Why the Future is Not Solar

The future is solar—apparently. This was the argument advanced by Andrew Blakers and Matthew Stocks (2018) of the ANU on the blog The Conversation, which tells us on its masthead it combines academic rigour with journalistic flair. Unfortunately, the piece in question demonstrates considerable flair and enthusiasm for solar, but academic rigour is rather harder to find.

An energy future without fossil fuels is appealing, and people frequently imagine that solar energy will allow us to achieve that at little or no cost. Blakers and Stocks had earlier joined with a colleague to tell us that Australia could meet its target under the Paris Agreement with zero net cost (Blakers, Lu and Stocks, 2017). Blakers and Stocks’s piece reinforced the current belief among many that renewables are now competitive with coal-fired electricity generation, just as their enthusiasts said they would become—if only governments supported them with the right policies.

The Australian electricity market might be in a mess, and wholesale prices might have doubled, but—the Blakers and Stocks meme would have it—we are about to reap the rewards of a solar-based cornucopia that will make it all worthwhile. The Blakers and Stocks piece is, unfortunately, based upon an uncritical view of the place of solar energy that is far too sanguine about the prospects for solar, making several errors and glossing over some inconvenient truths.

This essay appears in the latest Quadrant, now on sale.
Click here to subscribe and gain instant access to every article, essay, poem and column the very moment they are uploaded

Blakers and Stocks are leading scientists in the development of solar cells, but they do not seem sufficiently endowed with the scepticism that should accompany any technology. They remind me most of all of the engineers I studied in electric utilities in Tasmania, Victoria, New Zealand, Ontario and British Columbia (Kellow, 1996) who found ways to ensure that their evaluations of alternatives always managed to support their preferred project. Langdon Winner (1978) referred to this as “reverse adaptation”, or the adaptation of ends to suit preferred means, but Abraham Maslow perhaps put the problem most elegantly when he remarked that when the only tool you have is a hammer, everything starts to resemble a nail. Experts, in other words, tend to favour the things they are expert in.

I will return to the economics of solar in Blakers and Stocks’s analysis below, but first I point to some of the other statements where they simply gloss over the faults and limitations of solar (and wind) energy—including their environmental limitations.

Some inconvenient solar (and wind) truths

Blakers and Stocks make the following statement:

PV [photovoltaic cells] and wind have minimal environmental impacts and water requirements. The raw materials for PV—silicon, oxygen, hydrogen, carbon, aluminium, glass, steel and small amounts of other materials—are effectively in unlimited supply.

Most of these raw materials require energy to produce. There is a debate over whether the energy embodied in various technologies is large enough to offset that which they produce (Fthenakis and Kim, 2007; Ferroni and Hopkirk, 2016; Raugei, et al, 2017; Ferroni, Guekos and Hopkirk, 2017). After dealing with criticisms of their original paper, Ferroni, Guekos and Hopkirk (2017: 498) conclude:

Any attempt to adopt an Energy Transition strategy by substitution of intermittent for base load power generation in countries like Switzerland or further north will result in unavoidable net energy loss.

Australia has better insolation, but there is a global concern here.

What Blakers and Stocks also gloss over with their dismissive “small amounts of other materials” is that the manufacture of PV panels requires the use of small—but still significant—amounts of solvents that have Global Warming Potential numbers around 20,000 times that of carbon dioxide. Nitrogen trifluoride was not covered by the first commitment period under the Kyoto Protocol, but is 16,000 times more powerful a greenhouse gas (GHG) than carbon dioxide, and sulphur hexafluoride is 23,900 times more powerful than carbon dioxide. This means that—on a life-cycle basis in Germany—Ferroni (2014) has suggested that PV solar is worse for climate forcing than gas or coal. Ferroni has calculated that lifetime (twenty-five years) emissions from solar energy in Germany (panels made in China, shipped to Germany, including transport and peripherals) is 978g carbon dioxide equivalent per kWh. For state-of-the-art coal the figure is 846g and for gas (CCGT) 400g.

