Quote:
Originally Posted by GreyWolf
The manipulation of the data to "hide the decline" was done in the base data set. The graph is irrelevant. The explanation later given for the use of the word "trick" smacks of post hoc rationalisation. The entire base data set has been manipulated, and almost certainly appropriately so, in order to try and establish a best-guess homogenous data set. So, the further discussion of replacing selected post-sanitization data points with the express intention of hiding a decline smacks of unethical behaviour. The director of the CRU has stepped aside while an investigation of the whole matter is on-going.
What is more troubling, is that the raw data set (pre-sanitization) is missing... thrown out. This alone makes any review of the current base data set difficult. Again, the rationalisation that it was for space considerations is shaky... microfiche is quite inexpensive, and digitized data can be stored anywhere.
I am not a climatologist, although I have perhaps a slightly more than layman's understanding of the subject. I AM an expert on computer time-series modeling. And climatologists need a modeler to be able to put the data into a format to produce a reasonable forecast. The form of the model significantly affects its predictive qualities and reliability. Univariate, multivariate, regressive, decompositive, dynamic regressive, event models... they all have their strengths, weaknesses, and idiosyncracies.
Deciding which model works best on a given data set is as much the realm of the modeler as it is the subject specialist. I stand by my assertion that I am qualified to comment on the possible effects of the disclosed data manipulation on the predictive quality of the model.
But... that's not been the point of my comments all along. I'm more concerned about the socio/political implications for science research. Scientific research must be about openness and peer review. This reveals systematic violation of that openness, and a predispositon to silence or discredit dissenting views. Healthy debate is good for science. Rejecting criticisms without addressing them, simply because they do not fit your world view, is anathema to proper research.
And I think Ace has hit the nail on the head. If anyone is asking questions that can't be answered, or cause discomfort, then something is wrong. And it's not the people asking the questions.
|
The fact that you can't even keep straight what is in the emails speaks volumes about your willingness to consider the issue. The email about the "trick" to "hide the decline" is very explicit about this:
Quote:
Once Tim’s got a diagram here we’ll send that either later
today or first thing tomorrow. I’ve just completed Mike’s Nature
trick of adding in the real temps to each series for the last 20
years (ie from 1981 onwards) amd [sic] from1961 for Keith’s to
hide the decline.
|
He talks about creating a graph from mixing data from two different series, not altering the series themselves.
And the raw dataset was thrown out in the 80s, but ALL the raw data they had is still available. They threw out their own compilation, but the data that went into that compilation is available from the original agencies.
And a forecaster, you should know enough about parameter sensitivity and linearity to know how models differ widely from one another.