Last week I had the opportunity to attend a lecture at the University of Bristol’s Cabot Institute given by Prof. David MacKay, chief scientific adviser to DECC (UK Government Department for Energy and Climate Change). There were two main focuses of his lecture: firstly, a discussion of various sources of energy and secondly, an introduction to DECC’s 2050 pathways tool.
The pathway tool I like and I would encourage anyone interested in the UK’s future energy system and associated carbon emissions to have a play with it. All I would say is that, like many scenario analyses, it is too narrow. The whole point of carrying out scenario analysis is to explore the possibility space. The DECC tool assumes both population and GDP growth. These parameters may be “out of the scope of consideration”, but I would have liked the energy and emissions tool to allow the exploration of steady-state and also economic contraction scenarios. I got the impression MacKay would also have liked to include this flexibility but he said Westminster wouldn’t allow it.
I did not like his comparison of energy sources though. He promotes the use of a single metric to compare energy sources: power density. This means the amount of power delivered per unit area, expressed in watts per square meter [W/m2]. The lecture focused on wind and nuclear power but the analysis can be done for solar power, energy crops, or fossil fuelled power stations.
The headline results were that wind has a power density of 2.5 W/m2 whereas nuclear delivers 1000 W/m2. Sounds good for nuclear and not so good for wind! But the difficulty is that MacKay is comparing apples and oranges.
In order to compare things quantitatively as MacKay is attempting to, it helps if units are the same. MacKay’s m2 of wind farm are not the same as his m2 of nuclear power station. There are three main problems with this analysis:
- Layering
- Time
- Externalities
Layering
Cows in a wind farm
The square km of land underneath a nuclear power station is 100% used up. There is nothing else that land can be used for. However, with many renewables the land isn’t used up in a comparable way. Solar panels can be installed on the top of existing buildings requiring none of the underlying land to be used up. Wind farms use around 5% of the land under the turbines, leaving the remaining 95% available for other uses (such as livestock or crops). This 5% compared to 100% improves the power per unit area of wind turbines by a factor of 20.
Time
MacKay made no allowance for the time dimension. He just divided the power of a wind farm or power station by its area. This fails to consider that the nuclear power station took at least 10 years to build before its ~40 year generating lifespan, followed by a ~100 year decommissioning period. In contrast, the wind turbines are generating within months of build commencing and decommission can be similarly swift. This results in the nuclear power station using up the land for around three times longer than the period of time it is generating for, which effectively reduces its power per unit area by a factor of three.
Externalities
MacKay also made no allowance for the land requirements outside the perimeter fence of ether the nuclear power station or wind farm. This discounts the land required for the uranium mine, the uranium processing, the water required for cooling and importantly the waste storage. The wind turbines also required an iron ore mine, steel foundry and factory. I am not able to quantify the differences in land requirement but I expect the nuclear power station’s “invisible footprint” to be larger, especially when multiplying up the area used for waste storage by the duration for which the land is required (potentially many thousands of years) as described above. Finally, nuclear power stations have a non-zero probability of catastrophic failure, then requiring exclusion zones of hundreds of km2 for decades (Chernobyl, Fukushima).
A Comparable Analysis?
A comparable analysis would consider the fractional land use (layering) of an energy source, the total duration for which this land was used (time) and the land required beyond the immediate installation (externalities). That MacKay’s analysis doesn’t consider these aspects, and that they impact the final results by many factors suggests to me that this metric of comparison is oversimplified. I do not object to the use of the power density metric but would like to see it done properly; otherwise it is comparing apples and oranges and is not useful information.
I don’t doubt that MacKay has considered the points raised above. I am worried that the seemingly-deliberate omission of these factors is presenting an overly political bias towards one source of energy.
According to the above back-of-the-envelope estimates, I would therefore amend MacKay’s comparison of nuclear (1000 W/m2) and wind (2.5 W/m2) to the more realistic 300 W/m2 (accounting for time) and 50 W/m2 (accounting for layering). These adjustments reduce the difference between nuclear and wind from 400- to 6-fold. A further unquantified adjustment to account for externalities is likely to reduce this still further.
Of course, in the final analysis the total land area that is needed is reflected by the naive energy densities MacKay calculates – to generate most of our power from wind (or solar, or biomass) would indeed require vast proportions of the countryside or sea surface to be utilised, and this is an important consideration. However, given the above considerations, it is clear that the headline numbers MacKay is promoting are unfair to renewables, and overly generous towards nuclear.
