100% renewable supply? Comments on the reply by Jacobson and Delucchi to the critique by Trainer.


                                                                 Ted Trainer

I recently criticised the claim by Jacobson and Delucchi that renewable energy sources could meet world energy demand.  Jacobson and Delucchi replied defending their position.  This is a response to the main points made in that reply. The main issues are to do with intermittency of renewable energy sources and the implications for redundant plant and storage, vehicle to grid systems as a storage solution, the embodied energy costs of renewable energy, and overall system capital costs.  It is argued that Jacobson and Delucci do not provide satisfactory analyses of these issues and that they do not show that energy supply can be 100% renewable.  This discussion is intended to clarify some of the core issues in the debate about the limits of renewable energy.

Several impressive documents have been published claiming to show how world energy demand can be met by renewable energy sources, e.g., IPCC (2011), Greenpeace, (2010), Beyond Zero Emissions (Wright and Hearps, 2010), Zero Carbon  Britain 2030 (2007), Stern (2006),  WWF (2010), and Jacobson and Delucchi (2011a, 2011b).  There has been a strong tendency for these to have been accepted as settling the issue, especially among environmental theorists (e. g. Flannery, 2005), agencies (e.g., the Australian Conservation Foundation) and Green political parties.  Little critical analysis of the thesis has been published, especially in academic and technical literature.  Hayden, (2004), Bryce, (2004), and Moriarty and Honnery, (2010) have put forward critical arguments. Previous to Trainer (2007) Hayden’s book (2003, 2004) seems to have been the only one published offering a critical view.  Critiques of several of the above optimistic reports are given at Trainer, 2012c.  An attempt to develop a case regarding the global situation was given in Trainer (2010) but for a significantly improved case see Trainer (2012).

My analyses have been offered as uncertain attempts to explore the crucially important and neglected question of the limits of renewable energy, which is likely to remain unsettled for some time to come.  Apart from this reply by Jacobson and Delucci little or no critical comment on these analyses has been received despite informal circulation encouraging independent evaluation.    Unfortunately the reply by Jacobson and Delucci does not deal satisfactorily with the core issues.  This discussion of their reply enables further clarification and elaboration of the problems and limits.

Integrating intermittent sources.

Jacobson and Delucchi argue that problems of intermittency do not cause insurmountable problems for 100% renewable energy supply.  Their case in their reply rests on reference to three main studies.  Limited space and the insufficiency of the information given do not permit a detailed examination of these cases but it is appropriate to briefly note the kinds of problems they involve. 

The analysis by Lund and Mathieson (2009) begins, as renewable-optimistic analyses usually do, by making very optimistic assumptions regarding achievable reductions in demand .  (See for example Lovins, 2012,and Wright and Hearps, 2010.)  It is not explained why the 2030 business as usual demand is expected to be only 11% higher than 2004 demand, or why no further increase is expected between 2030 and 2050.  It is claimed that a 50% reduction in building heat, a 40% reduction in transport fuel and a 50% reduction in household electricity consumption can be made by 2030.   Further large reductions are assumed between 2030 and 2050, including another 20% reduction in building and heating and a further 10% reduction in electricity.  These add to an assumed a 60% reduction on the (questionably low) business as usual figure they state, despite assuming large scale transfer of transport to electricity, use of hydrogen which is an energy costly option, and the introduction of much dependence on electricity-driven heat pumps.  The target arrived at is only 420 PJ/y (although the quantities in Fig. 6 add to 376 PJ/y).

In recent years it has been commonly assumed that business as usual energy demand has been heading towards a doubling by 2050 (Moriarty and Honnery, 2010) and that GDP will more or less multiply by 3 to 4. (Recent price rises and the GFC have reduced trend rates at least in the short term.)  However this proposal implies that electricity demand can be cut to 20% of such a 2050 business as usual level, while the functions given to it are considerably increased. No reasons are given to indicate that this is technically reasonable or socially/politically plausible.  At least convincing answers and derivations that can be followed would be needed before the target adopted could be regarded as acceptable. 

Other but minor difficulties include the use of hydrogen and fuel cells with no evidence given on energy losses, embodied energy costs and capital costs, considerable use of biomass-gas-electricity generation with no discussion of the total system energy efficiency (which is likely to be under 20% given that generation of the gas from biomass is at best around 57% efficient according to Staubing, Zah and Ludwig,  2012, p. 169, and the energy costs of producing and supplying the biomass would need to be deducted from output), and considerable dependence on PV (1500 MW) and solar thermal sources with no discussion of how the PV input is to be stored and what contribution these two sources can be expected to make in winter when demand is at its highest.  Questionably high conversion efficiencies are evident in Fig.6, 84% for hydrogen from electricity and especially the 75% for fuels from biomass. 

