Global Warming Contrarians Part 2.1: Cooling Trend or Noise?12 May 2010
In Part 2 of this series of posts examining the (mostly bad) arguments against the reality of global warming, I debunked the oft-heard claim that global warming stopped in 1998. Basically, the people who make that claim are just joining endpoints, rather than looking at the trend from 1998 to now. But there is a slightly more sophisticated version of that argument, which does actually look at the trend. It goes something like:
Claim: There’s been global cooling since 2001 (or 2002), and no significant warming since 1995.
Fact: Firstly, the exact trend varies from one record to another. The NCDC record shows a slight negative trend for 2001-2009, and a more negative trend for 2002-2009. GISTEMP shows a very slight positive trend for 2001-2009, and a negative trend for 2002-2009. Both trends are decidedly negative in the HadCRUT3 record. But all three records show a positive trend for 2000-2009.
In earlier posts I’ve mainly cited the NCDC record when talking about surface temperatures; for the sake of consistency, I’ll continue to do so here. According to NCDC, the trends are +0.07°C/decade for 2000-2009, -0.03°C/decade for 2001-2009, and –0.08°C/decade for 2002-2009. So technically, we have seen about nine years of “cooling”. But is that enough to indicate a change in the long-term trend?
The temperature record is notoriously noisy. I discussed the reasons for this in Part 2. To recap, surface temperatures fluctuate from year to year due to internal variability — which basically means the exchange of heat between the atmosphere and the ocean. So a short-term cooling trend may mean nothing at all. In the long run, it’s the amount of heat entering and leaving the entire climate system that determines the average global temperature.
But just how long is the long run? It’s impossible to put an exact number on what timeframe is required to detect the underlying trend, but looking at historical trends might give us some idea. I’ve calculated the trend for each nine-year period in the NCDC record, and graphed the results. (Disclaimer: I’m no climatologist, and I know nothing about statistics, so I’m not under the delusion that my amateurish analysis below is in any way definitive. But I thought it was worth sharing anyway.)
True, 2001-2009 does stand out a bit; it’s the first nine-year period with a negative trend (–0.03°C/decade) since 1977-1985 (–0.01°C/decade). But as you can see, the nine-year trend can change very quickly, and in a way that is not indicative of the long-term trend. The last 30 nine-year trends varied from as low as –0.03°C/decade to as high as 0.36°C/decade (1975-1983), but they all average out to 0.18°C/decade. So on this timescale, even the trend can still be mostly noise.
What about a 15-year trend? Well, it’s certainly more meaningful than a nine-year trend:
So the 1995-2009 trend is also unusually low, only 0.12°C/decade. And the sudden drop in the 15-year trend looks dramatic; the 1992-2006 trend was 0.26°C/decade, the fastest in the entire record. This is probably not just noise; I’m guessing it’s because the 1992-2006 period began with the temporary cooling effect of the 1991 Mount Pinatubo eruption, and the 1995-2009 period ended with the lowest solar minimum in a century. In the long run, the continuing increase of greenhouse gas levels should overwhelm these short-term climatic effects. The slower 15-year trend probably doesn’t signal the end of global warming.
So even 15 years may not be long enough. The Copenhagen Diagnosis recommends using 25-year trends, so I’ll finish with a graph of those:
The 1985-2009 trend is 0.18°C/decade, fully consistent with continuing anthropogenic warming. In fact, it exceeds the 25-year trend at any time during the early-20th-century warming period. This graph also suggests that at the end of a warming period, the 25-year trend should drop to zero in about a decade. Well, there’s no sign that that’s about to happen.
As for the shorter-term trends, I suspect that a few years from now, the 9-year trend will be positive again and the 15-year trend will be faster. But by that time, no doubt there will be more short-term fluctuations in other climate indicators which can be used to distract the public.