Mime-Version: 1.0
Message-Id: <p06020409c187d87289a8@[129.116.190.133]>
In-Reply-To: <6.2.1.2.2.20061117175415.027671a8@pop.cls.fr>
References: <200611151125.kAFBP2q01543@cnes.fr>
 <6.2.1.2.2.20061117175415.027671a8@pop.cls.fr>
Date: Mon, 20 Nov 2006 16:24:08 -0600
To: ids.analysis.forum@cls.fr, Jean-Paul Berthias <jean-paul.berthias@cnes.fr>
From: "John C. Ries" <ries@csr.utexas.edu>
Subject: Re: [ids.analysis.forum] Scaling of DORIS solution in ITRF 2005
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit

At 6:20 PM +0100 11/17/06, Jean-Jacques Valette wrote:
>Dear Jean-Paul and dear colleagues,
>
>To complete your point on operational consideration and new DORIS 
>stations it seems interesting to add a comparison between the 
>scaling (+ scaling derive?) of ITRF2005 and the differences between 
>ITRF2005/IDS and ITRF2005/IGS stations velocities.
>
>For Toulouse (TOUL & TLSA), the difference is 1.1 mm/yr in amplitude 
>that is to say about 7 mm since 2000. For Easter Island (ISPA-EISL & 
>EASA) it is about 3 mm/yr or nearly 2 cm since 2000. Those stations 
>have a long standing operation and we can expect a difference 
>probably higher with recent stations. The order of magnitude of such 
>station coordinate errors (mainly horizontal) is at least similar 
>and probably higher than the one resulting from the scaling.
>
>Another approach of the investigation you request for could be the 
>final altimeter user approach. Why not making some CAL/VAL cy tests 
>at epochs separated by 6 months but over a 10 years period? Would 
>those tests significant enough to evaluate improvements on 
>geographically correlated errors, and long terms derives? Would it 
>be relevant, profitable and reasonnable effort?
>
>JJacques
>

Hi Everyone,

In response to Jean-Jacques' inquiry, I note that I had reprocessed 
the first 90 cycles both with ITRF2005 DORIS coordinates and again 
with the DORIS coordinates scaled by 1.2 ppb. This introduced a 
height difference in the two networks of about 7 mm. The effect on 
the orbit was completely negligible. The maximum radial difference I 
saw was under 0.5 mm (it was usually less than 0.1 mm) and the rms 
difference was less than 0.1 mm. Even multiplying these results by 3 
leads to insignificant orbit effects.

This is consistent with expectations...the more dynamical orbit 
approaches are very insensitive to a scaling of the network when that 
network is fairly well distributed geographically. For sparse 
networks, like SLR, the radius of the orbit does not change 
significantly, but the orbit does get distorted to some degree as the 
height biases 'tug' differentially on the orbit.

I imagine the reduced-dynamics approaches will be a little more sensitive.

Best regards,
JR


                           Cliquez sur l'url suivante 
https://www.mailcontrol.com/sr/vFRBwFvyflhrZanXRfpfu!tNbWdd5FpRNcmAzoaWpOs3WxjQ9U4hrYzv!zL22A5sOtV0esVFFcwzjn+hDGiI9zDihwMC5EWFpdgmKyf1ViP!pEehyW+OFnNusuMDxZWDXRt7yMaqVIltym+1ut!P5XKGkkYwwkEclFQb4gdnl7pTrYkZY9r8dF3YTitAaiQRndBtJRTESKkRYH0wRrjfudMhExsGfFrC  
                    si ce message est indésirable (pourriel).