<?xml version="1.0" encoding="utf-8"?>
<!-- generator="Kunena 1.0.11"> -->
<rss version="0.91">
    <channel>
        <title>hirlam - Forum</title>
        <description>Kunena Site Syndication</description>
        <link>http://www.hirlam.org</link>
        <lastBuildDate>Tue, 21 May 2013 22:37:43 +0000</lastBuildDate>
        <generator>Kunena 1.0.11</generator>
        <image>
	        <url>http://www.hirlam.org/components/com_kunena/template/default_ex/images/english/emoticons/rss.gif</url>
	        <title>Powered by Kunena</title>
	        <link>http://www.hirlam.org</link>
	        <description>Kunena Site Syndication</description>
        </image>
        <item>
            <title>Subject: How to turn on DFI? (config_exp.h not enough!) - by: Trygve Aspelien</title>
            <link>http://www.hirlam.org/index.php?option=com_kunena&amp;Itemid=160&amp;func=view&amp;catid=2&amp;id=1073#1075</link>
            <description>Hi Enda

Are you using AROME? DFI might work now with DFI (possibly cy38 or even cy37???), but didn't work earlier. Is it the first cycle that you test?

I see this test in scr/submission.db, which disables DFI for the first cycle:

# No dfi the first cycle if SURFEX and ASSIMILATION
   if ( $ENV eq 'CANARI_OI_MAIN' or $ENV eq 'CANARI_EKF_SURFEX' ) {
     $complete = 1 if ($SMSNAME =~ m~Forecasting/Dfi~);
   } ;
 
Is it the logfile for DFI that has NEINI=0 or the forecast? DFI is a separate task before the forecast (even though it is possible to do it as a part of the forecast I think).</description>
            <pubDate>Tue, 21 May 2013 05:37:07 +0000</pubDate>
        </item>
        <item>
            <title>Subject: Use gl to extract a sub-domain - nointerpolation - by: Eoin Whelan</title>
            <link>http://www.hirlam.org/index.php?option=com_kunena&amp;Itemid=160&amp;func=view&amp;catid=2&amp;id=1066#1072</link>
            <description>Ulf,

taking a quick look NCEP GRIB2 documentation (http://www.nco.ncep.noaa.gov/pmb/docs/grib2/grib2_temp3-30.shtml) and UCAR GRIB1 documentation (http://rda.ucar.edu/docs/formats/grib/gribdoc/) it seems GRIB2 uses 4 octets instead of 2 octets to store the latitude and longitude of the first grid-point of a Lambert grid. 

GRIB2 would probably help with grid accuracy.

Eoin</description>
            <pubDate>Thu, 02 May 2013 08:34:49 +0000</pubDate>
        </item>
        <item>
            <title>Subject: Cleaning of HARMONIE versions at ECMWF - by: Ulf Andrae</title>
            <link>http://www.hirlam.org/index.php?option=com_kunena&amp;Itemid=160&amp;func=view&amp;catid=2&amp;id=1070#1070</link>
            <description>To all  HARMONIE users at ECMWF

There are a number of HARMONIE versions available on ecgate, from cy35 to cy38. At the moment only cy37 and cy38 have been adapted to c2a and there is no plan to spend more efforts on the older cycles. There are a number of versions
under /home/ms/spsehlam/hlam/harmonie_release/ I would like to remove from ecgate. See the list below. If you are using any of these in any way please let me know.

Ulf


 /home/ms/spsehlam/hlam/harmonie_release/ 

35h1.2
35h1.3
35h1.4
35h13
35h1branch
harmonie-35h1

trunk876xr

branches/harmonie-36h1_9197
branches/phasing/cy38
vendor/cy38</description>
            <pubDate>Thu, 02 May 2013 08:06:14 +0000</pubDate>
        </item>
    </channel>
</rss>