On Wed, 4 Nov 2009 15:50:32 -0800 (PST), Frogwatch
wrote:
On Nov 4, 6:30 pm, John H. wrote:
On Wed, 04 Nov 2009 13:11:58 -0500, NotNow wrote:
Please read completely. Don't kill the messenger, don't give anecdotal
crap, but respond with good, solid science to refute each of the points.
http://www.commondreams.org/headlines05/0219-01.htm
Common dreams? You've got that right.
--
Loogy says:
Conservative = Good
Liberal = Bad
I agree. John H
The only "evidence" they present for it being human caused is a model,
that really is all they give there as evidence. By varying paramaters
in a model, I can make it prove ANYTHING.
Er..that's what a mathematical/statistical model is supposed to do. By
manipulating variables, you obtain different results - that's why it's
called a model. You're taking a given set of parameters and varying
them to obtain a result.
Now can you develop a model that will produce the results you want?
Certainly - it's easy enough to do if the parameters and data sets are
limited and confined to already established results.
A good exampe is Dr Michael Mann's GRL paper (the infamous "hockey
stick"), which, in one scientific coup, overturned the whole of
climate history.
It was an essentially overlaid "graph" based on past temperature
max/mins and a set of tree ring data that was tightly controlled.
Within these limited data sets, Mann purportedly used a standard
analysis methodology called Principal Component Analysis which is a
fairly standard type of evaluative tool. PCA utilizes a technique
called normalization in which all data sets are normalized within
certain parameters. What Mann did was supress the data that did not
support his theory and enhanced the data that did. It was totally
improper, unethical and unscientific.
When the data used, even as limited as it was, is normalized within
accepted parameters, the hockey stick goes away.
So my point is that you can build a model using standard techniques
that will produce a unknown result or you can build a model using
parameters outside the accepted technique pool to produce a wanted
result.