Date: Sun, 7 Dec 2003 16:44:34 -0600
From: "John C. Ries" 
To: Martine Feissel-Vernier ,
    Ramesh Govind ,
    Pascal Willis ,
    Laurent Soudarin ,
    Jean-Michel Lemoine ,
    Suryia Tatevian ,
    Serguei Kuzin 
Cc: Gilles Tavernier ,
    Jean-Jacques Valette ,
    Jean-Pierre Granier ,
    Carey Noll ,
    Edouard Gaulué , feissel@ensg.ign.fr
Subject: Re: Participation in the IDS-GRACE Campaign

At 10:13 AM +0100 12/3/03, Martine Feissel-Vernier wrote:
>Dear colleagues
>
>We had a meeting with the IDS Central Bureau to prepare our contributions
>to the 2003 Analysis campaign: products collection, comparisons,
>evaluations. Our discussions led us to clarify the following aspects that
>we ask you to take into account.
>
>- Exact version of gravity fields considered:
>   . the version of GRIM5 to be used is GRIM5-C1
>   . the version of GGM01 to be used may be either GGM01C or GGM01S.
>
>- SAA effect on the Jason data: Laurent Soudarin transmitted to me the
>   following list of stations that are affected and should be deweighted or
>   not considered in the analyses. He originally obtained it from Pascal
>   Willis. The stations are the following.
>
>         KRUB  AREB  ASDB  CACB  EASB  HBKB  HELB
>         LIBB  SANB  GALA  PAQB  MAHB  PDMB  TRIB
>
>   This list can be further discussed with other analysts if necessary.

Hi Everyone,

I probably won't have the time to participate in the analysis 
campaign other than through some limited orbit comparisons. But that 
doesn't stop me from offering some comments on the proposed analysis. 
:-)

I strongly recommend the use of GGM01S rather than GGM01C. Depending 
on your background models (esp. ocean tides) you may find better 
results with EIGEN-GRACE01S or GGM01S, particularly for the sun-synch 
satellites. The EIGEN solution is very good even in our own tests. 
Hopefully, good results are obtained with either of these GRACE 
models.

I also do not see any point in including Jason in any reference frame 
solution. Eliminating the SAA stations does NOT remove the indirect 
effect from the CNES frequency polynomial. As I showed at the last 
SWT, all the non-SAA stations are currently affected at the ~5 cm 
level in height (that's almost 3 cm/yr height drift error), and it 
will only get worse.  You can recover this with the proposed global 
bias drift parameter, but that simply weakens the global scale of the 
network as determined by Jason. Between tossing a third of the 
network and weakening the remainder (or worse, living with the 
biases), Jason brings no benefit to this analysis. Unless people will 
do two solutions, with and without Jason, I think the effort will not 
lead to any meaningful results. It will be too sensitive to the 
handling and editing of the Jason data.

Best regards,
John

ps. For those who didn't already receive word, the POD/GEOID splinter 
presentations from the Jason SWT are available at

ftp://ftp.csr.utexas.edu/outgoing/jason/SWT_03_POD_GEOID

This includes the available presentations about the SAA effect.
-- 


John C. Ries                      Senior Research Scientist

Mail Address:                                   
  Center for Space Research        Phone: (512) 471-7486      
  Mail Code R1000                  Fax:   (512) 471-3570     
  The Univ. of Texas at Austin     E-mail: ries@csr.utexas.edu
  Austin, Tx  78712   USA          WWW: http://www.csr.utexas.edu 

Street Address:
   University of Texas at Austin
   Center for Space Research
   3925 W. Braker Lane, Suite 200
   Austin, Texas   78759-5321  USA