Jökull - 01.12.1994, Page 22

wavelength phase measurements on L2.
At the time of our survey there were two periods
of about 9 hours per day during which at least four
satellites were visible, from 9:40 to 18:40 GMT and
then from 22:30 to 8:30 GMT. We recorded for a sin-
gle, long period daily which averaged about eight
hours during the period 9:40 to 20:30 GMT. Data
were recorded at intervals of 15 seconds.
Point OS-1985-7134 (Núpafjall) was measured
during each session and the directly-measured lines
within the Hengill network varied in length from
about 6 to 33 km. Four regional, pre-existing GPS
points in South Iceland, points S001, S027, S056 and
S061 were also measured. These were located up to
75 km from Núpafjall and were measured to tie our
new network to the rest of the Iceland GPS network
and to neighbouring crustal blocks. These regional
points were measured for sessions that lasted about 20
hours.
DATA PROCESSING
The University of Durham implementation of the
Bernese V 3.2 Software was used to process the data
(Rothacher et ai, 1990). Point coordinates were calcu-
lated relative to point Núpafjall for which the GPS co-
ordinates derived in the 1986 survey were used.
Two-day satellite orbits were calculated using
broadcast orbit information and the modelling pro-
gram DEFSTD. Satellite clock corrections and a-pri-
ori point coordinates were calculated using the C/A
code data and program CODSPP. An ionosphere
model for each session was derived using the mea-
sured phase data and program IONEST. A standard
tropospheric model was used since the troposphere
could be considered uniform over the area.
For each daily measuring session cycle slips were
removed from between-point phase single difference
files using program MAUPRP which is reliable for
slips greater than about 5 cycles and different in L1
and L2. Point coordinates were then calculated using
the L3, ionosphere-free phase combination and pro-
gram GPSEST, solving for the phase integer ambigui-
ties as real numbers. The phase residuals were
browsed to detect slips smaller than 5 cycles previous-
ly undetected. These were added manually to the sin-
gle difference files.
The clean single difference files were then used to
calculate the so-called L5 ambiguities (the difference
0 10 20 30 40 S0
mm
Fig. 2 The effect of ocean loading in Iceland,
computed using the method of Wu- Ling and
Morgan (unpublished computer program).
Shading shows maximum vertical difference
with respect to stations at (a) Akureyri, (b)
Borgarfjörður, (c) Reykjavík, and (d) Búrfell
2. mynd. Utreiknuð áhrif sjávarfalla á
landhœð. Sýndur er mesti hceðarmunur
miðað við a) Akureyri, b) Búrfell, c)
Reykjavík og d) Borgarfjörð.
20
JÖKULL, No. 44