Climate change solutions: false, fictional and dangerous

[responsivevoice_button]

Almuth Ernsting from Biofuelwatch explores why sci-fi climate change solutions are dangerous even though they will never be realised and why it’s important that they are specifically challenged.

 

I wrote this article en route to a public meeting about a very peculiar planning proposal for Milford Haven, Pembrokeshire, the UK’s largest energy port. I have been involved in many campaigns against biomass or biofuel projects falsely presented as ‘climate change’ solutions, but which would in fact cause significant harm to forests and other ecosystems and to communities, whilst often causing as many if not more CO2 emissions than the fossil fuels they are said to replace. The planning proposal in Milford Haven, however, is different because there seems to be virtually no prospect of it being technically feasible. Failed attempts to operate a technology that doesn’t work can still cause significant local air pollution and noise, but such a scheme is fundamentally different from a conventional biofuel refinery or biomass power station.

At the public meeting, a Welsh-Cypriot company, Egnedol, tried to convince people that they can successfully operate a whole set of ‘low-carbon technologies’, which nobody has ever made to work in the UK, or even worldwide. The technologies include: a state-of-the-art gasification plant which turns waste and biomass into a gas that burns as cleanly as natural gas and that generates both electricity and heat; an indoor micro-algae farm which will produce not just food for an adjacent fish and prawn farm but also biofuels; and a biofuel refinery which will turn some of the purified gas from waste and wood burning into drop-in transport fuels (biofuels with identical properties to conventional diesel or petrol). In addition, they promise that their power plant will provide enough spare heat for a host of food production enterprises, including a perfectly closed-loop fish and prawn farm that will generate no effluent whatsoever.

In a previous article, I illustrated how the UK’s ‘pro-development’ planning and permitting systems discourages any scrutiny of improbable technology claims made by companies such as Egnedol. Laissez-faire developer-friendly policies are supposed to encourage cutting-edge innovation but all too often end up facilitating doomed start-up ventures which can waste millions of pounds worth of investors’ money. Any fraud is unlikely to be uncovered. The Serious Fraud Office’s budget has been so severely cut that, during 2014/15, only sixteen out of 2,832 reports of suspected fraud or corruption resulted in investigations.

Egnedol’s improbable plans bring together three separate technologies which are being widely promoted as climate change solutions. Between them, these technologies – biomass/waste[1] gasification, algal biofuels, and drop-in biofuels from wood or waste – have attracted billions of dollars in subsidies worldwide. Two of them – algal biofuels and ‘cellulosic’ biofuels made from wood, agricultural residues, whole crops, or biomass contained in waste – have been stuck in the Research and Development stages for many decades, with no breakthrough in sight. A small number of waste and biomass gasification plants have been operated successfully in other countries, but they have been expensive to build and operate, and have generally required at last a year of modifications and repairs. Even then, they provide at best minor efficiency and air emissions advantages when compared to standard combustion plants and waste incinerators. In the UK, all attempts to build and run such gasifiers have ended in failure.

Each of the three technologies represents false solutions to climate change for the same simple reason: climate science shows that greenhouse emissions must be cut as quickly and as steeply as possible if we want to have any hope of avoiding the worst impacts of climate change. Technologies which are nowhere near commercial deployment, or which are far too expensive and difficult to implement at scale, even with public subsidies, cannot contribute to this aim.

I call these non-existent ‘climate change solutions’ sci-fi solutions. Sci-fi solutions are very different from existing technologies which have been widely criticised as ‘false solutions’, such as biofuels made from sugar, cereals or plant oils, nuclear power, or fracking (which even the International Panel on Climate Change classes as low-carbon). Biofuels, together with biomass electricity, have by far the highest land footprint of all energy sources.[2] The International Energy Agency reported in 2011 that 30 million hectares of land worldwide were used to grow biofuel feedstock, but that those supplied only around 2% of the world’s transport fuels. This land-hunger has turned biofuels into a major driver of deforestation, biodiversity loss and land-grabbing, as well as food price volatility and loss of food sovereignty in many parts of the world. It also makes them responsible for greenhouse gas emissions which are commonly higher than those from the petrol or diesel they replace. But despite their disastrous social, environmental and climate impacts, first-generation biofuels are the product of perfectly viable and commercially attractive technologies. Cutting down rainforests for palm oil for biofuels might be madness, but cars can drive just fine with palm oil biodiesel blends.

