Mime-Version: 1.0
Message-Id: <p06200701bf1c716f263b@[192.168.0.4]>
Date: Sun, 7 Aug 2005 21:32:03 -0500
To: ids.analysis.forum@cls.fr
From: "John C. Ries" <ries@csr.utexas.edu>
Subject: Consider using Eccentricities for DORIS processing
Content-Type: text/plain; charset="us-ascii" ; format="flowed"

Dear Colleagues,

In DORISMail 395, Herve provided this information:

>Here is a summary of the different positions of the 2 GHz phase 
>center, on which the DORIS positionning measurements are actually 
>carried out:
>
>Time period			East offset	North offset
>26 Feb 2003 - Oct 2004		0		0		(ref. position)
>4 Oct 2004 - 29 Jun 2005	0.018 m		0.066 m		(loose foot)
>30 Jun 2005 -->			0.007 m		-0.003 m 
>	(new position)

This prompts the question regarding how to best implement this kind 
of information in the DORIS processing. I note that this is exactly 
the kind of known offsets that the use of "eccentricities" is 
designed to handle. These are common in SLR processing, so that many 
occupations of the same site are rigorously tied together without the 
need for breaks in the time series (unnecessarily weakening the 
solutions).

The alternatives are to simply live with the errors, for POD 
applications, or to create a sequence of artificial site names (MAN1, 
MAN2, etc), each with its own position vector. This inflates the set 
of network coordinates for no useful purpose (as long as the 
eccentricity is indeed know precisely).

My recommendation would be to consider the creation of a set of 
eccentricities, just as maintained for SLR, for sites where there are 
known shifts or offsets and where the data continues to have the same 
identifier. However, I don't know if the use of eccentricities is 
common for those people processing DORIS data.

We do not need to sort this out immediately, but at some point, we 
should consider what sort of recommendation would be best. I suspect 
we will need to be able to handle similar situations at other 
locations.

Best regards,
John