[IGSMAIL-4281]: IGS Reference Site Candidates

[Michael B Heflin]

******************************************************************************
IGS Electronic Mail      28 Feb 09:29:15 PST 2003      Message Number 4281
******************************************************************************

Author:  M B Heflin

The goal of this mail is to offer improved reference frame stability
by choosing a new, larger set of IGS reference sites which can be used
right away without any equipment upgrades or new installations.  Use
of the new list would offer improved stability, better coverage, and
additional redundancy.  The following list was chosen by keeping only
those GPS sites (IGS and others) which met the following criteria:

1.  located away from seismic zones and plate boundaries
2.  >= 3 years of historical data
3.  currently operating
4.  constant velocity
5.  <= 2 position breaks in the time series

There are 83 sites (all IGS, it turns out) which meet all five requirements:

ALGO ALIC AMC2 ARTU ASC1 BAHR BOR1 BRAZ BRMU BRUS
CAS1 CEDU CHAT CHUR DARW DAV1 DUBO EBRE EISL FLIN
FORT GLSV GODE GOUG GRAS GRAZ HERS HNPT HOB2 HOFN
IISC JAB1 JOZE KARR KELY KERG KIRU KOKB KOSG KOUR
KSTU LAMA LPGS MALI MAS1 MAW1 MCM4 MDVO METS MKEA
NALL NKLG NLIB NOUM NRC1 ONSA PENC PERT POTS PRDS
RIOG SCH2 SOL1 STJO STR1 SUTH SYOG THTI TIDB TOW2
TRO1 TROM USNA USNO VESL VILL WES2 WSRT WTZR YAR2
YELL ZIMM ZWEN

The rms isolation, zeta, is one way to characterize how well a given set of
sites covers the surface of the Earth. Roughly speaking, if one picks a
spot at random on the Earth's surface, the nearest site will be about zeta
away.  For uniformly distributed sites, the relation between site
separation d and zeta is zeta = d/sqrt(6). Lower values of zeta indicate
better coverage. Physical geography puts a practical lower limit on zeta; a
best case is zeta ~ 1300 km for ~ 90 strategically located sites. [For more
discussion see Zumberge, J F, Ruth E Neilan, Ivan I Mueller, Position Paper
1: "Densification of the IGS Global Network" from JPL Publication 95-11,
Densification of the IERS Terrestrial Reference Frame Through Regional GPS
Networks, Zumberge, J F, and R Liu (Eds), Pasadena, CA, 1994.] Zeta is 1894
km for the 83 sites above. Recall that the 83 sites were selected solely on
the basis of the 5 criteria above, not on uniformity of coverage. It is not
surprising, of course, that the achieved zeta is larger than what one could
do with 83 more uniformly distributed sites.  If the sites GALA, KWJ1, OHIG,
SEY1, and TOUL could somehow be upgraded to meet requirements 3 and 5, zeta
would be reduced to 1727 km.  The monument at GALA has been removed but
there is a new one at nearby GLPS.  We intend to upgrade a number of sites to
ensure reliable operations.

For comparison the old set of 54 IGS sites was:

ALGO ASC1 AUCK BAHR BRMU CAS1 CEDU CHAT DAV1 DGAR
FAIR FORT GGAO GOL2 GRAZ GUAM HOB2 HRAO IRKT KERG
KIT3 KOKB KOUR KWJ1 LHAS LPGS MAC1 MALI MAS1 MATE
MCM4 MDO1 NALL NLIB OHIG ONSA PENT PERT PIE1 POTS
RIOG SANT SHAO THU1 TID2 TROM TSKB VILL WEST WSRT
WTZR YAR1 YELL ZWEN

Zeta for the old set of 54 sites was 1942 km and three of them have had
significant offsets:

POSITION OFFSETS AND ERRORS IN MM.
NAME     N      E      V        SN     SE     SV
FAIR  -42.5   25.2   27.6      6.5    4.6   10.1
GUAM  -14.2    1.8  -37.8      4.9    8.8    9.4
KOKB  -19.1  -44.2  -30.3      6.8    7.0   10.1

The new list of 83 sites offers improved stability using sites with a
long history, which are far from seismic zones, and have minimal historical
position breaks.  The coverage is also improved with zeta reduced from
1942 to 1894 km.  Use of 83 sites also provides additional redundancy to
reduce the impact of missing data.

Mike Heflin
Don Argus
Dave Stowers
Frank Webb
James Zumberge

The densification paper is available at:
ftp://igscb.jpl.nasa.gov/igscb/resource/pubs/zumberge_94a.pdf