date: Fri, 20 Jan 2006 15:44:05 +0000 (GMT) from: Martin Juckes subject: MITRIE -- a new draft to: mitrie -- Anders Moberg , Eduardo Zorita , hegerl@duke.edu, Jan Esper , Keith Briffa , Martin Juckes , Myles Allen , Nanne Weber , t.osborn@uea.ac.uk Hello, here is a new draft -- I've tried to incorporate comments -- please accept my apologies if something you suggested is still not in. From the new reconstructions I've done the conclusion is that the compositing approach is more robust than inverse regression. I think I need to add a little more to back up the interpretation -- this comes from looking at the regression coefficients: with inverse regression the reconstruction is dominated by a small number of proxies. cheers, martin Attachment Converted: "c:\documents and settings\tim osborn\my documents\eudora\attach\mitrie_04.pdf" Attachment Converted: "c:\documents and settings\tim osborn\my documents\eudora\attach\mitrie_04.tex"