Generating electricity from nuclear fission is an equally proven technology, albeit a highly expensive and dangerous one, with no proven safe methods for storing nuclear waste, and one dependent on toxic mining of uranium. Fracking in the US has been so successful that it has significantly reduced energy costs and the US government expects the country to become a net exporter of gas in 2017, for the first time in sixty years. It also spews so much methane into the atmosphere that some scientists estimate that the increase in US methane emissions from gas drilling since 2002 is responsible for up to 60% of the widely reported rise in methane levels in the global atmosphere, making a unit of energy derived from fracking even worse for the climate than one derived from coal.[3]

Why do governments and corporations support sci-fi technologies?

It is easy to see why corporations love these profitable false climate change solutions – even if (as in the case of nuclear power) they are only profitable with very substantial public subsidies. It is also easy to see why governments like such technologies which allow them to reduce carbon emissions on paper, and which boost energy production without requiring any change in the prevailing economic and social model, nor much change in infrastructure.

Support for sci-fi technologies, especially amongst governments, appears more baffling at first sight. Why would anyone choose to sink millions or even billions into energy technologies which are unlikely to ever generate much, if any, energy? And why should this worry us, given that fossil fuels are being subsidised to the tune of $2 trillion a year worldwide?

To understand both the reasons why sci-fi solutions are being funded and the dangers they can pose, we need to first look in more detail at some of the technologies and players involved. The three examples I have chosen are coal power stations with carbon capture and storage (CCS), Bioenergy with Carbon Capture and Storage (BECCS), and cellulosic biofuels.

Coal power stations with CCS[4]:

The promise of ‘clean coal’ – i.e. coal burning without carbon and toxic air emissions – offers a lifeline to the coal industry. Coal power stations have been approved as ‘CCS ready’,[5] even if the prospects of their carbon emissions ever being captured are approximately zero. The US government has made $6 billion available to the development of CCS. The concept of low-carbon fossil fuels with CCS has been endorsed by the IPCC, the International Energy Agency and by governments.[6] Time and time again, CCS is held up as the reason why the world doesn’t need to turn away from fossil fuels.[7] Hype about CCS thus plays an important role in legitimising ongoing coal burning. Yet energy corporations and governments have been quietly abandoning coal with CCS as a viable prospect worthy of funding.

To date, only a single such project has been implemented at commercial scale: in 2014, the Canadian company SaskPower inaugurated a carbon capture facility for one unit of their Boundary Dam power station in Saskatchewan, which had originally been built in 1969. Industry media hailed the announcement as a breakthrough for CCS worldwide. One year later, the opposition New Democratic Party obtained a Freedom of Information request which revealed a very different picture: far from capturing 90% of the unit’s CO2 emissions as promised, SaskPower struggled to capture 55% throughout the year, and the carbon capture unit had been shut down for weeks at a time. A report commissioned by Community Wind Saskatchewan had already shown that the project would not have broken even during its lifetime if it had worked as intended. This analysis took account of the fact that SaskPower are selling the captured CO2 to an oil company, Cenovus, which uses it to pump additional oil which could not otherwise be recovered. Failure to capture as much CO2 as anticipated has turned the project into a serious liability for SaskPower: not only have they earned less money from CO2 sales but they have ended up paying millions of dollars in fines to Cenovus.

Perhaps most damaging to any future CCS ambitions have been the revelations about the energy costs of carbon capture. SaskPower has had to admit using 30-31% of the plant’s energy just to capture and compress CO2. This means that the equivalent of one new coal power station would be needed to power carbon capture units for two others. It is just as well that the Boundary Dam power station has been such a failure: the carbon footprint emitted during ‘Enhanced Oil Recovery’ (i.e. getting more oil out of the ground), plus the CO2 that cannot be captured, plus the CO2 from burning the additional recovered oil add up to significantly greater carbon emissions than those from an unabated coal power station unit.

