Coal vs Solar – Which is more space efficient?

solar and a coal fired power station

This is a guest post by Greg Bell – which was conceived in the comments section of another post on this blog.

Take it away Greg:


Some things we can say we know for sure about photovoltaics – the fuel is free, it’s quiet, and there are no moving parts. But other benefits are more controversial. Recently a claim was made to me that solar PV is actually more area efficient – that is, it generates more energy per hectare – than coal.

Specifically, the claim was

“The 1,600 megawatt brown coal Hazelwood Power Station and associated mine in Victoria covers 3,554 hectares. In a sunny location in Australia 20% efficient solar panels covering that area would produce more kilowatt-hours than Hazelwood’s average output.”

Let’s see if that’s true.

Why energy instead of power?

Why are we looking at energy instead of power? Well, coal-fired power plants are rated at their peak power generating capacity, however this pace is not kept up 24/7, 365 days per year. Other factors come into play, like time down for maintenance, and degrading efficiency over time. So we really want a power plant’s track record over a whole year. Power times time equals energy. So we’ll look at how many GWh of energy are produced per year (which, since that’s energy divided by time, is actually power again, but averaged throughout the year). We’ll take that number and divide it by the number of hectares the power plant takes up, and that will give us a figure of GWh/year/hectare – our “area efficiency” unit.

Why area efficiency?

Why do we care about area efficiency instead of economics or the environment? It’s not a case of “instead”, rather “and”. Area efficiency is important because whatever energy generation technology we use competes for land with other important uses – like agriculture, or (God forbid) land simply left wild. In the case of wild land, converting it to a coal or solar plant means it’s lost for good, or at least a very long time.

And, if we’re converting from a less area efficient source to a more area efficient source, well then as a society, we might be freeing up some agricultural land, which if the UN is to be believed, is very important indeed. Of course, the previous energy source can’t have polluted the land or the potential for re-use is limted.

The Area Efficiency of Hazelwood

Hazelwood generates 12.1 TWh/year. That’s terawatt-hours – 1 terawatt-hour is 1,000,000,000,000 watt-hours! At 3554 hectares, it’s energy efficiency is 0.0034 TWh/year/hectare. Converting that to gigawatt-hours for a friendlier looking number and we get 3.40 GWh/year/hectare. So for every hectare of land occupied, Hazelwood produces 3.40 GWh of energy per year.

The Area Efficiency of Solar Photovoltaics

So what if we levelled the whole site, and plastered it with solar panels? It might look like this:

Royalla Solar Farm

Image from RenewEconomy (click image for original)

Not just any solar panels, but let’s use SunPower’s record-setting 345 W panels. And to really give solar an advantage, let’s assume we’re able to completely cover the site – no space between tilt frames, and perfectly tilted. We’ll assume 3.6 sun-hours/day – a weird unit that wraps up the solar energy received by a site into one convenient number. If you multiply the sun-hours/day of a site by the power of a solar power system in kW, you get the kWh/day that system will produce. So “Solar Hazelwood” would produce 27068315 kWh/day. Sounds like a lot! Is solar winning? Unfortunately, no, because when we multiply that by 365 to get annual energy production, and divide by the area, we get 2.78 GWh/year/hectare. Solar is only 82% as area efficient as coal.

The Area Efficiency of Loy Yang B

Holding up Hazelwood as an example from Team Coal is a bit unfair since Hazelwood is sort of famous for being inefficient – at least as far as how much energy produced vs CO2 output. It’s old. Let’s guess that CO2 inefficiency will mean it’s inefficient in other ways too and look at a different coal-fired power plant.

The new kid on the block is Loy Yang, also in Victoria. This bad boy burns through 60,000 tonnes of brown coal a day, belching 34 kg of mercury into Australia’s air and water. But, by committing these environmental crimes, it produces 1/3 of Victoria’s electricity – all of which I’m hoping is used frugally and responsibly by citizens ever-mindful of coal’s environmental impact.

In the units we’re interested in, it’s producing 8 TWh of energy per year. That works out to 10.2 GWh/year/hectare, and that’s including the nearby coal mine which feeds it (but because I couldn’t find figures for it, this doesn’t include the other station, Loy Yang A, which is also fed by the mine).

So where are we at? Better-than-real-world solar photovoltaics gives us 2.78 GWh/year/hectare, while a modern coal fired plant, including its mine, produces 10.2 GWh/year/hectare – coal is 3.7 times more area efficient!

The Area Efficiency of Solar Thermal

But everybody knows large scale solar power is usually solar thermal – where large mirrors focus the sun to boil water, which turns a turbine, etc.

There’s one in the Mojave Desert in California, serving 54,000 homes, it sounds like a monster. A fair trade for that large area?

The first clue is that it only produces 250 MW of power. So it’s not even a quarter of a typical coal plant. Fine, but we’re interested in it’s annual energy output, and how efficient it is for the amount of land it takes up.

Turns out, in that super-sunny location, it only generates 0.864 GWh/year/hectare. Loy Yang is 11 TIMES more area efficient!


So, this is all very depressing. Solar is nowhere near as area efficient as coal. This is the conclusion we reach a lot of times when we compare the two – the energy density of fossil fuels is impressive. Clean energy sources – tidal, wind, solar – are all “diffuse”, that is, spread out over a large area. That means we have to collect it to concentrate it into useful quantities. Collecting requires land area.

A more holistic analysis would include the area required to mine the coal. And even the area required to mine the materials to make the trucks which mine the coal. And on and on. PhDs are written about this sort of thing. We’d have to do that analysis for solar too, which is a high-tech product requiring all sorts of advanced civilisation behind it to make it happen.

