A fundamental requirement that any energy supply system must satisfy for economic viability is a sufficiently high energy return on energy investment (EROI) for manufacturing, installing, operating and maintaining the system over its operating life. The question of what constitutes a sufficient return depends on the nature of the economy and society that the energy supply system is intended to support—while an EROI <1 implies a net energy sink, an EROI >1 does not automatically entail viability. Consider the limiting case in which net energy supply is zero, i.e. EROI =1. This would entail an economy consisting entirely of an energy supply sector that supported itself, but allowed for no economic activity beyond this. It’s certainly possible to imagine a functional economy along such lines, but it implies that every person living in such a society must dedicate their life to and focus all of their attention and effort on providing for the subsistence energy needs of their economic system. Such an economic system would serve no purpose beyond its own perpetuation; citizens of such a society might very well consider their lives to constitute a form of slavery to their economy.
Here, though, we start to move into a different area of inquiry, as it would be difficult to mount a case that a world in which all people did happen to find rich and enduring meaning in lives dedicated to the service of their energy infrastructure was not viable. Suffice to say at this point that, while it’s possible to imagine such a world in the abstract, the historical and anthropological record suggests as unlikely any real society subsisting for long in such a manner. By considering the limiting case—what we might regard, at least in the abstract, as “minimal viability”—we can see that for the pursuit of lives along the lines that humans tend to find meaningful, lives guided by purposes other than hand-to-mouth subsistence in which all effort and attention is focused on providing sufficient energy for basic survival alone, an EROI greater than 1 is required.
How much greater, though, is a matter of the purposes that guide a society’s members, and the ways of life that are meaningful for them. Where these entail an extensive and diverse material basis, higher EROI will be required than where these are materially more modest. Consumer societies based on global trade of commodities and manufactured goods unavoidably require higher EROI than societies where people are content to spend their time engaged in locally-situated contemplative and aesthetic pursuits. So ultimately, the question of sufficiency in relation to EROI is inseparable from questions of how we wish to live, and of what we find meaningful. In this respect, it’s noteworthy that so many popular sources of meaning in modern industrial consumer societies are themselves products of modern industrial economies, and hence of fossil-fuelled civilisation.
What many of us see today as essential to living worthwhile lives has arisen only in the context of the availability of what is now apparently essential (note the circularity). Given this situation of context-feedback in relation to human aspirations, one obvious way to approach the question of how high EROI must be in order to consider an energy supply system economically viable is to use an existing way of life as the benchmark. This is the approach that Hall, Balogh and Murphy take in their preliminary research that I reported on previously. In this sense, at least in principle, it is possible to establish a reference point for the minimum overall EROI that a society requires.
Solar PV through the EROI lens
With this in mind, recent studies by Pedro Prieto & Chales Hall  and Graham Palmer , suggest it would be unwise to assume that the forms of industrial consumer society that have developed in the context of fossil fuel energy systems can transition to renewable energy systems, at least while these forms of social organisation and their associated life-ways remain unchanged. Using actual performance data, Prieto & Hall found that for utility-scale solar photovoltaic (PV) electricity generation in Spain the EROI may be as low as 2.45. Their analysis indicates that the overall EROI for PV is strongly limited by upstream “balance of system” (BOS) energy costs for infrastructure additional to the PV modules and related equipment, and by downstream losses specific to PV (i.e. losses not incurred by other electricity generation technologies). Palmer’s study considered the situation for PV generation in Melbourne, Australia, and focused in particular on constraints associated with its intermittency. Intermittency strongly limits the proportion of PV generation that can be introduced to existing grid-based electricity systems developed in the context of conventional thermal and hydro generation.
This limitation can be reduced by introducing some form of energy storage—but at the cost of further energy investment that significantly reduces the overall EROI. Overall EROI is strongly dependent on the amount of storage capacity added to the supply system. Taking into account the BOS energy costs, including downstream losses similar to those included by Prieto & Hall in their study, and adding four hours of lead-acid battery storage (he characterises this as “a moderate amount”, in terms of its impact on the overall PV penetration that can be accommodated), Palmer arrives at an indicative EROI of 2 to 2.3, depending on the PV cell technology. While his analysis is subject to significant uncertainties, it highlights the extent to which EROI is dependent on factors other than the energy costs associated with emplacing the basic equipment required by grid-connected solar PV electricity generation facilities.
An aside on intermittency
It’s important to note here that Palmer’s assessment of intermittency issues is offered with full awareness of the arguments against their significance from proponents of a “smooth transition” to high penetration renewable electricity generation. It’s now common to encounter outright dismissal of concerns relating to intermittency, with justifications for this position advanced on both theoretical and empirical grounds. Theoretically, proponents typically argue along the lines that “the wind is always blowing (or sun is always shining) somewhere”, and so with a sufficiently wide geographic distribution of generators, electricity will always be available. This is the general stance adopted by Mark Jacobson & Mark Delucchi in a series of articles published in the journal Energy Policy assessing the feasibility of providing all global energy from wind, water and solar power., This theoretical approach typically neglects (or dismisses) an obvious implication: sufficient generating capacity must be located in every region from which all demand might need to be met at any particular time. This translates to very high levels of redundancy, and commensurate capital cost. I will discuss this in more detail in concluding Part 3 of this series.
Empirical arguments tend to focus on the experience in regions with high levels of renewable penetration, especially Germany, and more recently the Australian state of South Australia. These arguments often neglect the fact that it is the level of penetration for grids as a whole which is important, rather than the level of penetration for a given political region or geographic area. For example, around 30 percent of electricity generated in South Australia annually is from wind and solar PV. South Australia, though, is a participant in Australia’s National Electricity Market and is connected to the east coast grid, stretching from Queensland in the north, through New South Wales and Victoria, to Tasmania in the south. Two interconnectors link South Australia with the adjacent state of Victoria, allowing the state to both import and export electricity. As such, South Australia’s large proportion of wind and solar electricity needs to be considered in relation to the generation mix for the entire grid. The high proportion of gas thermal generation that can be brought on-line in South Australia is also an important consideration here. That the high proportion of wind and solar generation for the state is viable without storage does not confirm that intermittency is not an issue. This would be established empirically only if the overall level of intermittent renewable generation for the grid as a whole matched that of South Australia viewed in isolation.
EROI implications for renewable energy’s technical potential
Previously I’ve discussed how neglecting EROI when thinking about the availability of fossil energy sources significantly overstates resource potential (see posts here and here). As the comparatively low EROI figures for electricity supply from solar, wind and biomass suggest, the overall situation for the major non-hydro renewable energy sources is even more pronounced. This has major implications for estimating the technical potential of renewables—the upper bound on the rate at which they could in principle provide end-use energy to enable economic activity other than energy supply itself, taking into account constraints associated with system performance, topographgy, land use and environment.
Technical potential will necessarily be significantly less than overall resource potential—the energy ultimately available based on physical considerations. To illustrate the difference, the incident solar radiation at the Earth’s surface is approximately 3,900,000 EJ (exajoule, 1018 J) per year, compared with conventional estimates of technical potential for solar electricity generation ranging from around 120 to 2,600 EJ per year.[7, Table 3] For wind, total Earth energy flow is around 28,400 EJ per year, compared with conventional estimates of technical potential between 48 and 600 EJ per year.[7, Table 3] For reference, world primary energy supply in 2011 was 13,130 Mtoe (million tonnes of oil equivalent) or around 550 EJ.
The ratio of input to output energy over a system’s lifecycle is a central consideration in relation to system performance and hence for estimating technical potential. But as Patrick Moriarty and Damon Honnery discovered in drawing together the figures above on technical potential for solar and wind generated electricity, it’s rarely apparent that this is adequately taken into account. This is reflected in the order-of-magnitude range in estimates.
Variation in EROI from the first to the last unit of production plays a significant role here. Moriarty & Honnery point out that the energy ratio (their preferred term for EROI) will change as a resource is exploited. Initially, as the best sites are developed, EROI is typically high, but declines as production expands to more marginal sites. If exploitation of the resource expands far enough, eventually the remaining sites will have EROI less than 1. At that point, overall production has reached its maximum, and any further expansion will reduce the net rate. Consequentially, overall technical potential must take into account the way that EROI varies from the most to the least productive sites. It is not appropriate to take the highest attainable EROI as typical.
Moriarty & Honnery illustrate this for global wind potential, using an analytical approach based on siting a 2 MW turbine on each square kilometre of the Earth’s available land area. The energy ratio across all sites ranges from around 22:1 down to zero, but around half of the annual gross electricity output is delivered at energy ratios of less than 8:1. Similar distributions of gross energy return versus energy ratio can be determined for any desired generation scenario. For instance, Moriarty & Honnery also consider a land constrained scenario, in which environmentally sensitive forests and wetlands, urban areas and irrigated farm land are excluded. Correcting for the marginal energy ratio as further generation capacity is added from highest energy return sites to lowest, the net energy can be determined for a given generation scenario. In the unconstrained case (where all land area is treated as available for turbine installation), this leads to net energy return reaching a maximum value of around 420 EJ/year, corresponding to gross energy supply of around 500 EJ/year—that is, a reduction of 16 percent. This can be compared with the difference if we simply assumed that all wind generated electricity was available at the best energy ratio of around 22:1. In this case, net energy would be around only 4.5 percent less than gross energy. For the land constrained case, net energy peaks at around 290 EJ/year net, against 350 EJ/year gross—a reduction of 17 percent.
These scenarios exclude the effects of energy storage. As discussed earlier, storage becomes increasingly necessary for system reliability as the proportion of intermittent generation included in the total electricity supply expands. Storage, or buffering, allows a degree of decoupling between generation and demand, by diverting excess electricity towards an appropriate storage system during periods where available supply exceeds demand. When instantaneous demand exceeds generation capacity, stored energy, if available, can then be converted back to electricity to make up all or part of the shortfall.
Recognising the importance of storage for electricity systems with high proportions of intermittent renewable generation, Moriarty & Honnery also consider a third scenario in their study of global wind energy potential. For this situation, they start with the land constrained case and consider the implications of adding energy storage in the form of hydrogen, produced via electrolysis of water. Storage of any type comes at a further energy cost. This includes the embodied energy associated with the additional equipment and infrastructure, and the losses associated with the conversion cycle from electricity to the stored energy form, and from this back to electricity again. This additional energy cost will vary with the type and scale of storage employed, but in all cases it will reduce the overall energy ratio for the system, and hence the net energy.
For the scenario that Moriarty & Honnery consider, the net energy output reduces to 105 EJ/year, for a gross output of 170 EJ/year. Notice that in this case, the addition of storage entails a 38 percent reduction in net output compared with gross output, compared with a 17 percent reduction without storage. While the absolute figures involved here are dependent on a wide range of variable parameters associated with the specific technologies involved and the way they are deployed, it is the relative difference between the net output with and without storage that tells the important story here. From the initial unconstrained gross output of around 500 EJ/year (recalling that this is based on the assumption of situating a 2MW turbine for each square kilometre of the Earth’s available land area), and accounting only for a series of realistic constraints associated with land use and supply reliability (i.e. ignoring other plausible constraints), the net energy available may be reduced by as much as 80 percent.
While the specific example here focuses on wind, the general principles involved apply to all renewable energy sources. It’s particularly noteworthy that for all three scenarios considered, adding further capacity beyond the maximum or peak level—that is, adding sites with energy ratios less than 1:1—results in a net decrease in energy output. In other words, the reduction in energy ratio at the margin implies that as renewable energy resources are developed, if this development proceeds far enough, we should expect to converge eventually on an optimum or peak supply level. If the aggregate optimum level with all sources added falls within the range of anticipated or desired future energy supply, then this implies that humanity faces fundamental energetic limits.
Moriarty & Honnery argue that the maxima for all sources do indeed sum to a level within the range of typical expectations for global growth in energy use out to 2050. Furthermore, when a range of other constraints associated with environmental (especially climate change related), political, social, economic and transition timing factors are taken into account, it appears highly unlikely that renewable energy sources have sufficient potential to meet even current global energy demand.
Next week, in Part 2 of this three part series, I’ll take this extended look at EROI a step further by examining dynamic consequences related to the rates of energy expenditure and return, and how these are staged throughout the life-cycle of an energy supply installation. These dynamic effects usually remain hidden from view when we take the overall life-cycle as the basic unit of analysis. Following this line of inquiry, I will build the case for introducing power return on investment as a complementary indicator to EROI for assessing and comparing the viability of different energy supply systems.
 Palmer, G. (2013), “Household Solar Photovoltaics: Supplier of Marginal Abatement, or Primary Source of Low-Emission Power?”, Sustainability, Vol. 5 No. 4, pp. 1406-1442.
 Jacobson, M. Z. & Delucchi, M. A. (2011), “Providing all global energy with wind, water, and solar power, Part I: Technologies, energy resources, quantities and areas of infrastructure, and materials”, Energy Policy, Vol. 39 No. 3, pp. 1154-1169.
 Delucchi, M. A. & Jacobson, M. Z. (2011), “Providing all global energy with wind, water, and solar power, Part II: Reliability, system and transmission costs, and policies”, Energy Policy, Vol. 39 No. 3, pp. 1170-1190.