Oh, the The Mail on Sunday. Is there no limit to your denial of climate change?
The Mail on Sunday is a sister publication of the UK tabloid Daily Mail, and has a history of running ridiculously misleading claims downplaying the reality of climate change. Probably the worst offender is David Rose, who has been constantly hammering the ideaâdespite all the evidence against itâthat the Earth has not been getting warmer for the past 16 years. To make this claim he has to egregiously cherry-pick his data, choosing where to look on a graph of temperatures to make it look like warming has slowed.
Iâve shown just how Rose is so fast and loose with reality in previous posts, when he first came to my attention for claiming warming had stopped (and then tried to show that the Sunâs lack of activity would cool the Earth, a claim for which there is essentially no good evidence), and then again when he posted a graph so wrong it would mean getting an F in ninth-grade math. You can also read debunkings of Roseâs ridiculosities from the UK Met Office, the national weather service for the United Kingdom, which regularly has to issue articles debunking the nonsense posted in The Mail.
Unfortunately, because he is so loud and given a venue in the The Mail, people who prefer fiction to reality use Roseâs claims to bolster their own. Big names in climate denial then write fact-free OpEd letters to venues like the Wall Street Journal, which get read by even more people, fooling them into thinking climate change isnât happening.
But it is happening. So Rose is still denial mode, writing yet another error-laden article for The Mail, again claiming warming has stopped, and again relying on grossly misinterpreting real data. His point this time rests on this graph:
The graph shows the temperature anomalyâthe difference in temperature from an average temperatureâcompared to predictions of surface air temperatures made by climate scientists. The graph is actually real; Rose claims itâs from the UK Met Office and the IPCC reports, though it looks very much like it was taken from climatologist Ed Hawkinsâ blog. Either way, Rose starts flinging accusations that are remarkable only by how much they miss their marks.
First, the graph is showing air temperatures. These are not the best representation of global warming, since a huge amount of the extra heat is going into the oceans. So right away, you have to be very careful extrapolating this graph to global heat content overall.
Also, you have to note what the graph Rose uses is actually showing you, and not what Rose is claiming. The light red band is the temperature prediction with a 90 percent confidence intervalâthis is a way of showing how much you trust measurements or predictions. In this case, it means there is a 10% chance (5% off the top or 5% off the bottom) that the measurements are likely to fall out of the predicted region.*
Now look again at the graph, and note the measured temperatures are still within that band. Sure, itâs at the low end, but even if the temperatures fell outside the band it doesnât mean âthe world isnât getting warmerâ as Rose so incorrectly claims. It just means the temperatures werenât quite as high as predicted. They are still within the expected range, though, and still running at a high confidence level.
Incidentally, as the site RealClimate shows, a different set of models has a wider 90 percent confidence interval, meaning the measured temperatures are even farther inside the expected range:
So, as usual, what Rose is showing you is misleading. But weâre not done.
What we've seen is that Rose isn't interpreting the graph correctly, and even if he were his claim of global warming having âstoppedâ is wrong because heâs only looking at surface air temperatures. But his problems go even deeper than that: Heâs also only looking at the past few years of data. But you have to be very, very careful when looking at short-period fluctuations! They can fool you, and extrapolating them is very dangerous if you donât know what youâre doing.
Thatâs because you expect to see ups and downs from year to year. Looking over the graph makes that pretty obvious. Sometimes you get more downs than ups, sometimes itâs the other way. Itâs a bit like throwing a pair of dice: Do it enough times and youâll get a few snake eyes in a row just by random statistics. But if the dice are off-balance, just a bit, then over time youâll see there are way more bad throws than youâd expect if the dice were fair.
Thatâs our current climate. Because human-generated global warming is loading the dice, we see that upward trend over time. You canât look at one throw of the dice, or even three or four; you need to look at dozens to understand the trend. What Rose is doing is showing you the past three throws and ignoring what was going on in the last 50.
Perhaps even more importantly, Rose is also not accounting for global climate effects due to La NiÃ±a. This is the name given to a seasonal cooling of ocean surface temperatures in the Pacific, which affects temperatures and weather across the globe. It provides a cooling trend on average, so temperatures measured in La NiÃ±a years will be lower. Itâs important to know that while the surface of the oceans cool, the heat from global warming goes into deeper waters, so the temperatures measured in La NiÃ±a years are doubly misleading when it comes to warming.
2010 and 2011 were La NiÃ±a years, and it even persisted into 2012, longer than usual. That cooled those yearsâ measurements overall, making it look like warming has slowedâbut thatâs not true; the underlying trend to higher temperatures certainly still continues. Couple that with the normal unpredictability of short-term trends and you have a denierâs dream graph.
But a dream is all it is. When you do the analysis correctly (like here and here and here) you see what happens when you wake up to reality. Clearly, Roseâs article stating âThe Mail on Sunday today presents irrefutable evidence that official predictions of global climate warming have been catastrophically flawedâ is just a load of fish-wrapping.
He also says, âThe graph confirms there has been no statistically significant increase in the worldâs average temperature since January 1997 â as this newspaper first disclosed last yearâ which again was wrong last year and just as wrong now.
After all this I havenât even said anything about the text of his article, which pretty much follows the same misleading path. I know a few other climate bloggers will be tackling this, so Iâll leave it to them to debunk, and Iâll update this post with links as they go online. [UPDATE (Mar. 19): As promised, Skeptical Science now has a takedown of David Rose's claims.]
And while deniers like David Rose publish misleading articles in The Mail on Sunday, in the real world temperatures keep going up.
Remember: nine of the ten hottest years on record have been in the past decade, there have been vast declines in Arctic and Antarctic sea ice, a corresponding rise in sea level, glacial retreats, more devastating wildfires, wilder weather, and, of course, an increase in denial in all these things.
The truth is, the planet is warming up and weâre the cause. As I have to state over and again, there is no scientific controversy about this. The only arguing going is manufactured ideologically by people who cannot back up their claims with real science. They rely on smoke and mirrors, and where thereâs smokeâ¦well.
* Correction (Mar. 19, 2013): It's come to my attention that my original phrasing about statistics in this graph was technically incorrect. This was due partially to a difference between the way astronomers and climatologists talk about statistics, but also because I was trying to simplify a complicated mathematical discussion to make it understandable, and I oversimplified (not only that, but Rose himself messed this up in his article and I followed what he said; a mistake on my part). I originally said, "The light red band is the temperature prediction with a 5 percent â 95 percent confidence. Confidence levels are a statistical way of measuring how well you can trust the data you see. Something at the 95 percent level means thereâs only a 5 percent chance the numbers are due to random noise, for example." There is a difference between confidence intervals and certainty ranges, and that's what I oversimplified. I corrected the affected text above. The last sentence about random noise was just an error on my part, though; it's also fixed in the text above. My thanks to Dave Cade at Oregon State for straightening me out, and any remaining errors are my own.
Please note this doesn't really change anything in my post; Rose is still wrong, and the Earth is still suffering global warming. I may have more about this later; the modest slowing of warming is interesting, even if it is not an indication that global warming has stopped.