cc: "Clare Goodess" <C.Goodess@uea.ac.uk>, <c.harpham@uea.ac.uk>
date: Fri, 21 Dec 2007 12:05:14 -0000
from: "Jake Hacker" <jake.hacker@arup.com>
subject: RE: FW: Invitation to Tender - Design Summer Guidance for    London
to: "Phil Jones" <p.jones@uea.ac.uk>

   Hi Phil,



   Yes, great news about the EPSRC proposal.  It looked a bit iffy when we got the referees
   reports, but they clearly saw the light.  I thought it was a strong proposal and definitely
   the strongest team!



   On the GLA proposal that all sounds great.



   For the sites I'm thinking we need two, maybe three: a central urban, a peri-urban, and
   something in between, to make allowance for UHI.   Depends a bit though how much extra work
   that will mean for you and if you have time to do it.



   I think though that the return period analysis should be just done for one site, and the
   same year(s) used for the other sites.



   Heathrow - what we have at the moment - is something in between.  LWC is the obvious choice
   for the central site. I'm not sure yet on the peri-urban site.



   We have the hourly BADC data files here used for our last project with the GLA up to 2006,
   although have only looked at temperature.  We had approval to use this for the last
   project, but we (Arup) may have to ask for permission from the Met Office to use them -
   they're a bit funny about us using the BADC data as we're 'commercial'.



   The sites we have are:



   Longest records:

   Heathrow and LWC, 1975-2006 (pretty complete, at least for temperature).



   Shorter records, with some missing data:

   St James' Park: 1996-2006 (this is central urban park, we guess, but haven't been able to
   find out exactly where the station is!)

   Beaufort Park: 1988-2002

   Wisley:  2001-06

   Gatwick: 1971-1988



   As far as I know, that is all the hourly weather data for the London area that BADC/UKMO
   hold.



   Talk more in the New Year.  I think I'll be back on the 3rd.



   Have a great Christmas.



   Best wishes,

   Jake


     ______________________________________________________________________________________

   From: Phil Jones [mailto:p.jones@uea.ac.uk]
   Sent: 21 December 2007 10:20
   To: Jake Hacker
   Cc: Clare Goodess; c.harpham@uea.ac.uk
   Subject: RE: FW: Invitation to Tender - Design Summer Guidance for London

    Jake,
       Good news re the EPSRC project!
     Have Chris in some advisory role would be useful.
     Happy to be as sub-consultants to you.
    Fine re IPR.
    On the DSYs presumably we'll just be doing Heathrow, or are other London
    sites (say LWC) being considered?
     Happy for you to do option A. Colin Harpham will be downloading the most
    up-to-date data for Heathrow and other London sites. This will be daily, but
    we can get what we can for hourly - through BADC - for Heathrow. If LWC
    has hourly then we'll get that. Most of the other sites don't have hourly
    except Gatwick.  Reattached Heathrow, so Colin has the details of what
    needs to be in the file(s).
     Option A for the future, may end up with something mixing 2003 and 2006/7.
    Getting up to date hourly data is an issue. I don't think that 1975/76 for Heathrow
    are available hourly, so availability may dictate choice somewhat.
     Option B would essentially be based on the WG for UKCIP08, but the
    same type of return period analysis as A .
      I probably won't be checking email after this coming Sunday am, until
    maybe New Year's Eve/Day. Should be back in on Jan 2.
    Cheers
    Phil
   At 17:41 20/12/2007, Jake Hacker wrote:

     Phil,

     That's good to hear that you'd like to go in with us.  I'm happy to co-ordinate the
     proposal and to take on the general project management/admin issues.   I think the
     consortium would just be us, maybe with some formal advisory input from Chris Kilsby.

     On contracts, we can sort this out later on if and when we're appointed.  For the last
     project we did with Glenn for GLA, King's acted as subconsultants to Arup, and the
     previous one we had two separate contracts with GLA; it's largely a matter of who
     carries the risk.  On the IPR issues, all the background IPR (such as the WG) would rest
     with relevant parties under standard GLA contracts.  We have also negotiated a clause
     previously that all methodological IPR generated on the project should rest with the
     consultants not the GLA, only the results.

