date: Mon, 21 Apr 2008 08:56:29 +0000
from: Gerard van der Schrier <schrier@knmi.nl>
subject: Re: CRU TS3.0
to: t.osborn@uea.ac.uk

   Hi Tim,
   I'm slightly confused by the data on the ftp site. Are the .climgen files just unzipped
   versions of the .tar.gz files?
   Cheers, Gerard

Hi Gerard and Ed,

I didn't think that I'd linked my climgen data page from anywhere else
(yet) because it isn't ready for public release (yet), though I have
forwarded the URL to a few "private" users -- maybe they have added it as
a link to their websites?  May I ask where you stumbled across it?

There have been many problems with CRU TS 3.0 data files so far.  I
probably wouldn't use anything from Sep 2007 -- though perhaps mean
temperature (tmp) is ok; certainly min and max temperature (tmn and tmx)
and precipitation (pre) are rubbish and aren't salvageable by avoiding a
few regions!  If you got vapour pressure and/or wetday counts, then those
too were almost completely wrong.

However, I'm much more confident about the accuracy of the files on my
climgen website, which Harry has re-calculated over recent weeks.  At
least for data since 1950.  There's probably some problems during
1901-1950, though probably fewer that in the CRU TS 2.1 version (though
some problems are new!).  No doubt there are some problems post 1950 too,
but again there were probably more in CRU TS 2.1.  Still waiting for
vapour pressure and wetdays, and cloud cover updates are some months away.

The question is, are they ready for use?  Nobody has really used them in
anger yet.  I'm distributing them to the QUEST-GSI project, who will
probably be the first users -- they may find problems that we may correct.

If you're willing to use them with the expectation that there may be some
updates to correct certain problems (more so before 1950 I'd guess), then
I'm happy for this.  Harry sounds as if he isn't -- or at least that he
wasn't so happy for them to be distributed back in Sep 2007.  I've cc'd
Phil to see what he thinks -- also Phil may want to say how you might cite
the data set (e.g. provisional paper authors/title?).

If you do use the data, then the ascii/text files on my climgen website
are in a fairly self-explanatory format (time series for each land grid
box), though Harry has also produced netCDF files containing grids of
monthly data.

Cheers

Tim

Regarding data versions,
On Sun, April 20, 2008 10:44 am, Edward Cook wrote:


Hi Gerard,

Perhaps you already know it, but apparently the precip problem over
Bangladesh has been corrected now (Harry indicated that was in
progress) as indicated by a website description of Tim Osborn ([1]http://
[2]www.cru.uea.ac.uk/~timo/climgen/data/questgsi/). He even provides an
ftp link for the TS 3.0 data files ([3]http://www.cru.uea.ac.uk/~timo/
climgen/data/questgsi/globobs/), perhaps not thinking that I would
stumble across it. Anyway, if not already too late, perhaps you will
want to use the newest version of the TS 3.0 precip data to generate
the global scPDSI data.

Cheers,

Ed
==================================
Dr. Edward R. Cook
Doherty Senior Scholar and
Director, Tree-Ring Laboratory
Lamont-Doherty Earth Observatory
Palisades, New York 10964  USA
Email: [4]drdendro@ldeo.columbia.edu
Phone: 845-365-8618
Fax: 845-365-8152
==================================

On Apr 15, 2008, at 1:18 PM, Gerard van der Schrier wrote:


Hi Ed,

Given Harry's reply, shall I just use the data I have (version of
September 2007), and run some quality control checks?

Cheers, Gerard



Hi Gerard,

On 14 Apr 2008, at 8:38, Gerard van der Schrier wrote:

Hi Harry,

Sorry to bother you again about the CRU updated data.

Ed Cook and myself were wondering about the status of the CRU TS
3.0 data. Is it ready or not yet? We are slightly confused about
the status of this update.

Sorry about that. If it was up to me, there would be no confusion
because nobody would have seen it at all.. but in real life there
are  shifting priorities..


Some time ago, I received an ftp address for this data and I
downloaded the data (both Temp - tx, tg, tn- and Precip). But
later, I learned that CRU withdrew support for these data - so I
stopped working with these data.

Can you tell me where (which continents or which time periods)
you  suspect most in the dataset I downloaded? (downloaded it at
the end  of September last year). Or would you advise me not to
use the data  at all before there is a properly tested dataset?

I *think* temperature is OK. Precip has a few glitches, (which
I've  been investigating over the past two weeks), the major one
is  Bangladesh in the 1990s. Here, updates were added to the
database a  year or two back but were 10x too low.. I've now found
and corrected  that. Precip *currently* being regernerated (as I
type!).

So - use Temp, and Precip but not Bangladesh. Or anywhere else
that  looks odd when you examine it ;-)

Cheers

Harry



Cheers, Gerard

--
----------------------------------------------------------
Gerard van der Schrier
Royal Netherlands Meteorological Institute (KNMI)
dept. KS/KA
PO Box 201
3730 AE De Bilt
The Netherlands
[5]schrier@knmi.nl
+31-30-2206597
[6]www.knmi.nl/~schrier
----------------------------------------------------------


Ian "Harry" Harris
Climatic Research Unit
School of Environmental Sciences
University of East Anglia
Norwich NR4 7TJ
United Kingdom




--
----------------------------------------------------------
Gerard van der Schrier
Royal Netherlands Meteorological Institute (KNMI)
dept. KS/KA
PO Box 201
3730 AE De Bilt
The Netherlands
[7]schrier@knmi.nl
+31-30-2206597
[8]www.knmi.nl/~schrier
----------------------------------------------------------






--
----------------------------------------------------------
Gerard van der Schrier
Royal Netherlands Meteorological Institute (KNMI)
dept. KS/KA
PO Box 201
3730 AE De Bilt
The Netherlands
[9]schrier@knmi.nl
+31-30-2206597
[10]www.knmi.nl/~schrier
----------------------------------------------------------