However these have not been the main problems in this study. As is usually the case it deals only with annual average demand and supply quantities and does not deal with the variations in these, especially winter averages and extremes, and the implications for redundant plant.  This is the crucial issue for renewable supply and will be elaborated on below.

The study by Mathieson, Lund and Karlsson (2011) seems to use essential the same data as that by Lund and Mathieson, although it reveals that 100 PJ/y is assumed to come from algae biomass.  This is at least an uncertain prospect given the unsolved technical problems, such as removing water from the biomass, and providing the required very large carbon inputs in an economy that does not have fossil fuel power generation supplying carbon.  Again the crucial issues of guaranteeing supply during conditions of extremely low renewable energy availability, required redundant plant and consequent capital costs are not dealt with.

The third major study is by Hart and Jacobson, 2011, including a modeling study purporting to show that 99.8% of Californian delivered energy could be produced from renewable sources.  However an examination of the information given shows that this initially impressive claim is highly misleading.  As I have often pointed out it is not difficult to explain how renewables can meet 100% of demand; what is difficult is to explain how the huge amount of redundant plant required could be afforded.  This study makes no reference to capital costs.

Hart and Jacobson do admit that the amount of capacity required would be very large, but attention is not drawn to its magnitude or significance.  Table 2 on p. 2283 reveals that to meet a 66 GW demand with low carbon emissions no less than 281 GW of capacity would be needed.  Included in this is 75 GW of gas generating capacity, but this will be needed very rarely to plug gaps in renewable supply.  In fact it is stated that its capacity will be 2.6% (p. 2283) and it will provide only 5% of annual demand.  This means 75 power stations will sit idle almost all the time.

In other words, to show that over a year renewable sources can generate more than 95% of annual electricity requirements is not so impressive or important.  The important questions are what amount of redundant plant would be needed to meet electricity demand at every point in time, and what might the capital cost of this be? Among other things, this would require detailed examination of supply requirements through the most difficult periods of solar and wind availability.

Hart and Jacobson provide graphs (Fig. 4) showing how demand might be met on four days during the year.  These are far from difficult days for renewables, all four showing large wind, PV and solar thermal contributions.  It is remarkable however that the average gas contribution over these four days is around 40% and greater than 50% on one, again making it difficult to understand the Fig. 3 claim that gas would provide only 5% of electricity demand, or to see how this proposal could be acceptable with respect to emissions.

Thus Fig. 3a reveals the way renewable energy proposals are typically misleading by setting out annual average contributions.  It shows that (for minimum emissions) solar is to provide 100,000 of the total 410,000 GWh required over a year.  This gives the impression that only enough solar plant will be needed to produce 100,000 GWh in a year.  But when winds are low far more solar input (or gas input) will be needed to plug the gap, and vici versa, meaning that far more solar plant will be needed than Fig. 3a might suggest.  Fig. 3b shows this, and as has been noted the total comes to a remarkable system requirement that is four times as much generating capacity as would be needed if it was all in the form of coal, gas or nuclear plant.

Among the less important problems evident in the account is the fact that the 2050 low carbon scenario assumes wind will provide 50% of electrical supply.  Although the Danish and UK governments have announced 50% integration targets, the reviews (e.g., Lenzen, 2009) conclude that problems of integration might limit the percentage to as little as 20%, except in unusual circumstances such as in Denmark (below.) 

It should also be noted that the debate is about whether or not renewables can meet total energy demand, but the Hart and Jacobson paper is only concerned with electricity demand, which makes up only around only one-fifth of total demand.

Finally the 2050 scenario does not achieve a satisfactory emissions regime. On p. 22 a 34% reduction is claimed but this is not likely to be sufficient given that emissions from the other four-fifths of the economy would have to be added, and given the probability that by 2050 all emissions should be eliminated. (Meinshausen et al., 2009.)

It is clear therefore that this paper does not provide impressive support for the claim that renewables can meet 100% of demand, let alone that this could be afforded.  Again the paper makes no reference to capital costs and the discussion below shows these to be formidable at best.

The “big gap” weather event problem.

