Solar footprint

by Mark Heinicke

Proponents of direct solar like to tell us how many "homes" solar installations power, as in the typical statement "enough to power 1,000 homes".  What does powering so many homes mean in terms of land use, or “footprint?”

Thanks to solar industry PR efforts, we now have in the news some examples of new solar facilities with impressive claims for generation.  They are operational, and wired into the grid.  In that sense, they are real-world.   Whether the claims as to their output are just as real remains to be seen.

For the sake of argument, let's take the industry claims at face value, and see what it means for footprint, at regional and national scales for the U.S.  I present two examples below, one photovoltaic and one concentrated-solar.  Extrapolating from those examples, we can estimate what additional area would be required to replace, with solar, the current 68% contribution of fossil fuels to electrical generation in the U.S.   (See here for an overview of footprint compared for all energy types including biomass, wind, etc.)

A caveat: here I am using aggregate figures, without distinguishing between base and peak loads.  That is in keeping with the sweeping solar industry language, which also fails to distinguish between base and peak.  Second caveat: since the solar industry PR hype rarely mentions intermittency, I am mostly ignoring that as well.  As the reader knows, intermittency is by far the larger problem with solar and wind, but we will take that on in another essay.  

Average home consumption and solar footprints

How much electrical energy does an average home consume?  The U.S. Energy Information Administration (U.S. E.I.A.) gives a countrywide average of 10,837 kilowatt-hours per year, ranging from 6,367 in Maine to 15,046 in Louisiana.  For 1,000 homes, the total demand is 10,837 kW-hr/yr/home x 1,000 homes = 10.8 gigawatt-hours/year.

For general case of footprint, we can use estimates developed by the National Renewable Energy Lab (NREL).  Using empirical data, the NREL has calculated the approximate area required, in principle, for a solar installation to deliver a year's worth of electricity.  There is a range depending on size and type of the plant, but it comes out to roughly 3-4 acres per gigawatt-hours per year (for the entire plant, not just the collectors). Using 3.5 acres per gigawatt hour per year, coupled with the USEIA demand figures of 10.8 gigawatt-hours, we get a round number of 3.5 x 10.8 = 37.8 acres for 1,000 homes for a year.  To extrapolate to a U.S. total, the USEIA says there are roughly 126,000,000 residential "customers" in the U.S., so that gets us 37.8 x 126,000 = 4,762, 800 acres, or 7,442 square miles of solar for the total U.S. residential use. 

However, we only have to replace the 68% of generation now produced by fossil fuels (assuming we do not lose more nuclear power plants currently representing 20% of U.S. power generation, and 70% of the emission-free power in the US).  So we end up with 0.68 x 7,442 square miles = 5,060 square miles, or a little more than the land area of Connecticut.

That is the general case. Now for a couple of "real world" examples.  In each of the following cases, we know the total (claimed) annual output.  We also know the area. We will divide the total residential demand in the U.S. by the plant output to find how many plants of that size would serve the entire U.S., and then multiply that by the area of the plant to yield an approximate area of that type of plant to serve the entire U.S.


Blue Wing Solar Plant

Example One is the Blue Wing Solar Plant in San Antonio, billed as a "14.4 megawatt installation" producing 26,570 megawatt-hours per year, "enough to power 1,800 households".   The site occupies 140 acres.

This was the estimated production; actual production numbers were obtained for the plant from the EIA.  The total production for 2012 was found to be under 24 GWh.

Total U.S. residential demand in 2012, per USEIA, was 1,374,515 thousand megawatt-hours (units they use).  More simply, 1,374,515 gigawatt-hours.  Of that, 68% is produced by fossil fuels, so to replace the fossil fuels we need a total of 934,670 gigawatt hours.

Blue Wing's 24 gigawatt-hours per year gives 39,000 plants to serve the entire U.S. residential demand met by fossil fuels. At 140 acres per plant, the total area comes to 140 x 39,000 = 5.5 million acres, or 8,600 square miles.  A bit larger than the land area of Massachusetts.

Blue Wing gets 24 gigawatt-hours/yr on 140 acres, giving us 5.8 acres per gigawatt-hour per year—which requires more than 65% more land than the NREL calculations for average production.  Since San Antonio's solar harvest, at 29.4 degrees N. latitude with many cloudless days, must beat the national average by a Texas Mile, we might regard the NREL figures as not entirely reliable.

A side note: it is instructive to look back at the "14.4 megawatt” installation number.  If 14.4 megawatts were being generated every hour of the year, we should make 14.4 x 24 x 365 = 126 gigawatt-hours, not the 24 gigawatt-hours the plant is producing.  This is equivalent to a 19% capacity factor, for a practically new plant (compare this with 90% average capacity factor for nuclear plants).