Tony Abbott: ‘The Government is Kidding Us‘

This is, of course, partly a reflection of the poor insolation in Germany, and Mexico or Australia are more propitious sites. The advantage, in terms of virtue, is that only some of these were covered by the Kyoto Protocol, and the emissions (from inefficient fossil fuel electricity) related to manufacture mostly occur in China and those stemming from energy expended in transport are not charged to Germany.

Moreover, it cannot be assumed that covering vast areas with solar panels has “minimal” environmental impacts. While rainforest has greater aesthetic appeal (especially for environmentalists), the deserts often favoured for PV or solar thermal installations are not without significance. Indeed, there is research that suggests deserts have greater biodiversity than rainforests (Fierer and Jackson, 2006), and covering them with solar arrays does not constitute a minimal impact.

Blakers and Stocks also state that “Wind energy is an important complement to PV because it often produces at different times and places, allowing a smoother combined energy output.”

This is nonsense on stilts.

Often it does—but often it doesn’t. And often both produce negligible amounts—simultaneously. And while the sheer length of the grid in Australia is often used to suggest that the sun is likely to be shining or the wind blowing somewhere, there are substantial transmission losses to be considered. The sun shining in North Queensland is not much help when it is cloudy and calm in South Australia.

Indeed, the Australian Energy Market Operator recently slashed the “marginal loss factor” (MLF) (which reflects transmission losses) for renewables by up to 22 per cent after finding that the contribution of solar and wind to the market was less than expected (Parkinson, 2018). (The MLF calculates the difference between how much is produced by the generating facility, and measured at its meter, and how much is estimated to be delivered to customers.)

There are times in Germany, particularly in winter, when the output from solar and wind has been close to zero. Calm and mists and fogs often go hand-in-hand, as any meteorologist will tell you. This is why Germany continues to use coal and looks to continue to do so in the future. It has little prospect for pumped storage (which is only around 80 per cent efficient, let’s remember), and inquiries to use the more favourable geography of Norway and Switzerland have been met with polite refusals. (Blakers and Stocks have been advocating for pumped storage in Australia, identifying numerous potential sites; the costs—and environmental impacts—of these would have to be charged to an all-renewables future.)

The result of Germany’s Energiewende has been essentially no reduction in GHG emissions because thermal plant often runs at reduced and less efficient loads to accommodate the variability of renewables, and prices that sometimes turn negative, with excesses dumped on neighbours in the European market (undermining their own renewables generators) and increased prices overall for consumers. Germany and Denmark, with the highest proportion of renewables, have the highest prices in the developed world—although South Australia eclipses them both on a pre-tax basis. Subsidies are now ending, and the solar industry in Germany in particular is tottering.

One of the more remarkable claims Blakers and Stocks make is: “Complete replacement of all fossil fuels requires solar and wind collectors covering much less than 1 per cent of the world’s land surface area.” They provide a link to their authority for this statement. It is not to any peer-reviewed science, but to an article in Forbes magazine—and the article would certainly not pass peer review in any reputable scientific journal. The problems with this article, one would think, should have been obvious to Blakers and Stocks, yet they cite it as if it is an authoritative source.

The author of the Forbes paper, Mehran Moalem, actually made the bolder claim that we could power the entire world by harnessing solar energy from 1 per cent of the land area of the Sahara (Moalem, 2016). The first error (which should have rung alarm bells for Blakers and Stocks) was his assertion that “the power density can be as high as fourteen hundred watts per square meter”.

As even a quick check of Wikipedia reveals, the solar constant is 1361 watts per square metre (W/m²) at the solar minimum and approximately 1362 W/m² at the solar maximum—at the top of the atmosphere. Moalem added, “The site proposed here as an example in African Sahara is on the Equator and there are very few if any cloudy days per year.” Actually, the Equator runs through the Congo, a long way south of the southern edge of the Sahara, most of which lies to the north of the 15th parallel and only about half of which lies south of the Tropic of Cancer. (It extends north to about the 35th parallel.)