The discussion to this point has only been carried out in terms of average or typical renewable energy availability, including average winter values.  What matters most, as I stress in my analyses, is whether renewables can meet demand during “big gap” events, i.e., periods of many days in which whole continental areas can experience of calm and cloudy conditions.  As my critique noted, Oswald (2008) documents a two week period in February 2006 in which Western Europe had negligible sun and wind and UK electricity demand reached its highest peak for the year.  Several similar documentations of the magnitude of this phenomenon have been made in the renewable energy literature, e.g., Soder et al., (2007), for West Denmark, Sharman, (2005, 2011), Mackay, (2008), and Sharman, Layland and Livermore, (2012) for the UK, E On Netz, (2004) for Germany, Davey and Coppin, (2003), Elliston (2012) and Lawson, (2011) for Australia, Bach (2011) for five European countries and Flocard and Perves (2011) for seven, and Lenzen, (2009). Proposals claiming that a region or a nation can depend entirely on renewable energy are of little value unless they provide information on the long term availability of wind and solar energy with special attention to the frequency and severity of these ‘big gap” events in the region.

The general response Jacobson and Delucchi make to the kind of evidence Oswald and others provide on these big gap events is to say that their vision is for systems which extend beyond continental boundaries. “…we have not claimed that WWS systems must be self-contained within Europe; rather, we have explicitly talked about much larger super-grids.”   Reference is made to studies of such grids, for instance by Czisch (2004).  There is no discussion of whether these studies show that very large scale weather events can be overcome by such grids or at what cost.  The review Flocard and Perves report on the intermittency problem in the seven main European wind nations  (2012) leads them to say, “Huge grid interconnectors between these seven countries will not solve the problem resulting from insufficient wind production during large climatic intervals.” 

The most serious problem here is evident in the common argument that “…the wind is always blowing somewhere.”  The question is, where is it blowing right now?   Are we to have built enough wind turbines to meet all demand in the place where it is blowing today, and also to have built enough in the different region which is the only place where it will be blowing next time there isn’t much wind anywhere else, and so on...?  When the wind is blowing somewhere but there isn’t much wind anywhere the system capacity is very low, and even if turbines have been built everywhere demand can be met only if some non-wind sources of redundant capacity plug the gap.  Obviously, the capital cost of building enough turbines at all the places where the wind might be strong when it is not blowing anywhere else would be unaffordable.

In addition such multi- or intercontinental proposals involve significant transmission costs. If for instance Europe’s c. 340 GW demand was to be met on some occasions such as Oswald documents from Kazakhstan, and on some occasions from the Algerian coast and on some other occasions from around Crete, the total length of transmission lines required (at $.5 per km-kW, Harvey, 2011) would cost in the region of (at least) 340 million kW x 3 source regions x 5000km each x $.5) = $2.5 trillion, or around four times the cost of 340 GW of local gas-fired power stations. To say the least Jacobson and Delucchi do not (attempt to) show or report that intercontinental supply visions are realistic or affordable.  They refer to Czisch’s (2004) proposals, although these do not provide convincing answers to these questions.  Czisch notes that the cost of transmission lines from one solar thermal field in North Africa to southern Europe could be equal to one-third of the field’s cost.

Again it is evident that the main problem is not explaining how to meet large fractions of energy demand via renewables; it is whether the amount of plant required could be afforded.

            The significance of Denmark.

Jacobson and Delucchi argue that Denmark’s wind achievements show that these kinds of intermittency problems are not very serious and do not set inconvenient limits to the integration of renewables,.   “...studies of Denmark alone at large penetrations of renewables ( [Lund and Mathiesen, 2009] and [Mathiesen and Lund, 2009]), do not indicate the problems suggested by TT.”  Of course they don’t, because Denmark has nothing like 100% use of renewables and the modeling studies do not assume such a situation.  The debate is about whether 100% is possible and little light is thrown on it by referring to a country where intermittency problems have not been encountered as wind has risen to generating electricity equivalent to only one fifth of power used and c. 5% of energy used. 

Lenzen’s review (2009) makes the commonly recognised point that above a 20% contribution from wind integration problems accelerate and in general the limit seems to be well under 30%.  Denmark and the UK have stated that they aim to raise wind’s contribution to 50%, but this does not mean that the goal is achievable.  In Denmark’s case it might be given its atypical circumstances.  It is a very small nation (5 million) with a low demand (c. 4 MW) close to large neighbouring countries (e.g., Germany 82 million) which are capable of absorbing surpluses.    The country uses considerable amount of energy for urban heating, enabling some surpluses to be economically used.  As the IPCC (2011, p. 29) says re wind energy, Denmark has partly solved integration and “curtailment” (dumping) problems “...by increasing flexible operation of CHP ...”  Most importantly its neighbours Norway and Sweden possess large hydroelectric capacity that can in effect be used to store Danish electricity exports (by phasing down hydroelectric output when wind imports from Denmark are high.) Lund et al. (2010) argue that at present this does not have to be done and that the reason why exports correlate almost perfectly with high winds, as Bach (2012) and Flocard (2012) show, is because it pays to keep Denmark’s efficient thermal power sources operating and to export the surplus.  Lund et al. are saying that if the thermal sources were not that efficient they could be phased down at times of high wind and the wind energy could be used rather than exported.  But this does not establish that the integration problems associated with greater than 20% wind supply can in fact be dealt with; it only makes clear that they have not had to be dealt with. 

