[IGSMAIL-0511] Impact of AS

[J.]

******************************************************************************
IGS Electronic Mail      Fri Feb 25 16:42:45 PST 1994      Message Number 0511
******************************************************************************

Author: J. F. Zumberge
Subject: Impact of AS

Based on the periods 1994 Jan 01-30 and 1994 Feb 01-20, corresponding
respectively to anti-spoofing (AS) not in effect and AS in effect, we
have used the daily repeatabilities of (non-fiducial) station
coordinates as a measure of the degree to which AS has impacted the
quality of results from the Global Network of Rogue and TurboRogue
receivers.  The observed ~30% degradation is reasonably explained as a
consequence of changes in our estimation strategy designed to
effectively deal with AS.

For each non-fiducial station in the global network, we calculate the
weighted (using formal errors) rms deviation over time from the average
value during the period.  The results are tabulated below, expressed as
median (over all non-fiducial stations).

         median daily repeatability
                                          observed      expected
         94 Jan 01-30   94 Feb 01-20      increase      increase

north        3.1 mm         4.0 mm          29%           25%
east         5.4 mm         6.9 mm          28%           26%
vertical    10.8 mm        14.2 mm          32%           53%

The last column indicates the expected increase, based on the median
formal errors.  That is, we expect a degradation in daily repeatability
for two reasons: (i) we reject data below 20 degrees elevation during
AS (compared with 15 degrees during non-AS) and (ii) our AS strategy
adds additional carrier-phase bias parameters, based on observed
discontinuities in post-fit residuals (see igsreport.782; we suspect
that the TurboEdit algorithm, used with much success during non-AS, is
less successful during AS).  A third reason may have to do with
less-than-expected volume of data, although we have not investigated
this quantitatively.  Nor have we made any distinction between Rogue and
TurboRogue receivers in this current preliminary report.