Moalem’s vision of a well-insolated cloudless Sahara is appealing, but as Wikipedia also tells us, at most about 75 per cent of the solar energy at the top of the atmosphere reaches the earth’s surface because, even with a cloudless sky, it is partially reflected and absorbed by the atmosphere. Light cirrus clouds reduce this to 50 per cent, stronger cirrus clouds to 40 per cent, so that the solar energy actually arriving at the surface can vary from 550 W/m² with cirrus clouds to 1025 W/m² with a clear sky. Even at the Equator, with twelve-hour days, this potential can only be realised if the PV solar panels track the sun, which adds to costs. With a conversion efficiency of perhaps 25 per cent, the maximum yield from PV solar (with tracking) is currently a little over 250 W/m² at the Equator—or about 3 kWh/m² per day.

Tracking also requires that the panels be more widely spaced, because otherwise they cast a shadow on their neighbouring panels and diminish output, so that the yield per square metre of land is even less.

The Sahara has some additional disadvantages. For a start, it is where electricity consumers are not, so the losses in transmission would be considerable. The shifting dunes of the Sahara might also prove a challenge. These figures also assume that the panels are kept in a constant state of cleanliness, and this is difficult to achieve—especially in desert conditions where there are considerable quantities of sand and dust. Output can be reduced by dirt and dust (if not cleaned)—by as much as 15 per cent (Molki, 2010)—but some have argued that the costs of cleaning may, in fact, be greater than the costs associated with a lower degradation rate.

Deserts are often preferred locations, because land there is cheap—but they are also sandy and dusty. Regular cleaning ensures that the panels get as much light as possible, but this is labour intensive. At the Ketura 5MW array in Israel, panels were cleaned manually ten times a year, and so the operators installed a robotic cleaning system developed by Ecoppia. The system only works on fixed arrays, not tracking arrays (Lombardo, 2014).

Output from a PV panel can be about 25 per cent higher with a single-axis (east-west) tracking system, and 30 per cent higher with a dual-axis tracking system, but a tracking system nearly doubles the cost of the array, so it is often more cost-effective to simply add more panels at a fixed angle. That increases the land area required. But as a panel tracks the afternoon sun, it casts a shadow to the side, and an empty space must be left to prevent shading of neighbouring panels. For this reason, and the added cost of a tracking system, many solar farms use a fixed-tilt mount.

Dirt is just one problem affecting performance. Panels also degrade over time. Solar panel performance declines by an average of 0.8 per cent to 0.9 per cent each year (Jordan et al, 2016), with first-year, short-term degradation of 2 per cent, so it would be operating at 98 per cent of its potential efficiency going into the second year. From that point, assuming it degrades at 0.8 to 0.9 per cent per year, it will be operating at a considerably lower percentage of its initial efficiency after twenty-five years.

Numerous factors affect degradation rate, including high winds and humidity, high temperatures, snow and annual freeze/thaw cycles. The degradation rate from high winds can be particularly rapid—as they found in Puerto Rico when a hurricane totally destroyed several solar arrays.

Solar and wind came to blows in Puerto Rico. As the wreckage demonstrates, wind won.

There are similar issues with wind turbines. There are some awe-inspiring photos and videos to be seen by googling “wind turbine fail”: towers toppling, ice being thrown, blades being thrown, and fires. One group opposed to wind power documents 184 deaths from wind farm accidents (Caithness, 2018)—more than died from acute radiation exposure at Chernobyl and many more than the single direct death at Fukushima—a heart attack suffered by a first responder. (Total long-term deaths from radiation exposure at Chernobyl have been estimated by the WHO at around 4000.)

Recently, it was also found that erosion of the leading edge of off-shore turbine blades was reducing their economic life—perhaps to only ten years. Early studies indicated that even insects on leading edges could reduce output by more than 25 per cent (Corten and Veldkamp, 2001), although more recent studies suggest a 9 per cent drop in performance (Schramm et al, 2017).

In February, 2018, it was announced that blades would have to be removed and repaired at the Anholt wind farm off coastal Denmark that was commissioned only in September, 2013. Supplier Siemens Gamesa would not comment on the cost, but the company’s Danish subsidiary provided 4.5 billion krone ($750 million) for the repairs. Similar problems were reported with British and German wind farms.