More important than the export issue would seem to be the fact that times the Danish wind system is operating at around 3% capacity, (as are the far bigger systems in six other European countries, see Batch, 2012, and Flocard and Perves, 2011.)  Again this means that much alternative generating capacity must be retained, and the capital cost implications of this redundancy need to be taken into account when discussing the generalisabilty of the Danish system.

Another reason why Denmark is atypical is to do with its access to large quantities of biomass.  The proposal by Mathieson and Lund assume up to 333 PJ/y, (the recommended scheme assumes 260 PJ/y.)  This would be about 50% higher than would be possible per capita for all the world’s anticipated 2050 population if the average of the estimates of potential plantation plus waste biomass reported by the IPCC is assumed (and it can be argued that biomass use should be kept far below this level; see Trainer, in press.)  

Mathieson and Lund also assume much use of electricity to drive heat pumps in Denmark’s future, 445 MW.  Mackay (2008, p. 152) shows that these could not meet UK heating requirements in the UK due to the relatively low annual rate of heat recharge from solar radiation.

To summarise, the reference to Denmark provides little support for the 100% renewable claim, and their overall case regarding intermittency falls a long way short of showing that this problem can be solved, let alone at an affordable cost. 

Storage in vehicle batteries.

The problem of intermittency would be significantly reduced if not eliminated if large scale storage of electricity was possible.  Jacobson and Delucchi argue that storage in electric vehicle batteries can make a major contribution.  They rightly point to some aspects of my critique which are challengeable or mistaken, for instance they point out that batteries would not need to be fully charged before use, nor to be recharged every day.  However these points do little to further their case and without adding to it they proceed to conclude, “Reasonable inferences based on available data and our general understanding of economics and consumer behavior suggest that V2G incentives, time-of-day pricing, the wide range of types of EVs and recharging opportunities available in a 100% WWS world, and consumer adaptation to new technologies will create opportunities for substantial amounts of potential V2G energy storage.”  This is an almost meaningless statement as although there is no doubt that vehicle batteries will provide some storage capacity everything depends on how substantial the amounts could be and no quantitative estimate is given.  There is reason to believe that it could not be very substantial.

The scale of the issue can be illustrated by considering the present Australian energy pattern, in which transport accounts for 33% of final use.  Let us assume that 60% of this can be converted to electric drives, and that this enables vehicle energy efficiency to be cut by a factor of 4 as is commonly claimed.  That would mean that the amount of energy required to power electric vehicles would be around 5.5% of final energy consumption. If on average battery capacity enabled (an unnecessarily  generous) 3 days of car use storage capacity would be c. 16.5% of final energy use.  If on average batteries were kept two-thirds fully charged, storage capacity available for non-transport use would be around 5.5% of energy demand, i.e., c. 193 PJ/y or .53 PJ/d (and less when losses in conversion into and from 240 voltage, and charging and discharging losses are taken into account.)  Australian electricity demand is approximately 2 PJ/d, so vehicle batteries could make a significant but not large contribution to the 6 PJ storage task set by three days of poor weather.

Much the same conclusion comes from the study by Lund and Kempton (2008) which Jacobson and Delucci refer to in support of their position.  This is a valuable detailed analysis in which assumptions and derivations can be clearly followed and factors such as the time vehicles are in use or parked, or not plugged in, are taken into account.  However the amount of storage demonstrated is small in relation to potential storage need from the wider economy. They assume 30 kWh battery capacity, enabling 180 km travel. If we assume three times this capacity (which the authors consider but say is not viable) 171 GWh could be stored.  If on average half this capacity was available for storage use by the non-vehicular energy system it would constitute only 70% of the average Danish daily electricity demand, and a lower percentage of the winter average, let alone of a peak winter demand.  Again this would be helpful but would not make much difference in the kind of two week big gap problem Oswald and others document.

