From: Jan Kouba [mailto:kouba@geod.NRCan.gc.ca]
Sent: January 8, 2008 08:56
To: Jim Ray
Cc: Donahue, Brian
Subject: Re: EMR polar motion spectra]


Dear Jim,


Thank you for the emr pictures. It's encouraging that emr looks OK!


You're right that the fortnight PM signal is very small and well below the
noise of IGS PM (and usually it is neglected/not considered), so the
problem peak of the daily PM discontinuity is likely not caused by the
fortnight zonal tides, a more likely candidate is the subdaily (tidal) PM,
more specifically the O1 (25.8 hour period) wave, which beats (against 24h)
with an aliasing retrogate period of 14.2 days. The PM values - 24h averages
are insensitive to this aliasing signal, but the independent 24h rate
averages are quite sensitive to this (and other subdaily PM) alaising
signals. That's why the independent PM rates can be used to test subdaily
PM (and the peaks of daily break spectra are coming mostly from the PM
rates!). In fact, I've used the very same daily PM breaks to test various
subdaily PM model (see the attached 2003 paper).
 

So, in summary, the fortnight peak is likely due to problems in subdaily PM,
in this case the peak should be 14.2 or 14.8 days (see the table 1). There
are other possible subdaily PM (aliasing) peaks of 9.3 and 9.6 d. If it were
due to the fortnight tides, both rate and PM would show it and it would have
to be at 13.67 days, not 14.2 or 14.7 d. Finally, note that some peaks in PM
rates (in the week to fortnight bands) are not necessarily wrong, they may
be real and caused by the subdaily AAM (atmosphere) and OAM (oceans), but
they would have to be the same for all ACs and would have to disappear after
accounting for OAM and AAM. Another reason why the PM rate solutions should
be independent, in order to observe the aliased AAM/OAM  signals!
 

With kind regards
 

Jan
 

PS Brian (emr) had a 10day PM peak problem, which was caused by 30 hour
averaging of PM in the EMR Final solution (only one PM parameter was
introduced for a 30 hour arc), this has now been corrected and you're using
exactly 24h for PM, is this true Brian? Exactly since when the emr Final PM
are 24h averages?

 

cc Brian