The point is not that wind or solar should not be considered, but they should be considered, warts and all, alongside coal, nuclear and hydro with all their drawbacks. Too often, the drawbacks of renewables have simply been glossed over. But renewables have drawbacks, even at a technical level, and are unlikely to meet global energy needs.

Renewables alone cannot provide energy in sufficient density to meet the needs of future populations, with 9 billion people already in the demographic pipeline. Megacities such as Shanghai (currently 22 million) cannot provide solar or wind energy in sufficient quantities from their periphery (Kelly, 2016). Kelly argues that renewables technology might advance, but it is not yet at a stage where it can possibly replace fossil fuels or nuclear in either quantity or density. He draws attention to the many abandoned solar and wind sites as evidence of the folly of rushing to renewables prematurely.

Mistaking cost and price

It is the errors on cost made by Blakers and Stocks that are the most serious. They tell us:

The cost of PV and wind power has been declining rapidly for many decades and is now in the range A$55–70 per megawatt-hour in Australia. This is cheaper than electricity from new-build coal and gas units. There are many reports of PV electricity being produced from very large-scale plants for A$30–50 per MWh.

The authority they provide for the first set of figures is a link to the advocacy site reneweconomy.com.au that yields the result: “Sorry. Error 404: Page not found.”

The $55–70 per MWh estimate appears to come from the head of the Australian Renewable Energy Agency (ARENA), Ivor Frischknecht, who gave exactly these figures in answer to questions from Greens Senator Larissa Waters (Vorath, 2017). But these were estimates of the price at which contracts were being struck at that time—so they include the subsidies in the policy settings such as the Renewable Energy Target. They do not reflect the cost of generation, but the price at which PV solar is sold by generators, and so they include a discount for the sale of Renewable Energy Certificates. RECs have often been priced at $60–90/MWh.

The price of RECs varies, but at the end of 2016, electricity retailers ERM Power and Alinta Energy opted to pay more than $130 million in penalties ($65 per certificate not purchased), rather than meeting their REC obligations, because the current price was well above $65. Fifteen electricity retailers failed to surrender the correct number of renewable energy certificates, and eight retailers failed to surrender a single certificate (Norman, 2017).

The source for the second, rather optimistic price for future PV solar (A$30–50 per MW) is linked to another post in The Conversation by themselves and a colleague, where the source for their cost estimates is, again, Reneweconomy (Blakers, Lu and Stocks, 2017a). There they engage in boosterism for the future of solar, stating:

During the 2020s these prices are likely to fall still further—to below A$50 per MWh, judging by the lower-priced contracts being signed around the world, such as in Abu Dhabi, Mexico, India and Chile.

These contracts, of course, reflect both local policy settings (which can include subsidies and regulatory supports), as well as local insolation conditions. US installations attract a 30 per cent tax credit, for example. Again, Blakers and Stocks conflate the price at which generators contract to sell with cost; the price has little to do with the cost, so distorted can the policy settings be.

To give just one example, under Mexico’s General Climate Change Law, the Mexican government created a system of “Clean Energy Certificates” (CELs), with a CEL granted for each MW/h of electricity produced by a generator using clean energy technologies. Large consumers of electricity (mainly industrial and commercial, also known as Qualified Consumers) are then required to consume electricity generated from clean energy sources. The requirements begin in 2018 with a just over 5 per cent increase gradually over the next few years, reaching 30 per cent by 2021 and 35 per cent by 2024. Large consumers of electricity will obtain the CELs they need to comply with this requirement from Qualified Service Suppliers and then submit these certificates to avoid sanctions (Rodriguez, 2017).

Mexico has the largest domestic electricity subsidy among OECD countries—US$6.2 billion in 2017. It is planning a massive expansion in rooftop solar, with 80 per cent of the system’s cost initially covered by subsidies and the program then “self-funded” by “subsidies avoided” by new rooftop solar users. “Price” in this grossly distorted market has little to do with cost. Auction prices often reflect low bids that are then expected to be renegotiated later (as is also common in India), and bidders often seek to achieve a strategic place in emerging markets (Davis, 2017), but the Mexican contract price—as with the Australian prices they quote—includes the effect of renewable energy certificates.