It is important to briefly consider the logic of vehicle storage here. Electric vehicle batteries are quite expensive and use scarce materials, and their considerable weight detracts from vehicle energy efficiency.  The aim would therefore be to equip vehicles with as little storage capacity as possible.  The point of V2G systems is not to locate storage capacity for the general economy on vehicles; it is to take advantage of the storage capacity vehicles have for their needs but at times are not using.  The question then becomes how much capacity for non-vehicle use would an efficient fleet be capable of making available? 

The weight of petrol plus tank needed to store 1 kWh for a 6 km trip would probably be well under 1 kg, but the battery weight required would be around 7 kg.  If a vehicle was equipped with 90 kWh capacity the batteries would weigh around 600 kg, doubling the weight of a light EV.  As a petrol tank empties its weight reduces, but this is not the case with batteries.

There is therefore a high priority on equipping vehicles with just enough battery capacity to meet their typical needs.  Most car journeys are quite short, and if electric vehicles become the norm there will be recharging points at almost every parking place, enabling minimal battery capacity to be carried.  In addition there will be movement from car use to “micro-travel” options, such as battery powered bicycles.  Thus the majority of vehicle buyers might be inclined to purchase a storage capacity of around 3 - 5 kWh, making Jacobson and Delucchi’s 6 km/kWh assumption which would suffice to get to the office or supermarket where recharging could take place for the return journey.  (For the many very light vehicles used for short trips a much higher mileage would apply, meaning even lower battery capacity.) If however we assume that an average 10 kWh storage would suffice, the figures from Lund and Kempton indicate that available storage in a Danish a V2G system might meet electricity demand for only around 6 hours.

Vehicles needed for longer trips would not necessarily improve the situation as these are likely to use battery swap stations located at many points.  (Note that battery swap systems would more or less double the need for battery materials.)

Embodied energy costs.

Quite important for the discussion of capital costs is the issue of the embodied energy costs of renewable technologies, or the energy return they yield on the energy invested in their construction.  This is generally regarded as being a very important consideration, especially given marked declines in EROI for conventional fuels in recent decades, and the relatively low EROI for renewable energy technologies (except for hydroelectricity.)  Some are concerned that there might be a minimally viable EROI for high energy societies, and that this might be in the region of 10, which is around the figure commonly claimed for most renewables.  (Murphy and Hall, 2010.) It is conceivable that embodied energy costs alone could disqualify some renewable options, (see below.)   

Nevertheless the full statement on the issue in the reply by Jacobson and Delucchi is  “… discussion of embodied energy is irrelevant, because with an indefinitely renewable energy resource with no external costs, the full lifetime cost as we have estimated is the relevant factor—there is no additional pertinence to embodied energy per se.”  The meaning intended here is far from clear.  Suffice it to say that the contribution a renewable technology might make and its capital cost per unit of energy provided must at least take into account the amount of energy needed to produce the plant and deduct this from gross output to arrive at net supply.  Although the actual proportion with respect to PV and solar thermal sources seems to be far from settled, its importance is not in dispute.  Some analysts argue that when the appropriate boundaries are taken and all “upstream” factors are taken into account (i.e., the energy cost of the smelters that produced the aluminium for PV panel frames) the embodied energy cost of PV could be 30% and even 50%. (Crawford, et al., 2006, Crawford 2011, Lenzen and Treloar, 2003.)  If this is so the viability of these technologies is in serious doubt.

A dramatic illustration of the potential significance of this factor is given by recent evidence on the energy cost of storing energy via Ammonia dissociation.  In Trainer 2010 it was assumed that the most promising solar thermal technology would be Big Dishes storing heat this way.  However a recent discussion of the approach by those developing and advocating it (Dunn, Lovegrove and Burgess, 2012) reveals that for a 10 MW generator the storage system would require162 km of 30cm diameter steel gas pipe.  (This has been confirmed by personal communication.)  The embodied energy cost of such a large amount of steel, apart from construction, might approach 40% of plant lifetime output.  If so this would seem to completely disqualify the technology from competing with central receivers or troughs using conventional oil or salt tanks storage.  This is a graphic demonstration of the importance of taking embodied energy costs into consideration.

Capital costs.

The reply Jacobson and Delucchi give to my arguments re the crucial capital cost issue is given 243 words.  The brevity seems to be due to the extremely puzzling statement that they regard the issued as irrelevant.  They begin by saying, “Estimates of the total capital cost are relevant only if one argues that there are some constraints on the availability of capital not adequately reflected in the opportunity cost of capital. T11 makes no such claim, so this discussion is irrelevant.”

