Art said:
I never argued average global temperatures don't change.
I'm sorry, I misunderstood you.
However man-made emissions of CO2 don't appear to have much if anything to do with it else why has there been no increase in global temperatures for the past 10 years?
Has there not? Well, we need to look carefully to determine if "increase in global temperatures in the last 10 years" is even a meaningful quantity.
First, to extract a trend out of data, you need to have enough data that the extracted trend is meaningful. For instance, if the mean temperature some year is 0.1K greater than the mean temperature for the previous year, one can't use these two data points and say there's been a warming trend of 10K/century. That's sounds ridiculous. Surely, 2 data points is insufficient to extract a trend. So how many data points do you need?
The condition that determines this is that the trend extracted from a data set should not change significantly by adding a small number of data points to the data set. In the above example, if we go back an extra year and find that the temperature has dropped by 0.2K since that year, it drastically changes the "trend" from +10K/cent to about -10K/cent. That's not acceptable. We need to keep including more and more data points in the data set so that additional inclusions make sufficiently small changes to the extracted slope.
Now if you look at global temperature data (
from GHCN) for the last 9 years, do a linear least squares fit and extract a slope, you get a number of +3.40K/cent. That is a huge warming trend...but is it meaningful? Adding one more year (1998, which was a really warm year) to the dataset, reduces the trend to +1.94K/cent. That's still a large, positive warming trend, but the number has changed a fair bit. If you include 1997, the trend becomes +2.49K/cent. A smaller change, but still not small enough for a good estimate of the trend. The lesson here is that one can talk about temperature changes over a decade, only if one is willing to admit the kind of errors we see above.
If we approximate the noise in the global temperatures as Gaussian, then the condition to be met for "good data" is: \Delta T =sn\gg \sigma/\sqrt{n}, where \Delta T is the change in temperature over the period spanned by the data set, s is the trend in K/yr, n is the number of years in the set and \sigma is the RMS noise in the data. From the GHCN data, I estimated the noise to be about 0.1K. So, if you are extracting a trend which is smaller than 1K/cent (or 0.01K/yr), you need a data set with n \gg 10^{2/3} \approx 5.
One can also work backwards and determine the goodness of a data set. With n=10, s=0.0194K/yr (the numbers I get from the GHCN data for the last decade), the signal to noise ratio for the extracted temperature change is roughly given by S/N\approx (s/\sigma)n^{1.5} \approx 6. At a 68% (1 standard deviation) confidence level the error bar on the data is just the inverse of this number, or about \pm 15%, which is not great, but not completely terrible either. Naturally, if you want a greater confidence in the data, the error bar increases.
So, at this confidence level, we can say that the increase in the global mean temperature over the last decade, according to GHCN data is 0.19K \pm 0.03K. That trend is significantly larger than the trend for the 21st century, which I think is about 0.5K/cent. So I'd say, with a fairly high confidence, that based on GHCN data, the global temperatures over the last decade have not only been rising, but have been rising significantly faster than the average over the last century.
For AGW to be true then isn't a rise in temperatures a fundamental requirement?
Actually, it's not.
Evo said:
Talking about a single data point is meaningless. As shown above, even a decade worth of data points is only moderately useful.
Evo said:
What about the cooling oceans that they just discovered that was a complete surprise? They said the oceans were getting warmer, but after actually testing, found they're getting colder.
I don't think that's accurate. If I'm not mistaken, sea surface temperatures have been increasing as expected, but deep sea temperatures in recent years have dropped due to the slow migration of Arctic deep waters resulting from increased melting of polar ice.
Art said:
How would this settle the issue? If temperatures rose then AGW proponents would claim vindication and cry too little too late and if they fell they would also claim vindication.
If I recall correctly, the last IPCC report included very specific predictions based on different scenarios for CO2 emissions. I think, at the very least, if we follow one of those scenarios and observe temperatures that are very different from their predictions, they will have to admit that there are flaws in their model.
