Nuclear Has One of the Smallest Footprints

From Fuels to Building Materials, the Atom is Antidote to Sprawl


The footprint of any energy source ought to include more than its carbon emissions or land use, but also the footprint of its building materials, the energy density of its fuel, and its overall emissions intensity. Using these metrics, a more nuanced and compelling case for nuclear emerges: it uses less land, significantly less concrete and steel, has a low emissions intensity, and generates less expensive electricity than solar and wind alternatives. If we want to protect our environment and reduce our impact on Earth, we must consider nuclear as part of the energy mix.

September 20, 2013 | Martin Nicholson,

When evaluating the footprint of nuclear, writers and analysts tend to focus on its near-zero carbon emissions. Yet, there are many other areas where nuclear power consumes fewer resources than other electricity-generating technologies. In fact, when compared to coal, natural gas, and renewables, nuclear is the most land efficient, energy-dense source of power, with the lowest use of building materials per unit of energy generated per year, and one of the least expensive in terms of levelized costs. Evaluating these different aspects of its footprint demonstrates that nuclear is one of our most viable solutions to readily decarbonize the economy.

Land Footprint

Advocates of a particular generating technology will often use land use as an argument against competing technologies. With some technologies, like wind, there is the risk of apples-and-oranges comparisons in terms of land use. Do you count the whole area of the wind farm in the calculation or just the footprint of the wind turbine plus access roads? The difference between the two can be a factor of 50.

To avoid apple and orange mix-ups, the table below identifies land use per GWh per year to ensure a like comparison between the technologies. From this comparison, nuclear is one of the most land-efficient sources of energy. Furthermore, in the United States, the average nuclear plant uses 450 m2 per GWh per year. This is less than half the Massachusetts Institute of Technology (MIT) estimate shown below. Although this data doesn’t account for uranium mining, we will see later that it is not significant per GWh when you look at its fuel footprint. 

From MIT except wind: NREL, hydro: author’s calculation, and biomass: Minnesota

From a land perspective, biomass is an extensive challenge except for countries with plenty of spare arable land. Solar thermal is often criticized for requiring vast tracks of land, but it’s still more land efficient than surface strip coal mining, according to MIT. Hydro is a special case because the land use depends on the head height, the average reservoir depth, and the flow rate through the reservoir; the table above uses typical numbers.

Building Materials Footprint

Nuclear power is often criticized as a huge consumer of building materials. This is true if you just look at the materials used to build a power station without considering the amount of energy the power station generates over its life. As such, building materials are often quoted in tonnes per MW (power plant size) rather than tonnes per MWh (the power plant’s energy generation). This can mislead us into thinking that nuclear power uses more resources than solar panels, when the opposite is true.

The table below shows the concrete and steel used in some plant constructions expressed as tonnes per GWh per year. The capacity factors shown are the ones used in the referenced reports. Of these plants, nuclear power uses the least amount of concrete and steel per unit of energy generated in one year. If the full lifetime of the plants had been considered, then the nuclear plant’s use of concrete and steel would be even less because nuclear plants have some of the longest lifespans. Compared to a nuclear plant’s lifespan of 40 years, for instance, a solar panel may last less than 20 years. A true comparison could significantly increase the materials required for the solar plant.

From ISA except for solar thermal, which is from NEEDS

Opposition to nuclear in part stems from a fear of radiation exposure. UNSCEAR demonstrated that steel production exposes the general population to more radiation than nuclear power plants. Installing solar thermal and wind in preference to nuclear power, and so using significantly more steel, may well be increasing the radiation risk, not reducing it.

Fuel Footprint

Fuels with a high energy density are able to store large quantities of energy in a smaller volume. The smaller the volume of fuel to be mined, the smaller the footprint on the Earth and the lower the fuel cost. Uranium has a very high energy density. When it is used in light-water reactors (LWR), the electrical energy density of uranium is over 30,000 times the energy density of black coal. So a 1 MW coal plant might use 3 million tonnes of coal per year, but a similar sized nuclear LWR plant might use only 170 tonnes of natural uranium (with some energy density loss in the enrichment process). Hence, the fuel cost for a nuclear plant is a fraction of the fuel cost of a coal or gas plant.

Wind and solar advocates often suggest using gas or biomass to balance the variability of the wind or solar plants rather than using uranium or coal. Both gas and dry wood have very low energy densities compared to uranium or coal and need larger volumes of fuel. The reason that natural resources like coal and uranium have been so successful at generating electricity is largely because of their relatively high energy densities – particularly uranium. Using uranium in fast reactors rather than LWRs would lower the fuel footprint even further by more than 100 times.

