Mime-Version: 1.0
Message-Id: <p0623090ec4f04de7f910@[207.42.54.35]>
In-Reply-To: 
 <2CD3B9CB2F1FD44586C120B4234B411503CE3653@cst-xch-001.cnesnet.ad.cnes.fr>
References: 
 <2CD3B9CB2F1FD44586C120B4234B411503CE3653@cst-xch-001.cnesnet.ad.cnes.fr>
Date: Fri, 12 Sep 2008 10:14:10 -0700
To: "Lemoine Jean-Michel" <jean-michel.lemoine@cnes.fr>,
        <ids.analysis.forum@cls.fr>
From: "John C. Ries" <ries@csr.utexas.edu>
Subject: Re: [ids.analysis.forum] Drift or no Drift ?
Content-Type: text/plain; charset="us-ascii" ; format="flowed"


>
>  >
>>  AGAINST:
>>  * The rates were determined over a short period (mid-2002 / 
>>mid-2006). Any error in those rates will have an increasing impact 
>>over the long term, in particular for very low degree terms ((3,1) 
>>for instance) which barely show in the map of drif> t rates but can 
>>have a noticeable impact on the orbit (because of the low-pass 
>>filter nature of orbital dynamics).
>>  * Some of the rates observed by GRACE, particularly hydrology in 
>>South America, Siberia and Canada, are likely to be inter-annual 
>>variability and not true long-term trends.
>  > * The Sumatra-Andaman event appears as a rate, whereas it is a 
>sudden event and ought to be treated as such (i.e. as a step in the 
>gravity field).

Hi Everyone,

I fully agree with Jean-Michel on this point. We know there are rates 
in the gravity field, as well as steps. But the GRACE mission is just 
not long enough to determine rates that can be extrapolated very far, 
because of the annual and inter-annual variability and the danger of 
long-period tidal aliases. And it would certainly be wrong to model 
the 'step' from the earthquake as a rate, which is the consequence of 
fitting the existing GRACE time series. At the moment, it is far 
safer to avoid adopting rates that have not been particularly well 
verified. After all, what independent analysis indicates that the 
rates are reliable? While this might provide somewhat improved 
results for cycles coincident with GRACE, I would be very worried 
about the application of the rates beyond that.

We are already discussing the next generation of GRACE models, 
including a more complete modeling of PGR (to replace the few rates 
currently used) and properly modeling the Sumatra earthquake. So I 
think it is premature to try to do this on our own using GRACE models 
where this was not incorporated.

Best regards,
John