     On the DSYs - they are composed of hourly data. I've attached the current one for London
     (you may already have this).   There are not 14 CIBSE sites across the UK (since 2006)
     and the data was produced by Geoff Levermore .  As far as I understand, this was done
     using BADC for the synoptic variables (including wind, I think) and the solar was
     synthesized from the cloud cover data using Tariq Muneer's algorithms. Prior to 2006,
     there were only three CIBSE weather year sites, and the solar was from observations.

     At the moment, I'm thinking that the diagnostic for summer warmth will be a combination
     of hours over a discomfort temperature threshold (e.g. 28C) and number of degrees over
     the threshold - a cooling degree hour type concept although the weighting might not be
     linear.  For the adaptative comfort theory CIBSE are advocating, the discomfort
     temperatures change from day to day according to the running mean outdoor temperature.
     So we'll need hourly data or at least an approximation of hourly data for the summer
     warmth diagnostic/probability analysis.

     For the future years/change factors, I think GLA would be happy if we take a
     deterministic UKCIP02 approach.  I don't think they're envisaging that the probabilistic
     nature of UKCIP08 needs to be incorporated until after UKCIP08 comes out; maybe informed
     by the EPSRC projects.  In a way, this project is a 'stop gap' until UKCIP08 is
     finalised and digested.

     On the scope, I think there are two ways of approaching it:

     OPTION A - using only observations:  estimate 'return' periods from historical data
     (e.g. last 30 years at the site) aauming stationarity; select 'present day' DSYs based
     on return periods, synthesising solar from cloud cover as in current DSYs; morph current
     DSYs under UKCIP02 (which supposes morphed years have same 'return periods' for the
     future timeslices.

     OPTION B - using WGs:  return period analysis and present-day DSYs based on WGs;  future
     DSYs determined in similar way using the WG, with deterministic change factors based on
     UKCIP08 for 2020s and 2050s.

     In either approach, I guess there's no way around assuming stationarity?

     I'd like us to pursue both options, as the first is 'safe', but a little pedestrian,
     whereas the second in more innovative and extendable to UKCIP08; and if both pan out ok
     as we expect we can compare and cross-validate.

     I'm quite interested to pursue Option A at Arup.  Would that be ok?  CRU would obviously
     need to do Option B.

     Let me know if you think that sounds ok.

     Jake
       ___________________________________________________________________________________

     From: Phil Jones [[1] mailto:p.jones@uea.ac.uk]
     Sent: 19 December 2007 14:03
     To: Jake Hacker
     Cc: Clare Goodess
     Subject: RE: FW: Invitation to Tender - Design Summer Guidance for London
      Jake,
         I don't think we want to submit our own bid, so would like to join your
      consortium. So we do need to think about project management and
      contractual issues.
         The WG within UKCIP08 is essentially up and running now, or very soon.
      We have to calculate a load of results for measures like Heating/Cooling degree
      days through the WG for the report - across the whole country. These will be
      quite time consuming, so we can't be changing the WG (at least the
      daily part) from what we have now. The PMG of UKCIP08 have to set a
      cut-off date fairly soon to begin this work - probably at the next PMG on
      Jan 17.
          Thanks for the reasons for wind direction and pressure data being required.
      Observationally, we can get these quite easily from gridded pressure data,
      from Reanalysis products in fact. These are on a 2.5 by 2.5 degree grid. This
      is fine for pressure - easy to show. For wind direction we can calculate the
      large-scale wind direction (at this grid scale) from the pressure data,
      which should be fine. This could then be added to the station observations for
      the other variables.
         I am assuming a few things
     1. DSYs are just based on daily average data. This would include Tx, Tn, but
      for other variables it is an average or a total for the day (the 09-09day). Wind
      direction would only vary from day to day.
      2. Selecting a DSY is based on say temperature. The rest of the variables
      are those that were recorded for that choice. This way you get a 'real' sequence
      that occurred. So we can add pressure and wind direction for the choice.
      If different sites across London were needed, most of the variables would be
      similar (the choice being year or season Y), but if we go with the way of
      getting Pressure and wind direction, then these wouldn't vary much spatially.
      3. If hourly were wanted we could use our hourly WG to disaggregate the
      daily DSY sequence to hourly. This is what UKCIP08 will be doing.  Just
      daily would be simplest.
      How to perturb this is the problem. One possibility is to take the perturbations
      from UKCIP02 - and select a season (or months) from those that have occurred
      to get the future DSY. If the DSY for now was the 3rd warmest year/season
      then the future one might be just like the warmest one.  With UKCIP08 you'd
      be able to do this for more possibilities. I think in all this the emphasis has
      to be on real and generated sequences. For the GLA, they will get extremish
      and very extremish 'real' sequences.
      Cheers
      Phil
     At 11:26 19/12/2007, Jake Hacker wrote:

