Jökull - 01.01.2021, Side 55

Bedrock and tephra layer topography within the Katla caldera
Figure 7. a) Ice thickness map of the study area calculated using the glacier surface in September 2019. b) Static
water potential at the glacier bed, revealed as a greyscale image and a contour map (100 kPa contour interval
equivalent to 10 m of hydrostatic head), calculated from the new bedrock DEM and glacier surface in Septem-
ber 2019. Blue lines show water divides between the labelled drainage basins, obtained from the potential map.
The locations of cauldrons within different water drainage basins are shown. The red coloured image on top
of the potential map indicates depth and size of closed depressions in the potential map forming local minima.
– Ísþykktar- (a) og vatnsmættiskort m.v. að jökull liggi á botni með fullu ísfargi (b) reiknað út frá jökulbotni
(6. mynd) og jökulyfirborði í september 2019. Jafnmættislínur á mynd (b) eru með 100 kPa millibili (jafngildi
10 m hárrar vatnsssúlu). Bláar línur (b) sýna vatnaskil milli helstu vatnasviða samkvæmt vatnsmættiskorti og
þar með hvaða sigkatlar falla innan hvers vatnasviðs. Rauðskalamyndin sem felld er ofan á mynd (b) sýnir
stærð og dýpi lokaðra lægða í vatnsmættiskorti.
layer was close to the glacier bed the migration of both
the 2D and 3D data was repeated with all reflections
likely to originate from the glacier bed masked out in
the input data before the tephra layer reflections were
traced. This was done due to limitations of the mi-
gration method, which in some cases produces weak
artificial signals propagating upwards from the rela-
tively strong bed reflections. This artefact was not a
problem when tracing the bedrock reflections, but in
some cases, it obscured the much weaker internal re-
flections near the bed. The masking was done by cal-
culating for each survey point the distance from the
transmitter to the nearest point in the final bedrock
DEM and back to the receiver and the corresponding
travel time, tbed_nearest (with cgl=1.68× 108 m s−1)
and then replacing all backscatter values with travel
time >(tbed_nearest -1× 10−7 s) with 0. Assuming that
the final bedrock DEM is accurate this should only
leave internal backscatter originating >∼10 m above
the glacier bed. When tracing reflections from the
deeper tephra layer and parts of the 1918 tephra layer,
the tracing algorithm described above proved ineffi-
cient due to a low signal to noise ratio, causing the au-
tomatic tracing to fail repeatedly even though it could
be traced from visual inspection. In those cases the
tracing was carried out with manual digitization.
The traced tephra reflections in the 2D migrated
data, were filtered and subsampled to values at 20 m
intervals along the profile while the traced reflec-
tion from the 3D-migrated data was used unfiltered.
JÖKULL No. 71, 2021 53