Their subsequent comments not only fail to connect with the core issue but misinterpret my critique.  They say, “TT claims that we do not justify our assumptions regarding capital costs…Our estimates of energy demand are presented in Table 2 of JD11 and explained in detail in Appendix A.2 of JD11. Thus, TT's claim that our estimate ‘…is not explained or justified’ is untrue.”   My point was not any failure to provide documented cost information on specific technologies or references to sources; it was to do with whether the conclusions drawn from that information are misleading and my critique was that they are seriously misleading mainly because they do not take into account the need for redundant plant when estimating total system cost.

Occasionally the studies referred to make brief reference to the crucial issue, the fact that that the systems envisaged would involve very large amounts of redundant plant for generation and transmission.  This is done by Hart and Jacobson in the study discussed above, where it is said, “The low-carbon systems described in this study require large capacities of dispatchable generation with very low capacity

factors... will also require significant investments in transmission and distribution infrastructure...”  (Hart and Jacobson, 2011, p. 2285.)   However despite stating a capacity required that is four times average demand no effort is made to assess the overall system capital costs. 

The focal concern in my attempts to assess the potential and limits of renewable energy is the total system capital cost that would be involved if the system is to be capable of dealing with the intermittency of inputs.  As has been shown above, that capacity requires substantial redundancy, i.e., much plant of some kind to turn to when solar or wind or both sectors are contributing little or nothing.  Estimating the total system capital cost therefore involves adding the capital cost of the quantities of all the various kinds of generating plant that would be needed to cover these below average contributions from the various sectors, along with the cost of other non-generation elements within the systems, such as long distance HVDC transmission lines, any equipment for hydrogen conversion, pumping, piping and storage, and the systems for growing and delivering biomass to generating plants. 

Jacobson and Delucchi do not deal with this issue; they do not attempt to estimate the amount of redundant plant required or its aggregate capital cost.  How, for instance, would the total capital cost of a system requiring 281 GW of renewable generating capacity compare with the total capital cost of a system requiring 66 GW of coal-fired capacity? Note that the calculation must take into account the fact that 1kW (peak) of coal-fired capacity generates on average .8 kW but 1 kW of solar thermal or PV (peak) capacity, costing at least twice as much, generates on average only .2 kW.

Fig. 3 from Hart and Jacobson shows that in their proposal 110 GW of solar and 75 GW of wind capacity would be needed.  These quantities indicate a capital cost of around $ 600 billion, for two thirds of the capacity needed.  To meet the demand via gas-fired plant might cost c. $70 billion.  (It seemed above that the proposal actually assumed/required sufficient gas capacity to meet all demand.) This aligns with my general finding that 100% renewable energy supply systems are likely to involve capital costs well in excess of ten times the present capital costs of energy supply.

These considerations show why it is highly misleading to analyse system costs in terms of levelised costs. The latter figures only indicate the cost of energy from a technology, such as wind or PV, when all its lifetime construction, operations, interest etc. costs are divided by its lifetime output.  Such a figure tells us nothing about how many extra turbines, or solar, coal, gas, oil or nuclear generating plants must also be built to provide a system with the back-up capacity that will enable the wind sector to meet its average contribution when there is little or no wind.  As Lenzen (2009) points out, the cost of a sector’s back-up plant should be added to the cost of that sector, just as the cost of a home PV system should include that of the emergency petrol generator.

It is puzzling that Jacobson and Delucchi seem dismiss the significance of this distinction.  Their full response on the levelised cost point is,  “T11 concludes that ‘the common practice of focusing on levelised costs in estimating total system capital costs leads to serious underestimation of system costs.‘ This is incorrect. Levelised costs are based on the estimated capacity factor, where the capacity factor is what would be obtained in an optimized system (i.e., the least-cost system that reliably satisfies demand). This is part of a correct and complete estimate of the average energy cost of the system; there is in principle no underestimation whatsoever.”  Apart from the problem of grasping what is being said here, there is no recognition of a redundancy problem or its implications for total system capital cost. Nevertheless Jacobson and Delucchi conclude, “In sum, T11 provides no valid criticism of the detailed methods or assumptions of our analyses of energy cost and energy demand.”

