The “Global Warming Policy Foundation” has released a GWPF Background Paper by Benny Peiser.
It begins by stating what they claim to agree with the scientific establishment about on the subject:
A. Matters where we agree with the dominant scientific establishment and can quantify the outcome
1. The greenhouse effect is real and CO2 is a greenhouse gas.
2. CO2 has increased in the atmosphere from approximately 0.029% to 0.039% over the past 50 years.
3. CO2′s greenhouse warming potential follows a logarithmic curve with diminishing returns to higher concentrations.
4. Absent feedbacks, and other things being equal, a doubling of carbon dioxide from pre-industrial levels would warm the atmosphere by approximately 1.1C.
5. Since 1980 global temperatures have increased at an average rate of about 0.1C per decade. This is significantly slower than forecast by the vast majority of GCMs.
I’ll agree with points 1 through 4. But what’s their explanation for point 5?
There are 5 major estimates of global average temperature, 3 for the surface, 2 for the lower troposphere (the lower layer of earth’s atmosphere). All make their data publicly available. We can use the data to estimate the warming rates for each record, as well as the uncertainty in those estimates (rates were estimated by linear regression, standard errors using an ARMA(1,1) model of the noise in monthly data). Here are the rates for each data source, together with error bars extending 2 standard deviations above and below the estimate:
None of the data sources indicates a warming rate of 0.1 C/decade as claimed by the GWPF, they all indicate a faster rate. For the three surface-temperature data sets, the GWPF claim is just plain statistically rejected.
So what’s their basis for that claim?
I have an idea. Suppose instead of fitting a line to the data (by least-squares regression or any other method), you used annual averages, then simply computed the slope of a line from the first annual average to the last one. This would give you a different estimate which we could call the “point-to-point” estimate, one which is far more uncertain than the linear-regression method. It would also give you a lower estimate, because the 1980 value is above the trend line while the 2013 value (or 2012 if you don’t want to use incomplete years) is below the trend line. It’s similar to the trick used here.
We can also compute the point-to-point estimated rates as well as their uncertainties. For instance, using data from NASA GISS the linear regression trend estimate is 0.16 deg.C/decade, but the point-to-point estimate is only 0.1 deg.C/decade:
In fact we can plot these estimates, together with error bars 2 standard deviations above and below, side by side with the linear regression estimates:
I suspect that the point-to-point estimates form their basis for the claim that “Since 1980 global temperatures have increased at an average rate of about 0.1C per decade.”
In my opinion, their claim falls far short of an honest portrayal, very far short, and is the kind of failure to engage honestly which makes it impossible to have a rational discussion about global warming.
Perhaps Benny Peiser will soon be giving testimony at a Congressional hearing on the subject.