Lies, damned lies and renewable energy statistics!

It's a terrible thing when people misuse statistics, because it gives the whole field a bad name. It feeds the anti-science cover-story that says: "People can prove any bloody thing with numbers!". In this little blog I want to take issue with a graph produced by Simon Holmes a Court (SHAC). Strictly speaking he didn't produce it ... but he tweeted it and the graph says clearly that he supplied the "concept".

Dodgy graphs are a particularly dangerous form of statistical abuse, because they carry an air of scientific credibility.

The king of picking dodgy graphs to pieces is Albert Cairo. He's a Professor of Visual Journalism at the University of Miami. Cairo has written a few books on the subject and spent years preparing data graphics for newspapers. His books have great titles like "How charts lie" and "The Truthful Art" (both are available from many on-line book sellers). But they are far more than just catchy titles, they are chock full of information about how to make your graphs or charts more accurate. He has plenty of case studies featuring dodgy charts from some very major news sources. He's also perfectly comfortable picking apart his own mistakes; after all, some mistakes are innocent; but others are either deliberate or culpably stupid. Where does SHAC's graph fit in? You be the judge.

So here's the chart. Put out on twitter last year (Jan 12th). I'd forgotten about it until somebody reminded me and I figured it was time to write about it.

According to the chart, non-hydro renewables during the decade 2008 to 2017 grew faster than nuclear power in the decade 1979 to 1988. What is being measured here is additions of electricity generation measured in terawatt-hours per year (TWh/yr). For example, if you build a 1,000 megawatt nuclear reactor it can probably generate between 6 and 8 terrawatt hours of energy per year depending on how much up-time it has. So the TWh/yr figure rises by 6 to 8. If you install 1,000 megawatts of wind power, then it can generate between 2 and 3.5 terawatt hours per year ... so that's the increase. If you know the total electricity output in various years, you can be more precise about the increase.

What's the first thing you might notice in SHAC's graph?

Here's a hint. Can you compare the growth of solar power (measured in additions of TWh/year) in China and Australia?

No?

Why not?

Perhaps because Australia is sunnier than China? Yes, that's true. But mostly because China has 1.4 billion people compared to 25 million in Australia. So of course China can can add more of any bloody thing she wants to.

So ... what was the population of the planet in 1979 at the start of the nuclear expansion?

According the Worldometers the global population in 1979 was just 4.3 billion. And in 2008? 6.7 billion. In fact, the average global population over the decade starting in 1979 was about 4.6 billion and in the later decade it was 7.1 billion. So the population was about 50 percent higher during the right hand decade than the left. So, to take this into account, we can shrink the right hand graph by a third or multiply the left hand graph by 1.5. Both will roughly transform the graphs into per-person increases. That removes SHAC's China-vs-Australia type mistake. It isn't the only mistake but it is the most blindingly obvious one.

Let's look at the impact of fixing this mistake on the two cumulative columns on the far right. Just multiply the nuclear column by 1.5 ... and you get 1911 ... rather higher than the non-hydro renewable total of 1639.

That was easy. There can't be too many people on the planet who don't understand that comparisons of regions (or times) with different populations need to be normalised. Actually, that's not true. SHAC's mistake is pretty common. Plenty of people reject Australia doing any thing about climate change because "we only generate 2 percent of the emissions". It's the same mistake ... so SHAC is in pretty common, if bad, company.

But wait there's more

If that was the only problem with this graph, it would be terrible. But the more you think about it, the worse it becomes.

Does anybody really think burning forests for electricity is good? By including biomass in with wind and solar, SHAC clearly does ... but I certainly don't. Adding in biomass on the right hand side like it's a benign source of electricity is a sign of desperation. I didn't bother to check whether he included biofuel energy or only burning wood for electricity. Either way. It should be left out.

And now a couple of bigger problems.

