cc: Keith Briffa ,i.harris@uea.ac.uk, Nanne Weber date: Fri, 19 Jan 2007 13:11:18 +0000 from: Phil Jones subject: Re: PDSI to: Gerard van der Schrier Dear Gerard, Let's discuss this more when you're here. I will be around. The PDSI paper (and the Penman part) will make a very good and well referenced paper. The whole issue of PDSI has come up in the chapter comments frequently. I reckon this paper will address all of these - so will be very useful for AR5 (presumably there will be one!). Can you use this as an argument? When I did the KNMI review some years ago the use by IPCC of references written by KNMI authors was a metric used. Whilst I don't think this is the best metric for research performance, it is widely used - the number of times we had comments (we ignored most by the way) saying reference these papers I've done, was startling! On the CRU High-Res dataset, Harry will have sorted out the problems of gaps, relaxing to zero anomaly etc which affected the Mitchell and Jones (2005) version. So I think we will be able to have real missing areas when there is no precip anywhere nearby. He's going to sort out the Sahara for example. Many of these are in areas where PDSI is likely meaningless anyway. Let's discuss this more in Feb. I'm away Jan23-Feb2 at the last IPCC meeting. I think I might know who the Dutch reviewer might be. Independently Albert has just sent me some comments from Geert van Oldenburgh. It might also be him as the comments are long the same lines. Cheers Phil At 10:19 19/01/2007, Gerard van der Schrier wrote: >Phil, > >Many thanks for the update. I haven't got a clue who the reviewer is you >complained about in the other mail. I think Albert is able to make an >informed guess, haven't discussed this with him though. > >About a new global PDSI dataset, or the "weighted" PDSI (no backtracking). >I agree that it would be very nice to have this dataset. With the Penman >parametrization. It should make a nice paper. > >The main problem is the lack of time. I have calculated the dataset I send >to Ed (and to Jurg Luterbacher: he was interested in PDSI for Northern >Africa) in the evening hours. Adding the Penmon parametrization, testing >the routines, comparing Penmon and Thornthwaite, doing the analysis and >writing the article will not be possible do in the evening. I need 3 >months for that (rough estimate). > >Nanne Weber and myself are trying to persuade the head of Albert Klein >Tank's department to find the funds for this. Nothing definitive on this >yet. Do you have any suggestions how we can twist his arm most effectively >(in case he is blind to the opportunity this provides)? > >I'll be at CRU on Feb 14th for teaching. Shall we discuss things then? > >There is also a technical concern: with the gaps replaced with, or >"relaxed" to, climatology in the CRU dataset, a possible problem in >applying the PDSI algorithm occurs (be it self-calibrating or not). Palmer >calculated "Climatically Appropriate For Existing Conditions" and PDSI >values are calculated as deviations from this. When the deviations are too >small, strange things happen. There are some other problems too, which I >haven't been able to trace yet. This results in a coverage with some >"gaps" , mainly over sparsely populated areas with rather extreme climates >as the Sahara desert or the Siberian Tundra. > >Cheers, Gerard >> >>> Gerard, >> I've been talking to Keith about PDSI and I showed him the >> figure attached. This is one from a Kevin Trenberth talk, so >> Aiguo Dai probably produced it. Don't pass this figure on. >> I talked to you earlier about the IPCC CH 3 and the comments >> we'd had on the drought section. These comments said one thing >> mainly. PDSI is no good because it calculates evap crudely. >> Harry will have a revised CRU TS soon - up to jun 2006. He will >> also have calculated Penman PET for each of the land squares. >> >> What I would want to see in a paper is the following: >> >> 1. Direct comparison of PDSI (with sc version) with Thornthwaite >> and Penman PET - to show it makes no difference. >> >> 2. Major PC patterns of scPDSI (now with Penman) >> >> 3. Like the ppt, showing the effect of temperature. Dai has >> got a bigger effect by - I think - basin the temperature on an >> earlier period (say 1951-80 as opposed to 1961-90). Is this >> effect too big? - probably >> >> Any paper will need a few new runs >> >> - with the new data >> - fixing T at some level (1951-80 or 1961-90). Latter better, but an >> earlier cooler >> level will enhance the temperature component >> - Penman instead of Thornthwaite >> - fixed T, so no temperature component (depends on the base) >> >> Perhaps we can discuss this over the coming few weeks. Harry >> should have all the data ready by the end of Feb. >> >> Need to et soils everywhere, but Keith tells me you have that and it looks >> good. >> >> Cheers >> Phil >> >> >> >> >> >>Prof. Phil Jones >>Climatic Research Unit Telephone +44 (0) 1603 592090 >>School of Environmental Sciences Fax +44 (0) 1603 507784 >>University of East Anglia >>Norwich Email p.jones@uea.ac.uk >>NR4 7TJ >>UK >>---------------------------------------------------------------------------- >> > > > >-- >---------------------------------------------------------- >Gerard van der Schrier >Royal Netherlands Meteorological Institute (KNMI) >dept. KS/CK >PO Box 201 >3730 AE De Bilt >The Netherlands >schrier@knmi.nl >+31-30-2206597 >www.knmi.nl/~schrier >---------------------------------------------------------- Prof. Phil Jones Climatic Research Unit Telephone +44 (0) 1603 592090 School of Environmental Sciences Fax +44 (0) 1603 507784 University of East Anglia Norwich Email p.jones@uea.ac.uk NR4 7TJ UK ----------------------------------------------------------------------------