 My initial attempt to frame an approach to assessing the viability of a global 100% renewable energy supply in view of these intermittency and redundancy issues (Trainer 2010) could only be based on data regarding solar thermal power that was not very satisfactory and probably arrived at a cost conclusion that now appears to have been considerably too high.  Since then more confident assumptions have been enabled by the publications by Hearps and McConnell, (2011), Lovegrove et al. (2012), AETA, (2012), and especially the NREL (2010, 2011) SAM packages.  (Contrary to my 2010 impression these seem to clearly establish central receivers as much more viable than Big Dishes with Ammonia storage capacity.)  Trainer (2012e) uses these sources in an improved exploration of four possible strategies for a global 100% renewable energy supply, and concludes that the annual capital investment required would be in the region of ten times as great as it is at present, even without including several important cost factors which cannot be quantified satisfactorily.


The assessment of the limits of renewable energy has been a neglected issue but it  is of the utmost importance as crucial and costly policy decisions will have to be made in the near future, e.g., between nuclear, fossil fuel with CCS and renewable paths.  Unfortunately the reply by Jacobson and Delucchi is not very helpful in clarifying the core issues.  In my view they do not put forward a satisfactory case for their claims and do not deal satisfactorily with the criticisms of their position I originally published.  Nevertheless in my view the exchange contributes to the clarification of the field.  Following are the issues on which further analyses might best focus.

AETA, (2012).   Australian Energy Technology Assessment, ABARES, Canberra.

Batch, D. F., 2011. Wind power in Denmark, Germany, Ireland, Great  Britain and France. Statistical Survey.

Bryce, R., 2010.  Power Hungry, Public Affairs, New York.

California Wind Energy Collaborative, 2003. California RPS integration cost analysis phase 1: One year analysis of existing resources. California Energy Commission; Final Report CEC-500-03-108C.

Crawford, R., 2011.  Towards a comprehensive approach to zero emissions housing, Architectural Science Review, 54. 4, PP. 277 – 284. 

Crawford, R, Treloar, G. J., and Fuller, R., J., 2006.  Life cycle energy analysis of building integrated photo voltaic (BIPVs) with heat recovery  unitRenewable and Sustainable Energy Reviews, 10, 559 – 576.

Davy, R. and Coppin, P., 2003. South East Australian Wind Power Study, Wind Energy Research Unit, CSIRO, Canberra, Australia.

Dunn, R., K. Lovegrove, and G. Burgess, 2012.  A review of Ammonia based thermochemical energy storage for concentrating solar power, Proceedings of the IEEE, 100, 2, Feb .,  391-400.

E.On Netz, 2005. Wind Report 2005, http://www.eon-netz.com

Elliston, B., 2012. Simulations of Scenarios with 100% Renewable Electricity in the Australian National Electricity Market, (http://www.ceem.unsw.edu.au/content/userDocs/presentation.pdf)

Elliston, B., M. Kay, M. Diesendorf, and I. MacGill,  2012.   Analysing Power System Impacts Using Solar Radiation Data For Australia.

Flannery, T., 2005. The Weather Makers, New York, Atlantic Monthly Press.

Flocard, H.  and J. P. Perves, 2011.  Wind Production intermittency. Cross border compensation: what to expect in Western Europe? Analysis of Winter 2010/2011.  www.sauvonsleclimat.org

Greenpeace,  2010. 2011.  World Energy (R)Evolution: A Sustainable World Energy Outlook.  International and European Renewable Energy Council,

Hansen, J., et al., 2008.  Target atmospheric CO2; Where Should humanity aim?, The Open Atmospheric Science Journal, 2, 217 – 231.

Hart, E. K., and M,. Z. Jacobson, 2011.  A Monte Carlo approach to generator portfolio planning and carbon emissions assessments of systems with large penetrations of variable renewable, Renewable Energy, 36, 2278 – 2286.

Hayden, H. C., 2004. The Solar Fraud, Pueblo West, Co, Vales Lake Publishing.

Hearps, P. and D. McConnell, 2011. Renewable Energy Technology Cost Review, University of Melbourne. http://energy.unimelb.edu.au/index.php?page=technical-publication-series

Intergovernmental Panel on Climate Change, Working Group 111, Mitigation of Climate Change, Special Report on Renewable Energy Sources and Climate Mitigation. June, 2011.  http:www.srren.ipcc-wg3.de/report

Jacobson, M. Z. and M. A.  Dellucchi, 2011a.  Providing all global energy with wind, water and solar power, Part 1: Technologies, energy resources, quantities and areas of infrastructure, and materials, Energy Policy, 39,  1154 – 1169.

Jacobson, M. Z. and M. A.  Dellucchi, 2011b.  Providing all global energy with wind, water and solar power, Part 1: Reliability, system and transmission costs, and policies,”  Energy Policy, 39,  1170 – 1190.