Evidently, the 14.4 megawatts is the peak power output, sometime in the early afternoon on a cloudless day.  That is why you always need to beware the n-megawatts-per-plant numbers tossed around by solar proponents.  Unless you know the output per unit time, such as gigawatt-hours per year, you have just a very rough idea of what the plant is actually producing.  Intermittency will be covered in a future article.


Solana Concentrated Solar Power Facility

Example Number 2 is the Solana concentrated solar power (CSP) facility recently brought on line in Arizona.  By means of parabolic mirrors concentrating sunlight on a heat transfer fluid, this plant, it is claimed, produces 944,000 megawatt-hours/year.  Touted to serve 70,000 customers, it dwarfs Blue Wing photovoltaic.  What's more, it has molten salt thermal storage, which somewhat solves the intermittency problem (the claim is "up to six hours").

Once again, the total production is available from the U.S. EIA site.  Although it does not provide data for a full year yet, it can be extrapolated that given maximum production Solana will be under about 700,000 MWh/yr.  That means Solana, in its first year, will already be more than 25% below estimated production values. The footprint of Solana is 780 hectares, or 1,927 acres (three square miles).  At 700 gigawatt hours per year, its footprint is a mere 2.75 acres per gigawatt hours per year—beating the NREL estimates, and far ahead of Blue Wing!

If we divide Solana's 700 gigawatt hours per year into the total U.S. residential demand, we get a mere 1,300 plants for the entire U.S.  At three square miles per plant, we find a total area of 4,000 square miles, about the size of two Delawares.



You could stick a gigantic facility—let's call it super-Solana—in a corner of Arizona, and hardly anyone would know the difference.  (Any human, that is: millions of non-human animals would definitely know.)

The trouble is that most U.S. residents do not live near any corner of Arizona.  With transmission losses of about 6% per 1000 km of line, by the time you sent the electricity to the places most people live, you will have lost upwards of 10% of the wattage.  From Phoenix to NYC is 3451 km (20.7% loss), to Chicago 2340 km (14.0 % loss), to Seattle 1795 km (10.8% loss)—keep in mind these are "as the crow flies" distances, irrespective of inconveniences such as terrain, municipalities, parks, airports, etc.  Therefore, we would have to make the super-Solana a minimum of 10-20% bigger to compensate for transmission losses, not to mention flinging multiple high voltage transmission lines from Arizona to every corner of the U.S.

But wait!  High voltage DC transmission (HVDC) lines reportedly might lose as little as 3.5% per 1000 km.   It is looking better for long distance transmission from our hypothetical super-Solana, if we are willing to overhaul our electric infrastructure to accommodate HVDC.  (A job whose cost would be difficult to calculate, but it would be quite a bundle.)

To be fair to Solana, the aggregate numbers we are using are for all 50 U.S. states plus D.C.  We would not, even in our exaggerated scenario, to build transmission lines to Hawaii and Alaska.  Nevertheless, the contiguous "lower 48" plus D.C. has 99.6% of the total population (U.S. Census), and likewise consumes 99.6% of the electricity (USEIA).  So there is a factor that gives Solana 0.7% (point-seven percent) wiggle room for the lower 48. 

Alternatively, we could pepper the country with Solana-like installations near their points of distribution. 1,300 of them.  To be anything like the Arizona Solana, they would require nice flat 3-square-mile plots of land that are not otherwise in use.  If they're in the northeast or northwest of the country—i.e. at higher latitudes with less intense sunlight, shorter winter days, and frequent cloud cover—then they'll have to have still more collectors, and bigger storage systems... meaning bigger footprint. 


Distributed Generation?

What about reducing the demand for utility-provided solar with rooftop solar collectors?  Good idea, except that most customers cannot afford installing rooftop systems that would come close to their household power requirements at $18,000+ a pop.  That is why solar buyers and the solar industry are getting multiple subsidies (such as tax credits) that the rest of the public is paying for.   Even with those subsidies, it would be difficult to imagine more than 20% of homeowners having enough discretionary income to put up or borrow that kind of money when looking at a payback interval of 10+ years and unknown degeneration factors over the years.


Reality checks.

So am I creating a solar straw man I like to kick around?  Portraying an absurdity that would never be played out in the real world, where we are adding wind power in big numbers, and improving the efficiency of photovoltaics, where decentralization and distributed generation across a smart grid will be the names of the future game, and home solar systems will cost less than a Scion IQ?

Actually, I am giving solar several benefits of the doubt. 

For one thing, I have touched only slightly on the more significant problem of intermittency, the subject of my next article on this site.  Intermittency in production is only further compounded by lack of production, such as at night, requiring long-term storage and release capabilities.

