Very Poor Climate Reporting – Do New Media Reporters Think We’re Stupid? @wireduk
Climatologists looking at temperature statistics for 2013 estimate that it was one of the ten warmest years on record, and 2014 is expected to follow suit.
Researchers at the UK Met Office have examined data for January to October 2013, seeing how far temperatures have deviated from the long-term average of 14.0C, calculated between 1961 and 1990.
Their early results indicate that the planet was between 0.39C and 0.59C warmer than that average figure. Using the mid-point of that range, 0.49C, would make 2013 the ninth warmest year since 1880. Every one of the top ten years has occurred since 1998.
Full analysis using data from the entire year, including November and December 2013, will be published in March 2014, alongside a comparison with the forecast from the previous year.
Meanwhile, the Met Office has also forecast the global temperature deviation for 2014. It’s also expected to join 2013 in the list of the ten warmest ever years, with a deviation above average of between 0.43C and 0.71C. Taking the midpoint of that, 0.57C, would make it the hottest year on record.
These estimates are made using an average of the three main global temperature datasets, collected by the Met Office and University of East Anglia in Britain (HadCRUT4), the US National Climatic Data Centre (NOAA NCDC) and the NASA Goddard Institute of Space Studies (NASA GISS)
Let’s break this down logically:
- Immediately, in the first sentence, we have “estimate” – I’m okay with people estimating, but when you contrast it with the headline, suddenly what was once a very definitive statement is now just an estimation.
- “…and 2014 is expected to follow suit” – where does the logic of this come from? The last time I checked, one year of results does not equal a pattern. (I know they go on to mention that ‘every one of the top ten years has occurred since 1998’ but this is based on spurious logic which you’ll see below.)
- “have examined data for January to October 2013” – of course, they don’t yet ‘have’ the data for November and December, likely to be two very cold months. You’re suddenly making claims for an entire year while missing 1/6 of the potential data. (This works out to be 16.7% of the data not present in this ‘estimation’.) I understand statistical sampling can be used, but conveniently leaving out data which can significantly alter the results of your statistical sample is considered to be very poor form by scientists.
- “calculated between 1961 and 1990” – anytime one reads an article like this, the entire history of the planet is ignored. If you were to compare the temperatures between say 1901 and 1990, you might find that the results are completely different – either in your favour or against. The point is that climatologists, who have access to ice cores which allow them access to the climate data stretching back hundreds of thousands of years, are only comparing with a stretch of 29 years, and only in this century, and only based on the data from two countries (US and UK, see below).
- “early results” – of course, these are an estimate. When the results are finalised (however spurious) they’ll be underreported unless they further confirm the already expected theory.
- “using the midpoint of that range” – why use the midpoint? If you use the largest number, surely the results would be ever more in the favour of the presupposed results? Of course, using the lowest number might not work as well – if might contradict the results. So let’s pick a random number in a range that’s devised poorly to begin with and use the middle one to make it seem more ‘reasonable’. Had the midpoint not worked out, you can bet that the range would have been adjusted or there would have been another average somewhere that made it work. Of course, using .43C instead of .49C points to different results.
- “would make 2013 the ninth warmest year since 1880” – again, with reference to point number 4 – we are ignoring thousands and thousands of years of history to make these points. What difference would it make if 2013 were the ninth warmest year since 1880 but only the 91st warmest year since 1700? A big difference! (I know that 1880 is often used as the marking point where anthropogenic events supposedly starting taking effect, but considering that there were no ice caps 30-odd million years ago then it must have been a lot warmer at some point in the past than it is now.)
- “Full analysis … will be published in March 2014” – meaning that this story about how 2013 is ‘so warm’ will be floating around for a full two months in the mainstream and new media before we get even the slightest possibility of any correction or evidence to the contrary.
- “expected… with a deviation above average of between .43C and .71C… midpoint … .57C… would make it the hottest year on record” – first, this is a guess at the future. So far, the US is having record lows (just Google “us record lows” and see what you get) and low temperatures not seen in 20 years and last year, the UK was forecast last year for its worst winter in 50, 60, or 100 years depending on which headline you read. (It actually was really cold last winter; I have pictures of the various snows around Canary Wharf, and there were few days one could go out without thermals.) Guessing at the future can be wrong, of course – just ask anyone who has ever played the stock market or gambled. Weather is a chaotic system and scientists know that while you can predict chaotic systems to a certain extent they’re only predictable in the overall sense (i.e., weather will continue and have variations) but they underlying variances are what are so hard to get right. Refer also to point number 6 regarding random ranges and midpoints.
- “estimates are made using an average” – this makes it seem like a very sensible thing is occurring – that the data is being “flattened out” and that only the averages are being used, and compiled across three datasets at that. This is particularly spurious because it’s only three data sets. Two are from the US and one from the UK. Where are the datasets of the Russians? The Chinese? The Indians? The Australians? The South Africans? If averaging datasets is such a great thing (and it can be) then why are we using so few? Also, if these data sets represent only the time periods previously disclosed, then they’re already presenting a skewed picture and no matter how much averaging you’re doing it won’t matter.
It’s unfortunate that stories like these are passed along as reporting. We as intelligent people should have access to better and complete information that doesn’t rely on estimates that don’t exclude potentially ‘exculpatory’ data (winter data that might reverse the end result). We should have a complete record listed of how warm a year this was as compared to the last four centuries, and then according to the ice core data. We get confusing pictures about how much ice is important (if it’s at all important) and when evidence that things like ice returning is contradictory to the prevailing theory we get a change in wording about how the theory is no longer ‘warming’ but ‘changing’. Well, everything is changing. Or is it? Apparently things may have paused changing – or more accurately, changed in such a way that the end result is no real change in the last twenty years. Of course, with the state of reporting as it is, we can’t really get a good idea.
Humans have a responsibility to protect the climate, and as stewards of the earth, a responsibility to do everything in their power to make as minimal an impact on the planet as possible. Part of this is being responsible in the reporting of data and the reporting of trends. Estimates, small ranges, meaningless comparisons, and wild guesses about the future based on already spurious data is not how to do that. We need to invest in alternative energies (performing cost benefit analyses to ensure that they are in fact worthwhile), ensure that our negative externalities are minimised, and behave as responsible people (and corporations are people!) for our own sake, not for the sake of some trumped up estimations reported as ‘fact’.