Lies, damn lies, and statistics — examining the claims for renewable technologies
What the plight of ‘The Flight of the Phoenix’ teaches us about current energy priorities
Proponents of sustainable energy arrangements (of which I am one) face a difficult challenge — reconciling the the fundamental incompatibility between the low grade¹, intermittency, and high cost of current renewable energy technologies with the requirement of industrial economies for energy sources which are high grade, continuously available, and inexpensive.
In our efforts to build confidence, mobilise support, and accelerate transition, we are inclined to refract the statistics of renewable technologies through glasses of the rosiest hue. Setting aside clearly recognisable self interests, this is largely well meaning. It is not, however, a luxury we can afford.
If the overwhelming consensus of the global community of climate scientists is to be relied on, the earth has already exceeded the safe working limit of the atmosphere to absorb climate change gases. Even if we stopped burning carbon based fuels today, they claim, our climate is on a trajectory with outcomes resembling the plot lines of Hollywood science fiction dystopia.
In the classic 1965 movie ‘The Flight of the Phoenix’, a cargo plane crash lands in the desert. The crew rebuilds a new plane — the Phoenix — from the wreckage of the old one. James Stewart has one engine starter cartridge left. If it doesn’t start the engine, they die. It’s gripping stuff.
So is energy policy in the 21st century. With time running out as the climate and environment degrade, energy policy has to assemble a new energy system from the wreckage of the old one, and start its engines with our small remaining stock of energy and financial starter cartridges. It’s essential that we don’t waste time, and that we don’t squander our starter cartridges.
Yet by causing the misallocating of finite resources, wasting time and wasting starter cartridges is precisely what we do when we contaminate energy policy formulation with statistical distortion, however well meaning that distortion is.
Take, for example, a recent article by Bloomberg New Energy Finance ‘The World Nears Peak Fossil Fuels for Electricity’², (June 13, 2016). BNEF claims to provide ‘unique analysis, tools and data for decision makers driving change in the energy system’ and are therefore typical of the quality of mainstream reporting in this area.
In the article, BNEF presents historical statistics claiming to demonstrate that renewable technology availability (‘capacity factor’) is improving dramatically. It presents future extrapolations intended to create the impression that retail prices and marginal construction costs are falling below thermal generation, and that the growth of electric vehicles required to tackle the storage problem is sustainable. On the basis of this analysis, it predicts that ‘There Will Be No Golden Age of Gas’.
A more considered inspection of the statistics appears to suggest precisely the reverse conclusion.
Going wild
BNEF claim that renewable power plant availability improvements is a ‘fast-moving story’ in which factors are ‘going wild’, citing factors of 50% as achievable. Yet, the graph supporting the claim reveals a different story.
In the fifteen years since 2000, average power plant unavailability has fallen by only ten percentage points, from 85% to 75% and, on current trends, will still exceed 60% in 2040. And, since investment is (presumably) allocated on a least-inefficient-first basis, the expectation that this trend can be maintained is at best unwarranted.
Evil math
BNEF claims for falling prices is more complex to deconstruct, and even more misleading. In a chart titled “The Beautiful Math of Solar Power”, the logarithm of the price of solar panels as plotted against the logarithm of cumulative generating capacity and a trend line constructed from which, it is claimed, “wind and solar power will be the cheapest forms of producing electricity in most of the world by the 2030s”. This is a critical claim, upon which the probability of achieving growth is highly dependent.
The so-called ‘log-log’ plot is an analytical device for interpreting phenomena which demonstrate exponential growth. If the phenomenon is exponential, it generates a straight line on a log-log plot. Extrapolation of that line allow future estimates to be easily computed.
Unfortunately, a straight line on a log-log plot can be drawn through the early data of a great many phenomena which are not exponential, and which produce extrapolations which are meaningless. Furthermore, the compression and distortion of the scale grossly magnifies the contribution of early time data, and makes interpretation of late time extrapolations highly unintuitive.
For example. Solar PV originated in the space program, and early solar panel costs were dominated by the expense of making them survive launch, 500 degree thermal cycling, and hard radiation. Any straight line coerced into passing through that data is going to have a higher (more flattering) slope.
There is no theoretical basis for assuming panel cost/cumulative capacity relationship is exponential. In fact, the cost price reductions since 2008 were caused by gross oversupply, not ‘learning’.
Finally, the math (properly understood) is evil, not beautiful. An order of magnitude increase of a small thing (the left of the graph) is easy. An order of magnitude increase of a large thing (the right of the graph) — particularly a thing that consumes quantities of energy, resources, and money — is very different. That innocuous final two centimetres, projected back into the time domain, implies colossal resource demand and several decades of effort. You can see why we prefer log-log.
Taken together, the claims that gas is obviated by a technology that, despite decades of effort, is still unavailable on average three quarters of the year, and will take decades and colossal quantities of material and financial resource to maintain necessary cost reduction trends, start to look unconvincing.
Furthermore, something has to keep the wheels turning when the renewable power plants sit idle during those long, regionally extensive periods of cold, dark, windless conditions that are associated with winter high pressure weather systems in the northern industrial regions. We can’t burn coal, oil supplies are proving remarkably difficult to sustain, and nuclear waste storage challenges present their own rich crop of Hollywood dystopia plot lines. There seem to be few viable alternatives to gas.
More problematic, it’s not entirely clear that we have sufficiency financial and energy resources to pursue science projects and white elephants. A view is emerging that the current forms of renewable technology are so hopelessly inefficient that they will never achieve the levels required to sustain economic activity, and that debt currently financing R&D and deployment should be reallocated to primary R&D into (currently unidentified) technologies with the potentials at least order of magnitude higher than current.
One cartridge. How are we going to use it?
¹ ‘grade’ here has a particular meaning in thermodynamics, referring to the concentration and therefore capacity to do work.
