[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: 19990503: Hi from the Canadian Meteorological Center (fwd)


===============================================================================
Robb Kambic                                Unidata Program Center
Software Engineer III                      Univ. Corp for Atmospheric Research
address@hidden             WWW: http://www.unidata.ucar.edu/
===============================================================================

---------- Forwarded message ----------
Date: Thu, 04 Nov 1999 10:41:36 -0700
From: Linda Miller <address@hidden>
To: address@hidden, address@hidden
Subject: Re: 19990503: Hi from the Canadian Meteorological Center (fwd)

Robb,

Here are the questions and answers that I received from Richard Hogue, Canadian
Met Center.  

Unfortunately, the answers are embedded with the questions that were asked, but
the information should be there.  I've provided lots of URLs in subsequent
communications with him on the LDM, etc.

Llinda

---------- Forwarded Message ----------
Date: Tue, May 4, 1999 12:48 PM +0000
From: Richard Hogue <address@hidden>
To: Linda Miller <address@hidden>
Subject: Re: 19990503: Hi from the Canadian Meteorological Center

Hi Linda,
Thanks for responding so quickly to my e-mail yesterday.

Most requests for GEM model data specifically from US Universities have
been for non-real time data (archive requests) for the purpose of
comparative studies with other models (ETA, MAPS, MM5, etc.).  These
requests have been made in the context of the GCIP project.  Examples:
Hugo Berbery (U of Maryland), Laura Hinkleman (Penn. State U).    

Concerning actual real-time model data for a US University, the only
current collaborative arrangement is with University of Washington
(Cliff Mass).  Actually, in that case, data is made available in
quasi-real time which means with a 4-6 hour time delay from its
operational availability.

1- What do you consider "quasi-real time"? : 4-6 hours time delay from
operational availability
2- What is the typical lag time for the data sets? This is function of
the requirement but typically our experience with Canadian and Europeen
universities is that a 4-6 hour time lag is OK for their needs. 
Essentially, there has been historical issues of ensuring that real time
operational model outputs was not available free of charge to external
users.
3- What method have you employed in making the GEM data available, e.g.,
FTP? : Yes, FTP is the way to either push the data to the user or the
user can come and pick it up from our site.  We have other means of
relaying the data to our internal users, such as satellite links but
these are not appropriate to specific point-to-point external users such
as Universities wanting GRIB or BUFR data.  Generic products (i.e.
charts ) are available from our public WEB page
(http://www.cmc.ec.gc.ca/) with a time delay.

4- Finally, can you give us some idea of the frequency of the model
data, the approximate area the model data covers and the volume of the
model data? :  The heart of the grid of the operational GEM model covers
North America at a horizontal resolution currently of 24km.  Direct
output is 2-3GB for each 48h forecast model run at 00Z and 12Z with 3-h
temporal output resolution.  From that raw output, everything is
configurable function of the needs.  This issue is similar to the type
of issue faced with the ETA model (i.e. different datasets are available
at different resolution to limit data volume and to try to meet bandwith
limitations...).  Typically, we have reduced datasets that we package to
meet general needs which amount to 1-2MB/6h timesteps (i.e. about
10MB/model run).  Smaller datasets are available to meet bandwith
limitations.

5- Based on your comments about the NWP archived outputs in GRIB and
BUFR, it seems that we can assume the GEM data is in BUFR or GRIB .  : 
Yes, that is the international data format in which we typically ship
the data.


Essentially, our goal is to ease access to our model data to
Universities for educational and research needs whithin an understanding
of non-redistribution of the data.  It is our understanding that the
Unidata program has similar goals in a well managed and controlled
context.  Some Canadian Universities have been subscribing to the
Unidata program (McGill U., U of Toronto, etc.) and giving them the
possibility to receive the GEM output as well would be beneficial for
their program as well.

Thanks for taking the time to examine and evaluate our general request
of collaboration,
Best Regards,
Richard.


Linda Miller wrote:
> 
> Hi Richard,
> 
> Please see below.
> 
> --On Mon, May 3, 1999 8:54 AM -0600 Unidata Support
> <address@hidden> wrote:
> 
> >> To: address@hidden
> >> From: Richard Hogue <address@hidden>
> >> Subject: Hi from the Canadian Meteorological Center
> >> Organization: .
> >> Keywords: 199905031216.GAA26672
> >
> > Dear Colleague,
> > As Canada's national Meteorological Center (http://www.cmc.ec.gc.ca) we
> > have been running operational NWP models since the 1960's.   In the past
> > few years we have had an increased interest from the University
> > community both in Canada and in the United States to access our
> > operational NWP outputs in quasi-real time for Research and educational
> > purposes.  In an effort to facilitate access to our operational data for
> > those specific uses it was suggested to us a number of times to examine
> > the possibility for Universities to access our NWP outputs (GEM regional
> > model) through your Unidata Program.
> >
> > We are already engaged in a number of projects with US universities and
> > Research Centers.  For instance, in the case of our participation to the
> > GCIP project (http://www.scd.ucar.edu/dss/pub/gcip/index.html) we have
> > been sending our NWP archived outputs in GRIB and BUFR format to NCAR
> > for the past 2 years.
> 
> You mentioned above that you are collaborating with several US universities.
> 
> - Have you been providing real-time model data to certain universities,
> based on a collaborative arrangement?
> - If so, could you tell us which ones?
> - What do you consider "quasi-real time"?
> - What is the typical lag time for the data sets?
> - What method have you employed in making the GEM data available, e.g., FTP?
> 
> Finally, can you give us some idea of the frequency of the model data, the
> approximate area the model data covers and the volume of the model data?
> 
> Based on your comments about the NWP archived outputs in GRIB and BUFR, it
> seems that we can assume the GEM data is in BUFR or GRIB.
> 
> >
> > Therefore, would you kindly put me in contact with the right person
> > within the Unidata Program with who I could discuss the possibility of
> > adding our GEM regional model outputs to your Unidata feed to
> > Universities.
> >
> Sorry about all of the questions, but your response will help us in
> evaluating your request.
> 
> Thanks so much for contacting us.
> 
> Linda Miller
> 
> 
> > Thank you in advance,
> >
> > Best Regards,
> >
> > Richard Hogue,
> > Chief, Implementation and Operational Services Division,
> > Canadian Metetorological Center
> > 2121 North Service Road                  Tel     : (514) 421-4662
> > Trans Canada Highway, suite 400          Fax     : (514) 421-4679
> > Dorval, Quebec, H9P 1J3                  address@hidden
> > Canada
> 
> Linda Miller
> External Liaison, Unidata
> University Corporation for Atmospheric Research
> P.O. Box 3000
> Boulder, CO 80307-3000
> 303 497-8646 fax: 303-497-8690
> URL:  http:/www.unidata.ucar.edu/staff/lmiller/un.act.html

---------- End Forwarded Message ----------



Linda Miller - address@hidden
External Liaison, Unidata
University Corporation for Atmospheric Research
P.O. Box 3000
Boulder, CO 80307-3000
303 497-8646 fax: 303-497-8690
URL:  http://www.unidata.ucar.edu/staff/lmiller/un.act.html