cc: "Phil Jones" date: Mon, 6 Jul 2009 17:02:48 -0600 (MDT) from: "Kevin Trenberth" subject: Re: Your visit to ECMWF and a question on the PREC/L dataset to: adrian.simmons@ecmwf.int I agree with Phil on that: poor or no documentation and none that is referencable does not help. Kevin > Kevin > > You'll see from later mails that it's a problem for PREC/L also, though > its apparently better analysis method (compared at least to that which > used to be used in the GPCC Monitoring Product) helps a bit. Also, Phil > Jones grumbles to me from time to time how difficult it is to get the > GPCC people to write up fully what they have done. He tells me he will > see you at the IPCC scoping meeting in a couple of weeks time, so > perhaps you'll be able to find a little time to discuss these matters > then. > > I think the real issue is to get the countries to send comprehensive > gauge records promptly to the data centres, not five or ten years after > the event. This is important in the climate-service context. > > Best regards > > Adrian > > > Kevin Trenberth wrote: >> Hi all >> My comment and concern is that when stations come and go it corrupts the >> record and this is especially an issue for GPCC. How that is dealt with >> whether by using anomalies or whatever is a big issue. In GPCC this was >> poorly done earlier (not taken into account) as I understand it >> Kevin >> >>> Aiguo >>> >>> I've just read the actual gauge numbers given with the PRECL data. For >>> North America there is a sharp drop in gauge numbers at the end of 1996 >>> (not 1997). But for Australia there is a dramatic drop at the end of >>> 1992 - from about 1400 gauges to about 90 gauges, and the number falls >>> after that so there are only 40-50 gauges being used in the latest >>> years. So it is not surprising that PREC/L appears to be the poorest of >>> all the datasets for Australia. I am inclined though to add it into the >>> paper, since it has been quoted in AR4 etc.. The Full Data GPCC v4 >>> product (and the forthcoming version of GPCC based on it) looks a much >>> better bet to me. >>> >>> PREC/L looks better in terms of coverage going back in time, with >>> ~13000 >>> gauges used globally for 1973, but the GPCC Full Data Product v4 has >>> ~28000 for this year. Globally, PREC/L is down to ~2000 gauges in 2007 >>> compared with ~8000 in GPCC v4. >>> >>> Best regards >>> >>> Adrian >>> >>> >>> Aiguo Dai wrote: >>>> Dear Adrian, >>>> >>>> Thanks for hosting our visit to ECMWF and we enjoyed talking to you >>>> and >>>> other colleagues there. >>>> >>>> The CPC analyses rely heavily on the NCDC GHCN2 station data for >>>> precipitation, and the GHCN2 has a smaller >>>> number of gauges since 1997. I normally use GPCP (after adjustment for >>>> mean difference) for years after >>>> 1997. >>>> >>>> Aiguo >>>> >>>> >>>> Adrian Simmons wrote: >>>>> Dear Aiguo >>>>> >>>>> It was good to talk with you and June when you visited ECMWF on >>>>> Friday. Thank you in particular for being willing to discuss matters >>>>> you raised in your review of our paper on near-surface humidity over >>>>> land. I'll amend it to make it clear that the GPCC full data product >>>>> version 4 is based on gridding anomalies, and add concluding >>>>> discussion of the evidence that leads us to conclude that the decline >>>>> in relative humidity in recent years is not dimply due to our picking >>>>> up shift in observational bias. >>>>> >>>>> After talking with you, I eventually found the PREC/L dataset on the >>>>> CPC website. I've produced versions of Table 5 and Fig 12 with it >>>>> included. Results are attached. To be honest, it does not look like >>>>> much of an improvement over CMAP, and it really seems to be the "odd >>>>> man out" when it comes to Australia. This is particularly the case >>>>> since about 1997, as can be seen in Fig. 12. Are you aware of any >>>>> particular problem with PREC/L after then, particularly with regard >>>>> to >>>>> Australia? I ask particularly because in your review you mentioned >>>>> that the gauge coverage for PREC/L may not be as good as for GPCC >>>>> after 1997. >>>>> >>>>> Best regards >>>>> >>>>> Adrian >>>> >>> -- >>> -------------------------------------------------- >>> Adrian Simmons >>> European Centre for Medium-Range Weather Forecasts >>> Shinfield Park, Reading, RG2 9AX, UK >>> Phone: +44 118 949 9700 >>> Fax: +44 118 986 9450 >>> -------------------------------------------------- >>> >> >> >> ___________________ >> Kevin Trenberth >> Climate Analysis Section, NCAR >> PO Box 3000 >> Boulder CO 80307 >> ph 303 497 1318 >> http://www.cgd.ucar.edu/cas/trenbert.html > > -- > -------------------------------------------------- > Adrian Simmons > European Centre for Medium-Range Weather Forecasts > Shinfield Park, Reading, RG2 9AX, UK > Phone: +44 118 949 9700 > Fax: +44 118 986 9450 > -------------------------------------------------- > ___________________ Kevin Trenberth Climate Analysis Section, NCAR PO Box 3000 Boulder CO 80307 ph 303 497 1318 http://www.cgd.ucar.edu/cas/trenbert.html