would need ‘only’ a few Terrawatt total capacity, but thinking about the future demand, a figure of 20 Terrawatt is probably closer to the mark in terms of normal energy demand on the total grid around the year 2030. Considering that one needs less on each bit of the line line, one is thinking of, say, 2 Terrawatt on an average bit of the 200,000 km grid.

      A very optimistic estimate of the current cost is 1m per km per 5 Gigawat DC line. Scale that up to this imaginary grid and you get (2000/5)*200,000=80,000,000 million dollars cost. That 80 trillion dollars is about twice current world GDP. If we take a more normal estimate of 5m per km per 5 Gigawat DC line, then the costs would be much higher, but of course 400 lines of 5 Gw close together are surely much cheaper to put down than 400 disparate lines, so lets run with the 80 trillion figure as an estimate of the cost of just the grid.

      The key thing to note is that 80 trillion would also buy one roughly 80,000 normal power-stations which would churn out about 60 Terrawat, much more than would be on this grid.

      In short, at current technology, a world grid that sends solar energy around the world is a clear no-starter, even if the solar energy generation itself were free. We’d have to get to about 20 million per kilometer for a 2 Terrawat DC line (which would take the cost of the grid to 8 trillion and would make the energy streams on the grid higher than what could be generated just for the price of the grid) to get into the right ball-bark for this world grid, which in turn is the major element needed to make solar viable.


Comments are closed.