Technological progress

The reactor designers in the 1960s and the engineers and builders of following decades worked without the advantages of cheap plentiful computing. This has impacts right through the supply chain; whether it's wind turbines, solar panel factories, logistics, mining. Every aspect of the wind and solar roll out of today has advantages over the preceding nuclear rollout, yet still the rollout continues to be glacial. I'll give an alternative graph below which I think more accurately allows comparisons of rollout speed. But let's continue with the problems.

Comparing many countries with a few

The earlier nuclear rollout involved a relatively small number of countries. Today's wind and solar deployment involves many more countries. It isn't just that there are more people on the planet now, but more of them are involved rolling out wind and solar than rolled out nuclear in the past. Why include China in the global renewable rollout figure when she had no part in the nuclear rollout?

According to the BP Statistical Review 2018, China produced 471 TWh of non-hydro renewable energy in 2017, and almost all of it was developed in the preceding decade. To properly compare the potential for climate saving renewable and nuclear rollouts, it would be reasonable to pick either the same countries in both aggregates, or at least the countries with a similar population size. In any event, China's 471 TWh shouldn't be included in the 1639 TWh. Including her rollout just increases the degree to which this comparison compares very different things.

The best comparison would be in countries with similar levels of technological development.

It would certainly be reasonable to compare China's current renewable rollout with China's current nuclear roll out ... if she was putting the same effort into both. China built 8 reactors in the 1990s but her first substantive rollout began after 2005, with some 30 reactors being constructed. Per-person, this is a dismal effort and certainly smaller than her renewable rollout. But both her renewable and nuclear rollouts have been dismal when compared to either her coal and gas developments or to the efforts of developed countries. Which brings us to my alternative graph.

An alternative

My chart uses the 2020 BP Statistical data. It doesn't deal with all of the difficulties of comparison, but it at least deals with the basics by doing per-person comparisons. The best comparison would be of a country during the past decade which has opted strongly for renewables but which is at a similar level of development to France or some other country in 1979 that chose nuclear. I don't think there is any such country, but I haven't searched exhaustively and if anybody has suggestions, then I'd be happy to plot the data for them.

What is clear is that nuclear was rolled out extremely rapidly by all the countries who tried back in the '70s and '80s and that the current renewable rollouts have been sluggish by comparison. Japan, post Fukushima, desperately rolled out plenty of wind power in an effort to replace the perfectly good nuclear plants that she took off-line, but still didn't match the rollout speed of countries which were much poorer during their nuclear rollouts.

My chart, as I mentioned already, shows how slow China's current rollout of renewables is today compared to the French rollout of the 70s and 80s. Is this because of China's lower level of technical development today relative to France in 1979? I doubt it. The GDP per person of France in 1979 was $11,179 (USD) compared to China's GDP per person today of $10,261 (USD). No. China's slower rollout of clean energy has been by choice. There's no question about how fast China can make wind turbines and solar panels ... but she's selling them rather than installing them. Here's the chart ... be careful, this isn't a per person chart; it's goal is to compare China's clean electricity deployment (nuclear, wind and solar) with her fossil fired electricity deployment.

The thing about China is that you can't easily predict the future by looking at a graph of the past.

China's recent coal fired power stations are "super critical", a fancy term meaning they run very hot ... which makes them more efficient than older coal technologies. But that's not the interesting thing. The interesting thing is China's HTR-PM nuclear reactors. These are designed to be "drop-in" boilers for those very same supercritical coal plants. Replace the boiler in one of those coal plants with a nuclear reactor and the investment in the rest of the plant is preserved. All the problems of site selection were dealt with during the building of the coal plant. If this goes well, it should enable China to clean up that huge fossil fueled grid at a pace that will make even the Swedish nuclear rollout of the 1980s look slow. Anybody concerned about the planet should be crossing their fingers and toes and wishing them the best of luck. It's not only a bold plan, but shows what you can do when you do actually plan, instead of using our Heath-Robinson market based chaos.