In the end, area efficiency is probably far down on the list of considerations. CO2 efficiency, sustainability, and possibly most importantly, nett energy production, which is what our civilisation actually runs on, would be more important considerations. We know fossil fuels are declining in their nett energy (EROI or EROEI if you want to Google it), so it’s only a matter of time until the crossover point happens and solar is more “energy profitable”. Then we only need to work on that fully solar-powered solar panel factory.

If you want the spreadsheet I used for these calculations (corrections welcome!), it’s here:

Download (XLSX)



About Finn Peacock

I'm a Chartered Electrical Engineer, Solar and Energy Efficiency nut, dad, and founder of My last "real job" was working for the CSIRO in their renewable energy division.


  1. Very interesting theoretically / academically. But – and absolutely no disrespect intended in any way because I love chewing the theoretical fat meself – … it’s sort of irrelevant on two counts:

    (1) as you say, it’s all about not polluting the environment we live in to the point that we cannot live here anymore; and

    (2) as far as I can tell by looking out the train window to/from work, most solar panels (that I can see anyway) are going on top of houses.

    And that’s even true in the cold windy rainy UK where I hail (pun intended) from. Houses there are increasingly having SPs fitted. So, if anything, despite the calculations etc, solar is more area-efficient than coal fired plants simply because we’re getting a 2-for-1 out of the land… somewhere to live / work AND a source of power too.

    Everytime I am in a high-up place, e.g. a commercial tower block, and look out of the window and see miles and miles of rooftops and can’t help but visualise what they would look like with solar panels on top. There’s plenty space for solar right there.

  2. Right-on jimmec – when I was growing up our roof did just one thing – kept the rain off. Now it does three – keeps the rain off, collects our drinking/washing water and generates some of our electricity. There is no additional land area required for energy generation so we average 12kWh/day [4380/year] which is 0.00438 GWh/year which doesn’t look like much until I divide by the zero additional hectares needed which gives me infinite area-efficiency.[That can’t be right – how can you have infinite area-efficiency!!]

  3. Infinite energy area efficiency hmmm???

    Remember you should never consume an energy field bigger than your own head.

    That’s Rule No. 22 on the Evile OverLord List that is.

  4. Come on, solar runs best in agriculturally useless desert land, residentally useless, with negligible natural value, because of its lack of biodiversity. A mere 25% of Australia’s desert, is enough to produce a trillion tonnes of liquid hydrogen a year. 1,000 square kilometers is a trillion square meters, without the need for water, except that from sea desalination plants, to be used in electrolysis, to make liquid hydrogen, straight to hyper tankers. Already we make more energy than is put into making the cells and installation of the power. Now we’re into pure energy profit, soon it will be cheaper than carbon power, currently the price of solar, is less than a hundredth the price it was, per kWh in 1977. We could produce 25 times as much energy as we currently use, just by using 25% of the world’s deserts.

    Australia is 75 % desert, with a small population, so if we were to use 25 % of our desert. At 2,500 times as much energy as we currently use, just a decade later. I know we’re unfamiliar with industrial revolutions, but a mere century ago one happened. They went from steam and horsepower, to cars and trucks and electrification, in a decade. They say that we can’t change the energy bases fast, but they just did in the United States, in just a few years, they’ve halved their energy imports. Contrary to common current dogma industrial revolutions, can be very rapid, case in point, the second industrial revolution. In 1915 there were 10 times less cars as in 1925, just a decade later. The fruit of Maxwells physics, electricity not just through the spark plug, but Edison’s light bulb, tesla’s alternating current and radio exploded, during that time. So a change in the physics, then half a century later comes the novelties, then another half century later comes the maturity, a decade of industrial revolution.

  5. Only problem Stuart is distribution losses – significant between the desert & the city. One of the great things about PV is it comes already distributed [geographically at least if not temporally] and I would think that most electricity is consumed inside roofed buildings. [This is not to say we don’t need some large-scale generation to cover industrial & high density residential usage.]

    • Stuart Brown says

      Ralf, true, solar panels on rooftops have less transmission losses, with manganese batteries, they can go off grid. Thus they don’t have to pay money, even before they use kWh one supply charge. Yet you’ve also got it that industry needs trunk line energy, less losses, because it doesn’t have to go through an elaborate network, like residential electricity. Normal carbon power, also suffers from transmission losses, so when solar soon becomes cheaper than cabon, they’ll still want power and more of it because it’s cheaper. As price of energy goes down, demand goes up, so the more people use residential solar power the better. This frees up trunk lines for industry, as I’ve indicated, the amount of potential energy is staggering.

      Globally 25 times as much as we currently use, in Australia 2,500 times as much, as we currently use. Trunk lines, will have to be massively upgraded, to meet industry and liquid hydrogen exports needs. Please remember solar power is 1/100th the price per kWh, as it was in 1977. We’re facing a ten year industrial revolution, 2015-25, just like last time this will result, in the roaring twenties recurring.

      Solar employs twice as many people as coal in Australia, more people in the US than automotive. We’ve therefore already started our move, on big carbon, soon big silicons huge capital reserves, will move into energy. Carbon is beginning to seem like, sub prime investment, that caused the GFC. Soon they’ll go cap in hand, begging bowl out, asking for a bailout. Energy has always been more important, than information, telegraph/steam, radio/internal combustion engine, virtual reality/solar. After all we’re taking about a reciever, for a massive fusion reactor, above our heads.

      This will power 3D printed construction, high rise agriculture, through hydroponics/LED’s, that makes concentrated nutrient production possible. It brings tens of thousands of square kms of desert sun, to the cities food production and processing. Big silicon, has so much spare capital, its funding a private space race, after all.

Speak Your Mind


%d bloggers like this: