In the previous post in this sequence, I developed the concept of power return on investment as a complementary indicator to energy return on investment (EROI) for assessing the viability of wind and solar PV as alternatives to thermal electricity generation. I used as my departure point for this an article in which Ioannis Kessides and David Wade introduce a dynamic approach to EROI analysis. Specifically, I drew on an illustrative example that they present, based on IEA data for coal-fired thermal and wind electricity generation in Japan, showing how the time required for coal and wind installations to provide sufficient energy to emplace additional generating capacity equal to their own can differ by an order of magnitude even where the EROI for coal and wind is identical. Given that the data on which this example was based was from prior to 2002, both the doubling time in Kessides & Wade’s example and the power return on investment in the extended analysis would likely be improved if up-to-date figures for emplacement energy and capacity factor were substituted for those from the IEA study. Unfortunately, this goes only a limited way to mitigating the central issue in terms of “real world” considerations.
Further implications of “uncontrollable variability”
The reason for this relates to the intermittent nature of wind as a primary energy source—a characteristic that also affects solar-generated electricity, which faces the added impediment of seasonally variable insolation. A key consequence of intermittency is that in order to maximise the proportion of total electricity from wind and solar sources in a supply system configured to meet typical demand patterns, some form of energy storage is required. In the previous post, intermittency and seasonal variation were taken into account simply by increasing the nameplate capacities of wind and solar installations, based on their respective capacity factors and operating & maintenance plowback rates, in order to match their lifetime average power outputs with that of coal. This averaging approach, however, provides only a very abstract way of thinking about the problem. It leaves aside the matter of electricity users, for the most part, expecting electricity on demand, rather than shaping their expectations to match irregular or seasonally variable availability.
As discussed briefly in Part 1, no amount of additional nameplate generating capacity can make up for situations where the wind is not blowing or the sun is not shining. Proponents of 100 percent renewable electricity systems argue that this need not matter if the supply system is distributed sufficiently widely, and if the system combines generating capacity from complementary sources. If the total generating region is large enough, then the wind will always be blowing (or the sun shining) somewhere. In the abstract, if the hypothetical generating region is large enough, this may be the case. However for actual electricity networks presently operating, historical data indicates that lengthy system-wide periods of overcast, calm weather (or windless conditions coinciding with winter minimum insolation) are not uncommon—see, for instance, Ted Trainer’s work relating to this.;;
But even if the argument was supported in practice—perhaps by the construction of continental scale super-grids—reliably meeting total electricity demand would entail having sufficient generating assets emplaced in every region that might be expected, at any given time, to carry the weight of the supply task to satisfy all demand. This would involve the construction and maintenance of an enormous quantity of redundant generating capacity—assets that for much of their operating life would contribute to an oversupply of electricity. We could reasonably expect then that any improvement in capacity factor and emplacement energy over the period since the IEA data used by Kessides and Wade was published would be more than off-set by the additional assets required to address supply reliability.
The nature of this problem is particularly apparent when considering solar PV generation. Due to the annual seasonal cycle, average insolation at every point on the Earth’s surface changes continuously from a mid-summer maximum to a mid-winter minimum, with the range of variation increasing with latitude (close to the equator, there are two maxima and minima each year). The generating capacity required to meet a given level of electricity demand at a particular location in winter is therefore unavoidably greater than in summer, and a supply system designed to meet this demand in winter will have excess output in summer. In places where the demand pattern matches well with the annual supply cycle—for instance, due to summer air conditioning load—this may not be problematic. But at higher latitudes where demand patterns are dominated by winter heating loads, the problem is in fact made worse: in a scenario where all energy supply was to be met by renewably-generated electricity, minimum output would coincide with maximum demand, and so the oversupply in summer would be increased.
It is often hoped that wind will complement solar PV in such situations i.e. that wind output will be maximised when PV output is minimised. But it is easily demonstrated that calm periods in mid-winter would lead to electricity supply shortfalls in regions with insufficient dispatchable thermal or hydro generation capacity to meet the entire demand task—in other words, without unusually high levels of backup capacity that would remain idle for most of its life. This has significant capital cost implications—not only in financial terms, but also in physical energetic terms i.e. in terms of energy that must be diverted from other purposes in order to emplace and maintain the backup capacity.
Energy storage presents an alternative (or, depending on context, a complement) to geographic dispersal of generating assets and dispatchable backup capacity, but provides no panacea in relation to the financial and energetic capital cost issue. Patrick Moriarty and Damon Honnery have specifically studied the implications for net energy return of combining hydrogen-based storage with wind turbines, and in particular the way that supply capacity growth rate affects this. For the model conditions that they consider, they found that, due to the high up-front energy cost of the coupled wind turbine and hydrogen storage installations, high growth rates lead to “boom and bust” cycles, both in equipment manufacturing and in net energy output. These cycles would occur with a period equal to the operating life of the equipment. If a “crash program” is instigated to roll out a large amount of capacity very quickly, then this will lead to a subsequent period around 25 to 30 years later (the typical expected operating life for the installations) where net energy available for purposes other than replacing electricity supply infrastructure drops precipitously. The depth of decline depends on the rate of initial expansion—the higher the expansion rate, the more pronounced the subsequent drop. For wind turbines only without storage, the drop during periods of end-of-life asset replacement could range from around 12 percent at low penetration of wind generation and low initial expansion rate, to over 60 percent at high penetration and high rate. But when storage is added, for the model conditions that Moriarty & Honnery assume, a high initial expansion rate could result in the net output declining to a level effectively below zero, even for low penetration—that is, the total net energy output from wind would be insufficient to supply the energy required to replace the turbines and storage systems as they reached their end of life. Supplementary energy would therefore be required from other sources, just to support the roll-over of assets.
In any of these scenarios, the supply gap could be filled by other electricity sources, but this again would entail emplacing large quantities of redundant generating capacity, with obvious capital cost implications. The higher the rate of initial expansion in wind, the larger the redundant capacity from other sources required to support it. Adding storage results in an extreme situation where it could even be necessary to temporarily provide capacity from alternative sources greater than the total provided by wind generation during normal steady-state operation outside the replacement periods. Moriarty & Honnery conclude that this places basic limits on the maximum rate at which we might transition to high penetrations of intermittent-source electricity generation. Avoiding large periodic declines in net energy available to consumers (and corresponding spikes in electricity cost) would require a carefully staged and gradual roll-out of supply assets.
Moriarty & Honnery also highlight a corresponding problem that arises where an initial expansion program for wind and solar generation is funded by fossil fuels. If the newly emplaced renewable energy infrastructure is to displace a large amount of the fossil fuel infrastructure that enables its manufacture and installation, then due to high up-front energy costs, this will entail an initial increase in fossil fuel use and associated emissions in order to continue satisfying existing energy demand from other economic activity. This is an issue that Tom Murphy explores via a concept that he calls The Energy Trap. He makes the very important observation that where fossil fuel production rates have plenty of headroom for further expansion, it may be possible to simply increase overall production in order to fund the emplacement of renewable energy capacity, without compromising ability to meet existing demand. But if, when the transition program begins in earnest, production is already pushing up against its ceiling—due to limits imposed by some combination of geology, technology and economics—then that program can proceed only at the expense of existing demand. This is not some distant hypothetical situation. With respect to petroleum, we may have been living within such a regime since around 2005, when global production reached its current bumpy plateau. End use energy efficiency improvements can in principle off-set some of this—but the extent of compensation potentially available is subject to considerable uncertainty.
Murphy’s central point is that under such conditions, the harder we push a transition away from conventional energy sources, the more we starve other economic activity of its lifeblood. This is the energy trap. Under conditions of energy scarcity, a crash program to shift to alternative sources—particularly where their exploitation involves relatively high up-front energy costs—makes that scarcity worse in the short term. The political implications, as he points out, are immediately apparent: leaving aside all other issues relating to the prospects for renewables to directly substitute for fossil fuels, anyone championing a long-term vision for renewable energy predicated on short-term pain must contend with the rather significant political problem that abandoning the program at any point will bring immediate relief. This presents proposals for transitioning energy supplies to 100 percent renewable sources at the global (or even national) scale with a formidable obstacle. Overcoming it would entail a remarkable—and likely unprecedented—degree of collective discipline at population-wide scales.
Feasibility assessment and the implications of high emplacement energy costs
The concerns raised by the research reviewed in this and the previous posts in the sequence strike me as compelling ones that deserve to be treated seriously. Conceptually, I don’t believe they are particularly difficult to grasp even for non-specialists. They don’t demand an unusually high level of technical or mathematical literacy. And yet, despite much of the analysis I’ve presented having originally been published in relatively mainstream peer-reviewed academic journals, the underlying ideas and their implications remain on the “attentional fringe” even amongst those who would consider themselves to be deeply immersed in questions of energy transition.
Indeed, some observers hold the view that proponents for a smooth and relatively straightforward transition to a global industrial society powered by renewables actively ignore the impediments presented by high initial energy costs, and primary source intermittency & seasonal variability. Some proponents are aware of these issues, but apparently refuse to engage in conversation about their significance and consequences. Others engage in the conversation, but claim that the concerns raised are of no relevance. This appears to be the situation with Mark Jacobson and Mark Delucchi’s response to Trainer’s critique of their case, mentioned briefly in Part 1, for meeting all global energy demand (not just electricity) from water, wind and solar sources.;;;
This particular debate—to the extent that the term applies, given the response seems largely to bypass the principal concerns raised—is especially noteworthy for the degree of conviction with which the view is advanced that the matters we’ve explored at some length in this sequence of posts are not important. I trust that the preceding discussion has established a context within which such a position, stated so boldly, would immediately reveal itself as somewhat narrow: from a physical, as distinct from entirely financial, economic perspective, how could these issues not be seen as worthy of close attention? But here are the terms in which Jacobson & Delucchi sum up their final response to Trainer:
[Trainer] raises the issue of embodied energy costs of a new energy system. Embodied energy is analytically relevant, but not in the way [Trainer] poses. First, we should dispense with the term ‘‘cost’’ and the notion of ‘‘Energy return on investment’’ (EROI). In a cost analysis ‘‘embodied’’ energy has no special standing, and in an otherwise complete economic, social, and environmental analysis EROI has no special relevance.[9, p.642-3]
As the need to move to renewable energy becomes more pressing, the need for sophisticated analyses of renewable energy systems becomes more urgent. Such analyses should focus on system optimization and on technical and economic details. In this respect, most of the list in [Trainer’s] conclusion should not serve as a guide: capital costs per se are not a relevant economic metric; the issue of system optimization is not usefully thought of in terms of ‘‘redundant’’ plants; embodied energy is not a useful concept in systematic energy-cost projections.[9, p.643]
Jacobson & Delucchi’s dismissal of EROI and related capital cost implications is based on these having no special status in life cycle assessment for conventional energy supply systems. Within such analysis, levelised cost of electricity (LCOE) is typically treated as the parameter of prime importance for determining viability of competing system configurations. For reference, the US Energy Information Administration defines LCOE as “the per-kilowatthour cost (in real dollars) of building and operating a generating plant over an assumed financial life and duty cycle”, key calculation inputs for which include “capital costs, fuel costs, fixed and variable operations and maintenance (O&M) costs, financing costs, and an assumed utilization rate for each plant type.”
The argument for taking this analytical approach as sufficient seems to be, in essence, that this is the established convention. An underlying assumption, though, is that such a convention necessarily holds in relation to system configurations fundamentally different from those for which the convention has previously been established. Kessides & Wade’s analysis is important here—they demonstrate why the relative size of upfront energy costs and costs incurred over the operating life of a supply system are highly relevant to overall system viability, not just in relation to each individual generator installation, but in terms of how a particular installation type supports the emplacement of new generators, and hence supports the expansion of that type of generation. Tom Murhpy’s Energy Trap concept deals with similar issues.
The underlying point that I want to emphasise in closing, though, is that there is a fundamental distinction between, on the one hand, conventional electricity supply systems where generators are demand responsive and hence electricity qualifies as dispatchable (and where use expectations have evolved in that context); and, on the other hand, supply systems based on intermittent and seasonally variable renewable sources that must be managed on the basis of supply-side factors (including, in addition to those specifically highlighted in this post sequence, resource distribution and low power density) that are largely beyond operators’ influence. This distinction needs to flow through to the analytical methods that we use for assessing viability, rather than simply relying on an assumption that the “correct” methods are those that suit the incumbent system.
I can see why it might be attractive to simply take the latter position as given: if renewably-powered electricity supply systems can be demonstrated to meet the viability criteria for conventional systems, then convincing those systems’ operators and financiers of renewable generation’s merits understandably becomes easier. In other words, it entails adopting the language of the incumbent decision makers, and meeting them on their terms. This is fine if the objective is solely to encourage the strong uptake of electricity generation from renewable sources. But this is surely inadequate if the objective is to comprehensively assess the viability of renewable sources for meeting a given supply task, or of establishing the type and scale of supply tasks that particular configurations of renewable sources can meet viably.
The conversation that is not happening here is one in which a distinction is made between the content and the context of the analysis. This may help with making sense of why Ted Trainer’s insights meet with such resistance rather than receiving the respectful consideration they properly deserve. If up-front energy costs, EROI and capital cost are presented as missing content in conventional analysis, their significance can be readily discounted, or neglected altogether, on narrow technical grounds. It is quite true that with the conventional framing, they have no special status. But it is not, in fact, the conventional analysis content that these factors really stand to affect, and that demands reconsideration; rather, it is the encompassing context of that analysis for which these factors are relevant.
What might we draw from this?
By bringing the conceptual tools discussed in this sequence of posts to bear on the task of characterising wind and solar PV electricity generation, we can see that in quite fundamental and highly consequential ways, these alternative means of generating electricity—especially at high levels of overall system penetration—are sufficiently different from incumbent technologies that the established conventions for assessing feasibility require a rethink. This shouldn’t surprise us—after all, solar and wind are frequently presented as game changers. If they change the very game we’re playing though, it’s hardly controversial to suggest that we’ll need to adapt the rule book to suit.
In the historical game—where electricity is produced predominantly by fossil fuelled thermal generation—a general rule of thumb is that the economic activity required to emplace generation assets is accommodated, fairly readily, within an economy’s existing resource, manufacturing and construction sectors. In other words, the scale of this activity is such that it has only marginal impact on the functioning and behaviour of other economic sectors that it enables, and that depend on it. This is not to say that building a large coal-fired power station, for instance, is an economically insignificant task—far from it. But it is a task that, within a developed industrial economy, is unlikely to distort existing labour, capital and resource relationships beyond the range of normal variation associated with other major infrastructure and service provision e.g. water supply and disposal, transport and communications. Furthermore, the manufacturing and construction sectors that service the maintenance and growth of conventional electricity supply systems have themselves developed in the context of the forms of electricity provision that comprise these systems. A defining characteristic of fossil fuelled industrialisation is the way in which it was, as Graham Palmer describes “able to bootstrap its own energy surplus and expand autonomously.”[11, p.vi] In other words, a modest initial investment (both financial and physical, i.e. in terms of energy and other resources) allowed for provision of surplus energy at rates sufficient both to satisfy current demand from non-energy related economic activity, and to grow capacity to meet increased future demand. Over industrial society’s historical course, this demonstrated ability to reliably satisfy growth expectations has in turn played a key part in stimulating future demand.
The new game is rather different. Modern renewable electricity generation technologies are inescapably products of a globalised and predominantly fossil fuelled industrial economy. At current emplacement rates, they can be treated more or less in the same way as any other product of industrial manufacturing, in which their economics is dictated by commercial and market considerations. Very generally, if their perceived use value to customers exceeds the price charged by suppliers, then they will be considered economically viable, regardless of broader considerations relating to the physical economics of supply. This is possible because, under such circumstances, their availability is not dependent on their physical economic viability, as this is underwritten by the same basic energy subsidy from fossil fuels enjoyed by all products of industrial manufacturing.
The fact that fossil fuels are themselves heavily subsidised is often advanced to counter claims that the increasingly attractive economics of renewable energy technologies is influenced by large financial subsidies. This line of reasoning neglects, however, at least two important considerations:
- As the comments above suggest, a sharp distinction needs to be drawn between monetary and energetic subsidy. The nature of the energetic subsidy from fossil fuels is brought to light by considering the difference between the rate at which they accumulated, and the rate at which we are now using them. Over a period in the order of hundreds of years, humanity is using natural resource stocks that took in the order of half a billion years to accumulate. This is fundamental to the affordability, in physical terms, of so much of the activity taken for granted in modern economies. Much activity that is affordable in the short term, as natural capital stocks are run down, would be rendered uneconomic if its viability depended on current rates of energy income. The discussion in Part 1 of the technical potential for renewables bears strongly on this matter, as this establishes (in the absence of significant changes to the role of nuclear fission) the limiting rate of energy income once fossil fuel capital is depleted to a point where its net supply rate is zero. What remains of current economic activity—including that directed towards the emplacement of electricity supply infrastructure—under such conditions is an open question. It appears increasingly plausible, however, that these constraints will translate to a global economy with overall reduced scale.
- Financial subsidisation of key inputs for any goods or services that also receive their own direct subsidies means that those goods or services effectively receive multiple layers of subsidy. The financial subsidies directed towards fossil fuels play an overall enabling role for all economic activity, and hence are built into the bottom lines for all modern renewable energy technologies. Removing subsidies for fossil fuels would increase the financial cost of renewable energy systems, as it would that of any other industrial products.
The sense-making narrative in which fossil fuels and renewable energy compete with one another can act to obscure these matters, as it implies that each represents a parallel and independent energy supply pathway stretching off into the future. Appreciating the extent to which contemporary renewable energy systems are in fact entirely dependent on fossil fuelled industrialism casts this relationship in a very different light. The rules of the game change fundamentally if the fossil fuel subsidy is withdrawn. When different energy supply scenarios are compared on the basis of levelised cost, this is hidden from view. Inputs to LCOE calculations will by default have the financial subsidies for fossil fuels built in, but they have no way of accounting for the energy subsidy. Nor do they include mechanisms by which this can be compensated, as the energy proportion from fossil fuels declines, when analysing large-scale, long term transition to renewables.
In closing here, I want to stress once more that this should not be construed as “anti-renewable energy”. What has arisen through this inquiry to date should not be read as implying that I’m somehow hostile to renewable energy; nor do I see reliance on energy from renewable sources as antithetical to human well-being. I am deeply interested in questions relating to how we might live well together within socio-economic circumstances shaped by energy flows from current solar, geothermal and tidal sources. At the same time, I appreciate the many previously unimagined benefits humanity has realised through three centuries of experimentation with the extremely high rates of energy use that fossil fuels afford us.
These benefits, however, cannot be separated from a host of co-consequences widely viewed as detrimental. To the more familiar detriments of environmental damage and disruption, social dislocation and cultural homogenisation, we need also to add what might be considered, after the tragic character of Greek mythology, a humanity-spanning Icarusian double bind: we have committed to a path of economic development that may by its very nature both block a return to previously viable life-ways, and foreclose options for charting a new course of our own choosing.
Imagine a people who, having heard tales of an abundant paradise on the far side of a great abyss, throw all their efforts into building a wooden bridge out over the void. But half way across, it starts to dawn on them that not only might all of the timber from every tree in their forests never span the divide; and not only might they destroy those forests, on which their livelihood depends, finding out; not only are their efforts to further stretch dwindling supplies of bridge-building materials making the entire structure increasingly precarious for those whose lives are now dedicated to it; but the paradise now coming into view on the other side is no richer than the home they have left. Are we these people?
The issues that this inquiry has explored suggest that we should take this question seriously. This points to the need for more circumspect consideration of the economic potential for renewably generated electricity than that evident in popular and mainstream discourse. There is a pressing need for the public conversation on renewable energy to embrace questions relating to what it means for us to live well together as a humanity. This is not a matter of whether renewably-powered human societies are viable (the historical record is clear on that front), but of what forms, natures and characters these might take.
With this in mind, the findings gleaned to date from this inquiry suggest to me that, at the very least, it would be prudent to take futures of declining energy supply rates, in the global aggregate, as a starting point for thinking about what options might lie before us, on any time horizon more than a handful of decades out (and possibly less). This is an area of inquiry towards which Beyond this Brief Anomaly has slowly been winding its way, and to which, as time allows, I’ll start to direct increasing attention.
 Delucchi, Mark A. & Jacobson, Mark Z. (2012), “Response to “A critique of Jacobson and Delucchi’s proposals for a world renewable energy supply” by Ted Trainer”, Energy Policy, Vol. 44, pp. 482-484.
 EIA (2014), “Levelized Cost and Levelized Avoided Cost of New Generation Resources in the Annual Energy Outlook 2014”, Energy Information Administration, U.S. Department of Energy, viewed 4 July 2014 at http://www.eia.gov/forecasts/aeo/electricity_generation.cfm.