Jökull - 01.06.2000, Page 13

Ice-thickness measurements on Sólheimajöhull
and receiver. At least 10 minutes of pseudorange
data were collected at each point, and positions were
computed using Magellan post processing software
(Magellan, 1991). Accuracy for the positions is ±2 to
5 m RMS for X, Y and ±6 to 15 m for Z co-ordinates
depending on satellite geometry and duration of data
logging at each location. Computed positions often
varied by over 100 m from uncorrected fixes, justify-
ing the use of differential GPS. Altitude checks were
made against the GPS results using a digital aneroid
barometer.
0 * 5.G0mV > l.OOusEBt 5.00mV > l.OOns
Figure 2. Downloaded oscilloscope screen for a
typical basal return signal (Point 2). The cursors are
located at the air and return pulses, and the two
way travel time is 3.16 microseconds. - Skjámynd af
sveiflusjá. Hver rúða jafngildir einni míkrósekúndu.
Lóðrétt lína liggur við endurkastið frá botni jökuls-
ins sem kemur 3,16 míkrósekúndum eftir að merkið er
sent afstað.
Ice depth was determined using a portable mono-
pulse ice radar constructed following the principles
of Watts and England (1975). Antennae systems for
both transmitter and receiver were resistively dam-
pened dipoles, and were adjusted to change frequency
depending on ice thickness (2.5 MHz for most rea-
dings and 5 MHz for ice thinner than ~150 m). The
antennae were oriented either parallel or perpendicul-
ar to the direction of flow depending on surface
geometry, degree of crevassing and signal strength.
The transmitter and receiver were separated by 50 m
for each sounding. The basal return signals were in-
terpreted in the field on a Fluke Scopemeter 96, but
were also downloaded onto a PC so that spurious
traces could be re-examined (Figure 2). The accuracy
depends largely on the time resolution of the digital
storage oscilloscope, where the corresponding points
on the transmitted pulse and the ice/bedrock echo
can usually be determined to the nearest time-step.
In practice these limitations resulted in an accuracy
for the thickness determinations of ~5%, and usually
less than ± 10 m. Ice thickness was calculated from the
two-way travel time allowing for the spacing between
transmitter and receiver, and time for the air pulse to
reach the receiver (Bogorodskii et al., 1985).
RESULTS
The long profile of the bed is notable for its lack
of large undulations (Figure 3). The most complex
topography occurs near the snout, where there is evi-
dence of a sub-glacial hill and an overdeepening. The
glacier bed remains close to sea level for 3.5 km in-
land from the snout, and would become a fjord if ice
free under higher sea levels (for example, during the
early Holocene). The greatest ice thickness of 433 m
was located at points 22 and 23, where the relief of
the valley sides is highest (Table 1). The ice is thinner
at the top of the survey line at the southern rim of the
Katla caldera. The results of two cross sections are
shown in Figure 3. Cross section 1 was taken at point
23, 4.5 km from the snout where the glacier reached
its maximum thickness (Figure 1). The trough approx-
imates a parabolic profile with steep sides and a gently
undulating floor. Cross section 2 was taken at point
34,1.25 km from the snout (Figure 1). Here the trough
was found to be asymmetrical, associated with diver-
gence of the snout around Jökulhaus, with the deepest
point occurring below the main (NW) ice stream.
The surface of Sólheimajökull has a gentle long
profile with few undulations. An exception occurs
JÖKULL No. 48 11