Emissions Footprint

On average, about 500 kilograms (kg) of carbon dioxide equivalent are produced per MWh of electricity generated in the world. This is known as the ‘emission intensity.’ To reach the emissions reductions needed by 2050, studies have shown that the average emission intensity needs to be reduced to as low as 50 kg CO2-e/MWh.

The emission intensity of various primary energy sources is seen in the table below. This data shows life cycle emissions, which means that they cover emissions during power plant construction, fuel mining and transport, operation, decommissioning, and waste disposal. Only the really low emitters, nuclear and renewable energy options, can deliver the 2050 emissions intensity target. The worst offenders are coal and oil, followed by gas. Carbon capture and storage (CCS) will help both coal and gas to reduce carbon dioxide emissions, but will still be unable to meet the challenge of 50 kg CO2-e/MWh target.

From World Energy Council

Cost Footprint

All the footprints discussed above – land, materials, fuel, and emissions – can impact the cost of electricity. Typical levelized costs for different power station types are shown in the table below. Based on these generation costs and the emission intensities shown above, a carbon price of say $30 a tonne of carbon dioxide could add 30 percent to the levelized cost of electricity from coal and 20 percent to the levelized cost from gas. 

In the end, nuclear power has one of the smallest footprints of any energy source: it uses less land, significantly less concrete and steel, and generates less expensive electricity than solar and wind alternatives. If we want to protect our environment and reduce our impact on Earth, we must consider nuclear as part of the energy mix.


Martin Nicholson is an Australia-based engineer and author who wrote the peer-reviewed book on low-carbon energy systems titled The Power Makers’ Challenge: And The Need for Fission Energy (Springer UK).

Photo Credit: Wikipedia Commons


  • Thanks Rainer. I hadn’t thought of mortality as a footprint on the Earth but you are quite right about the issue of lower mortality and morbidity - particularly compared to coal. The nuclear doubters probably find this more difficult to accept given all the fear, uncertainty and doubt coming out of Fukushima.

    By Martin Nicholson on 2013 09 21

    Reply to this comment / Quote and reply

  • By Craig Schumacher on 2013 09 21

    Reply to this comment / Quote and reply

  • Very nice article, thank you. Some additional thoughts….

    Power plants burn green wood, not dry wood, so energy output is less.

    The new EPA rules limiting CO2 emissions to 1100 pounds/MWh are about 500 g/kWh, to be compared to your next to last table, which shows that CCGT just meets the standard. The table omits open-cycle gas turbines, which will NOT comply, and such are necessary for backup for wind turbines because of their rapid ramp-up capabilities needed when the wind lulls.

    No power plants today have CCS, carbon capture and storage.

    By Robert Hargraves on 2013 09 22

    Reply to this comment / Quote and reply

  • Very informative article, thank you.

    I’ve tried to visualize the land use requirement here:

    Feel free to share if you find that useful.

    By J. M. Korhonen on 2013 09 22

    Reply to this comment / Quote and reply

  • Thanks for the link JMK. Translated into my units the numbers shown are wind = 7,500 km2/GWh/yr but this includes the total land not just the turbine footprint plus access roads so is 6 times higher than my figure.

    Nuclear = 490/GWh/yr which is close to my estimation of 450 for US plants excluding mining site.

    By Martin Nicholson on 2013 09 22

    Reply to this comment / Quote and reply

  • Most people have no concept about the size of a watt, a kw, a mw, or a gw. Hell even a combustion turbine (“CT”) has the biggest output for the smallest footprint. A combined cycle CT is even more efficent for land use and energy production/conversion They are quick to start 24/7 and do not require energy storage. The only negative is the combustion of natural gas and/or fuel oil and the resultant pollution. If natural gas is used, they can fit on a large residental property and produce upwards of 100 Mw. Actually, land acerage efficiency, natural gas use, energy production, and pollution should all be considered. Pollution can be controlled and reduced. Natural gas is too valuable a finite resource to be squandered on energy production. Land and open space are also too valuable. A mix of all types of energy conversion is wise, but the apportionment should be weight averaged according to the efficiency of the important parameters considered.

    By Joe Llaben on 2014 10 04

    Reply to this comment / Quote and reply

    • A “large residential property” is a bit small for a power plant, but it would fit in an industrial park. Also, the CT does indeed start quickly, but a more efficient CCGT (combined cycle gas turbine) takes about an hour; new ones are faster ramping.