For another, I have not mentioned the 63% of U.S. electrical consumption that is not residential.  Residential accounts for roughly 1,370 terawatt-hours annually; commercial, 1,330 terawatt hours; industrial, 986 terawatt hours, and transportation currently a negligible amount (these figures are all from the USEIA for the year 2012).  For plants producing like Solana, we would need roughly 11,000 square miles, or larger than all of Maryland, to serve all those sectors.

For a third, I have not accounted for degradation of production or other factors.

I am not knocking solar as far as it goes—it is just that it does not go very far.  We need to go very far, very fast, if we are to cut emissions to a level that will head off environmental catastrophe.

What will go far, and pretty fast if we make a hard commitment to it, is nuclear power generation.  Footprint-wise, it is a no-brainer.  Here's why: a single current design U.S. reactor typically produces a gigawatt-hour every hour, and an average U.S. plant produces 1.3 gigawatt-hours per hour.  In a year, the average plant generates 24 x 365 x 1.3 gigawatt-hours = 11,388 gigawatt-hours = 11.388 terawatt-hours.  Annual U.S. electrical consumption totals roughly 3,694 terawatt-hours annually, and here we're talking the entire consumption, not just residential.  The 68% generated by fossil fuels comes to 2,512 terawatt-hours.  That amount can be generated by 221 additional nuclear power plants of current average size and current design for the entire U.S.

The area taken by one nuclear plant as currently designed is one square mile including a comfortable buffer. So with nuclear we get 221 square miles footprint to replace all the electrical generation we now get from fossil fuel burning.  That compares favorably with the 4,000 square miles of Solana-type generating plants that would supply only residential. If we wanted to extrapolate Solana to include commercial, industrial, and transportation, we would have more like 8,500 square miles—back to the bigger-than-Massachusetts picture.

In reality, we would not need that many new nuclear plants because (a) new plants will probably average closer to 2.0 than 1.3 gigawatts; (b) nuclear would be used only for base load, with fossil fuels and hydro handling peak loads.  (Because of intermittency, you cannot rely on either solar or wind for base load or peak load, which precludes a direct comparison.)  If electrical storage reaches the level that renewable proponents dream of, nuclear-with-storage might handle peak loads as well.

Nuclear plants can be placed close enough to users in most instances—certainly within 1,000 km (620 miles)—to contain transmission losses without having to revamp the grid to install HVDC lines.  Nuclear plants run 24/7/365, in every kind of weather.  They produce the same output at 50 degrees N. latitude as at the equator.  They are designed to withstand earthquakes, and to shut down automatically in case of dangerous tremors. Nuclear power production is an established technology with a long track record, and despite ill-informed alarms about its allegedly catastrophic hazards, has never resulted in a disaster with thousands of lives lost.


No Comparison

In conclusion, when it comes to footprint, nuclear beats solar like the Seattle Seahawks beats your high school football team.  Here is where we can see that the overarching issue is fundamentally that of scale.  When dealing in thousands of terrawatt-hours, it is orders of magnitude beyond the capacity of current solar technology to address the desperate Anthropogenic Global Warming (AGW) problem.  We are in big
trouble, and it is no time to snatch at the promises of an immature industry when we already have a time-tested alternative.

One might say we are facing the Third Inconvenient Truth of fossil-fuel energy production. The first, and biggest, is of Al Gore's coinage: AGW.  The Second is the lowering of marine pH with unpredictable but very probably bad consequences.  The Third Inconvenient Truth is the incapacity of renewables to replace fossil fuels at the necessary scale and within the necessary timeframe to rein in our carbon emissions. 


Mark Heinicke is a freelance writer living in central Virginia.
Tags: × × × ×


  1. Excellent work, Mark. This is a well-researched article. Thank you for contributing! -Jacob

  2. Not going to be able to run subways off solar power, forget the borough of Queens -- and at the expense of persistently ravaging and spoiling natural landscapes and whacking local property values (like windmills do) just for the touchy-feely relief to cluck "we're nuclear-free!" like you deserve a medal for depriving humanity of safe clean low-environmental impact hi-density energy?

    Give me tiny footprint (or zero-footprint if underground) nuclear in a heartbeat!

    James Greenidge
    Queens NY

  3. Thanks. This is very clear and very current.

  4. Thanks, all (all so far!). We who believe in the importance of nuclear power at this critical moment in history must always bear in mind that much of our job is simply educational. So I encourage you to spread the word on footprint along with so many other relevant issues. Since the footprint issue essentially boils down to arithmetic, it is a good wedge issue for anyone concerned about habitat destruction. Greens concerned about habitat destruction? Have them look at the numbers.

  5. ok so Solar needs large tracts of land (much larger) and cannot be generating at full capacity - on call. During night and non sunny days capacity falls.

    Beyond these 2 already known issues, is there anything new here?

