I recently understood the elegance of H.T. Odum’s concept of transformity.
The transformity table presented at the link above shows that electricity has a transformity of 300,000 seJ/J, where seJ stands for solar-equivalent joule. It’s a weird unit, so I tried to digest it as follows: to produce electricity requires 300,000 solar-equivalent joules per joule. By contrast, sunlight is by definition 1 seJ/J. The numerator is the emergy input to the energy source / carrier in question. (Brief aside about emergy: the reason that emergy is measured in “solar-equivalent” joules, as I understand it, is so that there is a common basis for embodied energy comparisons. One trap that’s easy to fall into when doing such calculations—one I’ve fallen into—is to do a calculation based upon different sorts of energy inputs, and then to combine them together as if they’re equal, like oil and electricity. By using solar energy as the root, this issue is mostly resolved.)
Back to transformity: the confusing part is the measurement of the denominator. The joule in the denominator can be thought of as the useful work you can extract from the energy source. To put it more concretely, since a joule is a bit ugly as a unit (I don’t like thinking in newtons), we can translate all the transformities into calories (i.e. solar-equivalent calories / calorie). So to heat a gram of water 1 degree using sunlight takes 1 solar-equivalent calorie. But to do the same thing with electricity takes 300,000 solar equivalent calories. That’s a massive difference. Heating water using sunlight is not just a bit more efficient than using electricity or fossil fuels—it’s multiple orders of magnitude more efficient.
It seems the main reason for electricity’s high transformity is that the fossil fuels used to produce that electricity have high transformity. That in turn is because of the quantity of biomass transformed through geologic time (i.e. that biomass’s emergy is large) to produce fossil fuels.
From a practical (shortsighted) standpoint we don’t really care that biomass and geologic forces went into making fossil fuels, as long as we ignore that we’re running out of them. So you could think of rescaling the transformity of electricity to a fake unit we could call a “nature-equivalent joule / joule” (neJ/J) and treat seJ/J as equivalent. (The idea being that as long as a joule came from “nature”, I don’t care if it’s from the sun or from a chunk of coal.) Then we could rescale electricity by the value of coal to something like 4.5 neJ/J, which tells us even in this case using solar to heat water is 4.5 times more efficient than using electricity to do the same task. (However, using neJ/J, using coal directly is by definition as efficient as solar.)
There are some interesting ideas in the post linked above, such as Odum’s conjecture about how solar PV might be inefficient because it is trying to leap too far in transformity. The open question there, in my thinking, is whether electricity is fundamentally of high transformity. It would be interesting to see the analysis done again for CdTe cells, for example, since they may have high EREOI.
The reason to repeat that analysis is another reason that transformity is a great concept. EROEI, along with most net energy variants, have the weakness of conflating different sorts of energy inputs, and also of ignoring the fossil fuel subsidy provided to all industrial processes. Even if CdTe photovoltaic cells have high EROEI, that doesn’t mean that they have low transformity. And post-fossil fuels, it’s tranformity that will matter.
Hemenway made a good case for the resilience of the food system post-peak oil, and he argues, I think with good reason, that we ought to be concerned about the sustainability / stability of the numerous high transformity systems we depend upon in industrial societies before we worry about the food system. The transformity of today’s food system is low compared to banking and the electric grid and so on—and that’s today’s food system, which is more complex than it ought to be or needs to be.
In any case, I’m reminded of a great saying, one that I now understand is backed up by the idea of transformity: “let light be light; let heat be heat; let food be food.”