      By Robert Hargraves on 2014 10 05

      Reply to this comment / Quote and reply

  • Bob Hargraves, A large residental size property is certainly large enough for a power plant embedded in a residential or urban load center. Check with Public Service Electric and Gas Company (“PSE&G”). They were using simple CT for decades at all their power plants as black start/bootstrap units. They also purchased larger units for peaking. A simple Pratt and Whittney (“PW”) Turbo Power and Marine Division (“TP&M”) Power Pac using natureal gas will definitely fit on even a medium residential sized piece of property and produce about 20 Mw or so. All it is is a jet engine, an enclosure, a control area, a Worthington free turbine, an Electric Machinery (“EM”) generator, and a step up transformer about the size of my small house. A TP&M Twin Pac with C4 engines would be slightly larger than my house but still be within the confines of my property and produce about 60 Mw. Infact as far back as the 70’s PSE&G owned so many jet engines that Pratt and Whittney recognized the company as the world’s third largest airline.Simple cycle GasTurbines could be started, synchronized, and producing power in about 7 minutes. Combined Cycle Gas Turbines do take longer to start and are used for different purposes because of their higher efficiency. Such as base load, load following, cycling, etc.. I chose to compare their size to a residential property to emphasize in a familiar way the space required for this older technology. I thought this discussion was about Mw per acre per type of generation. CT’s should win hands down on this parameter alone.

    By Joe Llaben on 2014 10 05

    Reply to this comment / Quote and reply

  • “By 1976 all 55 miles of the Serpent River system were badly contaminated with acid generating, highly radioactive wastes. An official Ontario report noted that there were no living fish in the entire river located downstream from the mining wastes.” Did your land use calculation include this sort of land use (the mine in question produced ~93M tonnes of ~0.1% uranium ore)?

    By GeraldR on 2015 06 26

    Reply to this comment / Quote and reply

  • Why are you still believing EIA LCOE?  They NEVER get cost even close!
    search lazard energy version 8 if that does not work.

    Meanwhile you don’t use rooftop solar which uses no concrete pad, there are also pole mounted ground solar without concrete.

    You also didn’t count the massive amount of material used to get at the fuel.  average ore is .1%, so a thousand tuns per ton of yellowcake, then 8 tons per ton to refining to fuel.  That’s 8000 tons per ton of fuel.  Please overburden, and the like. 

    cherry picked all the way through.

    By Brian on 2015 08 08

    Reply to this comment / Quote and reply

  • Thanks Brian for reading my article of 2013. In my experience, LCOEs from EIA are not that different from IEA. Both being well accepted institutions for supplying LCOE information. Perhaps these LCOEs have changed over the last 2 years.

    By Martin Nicholson on 2015 08 08

    Reply to this comment / Quote and reply

  • For the record, I’m fairly pro-nuclear power, but I think the renewables boom has made it a lot less important.

    I actually had a discussion about EIA and IEA calculations for LCOE the other day. For awhile now, I’ve put a lot of stock in them. However, if you look at other estimates, it’s hard to square those numbers, especially for solar PV. For example, a Brattle Group report puts utility scale PV at around 6.6 cents- 11.7 cents, not accounting for transmission or a handful of other costs. Rooftop is more like 12-19 cents/kwh.

    It is a solar industry study, so maybe there is a bit of bias. Also, the costs they exclude relating to the grid could be significant. Still I find it funny that the EIA doesn’t even differentiate between rooftop and utility scale solar PV in its main table. That’s probably where some of the high cost comes in.

    However, it wasn’t just the EIA that made me think solar PV was costly. The IEA, and even an ecofys paper from around 2014 seemed to say so. However, if you look closely, while the paper itself is from 2014, the data for LCOE is based on 2008-2012 information. This makes me think that there is a huge data lag when it comes to cost estimates for solar PV. Normally a lag of a few years doesn’t make that much difference for calculations, but when the price of something is dropping at this speed, it can lead to some major errors.

    However, if you go to the other end of the spectrum, and look at solar PV for Lazard, they’re so much lower than anyone else’s that I can’t help but be suspicious.

    The German Fraunhofer institute should be pretty friendly towards solar PV, and their report for 2013 show solar PV as being more competitive than the EIA does, but still nowhere near projections by.Lazard or Brattle group.  Then again, their estimates are from 2013, and based on data from Germany, which is not famed for its solar insolation. If we apply the Fraunhofer estimates to a region like Colorado, which the Brattle group estimates are based on, I think they’re more or less on the same page. I still can’t make sense of the Lazard estimates though. They just don’t fit. It almost looks like they clipped off the middle and upper end ranges for solar PV cost. How, I don’t know.

    So all in all, it looks like in sunny regions, the range of utility scale solar PV costs overlap fairly well with the range of costs for fossil fuels and nuclear, and will likely be cheaper on average very soon. At this point, most sources seem to agree that they have at least become comparable. Onshore wind has been cheap for awhile now. The only problems are intermittency, transmission, high demand for materials,storage- things like that. Yet those aren’t likely to be serious issues until we reach much higher penetration levels. So for the time being, this pro-nuclear guy is cheering the renewables boom on.

    By Troy Kokoszka on 2015 09 04

    Reply to this comment / Quote and reply

Submit a comment