Join the conversation at Hot Topic

It appears that Ian Wishart is back on the climate beat, with a couple of posts in the last week attacking Hot Topic. One goes so far as to accuse me of incompetence and dishonesty, which is a bit rich coming from someone who was threatening to sue me for libel a few months ago… Anyway, his latest offering attempts to chastise me for stating in a comment that global temperatures were not falling. That gives me a welcome opportunity to post on the subject and introduce a nifty little gadget programmed by a Hot Topic reader. Here’s Wishart:

Virtually all the major datasets are now acknowledging atmospheric warming has slowed to a crawl or stopped over the past ten years, and even some leading climate alarmists scientists are publicly suggesting we’ve entered a climate shift and may not see warming return for a further decade or more. The data clearly shows temperature anomalies trending down despite CO2 emissions rising:

He appends a graph of UAH monthly temperature anomalies from 2002 to some point earlier this year, with a descending trend line. Lo and behold, “cooling”!

Over seven years, this data set suggests that global temperatures are falling. The problem, however, is that climate is not measured over seven year periods, and there’s a very good reason why. The climate system is ’noisy’ — natural variations in the ocean/atmosphere system such as the El Niño Southern Oscillation or volcanic eruptions mean that global temperatures go up and down a fair bit, so if you want to know about any underlying trends you have to look at a period of time long enough for the ’ups’ and ’downs’ to cancel out. That’s usually taken to be 30 years — the World Meteorological Association’s standard period for defining climate statistics.

Now for the nifty gadget. Hot Topic reader Colin Sharples (aka CTG) is a whizz with Javaâ„¢ programming. Prompted by a discussion on climate trends at ReaClimate, he took NASA’s GISS global temperature dataset (to be precise, it’s the GISTEMP series for the climatological year (Dec-Nov) and he’s estimated a figure for 2009 by ’completing’ the year by calculating the average anomalies for the last few months of the year over the last five years) and the Hadley Centre’s HadCRUTv3 series and produced an interactive graphic that shows how changing the length of the period you select for trend calculations affects the trends you see. Here it is:

The slider at the bottom starts at 30 years — the standard period for climate stats. Choose between data series with the button at the top. Blue lines show negative trends, red ones positive trends, calculated for each successive 30 year period. In the GISTEMP series you can see some falling trends in the 1880s and 90s, and again from the 1940s to 70s. The rest of the time, temperatures are rising. No sign of cooling in the 21st century.

Now try moving the slider to 20 years. There’s a bit more blue in the early years, but things look pretty similar to the 30 year trends. At 15 years, a couple of brief ’warmings’ appear in the previously cooling sections, but the most notable feature is that some of the trend lines become a lot steeper — there is faster warming and cooling appearing in the graph. With the slider at 10 years, not only do the trend slopes become even steeper (look at the ’warming’ in the 1890s!), but bits of spurious warming and cooling appear. There are a couple of blue lines in the late 1980s to mid-90s, the first episodes of ’cooling’ to show up over the last 40 years, but still no cooling to be seen in the last ten years. You have to wind the slider back to a seven year period before you get a cooling trend line in the 21st century. Just the one. And there’s still only one if you stick the slider at the minimum period of five years.

If you look at the big picture, using such a short period means that the trends oscillate up and down as global temperatures move and up and down. They’re reflecting the noise in the system, not the underlying trend — and when we’re discussing global warming we’re interested in the long term trend, not the ups and downs. Woodfortrees also provides a clear demonstration of the dangers of playing with short term trends here, using the UAH satellite temperature data.

That’s the central fallacy of Wishart’s assertion that warming has ’stopped or slowed to a crawl’ despite CO2 continuing to rise. It’s a claim you can only make by choosing very short timescales, and ignores the fact that the change we’re interested in emerges over periods of decades, not year to year.

The rest of Wishart’s long post is tortuous and unclear. He appears to be arguing that because one paper has found that aerosols and dust have an impact on the rate of ocean warming in the North Atlantic that this somehow downplays the effect of greenhouse gas forced warming. He says, for instance, that ocean warming is ’mainly solar driven depending on atmospheric clarity’. But all warming is ’solar driven’ — that’s where the heat’s coming from. Dust, aerosols and greenhouse gases change the rate at which it’s absorbed by the system. And by the way, the last time I looked at a map, the North Atlantic was only a small part of the global ocean, and can’t be said to be typical of the whole.

Even his conclusion is muddled:

Get real, Truffle. The atmospheric temperature anomaly is trending down, the oceans are not displaying significant signs of heating, and some of your cheerleaders are admitting it’s not so hot right now. I’m sure warming will return, but this is good evidence of strong natural cycles at play.

So warming’s going to return because it will overcome those strong natural cycles? Sounds reasonable. Perhaps Wishart’s been following the monthly temperature figures at NOAA, which has August 2009 as the warmest in the record, and September as the second warmest. Looks like the warming’s already back. Because, of course, it never went away.

My thanks to Colin for allowing me to embed his GISTEMP Java applet. It may also appear in other places and in enhanced form in due course, but remember where you saw it first… ;-)