MSimon wrote:Say I'm trying to find out the average change in length of copper, aluminum, iron, and wood vs temperature (for some stupid reason I need all four in series in equal length.)
Suppose I drop out the wood. And compensate for the missing wood measurement by averaging the copper, aluminum, and iron.
Am I really measuring the same thing any more? Do I still get the same average length change? Now suppose I failed to keep the humidity constant?
MSimon, it is increasingly hard to believe that you have any experience whatsoever with engineering.
Let's suppose you're building a bridge with those materials and you want to know the average expansion rate, because you want to compensate for summertime expansion, you cannot take any variables out and average over them.
But that analogy does not fit with what the scientists are doing.
First, as Vose pointed out, the records are retroactive, which is the reason so many existed in the 70s-80s. And we know that the trend still exists when you only use one continuious record:
http://www.yaleclimatemediaforum.org/pi ... ure-23.jpg
A better analogy (which is obviously stupid) is that you're building a bridge made only out of aluminum, and you want to know the overall expansion rate of said bridge during the summer. It is composed of several different alloys of aluminum, though, and your boss said you can only do the research on one of the alloys due to budget constraints. What do you do? You pick the one with the largest thermal expansion coefficient. Likewise the scientists pick the
longest records they can from the historical record (going from over 10,000 records to a few thousand).
But, of course, just as with your bridge, it is better to check each aluminum alloy (since some have amazing expansion properties), just as it is better to include proxies in with the data.
So your measurements suggest .19 C / decade warming, the combined measurements say .13 (with 90% likelihood).
You see getting the right answer with the wrong method is not real science. It is what we refer to in the sceptic trade as climate "science". A science so good that bad methods give correct results.
When you're talking about a chaotic system a "right answer" is non-existent. You are setting the goalposts so high that they're unachievable. What they can do is come up with a best estimate. I see nothing wrong with their best estimate. I suppose you can argue with Spencer about .13 C / decade. (This number is actually close to one that you reckoned, isn't it?)
And since vegetation matters in surface station measurements - is the vegetation response going to be the same on a mountainside as it is at the sea shore? In the jungle? One can hope.
GISS accounts for all relevant data at the sampling frequency we are capable of:
http://www.yaleclimatemediaforum.org/pi ... ure-23.jpg
Indeed, which is why when Spencer's satellite measurements didn't match the models, both the models and the data were analyzed in depth. It turned out in that case that the models were right. When GISS messed up some temperature measurements, that too was corrected.
There is
no evidence of scientific misconduct from the climatologist side.
But I showed once already where a meteorologist screwed up the analysis (11 days and no retraction).