Jökull - 01.01.2010, Blaðsíða 6

Geirsson et al.
planes, ships, and other vehicles. These sites are not
routinely included in our data processing of daily so-
lutions, nor do we count them in our numbers for
CGPS sites in Iceland. However, one such site in south
Iceland provided important high-rate data from the
May 2008 earthquake sequence, and helped constrain
source parameters and timing for these events (see be-
low). The data from the CGPS network has been uti-
lized for cadastral surveys as well, where the stations
have served as base stations for a variety of mapping
projects. In addition, the atmospheric delay imposed
on the satellite signals observed at the stations is being
used as an input to meteorological models.
Good coupling of the antenna to solid bedrock by
the GPS-monument is important to eliminate possi-
bilities of monument movement in the interpretation
of the results. Most of the CGPS stations in Iceland
have one of two kinds of monuments: (a) a one meter
high stainless steel quadripod anchored to bedrock by
drilling and cementing eight threaded rods typically
12–50 cm into decent bedrock; this we term the IS-
GPS monument; (b) the short-braced UNAVCO Plate
Boundary Observatory (PBO) monument, where four
2.5 cm thick rods are drilled and cemented typically
0.5 to 2 m into bedrock and welded together at the top
for the antenna mount. A few stations are on top of
buildings or have different kinds of monuments. The
ISGPS and PBO monuments each have their advan-
tages and both can be regarded as highly stable geode-
tic monuments when properly installed.
As the sites have been installed over more than a
10 year period, there is a wide range of instrumenta-
tion and communication types in use. The bulk of the
older sites nominally collect raw data every 15 sec-
onds as 24 hour-long files that are transmitted auto-
matically to IMO on a daily basis. Data collected at a
higher sampling rate (1 sample per second) are trans-
mitted from many of the newer sites, depending on in-
strumentation and communication bandwidth. Some
sites collect high-rate data that are not transferred un-
less investigation of seismic, volcanic, or other events
require the data. Many of the sites are co-located with
seismic stations, or transmit data to a nearby seismic
station through radio links, to make common use of
the existing infrastructure. Where the GPS equipment
has been upgraded an effort has been made not to
change or move the GPS antennas unless absolutely
necessary, to avoid creating artificial offsets in the
time series.
In addition to episodic and continuous GPS sta-
tions, there exist a number of semi-continuous sites
around the country, where data are collected repeat-
edly over shorter continuous periods of time (weeks
to a few years). These sites typically do not trans-
mit data, although communications were established
at sites around Eyjafjallajökull in 2010 for enhanced
monitoring capabilities of the volcano. Data from
semi-continuous sites is not included in this paper.
DATA PROCESSING AND SITE
VELOCITIES
To go from the raw data files to daily solutions for
site positions, here termed data processing, and esti-
mating and interpreting site velocities from time se-
ries of daily position solutions, is an important pro-
cess that can be performed in various ways. Once
the daily data files have been archived at IMO, they
are automatically processed and the daily results dis-
played on IMO’s website (www.vedur.is). Each of the
research groups that routinely analyze data from the
Iceland stations processes slightly different subsets of
the data set using distinctive processing software or
dissimilar processing options and external informa-
tion (e.g. satellite orbits, antenna calibrations, tidal
loading models) within each software package. Thus
far, the data have been processed using the Bernese
version 5.0 software (Dach et al., 2007), the GAMIT/-
GLOBK version 10.3 (King and Bock, 2005), and the
GIPSY/OASIS II version 5 software (Zumberge et al.,
1997). Geirsson et al. (2006) compared the outcomes
from these three different processing packages for Ice-
landic CGPS data collected from 1999 to 2004, and
found the results to be in a general agreement. The
purpose of this summary is to focus on specific geo-
dynamic processes observed by CGPS in Iceland. We
therefore present results from the different processing
approaches as fits each subject best, and leave more
detailed comparison of time series obtained from the
different processing approaches for separate studies.
6 JÖKULL No. 60