date: Fri May 16 10:39:23 2003 from: Mike Hulme subject: Re: Fwd: RE: HadRM3 paper and data to: Phil Jones Thanks Phil - this seems a bit of a fiasco (what is the literal meaning of that word anyway?)! I guess it means Dave is even gladder BADC are now handling data downloads. Mike At 09:19 16/05/2003 +0100, you wrote: Mike, You might be interested in these emails as they relate to HadRM3 and UKCIP. Also I see you are supervising an U/G on Scottish snow and the NAO - I'm the second marker. We've completed the reworking of monthly temperature series for the Scottish Mainland, NW/N islands and N. Ireland. There will be a paper submitted to IJC in the next few months and a report to SNIFFER. Better for the student to use these data than those in the old SNIFFER report. We've homogenized 8 long series (inc. Braemar) to do this. Cheers Phil Date: Thu, 15 May 2003 16:56:25 +0100 From: "Jenkins, Geoff" Subject: RE: HadRM3 paper and data To: 'Phil Jones' Cc: "Jones, Richard" , "Murphy, James" , "Mitchell, John FB (Chief Scientist)" X-Mailer: Internet Mail Service (5.5.2653.19) Phil I will leave Richard to answer most of tr email, but in item 5 you ask 5. You say UKCIP is OK, yet in a few months LINK will be distributing the new HadRM3H data, which will be different from that available through UKCIP02. We have meetings next week related to the EPSRC projects on impacts of future climate change over the UK. Here, we are developing tailored scenarios for different sectors. Should we base these on what we have here (i.e the HadRM3H used in UKCIP02) or should we use the new HadRM3H runs ? Some quick advice here would be useful as there are meetings on Monday and Thursday - both involving UKCIP/EPSRC and Stakeholders. The answer is: use HadRM3H as in UKCIP02. New HadRM3H run data is within the ensemble of UKCIP02 HadRM3H data over the UK, so there is no point using the new RM3H run. Any old-new differences pale into insignificance compared to the biggest uncertainty; that of the driving GCM. Cheers Geoff > -----Original Message----- > From: Phil Jones [SMTP:p.jones@uea.ac.uk] > Sent: 15 May 2003 15:36 > To: Richard Jones; d.viner@uea.ac.uk; jhc@dmi.dk; c.goodess@uea.ac.uk; > j.palutikof@uea.ac.uk > Cc: john.f.mitchell@metoffice.com; dave.griggs@metoffice.com; > geoff.jenkins@metoffice.com; james.murphy@metoffice.com; > simon.brown@metoffice.com; richard.jones@metoffice.com; > david.hassell@metoffice.com; dave.rowell@metoffice.com; > erasmo.buonomo@metoffice.com; david.hein@metoffice.com; > jonathan.gregory@metoffice.com; cath.senior@metoffice.com; > a.moberg@uea.ac.uk > Subject: Re: HadRM3 paper and data > > > Dear Richard, > Since calling you and Geoff we've had some more discussions in CRU, > > and a meeting > might be necessary to sort a few points out. In this email I'm going to > summarise briefly > a few points from the phonecalls, but also the concerns we have in CRU > and the implications > for some of our EU projects. > > 1. Your reason for rerunning HadAM3H/HadRM3H over Europe is to have the > > integrations > compatible with the PRECIS runs in other parts of the world. > > 2. I've talked to David Viner and he will get into contact with Bryan > Lawrence at BADC to > see how quickly they might be able to begin downloading the necessary > new > integrations. > Some have already been completed and most of the rest will be by June > 15, > when one > of T3Es is turned off. > > 3. I'm happy for Anders to rerun the programs for our specific paper. > Anders probably won't > be, but it seems the text will not have to changed that much as you > expect the results to be > little changed. I would appreciate comments on the current version when > you have some time. > With this in mind, though, it seems that you've not fully thought > through > the implications. If, > as you say, you expect the results to change little, how then has this > undermined our > justification for doing the study ? > > 4. The paper with Anders isn't the only paper affected. A number of > people have been working > with HadRM3H data for upwards of a year and have drafts of papers which > are near submittal > stage. One is for a project that has only 2 weeks to run, with the final > > report near completion. > Some in CRU read your email,as, all these studies need rerunning with > the > new HadRM3H > data - except possibly if they relate just to the UK. Is this a correct > reading of your (the HC) > intentions? This particular project is looking at storm tracks over the > > UK and another is > looking at extreme precipitation variation again across the UK. The > latter is part of an EU > project - can we go with our UK analyses ? - yet we have to tell our EU > partners who have > been doing similar work in Switzerland, Iberia and Germany that they > must > rerun their > analyses? It doesn't seem consistent. > > 5. You say UKCIP is OK, yet in a few months LINK will be distributing > the > new HadRM3H > data, which will be different from that available through UKCIP02. We > have meetings next > week related