On 04/12/2011 13:33, Terry Fields wrote:

I am answering Terry's post in two parts, the easy part first, so don't
think I have just ignored part of the message because I haven't.

How on earth do you come to such a conclusion? The models will use the
best data available and quite clearly the best temperature data
available are the direct measurements, not derived data that doesn't
match the recorded temperatures even in the limited area in which the
trees ring data was sourced.

You can not discard data, or sideline it with a reference to 'issues'
(the Wikipedia article you quoted from). You cannot just use 'the best
data'. Data has to be incorporated into the main data set, that's how
science works, and if the model can't cope with that, the model is
inadequate.

So if you had two rulers and one measured 12 inches to the foot and the
other only 11 inches you would average the two and come up with 11.5
inches? Of course not but that is what you appear to be saying. In the
subject under discussion we have a temperature series derived from tree
ring data and we have the actual temperatures as directly measured. That
the tree ring data are anomalous is interesting in itself but it has
absolutely no relevance to the temperatures to be used in a model for
the period since those temperatures have been directly measured. It is
not a matter of an inadequate model. This divergent tree ring data is of
no more relevant to a model of the climate than the colour of your eyes.
Neither play any part in it.

What the divergence has done is cast a certain amount of doubt on the
principles underlining tree ring data and/or the methodology used which
isn't good news for periods where direct temperature measurements are
not available.

--
Roger Chapman