Other coal CCS projects have been abandoned before they ever got commissioned: in the US, FutureGen was to have been the Bush Administration’s flagship CCS project, involving a highly complex Integrated Gasification Combined Cycle (IGCC) coal power station with carbon capture. It was scrapped after major cost overruns, at a $175.5 million loss to US taxpayers. President Obama revived the scheme as FutureGen 2.0, with a supposedly simpler technology: oxyfuel combustion with carbon capture. This time, the federal government spent $202.5 million before plans were abandoned after not a single private sector investor had come forward. Vattenfall and RWE, who previously invested in CCS Research and Development, have pulled the plug on CCS projects, citing prohibitive costs. The Norwegian government has given up on its flagship commercial CCS project in Mongstad, the EU is yet to co-fund a single CCS scheme (or rather, to find a single such project which they could possibly co-fund), and even the government of Alberta, hitherto one of the greatest enthusiasts for the technology, has announced that they would fund no more CCS because of the exorbitant costs involved. The UK government joined the global trend last year when they withdrew a $1 billion fund for CCS, though not before spending $44 million on feasibility studies for two now abandoned fossil fuel CCS plants.

Clearly, the main danger from CCS lies in the pernicious impact which the hype about it has on climate politics. And billions are being wasted on a hopeless technology which ought to be spent on proven ways to reduce carbon emissions, such as community wind and solar projects or home insulation.

Bioenergy with Carbon Capture and Storage (BECCS):

Compared to coal CCS, BECCS is an even more fanciful idea. It would involve capturing CO2 from biofuel refineries and biomass-burning power plants and sequestering it long-term. It is being promoted as a ‘carbon negative technology’ i.e. as a means of sucking previously emitted CO2 from the atmosphere. The idea is this: bioenergy is classed as inherently carbon neutral by governments and energy companies, based on the assumption that new plant growth will reabsorb the carbon emitted from burning existing trees or other plants. It is scientific nonsense: bioenergy releases carbon which has been stored in vegetation and which would otherwise continue to be sequestered by soils and ecosystems. Burning this carbon transfers it to the atmosphere. Furthermore burning biomass for heat or electricity results in greater upfront carbon emissions than burning fossil fuels (per unit of energy). But based on the idea that bioenergy is carbon neutral, BECCS proponents claim that capturing and storing some of the CO2 emissions from bioenergy will make it carbon negative. The 2014 IPCC report endorsed this bizarre idea, stating that the great majority of models showed that keeping global warming to within 2oC requires large-scale ‘negative emissions’ through BECCS.[8] The modellers had not actually studied the feasibility of BECCS, nor the climate impacts of procuring vast amounts of additional biomass – they had simply entered ‘carbon negative’ BECCS figures into their models. Unlike coal CCS, BECCS has never even been tested at a small scale, with one exception: there are a small number of ethanol refineries which capture CO2 from fermentation. This is relatively cheap and simple, but the amount of CO2 captured is less than the CO2 emitted from fossil fuel burning to power the refinery. Such a project thus cannot be considered carbon negative, not even if all the greenhouse gas emissions associated with converting land to ethanol production, with fertiliser use and with burning the ethanol are ignored.

Capturing CO2 from biomass-burning plants would be even more challenging, expensive and energy intensive than capturing it from coal: biomass feedstocks are chemically much less homogenous, which makes carbon capture more challenging. Even more seriously, generating energy from biomass results in greater CO2 than generating it from coal. This means that even more energy would be needed to capture that extra CO2, making the whole concept even less economically viable.

As with coal CCS, the real danger of BECCS stems from the hype around it. IPCC models are being used to falsely reassure policy makers that it is possible to ‘overshoot’ carbon targets, i.e. to burn enough fossil fuels to heat the temperature by far more than 2oC, yet to still reach the 2oC goal by scrubbing some of the carbon from the atmosphere later on. And given the growing hype and number of academic studies about BECCS, there is a distinct possibility that significant amounts of public funds could be squandered on non-viable BECCS projects.

Cellulosic biofuels:

Virtually all biofuels today are produced either from plant oils (or, in some cases, animal fats) and from starch in sugar crops or cereals, which is broken down into glucose and fermented to ethanol.

In 1910, Standard Alcohol Company started fermenting sawmill residues to ethanol in the world’s first ever cellulosic biofuel refinery in the US. Their plant, located in South Carolina, had a 5,000 gallon a day capacity. It was followed by a second refinery of the same type and size in Louisiana. Both refineries operated for a few years before they were closed down because they were unprofitable. The amount of ethanol produced would not have justified the energy inputs and costs. They had to steep the processed wood in dilute sulphuric acid at high temperatures to break down complex plant cell molecules into fermentable sugars.[9] They would have had to deal with equipment being corroded by the acid, with relatively low yields of fermentable sugars, and with chemicals forming during the process which inhibited ethanol fermentation. A significant proportion of the sugars would have been the wrong type of sugars for fermentation and significant energy would have been required to boil off the water in order to obtain pure ethanol.[10] Around twenty such plants were built worldwide during World War Two. All but a few in the Soviet Union were closed after the war and all would have faced the same limitations.