Lawson, M., 2011. Wind power; Not always there when you want it, On Line Opinion, 18th July.

Lenzen, M., 2009. Current state of development of electricity-generating technologies – A literature review. Integrated Life Cycle Analysis, Dept. of Physics, University of Sydney.

Lenzen, M. and G. Treloar, 2003. Differential convergence of life-cycle inventories toward upstream production layers, implications for life-cycle assessment”, Journal of Industrial Ecology,  6, 3-4.

Lund, H., and W. Kempton, 2008.  Integration of renewable energy into the transport and electricity sectors through V2G, Energy Policy, doi:10.1016/j.enpol.2008.06.007

Lund, H., et al., 2010.  Danish Wind Power Export and Cost, Coherent Energy and Environmental System Analysis, Dept of Environment and Planning, Aarlburg Univ., Denmark.

Lund, H. and B.V.Mathiesen, 2009.  Energy system analysis of 100% renewable energy systems — The case of Denmark in years 2030 and 2050, Energy, 34, 524– 531.

Mackay, D., 2008. Sustainable Energy – Without the Hot Air, Cavendish  Laboratory. http://www.withouthotair.com/download.html

Mathiesen, B. V., H. Lund, and K. Karlsson, (in press). 100% Renewable energy systems, climate mitigation and economic growth. Applied Energy. Vol. 88 (2), pp. 488-501, February 2011.

Meinshausen, M, N. Meinshausen, W. Hare, S. C. B. Raper, K. Frieler, R. Knuitti, D. J. Frame, and M. R. Allen, 2009.  Greenhouse gas emission targets for limiting global warming to 2 degrees C, Nature, 458, 30th April, 1158 -1162.

Moriarty, P., and D. Honnery, 2010. The Rise and Fall of Carbon Civilization, Springer, Dordrecht.

Murphy D. J. and C. A. S. Hall, 2010. Issue: Ecological Economics Reviews Year in review—EROI or energy return on (energy) invested, Ann. N.Y. Acad. Sci. 1185 102–118.

Sharjman,H. 2009. Wind Energy. The Case of Denmark, CEPOS (Centre for Politische Studier,  Copenhagen, Sept.

Sharman, H., 2005. Why UK wind power should not exceed 10 GW, Civil Engineering, 158, Nov., pp. 161 - 169.

Sharman, H., 2009. Wind Energy; The Case of Demark, CEPOS, Copenhagen.

Sharman, H., 2011. Renewable Energy; Vision or Mirage?, Adam Smith Research Trust.

Soder. L., Hoffman, L., Orfs, A., Holtinnen, H., Wan Y., and Tuiohy, A., 2007.  Experience from wind integration in some high penetration  areas.  IEEE Transactions on Energy Conversion, 22, 4 – 12.

Staubing, B., R. Zah and C. Ludwig, 2012.  Heat electricity or transportation?  The optimal use of residual and waste biomass in Europe from and Environmental perspective, Environmental Science and Technology, 46, 164 – 171.

Stern, N., 2006. Review on the Economics of Climate Change, H.M.Treasury, UK, Oct. http://www.sternreview.org.uk

Trainer, T., 2012a.  A critique of Jacobson and Delucchi’s proposals for a world renewable energy supply, Energy Policy, 44, 476–481.

Trainer, T., 2012b. Renewable energy – Cannot sustain an energy-intensive society.” http://socialsciences.arts.unsw.edu.au/tsw/REcant.html

Trainer, T., 2012c. Renewable Energy, alphabetical topic list at http://socialsciences.arts.unsw.edu.au/tsw/

Trainer, T., 2012d.  Critique of the proposal by Eliston, Diesendorf and MacGill,  http://socialsciences.arts.unsw.edu.au/tsw/RE.EDM.htm

Trainer, T., 2012e.  Can the world run on renewable energy?  A revised negative case.  http://socialsciences.arts.unsw.edu.au/tsw/CANW.htm

Trainer, T., in press.  Can Australia run on renewable energy? The negative case. Energy Policy.

World Wildlife Fund, 2010. The Energy Report; 100% Renewable Energy by 2050, WWF International, Switzerland.

Wright, M and P. Hearps, 2010.  Australian Sustainable Energy Zero Carbon Australia Stationary Energy Plan, Energy Research Institute, Melbourne University, Australia. http://energy.unimelb.edu.au/index.php?page=zero-carbon-plan

Zero Carbon Britain 2030, 2007.  A New Energy Strategy, Centre for Alternative Technology, Wales.  http://www.zerocarbonbritain.com/