Forecasts that the cost of solar technology will continue to decline rapidly should be treated with caution. According to what is known as Swanson’s Law, the price of solar photovoltaic modules tends to drop 20 per cent for every doubling of cumulative shipped volume. It is true that the average cost of solar cells dropped from US$76.67/watt in 1977 to just US$0.74/watt in 2013 and $US0.26/watt in mid-2016, but the historical cost curve is decidedly (declining) hyperbolic in shape—typical of most technological innovations—and it has levelled out. It would be heroic in the extreme to assume that past cost reductions will continue into the future.

Yet, despite the fallacy of taking contract prices as costs, Blakers, Lu and Stocks (2017) have had a paper accepted in a peer-reviewed journal of these very same optimistic forecasts of future costs, and repeated it in both their earlier Conversation post (Blakers, Lu and Stocks, 2017a) and their associated press release (Blakers, Lu and Stocks, 2017b).

There is a disconnect between the Blakers and Stocks cost estimate and that from other recent sources, and not just the 2015 cost of around $150/MWh in the CO₂CRC report (CO₂CRC, 2015). A report for AIG puts the (2017) cost of solar PV at $110/MWh—well above ultrasupercritical coal at $81.

The cost of PV solar has certainly come down since 2015, but that does not explain the discrepancy between the figure selected by Blakers and Stocks and that of the AIG, made just a few months before the figure they cite. The answer lies in the basis for the estimate, which is for “contracts being struck”. Contracts struck in the National Electricity Market (NEM), of course, reflect the fact that the supplier receives not just the revenue but also income from the sale of RECs.

When Blakers and Stocks state that the cost of PV solar is $70/MWh, what they mean to say is that this figure is the income being received by the generator—which is also receiving at least $65/MWh from the sale of the associated REC. The overall cost to the consumer of PV in this example is therefore $135/MWh (the contracted income plus the minimum $65 REC, paid by those purchasing RECs and passing costs to consumers). If the AIG estimate is correct and the true cost is actually $110, the PV generators are making economic rents (superprofits) of at least $25/MWh, since the $110 figure includes a fair market return on capital and other costs. And since the PV generators have an effective short-run marginal cost of zero, the thermal generators are simply priced out of the market, so that their capacity factor is driven down. (The situation is worse in the UK, where recent figures revealed that solar generators were receiving more in subsidies than they were receiving for selling electricity (Fernandez, 2018).)

The current policy settings mean that consumers are effectively paying renewables generators twice—once for the electricity and once for the REC—while being denied access to electricity that could be provided for $80/MWh on a reliable, dispatchable basis.

Ultrasupercritical coal (USC) technology is also an emerging technology, and while Blakers and Stocks do not consider it at all, there is a good chance it will come down in price and yield increases in efficiency, with vigorous competition in the marketplace from China, Japan and Korea. Blakers and Stocks are sanguine about improved prices and efficiencies for PV in the 2020s (arguing it would be surprising if Australia did not achieve the Mexican price by 2025), but fail to consider similar advances for coal—while citing prices for supercritical but not USC.

Claims that solar is now cheaper than coal are simply wrong, as are claims that coal globally is having some kind of near-death experience. There is plenty of scope for renewables—but at the margins of more reliable, dispatchable generation such as coal, gas and nuclear.

In their attempt to show that Australia could meet its Paris target at zero cost, Blakers and Stocks ignore the option of proven new technology currently commercially available at about the current wholesale price of the Australian Electricity Market that they cite ($80)—while assuming uncertain reductions in price for PV in the (unknown) future.

Existing average global coal plant efficiency is 33 per cent. State-of-the-art USC plant is achieving 45 per cent efficiency and better (CIAB, 2015), with each 1 per cent efficiency gain reducing emissions of carbon dioxide by 2 to 3 per cent. Indeed, such a “churning” of plant, possibly involving plant closures in Europe and Australia, and new plant construction in Asia (where more than 400 such plants are under construction and more than 1500 are planned), even if averaging current off-the-shelf supercritical plant efficiency of 40 per cent, has the potential to reduce carbon dioxide emissions by 2billion tonnes a year. This is only slightly less than the total emissions currently avoided through hydroelectricity or nuclear energy (each 2.2 billion tonnes), and more than three times the mitigation achieved in 2015 by all non-hydro renewables (CIAB, 2015: 2–3).