In recent decades, billions of dollars in public funds have been spent worldwide on trying to develop efficient cellulosic biofuel production, i.e. biofuels made from wood, grasses, or crop residues. But there is no evidence at all that the inherent problems faced by Standard Alcohol Company more than a century ago are being overcome.

Some of the methods being researched involve ‘thermo-chemical conversion’. This generally means exposing biomass to high temperatures either with controlled oxygen or none, cleaning the gas which results from this process and then converting it to chemicals which resemble conventional diesel, petrol or kerosene by using chemical catalysts. Others involve ‘biochemical conversion’. In most cases, this involves using enzymes secreted by microorganisms to break up plant cells into fermentable sugars, and then fermenting the different types of sugars to ethanol and/or butanol. Much of the research into biochemical conversion pathways focusses on genetically engineering micro-organisms some of which can secrete the right types and combinations of enzymes and others which can ferment all of the different types of sugars contained in plant cells simultaneously.

There is no evidence of any commercial breakthrough. There are currently three commercial-scale cellulosic ethanol refineries in the US, after another one was closed down last November having produced no ethanol.[11] Two of those ethanol plants are not producing any ethanol either. Operators of the third plant, which had been officially opened in September 2014, announced this May that they were finally ‘ramping up production’. Whether they will actually succeed remains to be seen. Three other cellulosic ethanol plants have opened, one in Italy and two in Brazil but there is no publicly available information to judge how – if at all – they are operating. Thermo-chemical conversion, in the meantime, has been all but abandoned by companies after a series of high-profile failures and bankruptcies. The challenges posed by biochemical conversion technologies remain formidable.[12]

The ‘promise’ of supposedly more ‘sustainable’ cellulosic biofuels has, time and time again, been used to legitimise biofuel subsidies and mandates which are responsible for the massive expansion in conventional biofuels in recent years, with all the disastrous impacts on biodiversity and forests, land rights, food and water sovereignty and on the climate which those entail. But, unlike most unsuccessful technologies, cellulosic biofuel research and deployment attempts entail high immediate risks. The great majority of projects use genetically engineered micro-organisms, which rely on synthetic biology.

Synthetic biology is extreme genetic engineering which involves more fundamental and aggressive changes to an organisms’ genome than traditional genetic engineering, and potentially even the synthesis of new organisms from artificially constructed DNA. Secure containment of GE micro-organisms inside industrial plants is an illusion. Industry magazine Biofuels Digest cited an anonymous ‘friend’ of theirs, speaking about biofuel company Amyris, who use GE refineries: “Having worked in nice university labs and clean room pharmaceuticals they did not know what was awaiting them in the down market dirty world of biofuel. You can’t make biofuels with anything you got to keep that clean.” There has been virtually no research into the potential environmental and public health impacts of accidentally released GE micro-organisms. But there are good reasons to worry: micro-organisms play a fundamental role in regulating the world’s carbon and nutrient cycles and in maintaining the earth’s support systems. They evolve more rapidly than higher life forms, which means that synthetic or trans-genes can mutate more rapidly than they would in plants or animals. There is no way of tracing escaped GE microbes. Finally, while GM crops can only breed with non-GM crops of the same species, many microorganisms routinely exchange genes with completely different species. There is even evidence of microbial genetic material having been passed to animals and plants.[13]

Why have sci-fi technologies come to dominate the debate about climate change solutions?

The discussion of coal CCS, BECCS and cellulosic biofuels above illustrates some of the interests that companies have in fuelling the hype around such sci-fi ‘solutions’: fossil fuel companies benefit from creating the illusion that clean coal is on the horizon and that carbon emitted today can be scrubbed from the atmosphere tomorrow. Biofuel companies can hide the very real destruction caused by their biofuels behind promises of future ‘sustainable’ cellulosic – and equally unlikely algal – biofuels.