     Phil,

     Thanks for your thoughts.

     I guess a key issues is whether one needs a WG to do this project or not.  I'm not sure
     one does - the aim is to redefine the basis for the DSY selection; potentially that can
     be done using obs, as for the existing procedures.  I also agree that WGs are the
     future, but using the WGs to actually produce the DSYs is an additional step. With
     regard to UKCIP08 - can we fully compatible with UKCIP08 prior to March, or is
     additional work needed later in the year?

     With regard to wind direction and pressure:  Wind direction is used to calculate (along
     with wind speed) wind pressures on facades; this affects infiltration and natural
     ventilation rates.  There are situations where naturally ventilated buildings are
     actually purposefully orientated (also streets) or have different sized ventilation
     openings on different facades to take account of prevailing wind directions (rightly or
     wrongly).  Atmos p is used in psycrometric calculations (dry bulb-wet bulb-moisture
     content relationships) - very important for designing and simulating air conditioning
     systems; the pressure affects are generally minor in the calcs but some engineers might
     object to having to use a standard reference value.

     Can you let me know how you want to take this forward? Do we want to form a consortium?
     - if we do we need to think about the project management and contractual structures - or
     do you want to submit your own tender?

     Best wishes,
     Jake
       ___________________________________________________________________________________

     From: Phil Jones [ [2]mailto:p.jones@uea.ac.uk]
     Sent: 18 December 2007 16:25
     To: Jake Hacker; Clare Goodess
     Subject: Re: FW: Invitation to Tender - Design Summer Guidance for London
      Jake,
         Both Clare and I received this overnight. Clare is in Singapore this week and
      will be back on Jan 4 - as you'll have seen from her automatic response system.
         She did email this morning and I gave her the deadline for submission and
      a few other things as she wasn't able to open the files.
         Some thoughts on your thoughts
      1. We will have downloaded from BADC all the weather data we can get for
      a number of London sites - by the first week of January. We will be getting/updating
      Heathrow, Gatwick, LWC, St James's Park, Wisley, High Beach and
      Rothamsted. This is all for another project(s) - Tyndall/SCORCHIO, so we will
      have what there is. We will be getting all the variables listed in the tender.
      2. We will get MSLP and wind direction, although the latter isn't often measured.
      Not sure why MSLP is needed. If it is to use Muneer's formulae for getting
      direct/diffuse it makes no difference. We have his formulae within the UKCIP08 WG
      and have verified that they work well against the only two long-term series
      of diffuse and direct radiation we could find in the BADC archives
      (for Hemsby and Eskdalemuir).  Formula uses sunshine and wet bulb temp.
      3. We both got the tender here as you did. I suspect Geoff will have got it
      and maybe Chris Kilsby and the Met Office also. There is no need to
      involve Glenn McGregor - he's leaving for Auckland in March, so ought to say no,
      even if asked.
      4. Useful to involve Chris Kilsby. He's not been responding to email in the
      last few days. Not sure what Stephen Belcher will do?
      5. CIBSE are correct that WGs are the future - because of UKCIP08. We have the
      UKCIP08 WG (at least for daily) finalised and not much to do on the hourly part.
      What we could do is use this, but perturb it with UKCIP02 factors instead of
      the latter ones. This doesn't make much difference - we will probably have to use
      the CRU rainfall generator as opposed to the Newcastle one to do this.
      WGs have to be the future - as there is no way to use the MOHC probabilistic
      projections otherwise.
      6. Your nagging in the back of your mind! The WG is conditioned (fit) to observed
