Jökull - 31.12.2001, Qupperneq 38

Thorsteinsson et al.
Table 3. Density difference between bubble-free ice layers and adjacent bubbly ice. Averages for adjacent ice
were determined for layers with a thickness comparable to the ice layers. - Eðlisþyngd í bólulausum íslinsum
og bólóttum ís á svipuðu dýpi.
Depth of bubble-free layer Dýpi íslinsu Player [kg/m3] Padjacent ice [kg/m3] Difference [%] Mismunur [%]
40.04^10.30 m 912 885 3%
64.21-64.59 m 925 911 1.5%
80.22-80.35 m 918 912 0.7%
98.22-98.32 m 912 907 0.6%
ray source (isotope 137Cs, energy = 661.660 keV)
with an activity of 3Ci and a photomultiplier detec-
tor unit. A gamma-ray beam collimated to a width
of 2 mm passes through the ice core at its maximum
diameter, whereby it becomes attenuated according to
the Lambert-Beer law
I = I0e~d^p (1)
where I is the intensity of the attenuated beam, Ia is
the intensity of the beam traveling in air in the ab-
sence of ice, d is the diameter of the ice core (mea-
sured manually at several locations for each piece of
the core) and p is the mass-attenuation coefficient,
which for ice is /rice =0.0085645 m2 kg_1±0.1%,
when 137Cs is used as a radiation source (Wilhelms,
1996). Equation (1) leads to the following relation:
p = -d-'p-HniI/Q (2)
which gives the density of the ice from core diameter
and intensity measurements. For a more detailed de-
scription of the gamma-ray densiometer see Wilhelms
(1996) and Gerland et al. (1999).
Results frorn the density measurements are shown
in Figure 7a. The transformation to glacial ice
{p— 830 kg/m3) appears to be completed by a depth
of 35 m. An interesting characteristic of the profile
is the nearly linear increase in density with depth in
the interval 0-40 m and sudden stabilization at values
near 900kg/m3 below 40 m. Variability above 40 m
depth is caused by the variation in the abundance of
ice layers, which have a density of about 900kg/m3.
There is no indication that the densification process
can be separated into different stages, dominated by
varying states of grain settling and creep, as observed
in polar ice (Paterson, 1994).
The conventional density measurements yield on
average 5% lower density values than the gamma-ray
method above 40 m but 2% lower values on average
below 40 m. Overestimation of core volumes in the
conventional measurement is the most likely expla-
nation for this discrepancy. In the softer firn cores
from above 35 m, pieces would sometimes break off
the ends and the sides of the core pieces, and below
that depth the core catchers would typically produce
grooves and thereby remove 1-2% of the volume of
each core piece.
The gamma-density results yield an average den-
sity of 910 kg/rn3 below 40 m, indicating that air bub-
bles comprise on average 0.8% of the ice volume
in the interval 40-100 m. A detailed comparison of
gamma-ray density values measured on bubble-free
ice layers and on adjacent bubbly ice, presented in
Table 3, shows the decrease in bubble volurne with
depth as the load of the overlying ice increases. The
average density of the bubble-free ice in the layers
shown in Table 3 is 917kg/m3, indicating that the
gamma-ray measurements provide an unbiased esti-
mate of the density. The deviation up to 925 kg/m3 in
one layer is not significant, since it cannot be guaran-
teed that the gamma rays pass through the maximum
diameter of the core at all times during the measure-
ment. An error limit of ±1% should thus be assigned
to the gamma-ray density results.
Figure 7b shows a more detailed profile of the
gamma density. The abundant peaks in density in the
36 JÖKULLNo. 51