Similarly, such false promises help governments with an interest in perpetuating the growth-based fossil fuel economy to appease public concern about climate change.

And when governments offer large grants to companies that promise to build, for example, the first ever commercially cellulosic ethanol plant, it is hardly surprising that start-up companies in particular will promise the impossible to get hold of that money. Hype around cellulosic and algal biofuels has helped synthetic biologists and start-up companies invested in extreme genetic engineering to attract large amounts of public funding and investment. Algal oil companies, such as Amyris and Solazyme (now calling itself TerraVia), procured much of their funding through claims about developing algal biofuels, before officially abandoning biofuels altogether and focussing on higher value niche food supplement and cosmetics products instead.

The IPCC’s endorsement of a whole range of sci-fi solutions such as BECCS can be understood as a result of demands put on them by policymakers: they are facing the conundrum that they must put forward scenarios which show how global warming can be limited to 2oC without admitting that this will be impossible without radical social and economic changes and an end to economic growth. As climate scientist Kevin Anderson has pointed out, their 2014 2oC scenarios rely not just on large-scale BECCS but also on a time-machine: they require greenhouse gas emissions to have peaked in 2010.

Companies and policymakers exploit techno-optimism (see Corporate Watch’s A-Z of Green Capitalism which explains techno-optimism and related concepts) in order to avoid meaningful action on climate change, to perpetuate harmful investments, and, in the case of some companies, to cash in on public subsidies. But the techno-optimism which is being exploited has much deeper roots. There is no doubt that many of those who work on sci-fi climate change solutions, whether for companies or academia, genuinely believe that they are helping to bring about transformative technological changes which will contribute to solving climate change and other crises.

Faith in ‘Technology Learning Curves’

The origins of techno-optimism can be traced back to the Enlightenment and the early Industrial Revolution. However, the most common justification for continuing to invest in technologies, which have not been successfully deployed despite decades of Research and Development, is the idea of the ‘technology learning curve’. This idea can be traced back to 1936 when Theodore Paul Wright, an aeronautical engineer in the US, observed how the cost of manufacturing aeroplanes was coming down as greater experience led to greater efficiency. He found that every time total production doubled, the requirement for and thus the cost of labour dropped by 10-15%. This learning curve – dubbed Wright’s Law – was to make first aerial warfare and later passenger aviation affordable.

Today, the learning curve idea is most commonly associated not with Wright but with Moore. George Moore was an electronic engineer who, in 1965, published an article called “Cramming more components onto integrated circuits“.[14] Moore’s article correctly foresaw that integrated circuits would “lead to such wonders as home computers”. It specifically predicted that by 1970, 65,000 units would be fitted on one chip, bringing the cost of each component down to one-tenth of what it was in 1965. He further predicted that the number of components on a chip (i.e. the power of computers) would at least double year on year for a minimum of ten years and very possibly beyond, bringing costs down at the same time. By 1975, Moore’s predictions had turned out to have been somewhat over optimistic, but the electronics industry was well on its way towards developing home computers. That year, Moore predicted that computing power would in the future double every two years rather than annually. Progress has since slowed down and Moore himself acknowledged in 2005 that it could not continue forever.

Gordon Moore of course was writing about one particular technology. Nonetheless, his optimistic forecast, which was followed by decades of real exponential progress in electronics, bolstered techno-optimism in general. The observation of learning curves in a small number of technology sectors morphed into faith in a Technology Learning Curve as a universal law – or at least as a law which applied to all but the most outlandish sci-fi ideas. In the energy sector, belief in a universal Learning Curve law was bolstered by the experience with solar PV: for several decades now, the unit cost of solar PV has been falling and efficiency has been rising. The idea that the success of solar PV is solely due to a Learning Curve is disputed: steep falls in raw materials, i.e. silicon prices, for example, have also helped to reduce costs.[15] Nonetheless, global investment in solar PV, once a very expensive and inefficient technology, has more than paid off.