      weather data at the site (in UKCIP08 we extrapolate between sites to get every
      5 by 5km squares as well). We then perturb the parameters based on the MOHC work
      to get future sequences. We need multiple generated sequences as there will be
      a range of possible futures from the MOHC pdfs (for temp/precip). Users define
      what they want in the way of futures (time period, emissions and rareness - 90th
      percentile).
      We're not using pdfs of observed weather data, but the observed data itself (to fit the
     WG). This
      may be a moot point, but all the data are better than the pdf. It also gives you
      sequences. I've been saying for a while that it would be possible to generate
      loads of sequences similar to the observed one - which could enable you to
      get better return period estimates for all manner of variables. This is, IF,
      you believed the climate was stationary!  In the pre- climate change days
      this is what a lot of hydrologists did.
      7. We're not doing wind direction. This is because the MOHC models don't do it
      very well, it isn't that related to the other weather variables that well, it
      isn't that well observed (at least in the BADC files) and as the models don't do
      it that well, the future is uncertain (it isn't in the MOHC variables). Finally, I
     don't think it
      will make a blind bit of difference to any building model either. I know
      people think they need it, but have they ever tried running the model
      without it - or just having an uncorrelated direction shifting gradually.  Wind
      direction is unlikely to change much in the future.
      8. My idea of developing the TRYs and DSYs would be to use the WG for 1)
      the observed, 2) the WG simulating the observed and 3) the WG simulating
      the future.  If you came up with a way of selection - say the 3rd warmest year
      in the observed, then you could mimic this in the WG runs. Say there were
      30 observed years, you then take the 3rd warmest - maybe combining different
      warm winters, springs etc .  With the WG you generate a number of 30 year
      runs and do the same thing. You can then compare how the building model
      responds to the observed TRY/DSY and the possible WG (based on observed) TRYs/DSYs.
      Selecting (how?) one of these from the WG is the only issue.
      Then you do a similar thing with the perturbed future WG simulations.
       The key thing in this is to show that the TRYs/DSYs (weather years) got from
      the observed data are no different from the same things got from the WG (observed)
     runs.
      This ought to give confidence in the approach with future (perturbed) WG runs.
      Just realised that there seems no way you can do this project without the WG.
      The only way you can make this compatible with UKCIP08 is to use a WG.
      So written in this WG way, it ought to be impossible for anyone without a
      WG to compete. All you can do without a WG is to modify the observed data
      in some way - and that doesn't seem that good these days.
      I'll be here the rest of the week, if you want to iterate on this further.
      Finally, only the output data (the TRYs and DSYs) will be the intellectual property
      of the GLA - not the WG. That IPR is essentially ours, but also DEFRA's who've
      funded the UKCIP08 work.
      Cheers
      Phil

     At 14:40 18/12/2007, Jake Hacker wrote:

     Phil - Clare,

     This is the spec for the Weather Years projects for the GLA.

     There are five work elements/outputs:

     1. diagnostic for summer warmth (this will likely relate in some way to 30-day running
     mean outdoor temperatures, the basis of the adaptive thermal comfort model)
     2. development of a probabilistic basis for selection of weather years
     3. selection of weather station(s) for London
     4. production of weather years
     5. production of climate change adjusted weather years

     I envisage we would take the lead on output 1.  Output 3 builds on the work Arup/King's
     did earlier in the year (which Phil reviewed), so I have some thoughts on this.