The problem is, of course, that no one will invest in USC technology when we have subsidised a gold (plated) rush to non-dispatchable renewables with close to zero marginal cost of deployment that is driven by regulation, not cost. (This seems to be a new, reverse form of the Averch–Johnson effect, which occurs when rate of return regulation encourages gold-plating of capital investments; with the AEM, regulation has encouraged investment in assets that have destroyed the capital value of existing assets and made reinvestment unlikely).

There are numerous other problems with the Blakers and Stocks analysis, including the constant use of the measure of the “Levellised Cost of Energy” (LCOE). They are not alone in this, but LCOE is only a useful measure in the context of a large, reliable system. LCOE encourages non-systems thinking, whereas electricity networks are systems. As South Australia learned to its cost, LCOE is irrelevant when the system has insufficient security. The LCOE of non-dispatchable renewables is meaningful if it comes at the margin of a secure system; otherwise, the cost of providing dispatchable back-up, such as pumped storage (only 80 per cent efficient) must be charged against wind or solar.

Non-dispatchable renewables have a relationship to the NEM that is essentially parasitic, rather than symbiotic: they draw benefit from the system, but are decreasing its well-being.

It should also be noted that Blakers and Stocks use an assumption that biases their analyses in favour of their preferred (capital-intensive) solar, wind and pumped-storage hydro technologies: a low discount rate of 5 per cent. This is the same device used by the Hydoelectric Commission in Tasmania when comparing coal-fired electricity with the Gordon Below Franklin Dam, where they not only used a 4 per cent discount rate, but assumed the price of coal would rise at 2 per cent real over the sixty-year life of the dam (Kellow, 1983). As economists pointed out at the time, the appropriate rate was the opportunity cost of capital in the private sector (about 8 per cent real) as lower rates would favour public sector investments and draw capital away from the private sector.

They also use a twenty-five-year “technical” life for both wind and PV. As noted above, there are reasons to be sceptical that either will have an economic life that long, as attrition of both will limit their output.

One issue on which one can agree with Blakers and Stocks is their preference for pumped storage rather than the celebrated battery revolution. Batteries in most household applications do not have good life expectancy. Tesla guarantees its lithium-ion batteries for eight years and there are some guaranteed for ten, but there is little historical data to back these up and some observers suspect the prices are front-loaded in the belief that gross profits now will cover future payouts under warranty. And it is not a matter of batteries lasting eight years and then needing replacement. Performance declines measurably during this time.

The future

As this critique shows, Blakers and Stocks are far too sanguine about the prospects for PV solar energy, especially that in large centralised “farms”. Solar is not competitive with new ultrasupercritical coal, which can achieve a reduction in GHG emissions in the ballpark of the Paris target for the existing coal-generating capacity it replaces. Solar, at up to $200/MWh, can compete without subsidy with retail electricity (my utility is charging me $259/MWh), and therefore should be encouraged in roof-top use—although, because it is competitive without subsidy, the justification for continued subsidisation is weak (and here in Tasmania it is displacing mostly hydro and wind, so the climate change impact is negative). Wind can probably provide electricity that is in the competitive ballpark if at the fringes of a reliable system, or (at additional cost) if co-investments in back-up generation are made.

Other recent analyses have raised scepticism about the prospects for a 100 per cent renewables future (Heard et al, 2017), and a recent statement by an electrical engineer at Cambridge University is worth noting (Kelly, 2016: 11):

humanity is owed a serious investigation of how we have gone so far with the decarbonization project without a serious challenge in terms of engineering reality. Have the engineers been supine and lacking in courage to challenge the orthodoxy? Or have their warnings been too gentle and dismissed or not heard?

If Australia sleepwalks into the future convinced that solar will become economically competitive with wholesale conventional energy if only we keep subsidising it, we are running a huge risk.

