Jökull - 01.01.2004, Page 39

Sediment thickness and erosion rates within Hvítárvatn, central Iceland
PHYSICAL SETTING
Hvítárvatn is 11 km long, 3 km wide, covers 28.9 km2
and has a maximum depth of 83 m. Runoff from its
820 km2 catchment is dominated by meltwater from
the two outlet glaciers of Langjökull; Norðurjökull,
which historical records indicate advanced into the
lake ca. AD 1800 and still is calving, and Suður-
jökull, which was calving into the lake in the early
1900s, and retreated from the lake about AD 1950
(Figure 2). One third of Langjökull is in the Hvítár-
vatn catchment. The largest meltwater rivers feed the
prograding Hvítárnes Delta on the northeastern mar-
gin of the lake; a smaller stream, Fróðá, has created a
smaller delta at the north end of the lake. Langjökull
has no surging behavior along its eastern margin, and
there has been no Holocene volcanic activity to af-
fect glacier fluctuations. Fluctuations in the two out-
let glaciers of Langjökull are therefore climate re-
lated. Prominent lateral moraines define the extent of
both outlet glaciers at their Little Ice Age (LIA) max-
ima, the most extensive limit of Langjökull since re-
gional deglaciation about 10 ka (Kaldal and Víkings-
son, 1990; Tómasson, 1993).
Hvítárvatn can be divided into two main regions,
a shallow and relatively flat southern basin, and a
deep northern basin that contains the major deposi-
tional centers of the lake (Figure 3). The deepest ar-
eas in Hvítárvatn (>80 m) are in front of the two out-
let glaciers. Distal of the LIA terminal moraines, the
deepest basin is located in the central part of the lake
with depths up to 79 m. A relatively flat shelf, 30 to
40 m deep, rises gently from the central deep toward
the northeast shore of the lake. The dominant feature
along the northeastern shore is the Hvítárnes Delta,
which forms a steep margin in water depths of 30 to
40 m. In the northern portion of the main basin a se-
ries of bathymetric highs form a loosely defined ridge
that rises approximately 10 m above the basin floor
and varies in width from 5 to 30 m at water depths of
25 to 35 m (Figure 3). The ridge, which is mantled
by 10 to 15 m of stratified sedimentary cover is seis-
mically opaque, and might be either an old moraine
or bedrock. Based on geometry, we presume the ridge
is composed of bedrock extending from an on-shore
hyaloclastite ridge to the south.
METHODS
Over 100 km of seismic reflection data were col-
lected in the summer of 2001 from the survey boat
Bláskel, equipped with a Geopulse Boomer system.
The boomer system operated at 175 J and the data
were filtered for frequencies between 0.75 kHz and
2 kHz. The boat was positioned with a Trimble 4000
DS Differential GPS system with ±1 m resolution.
Initial seismic profiles showed that the southern half
of the lake contained little sediment fill, mostly con-
centrated in hollows in the lake floor, and that the sed-
iment inside the LIA moraines was disturbed. Con-
sequently, we focused the detailed seismic survey on
the main, undisturbed depositional basin. Lines were
spaced 200 to 300 m apart in a grid across the basin
(Figure 2).
Seismostratigraphic units were defined by the na-
ture of the reflectors and the degree of acoustic trans-
parency. Reflectors vary in their thickness, density,
spacing, and continuity. Although reflectors across
most of the basin are essentially parallel, confirming
layer-cake stratigraphy, in some regions intersecting
reflectors were identified, indicative of ponded sedi-
ment. These differences helped to define specific seis-
mostratigraphic units.
Once the seismostratigraphic units were defined,
they were traced across each of the 21 seismic tran-
sects. Sediment thickness was calculated using fresh-
water sound velocity of 1456 m/s, representing a min-
imum estimate. The thicknesses of individual seismic
units was measured along each of the tracklines and
recorded with UTM coordinates logged by GPS.
Compiled data were entered into a Geographic
Information System (GIS) for further analysis, us-
ing ArcGIS v8.2. Raster and vector datasets were
assembled with a consistent projection (UTM Zone
27N) and datum (Hjörsey 1955). Datasets included:
point shapefiles for bathymetry and seismic unit thick-
ness; a raster version of the 1:50,000 Hvítárnes topo-
graphic quadrangle, from the Icelandic Geodetic Sur-
vey (DMA, 1989); the Hvítárvatn shoreline (digitized
from the Hvítárnes base map); glacier margins (also
digitized from the Hvítárnes base map); and shapefiles
for core locations and seismic transects (using DGPS
JÖKULL No. 54 39