|
I saw the article in the paper on the weekend, and in particular the graph which the OP has appended to the first post.
It is IMMEDIATELY obvious when looking at the raw data that there is a step-shift downwards in 1980 - obviously, "something" changed with the way the raw data was being collected in about 1980. You have to adjust the data for this "something" before you can attempt to analyse for underlying trends.
Neither of the segments left or right of the step shows any apparent downward trend. If you instead fit a straight line by eye to each of the two segments, what do you get? Two upward trending lines which are roughly parallel with the trend-line in the homogenised data.
It looks like the person who created the graph simply fitted a linear trend-line to the complete data set, including across the step-change, in order to support a claim that temperatures are dropping. The sole basis of the claimed downward trend is a spurious step-change in the data which is being "analysed".
Pseudo-science of the worst kind!
|