What if there are no substantial real cost reductions to be squeezed out of PV manufacture? What if $110/MWh is about as good as it gets—$30/MWh above USC, which (as a recent technology) might well be on the steeply declining part of the innovation cost curve?

Australia could find itself in 2030 as the RET comes to an end faced with the prospect of having to continue to subsidise solar, having forgone the opportunity of reducing GHG emissions at a cost lower by $30/MWh—or less, if the technology reduces in price. Unless the policy settings change, no investment will be made, existing plant will be progressively retired, and the lead time for constructing a new USC plant will mean that replacements cannot be built in time.

This is why it is important that the wishful thinking of Blakers and Stocks and others that has driven public opinion on electricity in Australia must be challenged. Forget the fact that the expansion of the RET by the Rudd government has doubled the real price of electricity in the wholesale market, destroying the competitive advantage Australia enjoyed with cheap, low-sulphur coal, natural gas, and—yes—good wind and solar potential. That milk is well and truly spilt. It is what might lie ahead that should really concern us.

References

Blakers, Andrew, Bin Lu and Matthew Stocks (2017) “100 per cent renewable electricity in Australia” Energy 133 471–482.

Blakers, Andrew, Bin Lu and Matthew Stocks (2017a) “What’s the net cost of using renewables to hit Australia’s climate target? Nothing” The Conversation 27 November. https://theconversation.com/whats-the-net-cost-of-using-renewables-to-hit-australias-climate-target-nothing-88021

Blakers, Andrew, Matt Stocks and Bin Lu (2017b) “Meeting Australia’s Paris greenhouse commitment at zero net cost” (Press Release, Australian National University). November 2017 http://re100.eng.anu.edu.au/research/re/assets/171122_Meeting_the_Paris_emissions_target.pdf

Blakers, Andrew and Matthew Stocks (2018) “Solar PV and wind are on track to replace all coal, oil and gas within two decades” The Conversation. 6 April. https://theconversation.com/solar-pv-and-wind-are-on-track-to-replace-all-coal-oil-and-gas-within-two-decades-94033

CIAB (2015) Coal Industry Advisory Board submission to the International Energy Agency to inform its forthcoming Special Report on Energy & Climate for UNFCCC COP21 decision makers. March 27.

Caithness Windfarm Information Forum (2018) “Summary of Wind Turbine Accident data to 31 March 2018” http://www.caithnesswindfarms.co.uk

CO2CRC Limited. (2015) Australian power generation technology report.

Corten, G.P. and Veldkamp, H.F. (2001) “Aerodynamics: Insects can halve wind-turbine power.” Nat. Brief Commun. 412: 41–42.

Davis, Lucas (2017) “Are Mexican Renewables Really this Cheap?” Haas Energy Institute Blog. 4 December. https://energyathaas.wordpress.com/2017/12/04/are-mexican-renewables-really-this-cheap

Fernandez, Colin (2018) “Solar farms receive more cash from green subsidies than selling the energy they produce” Daily Mail. 9 April. http://www.dailymail.co.uk/news/article-5592691/Solar-farms-receive-cash-green-subsidies-selling-energy-produce.html#ixzz5CGp2N6e4

Ferroni, Ferrucio (2014) “Solarstrom in Deutschland: ‘Klimakiller’ Nummer 1!” (“Solar power in Germany: ‘Climate killer’ number 1!”) Europäisches Institut für Klima & Energie 24 March. https://www.eike-klima-energie.eu/2014/03/23/solarstrom-in-deutschland-klimakiller-nummer-1

Ferroni, F. and Hopkirk, R.J. (2016) “Energy Return on Energy Invested (ERoEI) for photovoltaic solar systems in regions of moderate insolation”, Energy Policy, 94: 336–344.

Ferroni, Ferruccio, Alexandros Guekos and Robert J. Hopkirk (2017) “Further considerations to: Energy Return on Energy Invested (ERoEI) for photovoltaic solar systems in regions of moderate insolation,” Energy Policy 107: 498–505.

Fierer, N. and Jackson, R. B. (2006). “The diversity and biogeography of soil bacterial communities.” Proceedings of the National Academy of Sciences of the United States of America, 103(3): 626–631.

