cc: srutherford@virginia.edu, mann@virginia.edu date: Tue, 19 Sep 2000 10:11:30 -0400 from: "Michael E. Mann" subject: Re: my visit to: Tim Osborn HI Tim, Thanks for your message. Yes, the time is rapidly approaching indeed, but I haven't forgotten. I will put you up at the "Red Roof Inn" for the 10 nights. Its a nice enough hotel, but most importantly its right in the heart of the U.Va campus, and walkable to just about everything you'll need (even if you have a car, will be nice to be in the center of things I think, and on-campus parking is impossible, so being walking distance from our department will help out alot). Will have reservations made for you for the night of the 10th through 19th, checking out morning of the 20th... I'm cc'ing your message to Scott Rutherford (a research scientist working w/ me at U.Va) who will be working closely w/ us. This will be important, since I will be tied up w/ teaching responsibilities a good deal of the time, but you and Scott should be able to make progress while I'm tied up. You mention three very worthwhile topics. Items I and II in your list are worth doing, but they're a bit retrospective, particularly, because we've moved on to a significantly revised methodology, which we're in the testing stages with, and can show you some intriguing results from (as we speak, Scott is writing them up). Scott isn't working w/ my old code anymore so its become a bit of a relic. Would be preferable to do something based on the revised methodology (which makes use of regularization scheme employing a ridge regression technique, to avoid having to use a truncated EOF basis approximation of the data covariance). The topic of model/data comparison (iii) seems particularly worthwhile to me, because it would build on current efforts we have w/ the GFDL group, and efforts which you may have w/ the Hadley Centre crowd, etc? Two specific possibilities are: 1) Development of appropriate synthetic "pseudo proxy" networks. These would be derived from sparse gridpoints of model fields, and can be constructed to have a variety of systematic biases, additive noise levels, etc. which would be symptomatic of the perceived or established noise and biases in different types of proxy data. Since the model fields and effects of noise on the faithfulness of "paleoclimate reconstructions" could be estimated directly (be We are currently working w/ the GFDL model, but it would be wonderful if we could come up w/ synthetic datasets (perhaps a few different ones actually), which we might be generally happy enough that similar experiments could be done w/ the Hadley Centre model, etc. This would be very useful for comparison of methods, but would have very general usefulness... 2) Development of a multi-stage wavelet-based regression method for paleoclimate/instrumental calibration to calibrate different timescales of variability distinctly. I think this has application both to the "multiproxy" case and the tree-ring case. The idea is to only use those proxies which have a resonable level of power in a given frequency band, as predictors for the reconstructions in that frequency band, and then sum the wavelet reconstructions for a full reconstruction. Again, the idea would be to test on model data before applying to proxy data... Do either of these specific projects excite you? I think either would make for a very nice ongoing collaboration, and would pave some interesting new directions for the field too. Scott will be up in Rhode Island for the first few days of your visit, but I'm hoping he can join us by Thursday October 12th. Let me know what you think. Thanks, mike At 01:56 PM 9/19/00 +0100, you wrote: >Hi Mike, > >Hope all's well with you. My visit to Virginia is *rapidly* approaching, >so arrangements (practical & non-work) need to be made. > >My air tickets have just arrived, here are the details: >Tuesday 10th October: Depart Norwich 0620, Arrive Washington Dulles 1410. >Friday 20th October: Depart Washington Dulles 1805, Arrive Norwich 0940(+1 >day). > >I shall pick up a hire car at Dulles and drive to Charlottesville (I'll >keep the car for the duration). > >Have you been able to arrange accomodation for me (for the 10 nights)? >Could you let me know details/directions etc.? Cheers. > >I'll have about 7 and a half working days (I guess I'll have to leave >around midday on Friday to get to Dulles by the check-in time), which >should allow some useful work to be done. I have some ideas about >collaborative work that should prove useful, along the lines of (i) >comparison of methods (perhaps using your reconstruction approach with our >tree-ring-density-only data set, and comparing results with our >spatial/regional reconstructions); (ii) comparison of your warm-season >reconstructions with ours; (iii) development of methods for comparing model >output with palaeoclimate reconstructions. Before going into detail, do >these appeal to you? And do you have any additional or alternative >suggestions? > >I'm looking forward to my visit - it should prove useful for us all I hope. > >Best regards > >Tim > >PS. I'm not sure how things are organised over there, but I'd be happy to >present a seminar during my visit (on palaeo or non-palaeo topics). > > > >Dr Timothy J Osborn | phone: +44 1603 592089 >Senior Research Associate | fax: +44 1603 507784 >Climatic Research Unit | e-mail: t.osborn@uea.ac.uk >School of Environmental Sciences | web-site: >University of East Anglia __________| http://www.cru.uea.ac.uk/~timo/ >Norwich NR4 7TJ | sunclock: >UK | http://www.cru.uea.ac.uk/~timo/sunclock.htm > > > _______________________________________________________________________ Professor Michael E. Mann Department of Environmental Sciences, Clark Hall University of Virginia Charlottesville, VA 22903 _______________________________________________________________________ e-mail: mann@virginia.edu Phone: (804) 924-7770 FAX: (804) 982-2137 http://www.evsc.virginia.edu/faculty/people/mann.html