to the EPSRC projects on impacts of future climate change > over the UK. Here, > we are developing tailored scenarios for different sectors. Should we > base these on what we > have here (i.e the HadRM3H used in UKCIP02) or should we use the new > HadRM3H runs ? > Some quick advice here would be useful as there are meetings on Monday > and Thursday - > both involving UKCIP/EPSRC and Stakeholders. > > 6. The implications of rerunning much of the above work will be onerous, > > but they can be > achieved. Much more important though is the implications for the > PRUDENCE/STARDEX/MICE > projects. Here, the aims of the projects were to intercompare RCMs, > intercompare > statistical and dynamical downscaling and intercompare various impacts > sectors with > RCM and statistically downscaled inputs - all also with RCM downscaling > with > various models forced with near-perfect boundary conditions from NCEP. > Although you say > the alterations to HadRM3H and HadAM3H will not have major influences in > > Europe, the > whole experimental design of the many intercomparisons has been > compromised. The > modelling groups are supposed to continue with the boundary conditions > they have yet the > STARDEX and MICE groups and the impacts work in PRUDENCE are supposed to > > rerun > all their work with the new integrations. This seems, to all of us in > CRU, to have completely > scupperred the whole set-up of the three projects. Jens, would probably > > add that this was > already compromised by the different resolution of boundary conditions > supplied to the other > groups. > > 7. There is a simple way around all this - well to me anyway. Can the > current HadAM3H > and HadRM3H data we have within LINK and the various EU projects have be > > designated > HadAM3HU/HadRM3HU (the U referring to as used in UKCIP02) and the newer > versions > HadAM3HP/HadRM3HP (the P referring to PRECIS)? As you've already looked > > at how the > new runs compare with the old ones, can they for Europe be considered as > > a larger > ensemble, going from 3 to 6 for the A2 scenario and 1 to 2 for B2. If > the > differences between > the different integrations are within the ensemble noise this would seem > > possible. In the > paper with Anders the inter-ensemble variability seemed much lower than > I > would expect from > two independent 30-year observational periods. In other words, the > combination of two > versions of the models would give more realistic estimates of > within-model noise (natural > climate variability - call it what you like). > > > I've gone on for far too long, but I hope you've got a feel for our > concerns - and the need for > some quick responses on a few of the issues. There are a lot of people > and a lot of groups > around Europe involved in some form of HadRM3H analyses and a few less > for HadAM3H. > > Cheers > Phil > > > > > At 12:10 15/05/03 +0100, Richard Jones wrote: > >Dear Phil, > > Sorry for the long delay in replying to this but it has > >generated a lot of internal discussion on related issues which we have > >now resolved. On the paper itself we agree that it is an interesting and > >useful investigation into the performance of the RCM. However, we are > >going to have to ask for a fundamental, though hopefully not > >time-consuming, revision before submission. In brief, the reason for > >this is that we have undermined one of your main justifications for the > >paper in that we have upgraded HadAM3H/RM3H and are currently rerunning > >all the experiments. As a result (and the reason for the wide > >distribution of this email) we are going to withdraw all the data from > >the old experiments currently resident at LINK (except those relevant to > >UKCIP) and replace them. On the specific issue of your and Anders' paper > >this implies that you will need to regenerate your figures though our > >assessment of the new model is that little will change in respect of the > >paper's message (hence my prediction that the revision will not take > >long). > > OK so those are the headlines, now for the detail. HadAM3H was > >originally built as a model to overcome shortcomings in HadCM3 with > >respect to generating RCM predictions for UKCIP though with the > >important proviso that it was a model which performed globally as well > >as or better than HadCM3 (so we did not undermine its credibility for > >simulating climate or predicting climate change). Clearly, the European > >dimension of the experiments was also important. Another dimension was > >its use over other regions, specifically southern Africa and India and > >then, somewhat less critical at the time, as a source of boundary > >conditions globally (i.e. for PRECIS). In the meantime, our analysis of > >the model and its use over different regions has encouraged us to > >reformulate certain aspects of the cloud and precipitation physics which > >provide further improvements in surface climatology globally. The main > >motiviation here is now PRECIS with current users in India, China and > >Africa all using