    1. Anonymous: Scale, measurable scale.

      That's what's matters just the same as the amount of CO2 produced by any competing energy source matters. Without measurements, there's little meaning to the terms large or small. Measurement, for example, must be an essential part of the CCS discussion, but most of what I hear about CCS is all over the numerical map.

      If you don't want to consider amount of habitat that must be occupied by machinery in order to go full bore on solar, perhaps you would be content that solar collectors occupy an area equivalent to New York State? Georgia? Montana? Where do you draw the line?

      Advocates on whatever side of the nuclear vs renewables vs fossil fuels are apt to wave their hands with generalities such as large, small, reasonable, unsustainable, etc. Those are not sufficient grounds for making good decisions.

      Measurement is at the heart of most science. Dismiss measurement as a fundamental source of truth, then you're into philosophy rather than science.

  6. This is a tightly reasoned banner for nuclear, and I for one am convinced. Your elegant writing and competent massive insertions of data to back your proposals (namely: Renewables are immature and ultimately impractical; Nuclear, despite the storage prob, provides a much more global solution to the runaway train called AGW) makes your essay a breath of common-sense fresh air. Kudos and thanks! May those who leads us, many of whom are dozing at the switches, wake up and smell the ions.

  7. Good article that reminds of one of the great hypocrisies of solar power fans who would claim it can and should replace nuclear power. According to this line of thinking, nuclear power is unacceptable because a building or small plot of land may be crapped up, but covering tens of thousands of square miles of toxic PV panels is preferable or even desirable.

  8. Dude, so what you're saying is that if we cover 5% of Texas with solar panels we can provide all the electric needs of the entire Unites States with solar power? Let's do it! (Texas is actually about 250,000 square miles, and you say that less than 10,000 square miles is more than enough to supply the electric needs of the United states, right? 5% of Texas would be 12,500 square miles of solar panels.) Seriously, it's like NOTHING. It would cost a lot, which is why we should do it the way SolarCity is doing it. In fact, keep writing articles, and the 6,000 people at SolarCity will just get it done for us. There's really nothing to worry about. Those who get things done (like Elon Musk and his cousins) are going to solve the problem for those of us who don't . . . get anything done.

    1. Anonymous of June 29:

      You are leaving out the intermittency and transmission problems, as well as the degradation of solar facilities. Not to mention the environmental impact. None of which are serious objections to nuclear.

      We are doing a *comparison* here, intermittent and sprawling solar and wind facilities versus compact, around-the-clock at 90% capacity nuclear, with plants that can be located close to their users. If we had NO alternative to solar/wind to replace fossil fuels, taking 5% of Texas might make sense (ignoring the environmental fragmentation wrought by transmission lines to the rest of the U.S.). But we do have a time-tested, reliable alternative in nuclear, and the only reason it is being minimized is the largely baseless fear of radiation.

      Look at the amount, in watt-hours (don't be misled by watts alone), that Solar City projects are contributing to electrical generation at meaningful scale, and I think you'll find they come up short. Way short. It's great they're doing it, but it just doesn't match up with demand.


  9. How secure can a sprawling solar farm be if terrorists or a large gang for malice or hire went berserk overnight there with a couple of sledgehammers? Anyone ever seen gang-raids on jewelry stores in California?.

  10. A few lightening storms will put this problem to rest...

  11. According to your calcation 221 conventional power plants can generate 11.388TWh.

    Now if you want to stop global fossil fuels use and begin to build out infrastructure to actively remove the existing CO2 and Methane in the atmosphere so as to stabilize the climate, you need almost to replace about 80% of the world's total energy consumption that is coming from the burning of fossil fuels (0.8x104,426TWh=83,540TWh)

    Now using conventional light water reactor designs, according to you 221 sq mi of plants gets us 11.388TWh. So the total number of sq mi required for a global nuclear replacement of fossil fuels would be [(83,540TWh/11.388TWh)x221SqMi]=1,621,226SqMi or an area six times the size of Texas. Of course for the Solara type solar installations you are looking at 62,356,285 SqMi. There are only about 57,308,738 SqMi of land on earth...

    This does not account for additional infrastructure that might be required for generating synthetic fuels to supply the existing global infrastructure or the energy lost in converting nuclear power into them. So let's just say we want to double the total amount we generate to an area nearly that of the lower 48 states. Even if we spread that out all over the world it's a hell of a lot of land.

    Fortunately we already invested in a successful test reactor called the MSRE at Oak Ridge that could potentially reduce the footprint of a reactor site by 90% or more depending on the aggregate site design. So global fossil fuels production could be replaced with MSR nuclear totaling an area about 3/4 of the state of Alaska, perhaps much less.


Only comments that intelligently add to the discussion will be allowed.