Yet while some technologies have greatly advanced with investment and experience, there is no scientific basis for claims about a universal Technology Learning Curve. Moore’s predictions of future home computers were proven true in principle, but many past predictions about other technologies were not. Not only has nuclear energy not become “too cheap to meter”, as the Chair of the Atomic Energy Authority is said to have claimed in the 1950s,but the global experience of building hundreds of nuclear power stations has failed to bring down nuclear energy costs.[16] Costs have actually increased in several regions and a nuclear power plant without public subsidies remains as distant a prospect as ever.[17]

In 1984, a whole New York exhibition, called “Yesterday’s Tomorrow” entertained visitors with entire halls full of old futuristic predictions which seem bizarre today. 1950s and 60s fantasies about floating cities tended by robots and humans, or family cars doubling as private planes may seem like harmless science fiction today. But time for the urgent action needed to have any hope of avoiding the most catastrophic impacts of climate change is fast running out. In this context, fantasies about unproven technologies which will replace or clean up fossil fuels or scrub vast amounts of carbon from the air have become deadly distractions from the radical economic and social change that is required.

 

References

[1] Throughout this article, unless otherwise stated, the term ‘waste’ refers to Municipal Solid Waste.

[2] Energy Sprawl or Energy Efficiency: Climate Policy Impacts on Natural Habitat for the United States of America, Robert I McDonald et.al., PLoS One, 4(8), August 2009

[3] Methane emissions and climatic warming risk from hydraulic fracturing and shale gas development: implications for policy, Robert W Howarth, Energy and Emission Control Technologies, 2015:3

[4] Note that throughout this section, CCS is only discussed in the context of power stations, specifically coal power stations. There are certain industrial processes, including gas refining, from which CO2 can be captured far more cheaply and easily than from power stations, and there are a limited number of CCS projects involving such activities. The majority involves selling CO2 for Enhanced Oil Recovery, i.e. for getting more oil out of the ground.

[5] See www.bakermckenzie.com/-/media/files/people/rlms/rlm_amsterdam_weerokoster_20141015.pdf?la=en . RWE’s Eemshaven coal power station in the Netherlands opened in 2015 and it was specifically designed to be ‘CCS-ready’ even though RWE has announced no concrete plans for capturing any carbon.

[6] For example, www.iea.org/topics/ccs/ and www.globalccsinstitute.com/news/institute-updates/role-ccs-explained-latest-ipcc-report

[7] For example, Rajendra Pachauri, then Chair of the IPCC at the time, said after the publication of the latest IPCC Assessment Report: “With CCS it is entirely possible for fossil fuels to continue to be used on a large scale.” (www.theguardian.com/environment/2014/nov/02/rapid-carbon-emission-cuts-severe-impact-climate-change-ipcc-report)

[8] Climate Change 2014: Synthesis Report Summary for Policymakers, International Panel on Climate Change

[9] Acid-based hydrolysis processes for ethanol from lignocellulosic materials: A review, Mohammad J. Taherzadeh and Keikhosro Karimi, BioResources 2(3), 2007

[10] Dried wood contains 39-54% cellulose and 14-37% hemicelluloses. Cellulose consists of glucose, which is the sugar most commonly metabolised by organisms, including by yeast which would normally be used for ethanol fermentation. Hemicelluloses consist of different types of sugars, which cannot be fermented by any of the microorganisms used to ferment cellulose.

[11] This excludes bolt-on technologies introduced in some US corn ethanol refineries to extract some additional ethanol from cellulose in corn kernel fibre. Such technologies are classed as ‘cellulosic’ by the US authorities but do not involve any of the challenges faced by cellulosic ethanol plants.

[12] See Enzyme-based hydrolysis processes for ethanol from lignocellulosic materials: A review, Taherzadeh and Keikhosro Karimi, BioResources 2(4), 2007

[13] Expression of multiple horizontally acquired genes is a hallmark of both vertebrate and invertebrate genomes, A. Crisp et al., Genome Biology, 2015 AND Eukaryote-to-eukaryote gene transfer events revealed by the genome sequence of the wine yeast Saccharomyces cerevisiae EC1118, Maite Novo et al, PNAS, August 2009

[14] Cramming More Components onto Integrated Circuits, Gordon E. Moore, Electronics, April 1965

[15] Beyond the learning curve: factors influencing cost reductions in photovoltaics, Gregory F. Nemet, Energy Policy (34), 2006

[16] Note that others believe that he referred to future energy prices in general rather than nuclear energy in particular: www.thisdayinquotes.com/2009/09/too-cheap-to-meter-nuclear-quote-debate.html. Of course, the latter prediction would have proven just as mistaken.

[17] Historical construction costs of global nuclear power reactors, Jessica R. Loverin et.al., Energy Policy, 91, April 2016