cc: "Tim Osborn" date: Wed, 20 May 2009 08:46:46 +0100 (BST) from: "Tim Osborn" subject: Re: TMP: cruts3 extended to 2008 to: "Ian Harris" Hi again -- are you sure you've used the latest CRUTS3 that runs up to end of 2008, that I'm using. And are you using the netCDF 1901-2008 versions too, that were made on 8th/9th May? I've only looked at the netCDF files, perhaps some discrepancy could have been introduced coverting ascii to netCDF and/or combining pre-jul-2006 with post-jun-2006 into one file? In case they're useful, I've made similar seasonal country mean comparisions between CRUTS3(ver up to jun2006) and CRUTS(ver up to dec2008) for vap, dtr, tmn, tmx and wet. They're all here: /cru/u2/f055/to/cruts/ Most show some occasional weirdness/outliers, generally no worse than tmp and pre (some are better). WET is, however, considerably different. It looks to me like the data from actual WET obs is looking good, but that the WET anomalies that have been estimated from PRE anomalies (in the absence of WET obs) and scaled wrongly (e.g. 10 or 100 times too small?) before being combined with the other anomalies and the climatology. There are many parts of time series with reasonably looking mean levels but dramatically reduced variance (almost look like horizontal lines). It's only a guess, but seems possible. Hope that helps. I'm happy to take a look through the process with you if that might help to spot potential problems with intermediate files and hence narrow down any problems. Tim On Tue, May 19, 2009 5:51 pm, Ian Harris wrote: > Hi Tim, > > The GPCC comparisons are with the same data you're using. Of course, > we are using different regional definitions, but there still > shouldn't be major discrepancies. Again, I'll look at cases later > this week. > > Cheers for these checks - very useful!! > > Harry > > On 19 May 2009, at 16:38, Tim Osborn wrote: > >> Harry -- problems occur in the latest PRE data too -- attachment >> again shows seasonal country means. There are some big outliers >> here -- see Cuba on page 3! USA on page 4 is particularly >> worrying... the last decade is all way too high. Even though it >> USA is a big country and there are a number of years with >> erroneously high values, it might still be down to outliers (just >> lot's of them!). Cook Islands on page 25 is again interesting! >> >> Have you turned off some QC procedure when doing the last update? >> Phil showed me some plots you had made earlier with comparison to >> GPCC data that seemed to behave much better -- but were they made >> with an earlier version of CRUTS3.00? >> >> Sorry to add further problems! >> >> Tim >> >> >> >> >> At 16:15 19/05/2009, Phil Jones wrote: >> >>> Harry, >>> Hope you can easily sort out the outliers. Might they be values >>> where the flags say don't use? They seem like individual station >>> values that you could get rid of with an outlier check. >>> >>> The Cook Islands are interesting. >>> >>> Cheers >>> Phil >>> >>> >>> At 16:03 19/05/2009, Tim Osborn wrote: >>>> Hi Harry and Phil, >>>> >>>> attached are my comparisons of last year's CRU TS 3.00 seasonal >>>> country averages through to end of 2005 (I didn't use the Jan2006- >>>> Jun2006 data) in pink, with the extended CRU TS 3.00 through to >>>> end of 2008. >>>> >>>> I've just shown 1940-2008, otherwise things get too small to see >>>> clearly. >>>> >>>> There are differences throughout, but most are fairly minor. >>>> Small differences (e.g. +-0.1) might be related to my >>>> calculations, which I've done at full precision but I've only >>>> written the final country/seasonal means out with 1 decimal >>>> place. However it is likely that there are some cases that >>>> really have shifted by a small amount between the two Harry >>>> versions. Any idea why early data has changed? >>>> >>>> A number of big outliers have also been introduced. Some are in >>>> the update period but are clearly wrong. Some are in the >>>> pre-2005 period but are not in the previous CRUTS3 data from last >>>> year -- have you turned off any QC checking when making the new >>>> extended version? >>>> >>>> p.4 Guatemal SON 1995 >>>> p.5 Greenland DJF 1995/6 >>>> p.6 France/Germany/Denmark SON 2006 (these might be real?) >>>> p.10 Mauritania SON 1994 >>>> p.11 Ivory Coast (+p.12 Burkina.) SON 1999 >>>> p.12 Mali/Algeria/Niger JJA 2005 >>>> p.12 Mail (not Algeria/Niger) SON 1994 >>>> p.13 Ethiopia/Eritrea/Somalia JJA 2002 >>>> p.14 Kenya/Tanzania JJA 2002 (opposite sign to >>>> Ethiopia!) >>>> p.16 Mozambique/Zimbabwe MAM 2004 >>>> p.17 Madagascar MAM 1996 >>>> p.23 Tuvalu/Kiribati MAM 1997 >>>> p.23 Tuvalu JJA 1947 >>>> p.23 Tuvalu/Kiri./Marshall/Caro.) DJF 1995/6 >>>> p.24 Tonga SON 1994 >>>> >>>> I made anomaly maps for annual means for 2006, 2007 and 2008. >>>> They looked quite similar to the HadCRUT3 maps on Climate Monitor. >>>> >>>> Cheers >>>> >>>> Tim >>>> >>>> >>>> >>>> Dr Timothy J Osborn, Academic Fellow >>>> Climatic Research Unit >>>> School of Environmental Sciences >>>> University of East Anglia >>>> Norwich NR4 7TJ, UK >>>> >>>> e-mail: t.osborn@uea.ac.uk >>>> phone: +44 1603 592089 >>>> fax: +44 1603 507784 >>>> web: http://www.cru.uea.ac.uk/~timo/ >>>> sunclock: http://www.cru.uea.ac.uk/~timo/sunclock.htm >>> >>> 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 >>> --------------------------------------------------------------------- >>> ------- >> Dr Timothy J Osborn, Academic Fellow >> Climatic Research Unit >> School of Environmental Sciences >> University of East Anglia >> Norwich NR4 7TJ, UK >> >> e-mail: t.osborn@uea.ac.uk >> phone: +44 1603 592089 >> fax: +44 1603 507784 >> web: http://www.cru.uea.ac.uk/~timo/ >> sunclock: http://www.cru.uea.ac.uk/~timo/sunclock.htm >> > > Ian "Harry" Harris > Climatic Research Unit > School of Environmental Sciences > University of East Anglia > Norwich NR4 7TJ > United Kingdom > > > -- Dr. Tim Osborn RCUK Academic Fellow Climatic Research Unit School of Environmental Sciences University of East Anglia Norwich NR4 7TJ, UK www.cru.uea.ac.uk/~timo/