Fthenakis, V.M. and Kim, H.C. (2007) “Greenhouse-gas emissions from solar electric and nuclear power: a life-cycle study”, Energy Policy, 35(4): 2549–2557.

Heard, B.P., Brook, B.W., Wigley, T.M.L., and Bradshaw, C.J.A. (2017). “Burden of proof: A comprehensive review of the feasibility of 100 per cent renewable-electricity systems.” Renewable and Sustainable Energy Reviews, 76: 1122–1133.

Jordan, D.C., Sarak R. Kurtz, Kaitlyn VanSant and Jeff Newmiller (2016) “Compendium of photovoltaic degradation rates” Progress in Photovoltaics 24(7): 978–989.

Kellow, A. (1983). “Public project evaluation in an Australian state: Tasmania’s dam controversy.” Australian Quarterly, 55(3): 263–277.

Kellow, A. (1996) Transforming Power: The Politics of Electricity Planning. Cambridge: Cambridge University Press.

Kelly, M.J. (2016) “Lessons from technology development for energy and sustainability.” MRS Energy & Sustainability: A Review Journal 3: 1–13.

Lombardo, Tom (2014) “A Robot that Cleans Solar Panels” Electronics Design. 6 April. https://www.engineering.com/ElectronicsDesign/ElectronicsDesignArticles/ArticleID/7427/A-Robot-that-Cleans-Solar-Panels.aspx?e_src=relart

Moalem, Mehran (2016) “We Could Power The Entire World By Harnessing Solar Energy From 1 per cent Of The Sahara” Forbes, September 22. https://www.forbes.com/sites/quora/2016/09/22/we-could-power-the-entire-world-by-harnessing-solar-energy-from-1-of-the-sahara/#1659b92ed440

Molki, A. (2010). “Dust affects solar-cell efficiency” Physics Education 45: 456–458.

Norman, J. (2017) “ERM Power, Alinta Energy opt for penalties rather than meeting renewable energy requirements”, ABC News. 23 February. http://www.abc.net.au/news/2017-02-23/erm-power-alinta-pay-penalties-rather-than-meet-ret/8296680

Parkinson, Giles (2018) “Solar, wind farms hit as AEMO slashes output calculations.” Reneweconomy. 1 May. https://reneweconomy.com.au/solar-wind-farms-hit-as-aemo-slashes-output-calculations-18481

Raugei, M., Sgouridis, S., Murphy, D., Fthenakis, V., Frischknecht, R., Breyer, C., Bardi, U., Barnhart, C., Buckley, A., Carbajales-Dale, M. and Csala, D. (2017) “Energy Return on Energy Invested (ERoEI) for photovoltaic solar systems in regions of moderate insolation: a comprehensive response”, Energy Policy, 102: 377–384.

Reed, Tennant (2017) “Should we be looking at new coal-fired power stations?” 20 January. https://blog.aigroup.com.au/should-we-be-looking-at-new-coal-fired-power-stations

Rodriguez, Daniel (2017) “Mexico Seeks to Cut Power Subsidies, Launch 27.1 GW Solar Plan”. Platts. 9 May. https://www.platts.com/news-feature/2017/oil/commodities-in-mexico/cut-power-subsidies-solar-plan-050517

Schramm, M., Rahimi, H., Stoevesandt, B., and Tangager, K. (2017). “The Influence of Eroded Blades on Wind Turbine Performance Using Numerical Simulations.” Energies 10(9): 1420.

Shahan, Zachary (2014) “13 Charts On Solar Panel Cost & Growth Trends” CleanTechnica. 4 September. https://cleantechnica.com/2014/09/04/solar-panel-cost-trends-10-charts

Vorath, Sophie (2017) “Wind, solar cheaper than coal and gas, so let’s get on with it.” Reneweconomy (25 May) https://reneweconomy.com.au/video-wind-solar-cheaper-than-coal-and-gas-so-lets-get-on-with-it-67093

Winner, L. (1978). Autonomous technology: Technics-out-of-control as a theme in political thought. Cambridge, MA: MIT Press.