From: Yehuda Bock, Peng Fang, Jeff Behr
Date: 10 Nov 1994 20:00:47 
Subject: [SCIGN-0014] SCIGN processing update 

******************************************************************************
SCIGN Electronic Mail    Thu Nov 10 20:00:47 PST 1994      Message Number 0014
******************************************************************************

Author: Yehuda Bock, Peng Fang, Jeff Behr
Subject: SCIGN processing update

Scripps Orbit and Permanent Array Center
Date: November 10, 1994

We have implemented a stacking algorithm (S. Wdowinski et al.) to filter
the SCIGN time series of ITRF positions for common-mode errors. The
resulting position time series are considerably improved in all 
components, and the scatter in most of the series is quite good
(1-3 mm in horizontal, 4-7 mm in the vertical).
Also, these series provide excellent site diagnostics.

In estimating the stacking correction, we eliminate data from 
several problematic sites. The correction is later applied to all the
SCIGN position series. The following problems have surfaced:

(1) Oat Mountain, Mt. Wilson and Carr Hill (Parkfield) have significantly
higher noise levels than the other sites in all components (see SCIGN
mails 8-9 for a more detailed description).

(2) The SIO3 time series has an offset when switching from the old style
Ashtech Z-12 antenna to the new style (larger ground plane) antenna.
There is a clear offset in the vertical, but also a (smaller) offset
in the horizontal as well. More disturbing, we have found different
offsets for different antennas (of the new type), which was not
apparent in antennas of the old type.

(3) The JPLM time series has experienced a dramatic increase in noise level
in all components, starting on day 279 (Oct. 6). We checked the station
log from the IGS Central Bureau and found that the firmware had been
upgraded to 3.0 on day 285. Can anyone at JPL explain what is happening?

The filtered time series for 6 weeks of data can be viewed
on the SOPAC Mosaic: (http://igpp.ucsd.edu)
If you do not have access to Mosaic, let us know and we can make 
postscript files available.

Problems with other sites processed by SOPAC:

(1) The Mammoth Lakes (CASA) solutions are of poor quality.

(2) The Columbia site (CMBB) of the BARD array has a jump in the
vertical component for days 287-293 (Oct. 14-20), of the order
of 50 mm. Can anyone from BARD explain this?


[Mailed From: pgga@pgga.UCSD.EDU (Geodesy Lab account)]