This looks to be the e-mail exchange in which Mann first agrees to work with the UEA guys.
Note he’s happy to make his data, codes etc. available to Phil Jones — but certainly not to McKitrick and McIntyre several years later.
Also, note the National Science Foundation has so much money that the funding of duplicate research is not problem, according to Mann.
The e-mail is below.
###
cc: t.osborn@uea.ac.uk
date: Wed, 17 Jun 1998 12:03:13 -0400 (EDT)
from: mann@snow.geo.umass.edu
subject: Re: Something far more interesting
to: p.jones@uea.ac.uk
Dear Phil,
Of course I’ll be happy to be on board. I think the opportunity for some
direct collaboration between us (me, and you/tim/keith) is ripe, and
the plan to compare and contrast different approaches and data and
synthesize the different results is a good one. Though sidetracked
by other projects recently, I remain committed to doing this with
you guys, and to explore applications to synthetic datasets with
manufactured biases/etc remains high priority. It sounds like it
would all fit into the proposal you mention. There may be some
overlap w/proposals we will eventually submit to NSF (renewal
of our present funding), etc. by I don’t see a problem with that
in the least.
Once the collaboration is officially in place, I think that sharing
of codes, data, etc. should not be a problem. I would be happy to
make mine available, though can’t promise its the most user friendly
thing in the world.
In short, I like the idea. INclude me in, and let me know what you
need from me (cv, etc.).
cheers,
mike
____________________________________________________________________
Michael E. Mann
Adjunct Assistant Professor, Department of Geosciences
Morrill Science Center
University of Massachusetts
Amherst, MA 01003
Synthetic datasets are a method of stress testing reconstruction algorithms. Since the individuals involved wrote and published a paper on this technique in 2002, no one needed to steal their email to discover it. Furthermore, it just shows ignorance to claim that this is a smoking gun of some sort of conspiracy.
http://en.wikipedia.org/wiki/Pseudoproxy
For example, some people when they hear of a Monte Carlo technique think it convolves a result that is a random number. No, such techniques use random numbers as input that result in the non-random calculated result.
If this is the tip of the iceberg, it promises to be a whopper. Prepare thyselves for the screaming of woe from UPenn, as yet another mail is taken ‘out of context’.
“I remain committed to doing this with
you guys, and to explore applications to synthetic datasets with
manufactured biases/etc remains high priority.”