The book tells us we can run everything on renewable energy and get rich exporting it. Unsurprisingly it has been met with much fanfare and enthusiasm. But on the crucial issues it either provides flimsy and unconvincing analyses or fails to deal with them at all.
Whether or not everything can be run on renewable energy is hotly debated among experts in the field, and far from settled. Most people assert that it is possible, and indeed easily done, if only the politicians were not so stupid. Many studies and institutions also conclude this. But many heavy academic analyses conclude that it cannot be done at an acceptable cost in plant and dollars. I have published about 15 such papers explaining in numerical detail why I don’t think it can be done. (E.g., Trainer, 2017). I might be dead wrong, but at this stage no one has much justification for making confident assertions one way or the other.
The issue is complicated and technically difficult to clarify. Above all, the findings depend greatly on assumptions made, and in several areas we do not know what the correct assumptions are. Much, if not most, literature even in the academic arena is discursive, speculative, impressionistic, about what “could” be, and not capable of arriving at sound conclusions. There is only one approach that can eventually settle the issue, and unfortunately, there are few studies of this kind.
That approach takes the form of “simulation” studies using detailed, hourly solar and wind patterns for all locations in a region over a year, and attempting to work out what combination of renewable generating technologies and storage options will minimise the cost of energy delivered at a particular level of reliability. Conclusions can only apply to that region and can’t be generalised as a global finding. Maybe only two teams have tackled this in sufficient detail in Australia.
Because of the intermittency of sun and wind, these kinds of studies have found that the amount of generating plants needed to meet demand all the time, as there is no sun here today but there is wind over there, and neither in either place the day before, is several times the amount of plant needed in the form of fossil-fuelled generators. Some studies find that for a particular region it is three times as much plant as would be needed if all PV panels and turbines and so on were in ideal conditions all the time, and some find that it is up to 10 times. Lenzen et al. (2016) found that it would be six to seven times as much plant as could meet the Australian 23 GW demand via coal-fired generation, and they found that the production cost would be about seven times that of coal-fired electricity at the time. Again, these findings depend greatly on the assumptions made in the study.
What we need are many studies of this kind, making differing sets of assumptions, and only when consensus began to emerge on what the best combination could achieve could we move towards confidence about the answer.
Garnaut’s book makes no mention of this field and offers no derivations from such studies. It simply asserts claims and might-be’s, and enthuses about such things as falling costs and what carbon sequestration might do, without any effort to show that a particular combination of these technologies is collectively capable of doing the job at an acceptable cost. No technical case is given, no numerical analysis based on quantities for demand, generating capacities, energy conversion losses, transmission plant and losses, storage needed in what form, or costs for these factors, or the embodied energy costs of plant. Especially significant is the absence of a satisfactory discussion of the vexed EROI issue, the energy cost of producing various forms of energy.
He enthuses about biomass energy but does not worry about the fact that its EROI is very low, with some studies finding that biomass energy takes more energy to produce than it yields. EROI values for renewables are much lower than for fossil fuels, and recent studies of whole systems combining various renewable sources and storage provision indicate that the overall system EROI value might be in the region of six or, indeed, well under this. (My 2018 study of a hydrogen path estimated it at about six.) Some analysts think that an EROI under 10 cannot sustain an industrial society.
The book’s basic claim is that greatly increased quantities of electricity can be provided solely by renewables, reliably and at a reasonable cost. Most other sectors depend on this, notably transport. Here is Garnaut’s entire case in support of his electricity claim:
“After a dozen years of close acquaintance with the Australian and global energy transitions, I now have no doubt that intermittent renewables could meet 100 per cent of Australia’s electricity requirements by the 2030s, with high degrees of security and reliability, and at wholesale prices much lower than experienced in Australia over the past half dozen years.” (P. 71.)
In other words, Garnaut deals with none of the technical difficulties noted above and fails to indicate their magnitude. Consider, for example, the scale of the storage problem. Lenzen et al. (2016) found that to maintain Australian electricity supply through the worst week in 2010, we would have needed about half of the week’s demand in storage at the start of it. That means we would have needed to draw .5 x23 GW x 24 hours x 7 days = 1,932 GWh of electricity via stored hydrogen, and at the approximate hydrogen-to-electricity conversion efficiency of fuel cells we would have needed 4,830 GWh equivalent in the form of hydrogen stored in tanks, just to deal with electricity demand over that week. Snowy 2.0 will only store 350 GWh (and deliver less given its distance from most users.)
The biggest storage problem is set by inter-seasonal variation. For example, in Australia there is about twice as much solar energy in summer. To illustrate, if we had enough PV to meet average demand then we would have to store about one-quarter of summer generation to top up winter supply, which would be at least 4 months x 30 days x 24 hours x 0.25 x 23 GW = 16,560 GWh, meaning 48,900 GWh would have to be stored in the form of hydrogen to convert back to electricity. For electricity alone, 140 Snowy 2.0s? (Wind is somewhat better in winter than summer so a wind + solar system would reduce this figure.)
As for getting rich exporting hydrogen, consider the difficulties moving very light hydrogen gas. Even short distances within Australia would involve significant energy costs. According to Bossel and Eliason, a large, for example, 40 tonne, truck carrying a tank full of hydrogen at 200 bar pressure could deliver only about 400kg of hydrogen. These authors calculate that to move a unit of hydrogen energy this way would require about 32 times more truck fuel than moving energy in the form of petrol. Delivery to Asia by pipeline over thousands of kilometres might not be feasible because Bossel and Eliason say piping from North Africa to the UK would cost 30-40% of the energy delivered. “Too much energy is lost in the process to justify the scheme.” Note that the small hydrogen atom leaks easily and makes metals brittle, meaning that the pipelines would have to be lined. Again, add the costs for pumps, tanks and other infrastructures and subtract their embodied energy costs.
But all that is only for electricity, the easier problem, and 80% of energy demand is not for electricity. When you go on to the task of fuelling transport etc. via hydrogen the difficulty numbers escalate. If half of that presently non-electrical demand could be converted to electricity, and the rest could be met via hydrogen, the total electrical generation task would multiply by about 10, and the storage etc task would increase accordingly. Before you enthuse about running everything on sun and wind and getting rich on exporting the surplus, make sure my arithmetic is way out.
You had better hope that they don’t work out how to provide abundant cheap renewable or nuclear energy because if they do they will use it to accelerate consumer-capitalist society and hunt down the last shrimp in the sea.
There are many savage limits to growth, and energy is only one of them. The more energy they get the faster they will go through most of the other “planetary boundaries”. For 50 years the now vast “limits to growth” literature has made it clear that the range of global problems now threatening our very existence can only be solved by dramatically cutting resource use and environmental destruction and therefore consumption. As the DE-growth movement recognises, this means scrapping consumer-capitalism and accepting shifting down to a small fraction of present GDP and embracing very frugal and self-sufficient lifestyles and systems. (See thesimplerway.info/, and simplicityinstitute.org.) Garnaut’s Superpower reinforces the faith that there is no need to think about any of that.
Lenzen, M., B. McBain, T. Trainer, S. Jutte, O. Rey-Lescure, and J. Huang, (2016), Simulating low-carbon electricity supply for Australia, Applied Energy, 179, Oct., 553 – 564.
Trainer, T., (2017), “Can renewables meet total Australian energy demand: A “disaggregated” approach”, Energy Policy,109, 539-544. https://doi.org/10.1016/j.enpol.2017.07.040
Trainer, T., (2018), “Estimating the EROI of whole systems for 100% renewable electricity supply capable of dealing with intermittency,” Energy Policy, vol. 119(C), 648-653.