     I'm envisaging you would take the lead on outputs 2 and 4, and also 5 if we're not just
     morphing.  Something that is not included in the spec is a testing phase, where we check
     that the weather files give sensible results when run through a building model, and
     against the preconceptions forming the basis of the 'diagnostic of warmth' - we can do
     this.

     The tenders need to be in by 11th Jan, which gives a little time after Christmas, but it
     would be good if we can fix ideas before the hols.

     The project needs to be completed by 31st of March, which gives a tight timeframe. But
     we might be able to get away with leaving a few loose ends to be tied up after we get
     paid.

     There is no guide price and I don't know what they have in mind, but I would guess more
     than 20k and less than 50k is what they are looking at.

     Can you also let me know if you think it would be good to involve anybody other than
     ourselves. e.g.
     - Chris Kilsby - for output 2; he offered to contribute in an advisory capacity when we
     spoke about earlier in the year
     - Tariq Muneer - we could possibly involve him as a consultant on the solar data for 4
     and 5 if he is willing, which might go down well with CIBSE, but up to you if you think
     a good thing or not
     - Stephen Belcher & Glenn McGregor - we could involve them on the UHI side of things,
     but not necessary.

     I don't know who else has been asked to tender.  I suspect Geoff Levermore will have.
     Maybe also UKMO.  We would likely be well placed to win the work if the costs are ok.

     On the methodology, a couple of things:

     - we will need wind direction and atmospheric pressure for the weather years; does that
     affect decision to use observed years or WG output for the weather years themselves?
     - I'm a bit nervous about using the WG to create the weather years, as it's an
     additional methodological departure from current practice; that said, CIBSE are keen on
     WGs and see them as the future
     - on output 3, the probabilistic selection' methodology, we thought we would use long
     timeseries WG runs.  Just something nagging at the back of my mind - if the WG is
     preconditioned on an observed pdf for the site (?), does it provide additional
     information for the return period analysis, or could that have been done using the
     original pdf without employing the WG?

     All the best and hope to have caught you before you break up for Christmas,

     Jake

       ___________________________________________________________________________________

     From: Andrew Tucker [ [3]mailto:Andrew.Tucker@london.gov.uk]
     Sent: 17 December 2007 18:49
     Cc: Alex Nickson; Andrew Tucker
     Subject: Invitation to Tender - Design Summer Guidance for London
     Dear Colleague,
     On behalf of Alex Nickson at the Greater London Authority, please find attached an
     Invitation to Tender for the Development of Design Summer Guidance for London study.
     This invitation email has been sent to you as an invitee, and is not to be passed onto
     others unless part of your tender bid consortium.
     This email should contain the following documents;
     - Project Specification
     - Form of Tender
     - GLA's Access to Information Policy
     - Invitation to Quote Letter
     - Pricing Schedule
     - Procurement - Diversity Monitoring Form
     The Project's Terms and Conditions document will be sent to you over the next few days.
     Until Friday the 21st of December, I will be the point of contact for the project.  Alex
     Nickson will be Project's lead contact after the Christmas  break.
     <<DSY GLA's Access to information policy.doc>> <<DSY Invitation to quote letter.doc>>
     <<DSY Pricing Schedule.doc>> <<DSY Procurement-diversity monitoring form.doc>> <<DSY
     project spec.doc>> <<DSY Form of Tender.doc>>
     Best Regards
     Andrew
     Andrew Tucker
     London Climate Change Partnership Manager
     Greater London Authority
     City Hall
     Queens Walk
     London SE1 2AA
     Ph: 020 7983 4679
     M: 07766347776

     ABOLITION OF THE SLAVE TRADE BICENTENNIAL 2007 - IGNORED NO MORE

     GREATERLONDONAUTHORITY
     EMAIL NOTICE:
     The information in this email may contain confidential or privileged materials. Please
     read the full email notice at  [4]http://www.london.gov.uk/email-notice.jsp
____________________________________________________________
Electronic mail messages entering and leaving Arup  business
systems are scanned for acceptability of content and viruses

     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
   ----------------------------------------------------------------------------