the new model (the configuration of HadRM3H follows > >directly from that of HadAM3H). Thus, given that PRECIS is to provide > >the functionality for producing consistent high resolution climate > >scenarios globally we felt it necessary to regenerate our initial set of > >European experiments using the PRECIS RCM (i.e. HadRM3H). These are the > >new data which we will supply to LINK and thus will form the basis for > >European climate scenario generation from the current Hadley Centre RCM. > > Clearly there are a series of implications resulting from these > >developments, not least for your paper, which hopefully I will address > >below. These are mostly of common interest, hence the general email > >which I felt was the best way of disseminating this information, but are > >individually addressed to those most directly involved. I will start > >with the simplest first (and try and be brief). > > > >1) Dave, please ensure that no more of the existing HadRM3H data are > >released, similarly for HadAM3H for which no more extraction from our > >archives is necessary. We will need to discuss offline how to get the > >new data to you. > > > >2) Jens, the implications for Prudence I think can be split into > >categories. First is the use of HadAM3H to drive the other RCMs in > >Prudence and second is the use of HadRM3H within Prudence. On the first > >point there is no change, we are completing extraction of the third set > >of A2 6 hourly data for distribution. On the second point, the existing > >Hadley Centre RCM experiments in Prudence will be retained for the basic > >intercomparison work within WP2. However, for the use of daily data in > >the impacts areas of Prudence, we will provide only the new RCM data as > >this will then be consistent with others uses of the data in Europe via > >LINK and other uses of HadRM3H worldwide via PRECIS. All 50km > >experiments will be completed next month so this should imply little > >delay in getting data to the central archive. > > > >3) Clare and Jean, the implications for Stardex and Mice follow from the > >comments to Jens I think. Clare, your recent conversation with Simon > >Brown implied that little concerted use of HadAM3H/RM3H data has been > >made in Stardex. Any initial set-up of software etc. to handle the data > >will be immediately applicable to data from the new experiments and we > >will do all we can to get the relevant data to you so as not to > >compromise your milestones. As the new experiments are all being written > >into MASS (our new mass-storage system) and the teething problems with > >MASS have been resolved then this should not involve major delays. > > > > I am sorry if this comes as a bit of a shock to you and means > >disruption to your work. I am sure there are implications which we have > >not considered or are glossed over in my brief comments above. If you > >would like to discuss this in more detail then please phone me, today > >and next week I will only be available on my mobile (07855 822104) > >though I am in the office tomorrow and after that back in on Tuesday > >27th. > > Best wishes, > > Richard. > >-- > >Dr. Richard Jones Regional climate change research manager > >richard.jones@metoffice.com [1]http://www.metoffice.com > >Telephone: +44 (0)1344 856418 Fax: +44 (0)1344 854898 > >Mail: Met Office Hadley Centre, London Road, Bracknell, RG12 2SY, UK. > >________________________________________________________________________ > >| > >| Dear James and Richard, > >| Anders is keen to submit the paper and wants to move onto some > >|aspects of changes in extremes in the observational data, whilst he is > >|still a CRU employee. Will you be able to send any comments in the next > >|couple of weeks or are you happy with us submitting the paper ? I gave > John > >|a copy in Nice but haven't heard anything from him. > >| > >| Cheers > >| Phil > >| > >| > >| > >| Dear James and Richard, > >| I talked to John Mitchell at the EGS in Nice the other week and > gave > >|him a copy of this paper. He suggested I should send it to the two of > you > >|for any comments you might be able to make. This request stems from our > >|agreement to send you copies of papers before submission (our meeting > this > >|time last year here, when Richard, John and others came). > >| Anders and I are keen to submit this in the next month or so to > >|Climate Dynamics as their colour costs seem reasonable. Anders email is > >|a.moberg@uea.ac.uk, and if you have any comments can you cc them to him > as > >|well as me. > >| Have a good Easter - pity the hot spell isn't going to last ! > >| > >| 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 > >|------------------------------------------------------------------------ > ---- > >| > >| > >| > >| > > 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 > -------------------------------------------------------------------------- > -- > 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 ----------------------------------------------------------------------------