From the BBC…
Climate: Cherries are not the only fruit
Just about the most predictable event of the week was the tempest of opinion created by the analysis of global temperature changes published in the Proceedings of the National Academy of Sciences (PNAS) on Monday.
As we (and a number of other mainstream news outlets) reported, Robert Kaufmann and colleagues analysed the impact of growing coal use, particularly in China, and the cooling effect of the sulphate aerosol particles emitted into the atmosphere.
They concluded that with a bit of help from changes in solar output and natural climatic cycles such as the El Nino Southern Oscillation (ENSO), the growth in the volume of aerosols being pumped up power station chimneys was probably enough to block the warming effect of rising greenhouse gas emissions over the period 1998-2008.
For some commentators, such as the UK Daily Mail’s Christopher Brooker, this was further proof that the “climate scaremongers” had got it wrong…
Cherry in the pie
One thing that everyone in the climate blogosphere seems to agree on is that the best fruit in the world is the cherry, judging by the number that are picked.
And the Kaufmann paper has brought a few more down from the tree.
The Global Warming Policy Foundation (GPWF), the UK-based pressure group, said researchers “tweak an out-of-date computer model and cherry-pick the outcome to get their desired result”.
To which the opponents’ rejoinder is, and long had been: “well, choosing 1998 as the baseline is cherry-picking, to start with”.
To illustrate the point, I’ve been through a quick exercise using the approach that groups such as GPWF favour – and that Kaufmann’s research group adopted – of using annual temperatures rather than any kind of smoothed average, and looking for the temperature change over a decade.
I took the record of global temperatures maintained by Nasa’s Goddard Institute for Space Studies (GISS) which is one of the three main global datasets, and calculated the rate of change over each of the most recent 10 decades – ie, 1991-2001, 1992-2002, and so on up to 2000-2010.
I’ve summarised the results in a table on this page. What it basically shows is two things:
- the numbers vary quite a bit from year to year; and
- all but one give a temperature rise – the only one that shows a small drop being 1998-2008.
Seeing as it’s logically impossible that the world warmed between 1997 and 2007, cooled between 1998 and 2008, and warmed again from 1999 to 2009, one conclusion you might reach is that using annual temperatures is not a sensible thing to do as it gives you a set of answers that does not make sense.
… which is why most scientists use the running mean approach.
Mr. Black seems to be suggesting that Figure 3 from Kaufmann is a cherry being picked by climate realists…
While he thinks that GISTEMP is the “tree”…
Well, I say that Mr. Black is Goldilocks-picking. Mr. Black asserts that it is cherry-picking to use 1998 as a starting point and that the starting point must be 1880. What’s so special about 1880 (apart from it being the start of the instrumental record)?
First off, let’s have a look at a few “cherries.”
Here is the HadCRUT3 global temperature anomaly (GTA) for 1977-2010 plotted with the GTA for 1911-1944…
Here’s the HadCRUT3 Northern Hemisphere temperature anomaly for 1976-2010 plotted with a non-carbonated interval from the Medieval Warm Period (Moberg et al., 2005)…
In both examples, the slopes are statistically indistinguishable.
The 66-yr period from 1944-2010 is pretty well indistinguishable from the first 66 years of three different century-scale cool-warm-cool cycles from Moberg’s Medieval Warm Period reconstruction…
The peak of the Modern Warming is, at most, 0.1 to 0.2°C warmer than the peaks of three comparable, non-carbonated, intervals of the Medieval Warm Period, consistent with a net climate sensitivity of ~0.5°C. However, that difference is probably not statistically meaningful.
- The error bars of all of the data sets are greater than the differences between them.
- The proxy data show the MWP to be warmer than the late 20th century.
- The proxies invariably have a lower resolution than the instrumental data; thus the amplitude of the proxy time series is attenuated relative to the instrumental record.
This means that the late 20th century warming might have been slightly warmer than the peak of the MWP. Almost all of the potential error is in the direction of magnifying the warmth of the Modern Warming relative to the MWP, so the odds are that the modern warming is very comparable to the Medieval Warm Period.
Since Mr. Black would probably say that the Medieval Warm Period is another “cherry,” let’s go back another 1,000 years, or so.
What happens if I project the polynomial trend-line a few hundred years into the future?
It starts looking like a cyclical pattern doesn’t it?
One of the “problems” with the way climate data are handled is in the obsession with applying linear trend lines to non-linear data.
A Sine wave has no secular trend…
But… What happens if my data represent only a portion of a Sine wave pattern?
The r-squared of a linear trend line of this partial Sine wave is 0.88… 88% of the data fit the trend line. This implies a very strong secular trend; yet, we know that in reality Sine waves do not have secular trends.
If we take the entire HadCRUT3 series and apply a linear trend line, we get an apparent secular trend…
The r-squared is 0.55… 55% of the data fit the secular trend. This implies that there is a real long-term warming trend.
What happens to that secular trend if we expand our time series like we did with the Sine wave?
The apparent secular trend vanishes in a puff of mathematics…
How can such a clear secular trend vanish like that? The answer is easy. Each “up hill” and each “down hill” leg of a Sine wave has a very strong secular trend. Unless you have enough data to see several cycles, you don’t know if you are looking at a long-term trend or an incomplete cycle.
Using the GISP2 ice core data from central Greenland we can see that over the last 50,000 years, there have been statistically significant warming trends…
And there have been cooling trends of varying statistical significance…
What does all of this mean?
It means that the Earth’s climate is cyclical.
If you look very closely at the GISP2 core, you’ll find that the millennial scale cycle had a period of ~1,470 (+/150 yrs) years during the Pleistocene. Many people call this the Dansgaard-Oeschger cycle. This cycle appears to have sped up in the Holocene (~1,000-yr period +/-300 yrs). If you look at Moberg et. al., 2005, Ljungqvist, 2009 or even Mann et al., 2008, you can see that the last 2-3 cycles of this oscillation have had periods of ~1,000 years… Alternating 400-500 year periods of warming and cooling.
Embedded on that millennial-scale cycle is a multi-decadal or century-scale oscillation with a period of ~60 years. The most recent warming leg of that ~60-yr oscillation ran from ca 1976 to ca 2003. 1998 just happened to be a monster El Nino near the peak of both the millennial and multi-decadal oscillations. This is why it is the warmest year “on record” everywhere outside of NASA & NOAA.
The ~60-yr and ~1,000-yr cycles that are obvious in the HadcRUT3 data and on Ljungqvist’s reconstruction is also clearly present in the GISP2 ice core throughout the Holocene.
This means that the climate changes we’ve experienced over the last 150 years are not anomalous in any way, shape, fashion or form. And it means that the Mr. Black and the other warmists must “Goldilocks-Pick” their data. Too short of a time series yields no warming trend and too long of a time series also yields no warming trend. The time series must be “just right” in order to show an anomalous warming trend.