Jökull - 01.01.2021, Qupperneq 26

Gudmundsson et al.
June and 12 September 1919 (Figure 2) have been par-
ticularly useful.
The photo taken on 12 September 1919, about 10
months after the eruption ended (Figure 2a), shows
thick piles of tephra. In contrast, on the photo from
the glacier-covered southern rim of the Katla caldera
(Háabunga), taken in June 1919 (Figure 2b), and
other similar photos from the same trip (Gudmunds-
son and Högnadóttir, 2001) the surface is covered by
the 1918–19 winter accumulation of snow, which in
spring is usually close to 10 m thick in the caldera
(Ágústsson et al., 2013). Figure 2a shows the region
where the easternmost crater was located and the area
to the east of the inferred eruption site. The most con-
spicuous feature is the canyon that cuts across from
left to right (from west to east). A thick layer of
tephra covers the landscape and ice is exposed only
in few places, notably in the steep lower slopes of
the canyon. To the north of the canyon the surface is
covered by what remained of the snow layer from the
winter of 1918–19, expected to have been a few me-
tres thick at the end of summer. New snow, presum-
ably fallen hours or days before the photo was taken,
partly drapes the snow from 1918–19 and parts of the
tephra-covered landscape. Despite the chaotic land-
scape of ice and tephra in the foreground, crevasses
are not visible. The area in the upper right hand
(northeast) corner is the heavily crevassed steep north-
ern part of the Kötlujökull outlet glacier, where it
flows out of the caldera, about 3–4 km away. The pho-
tos taken by Kjartan Guðmundsson on 23 June 1919
(Figure 2b), when visibility was good, show that this
northern part of the outlet glacier was more or less in-
tact and not much altered by the eruption, apart from
being covered with tephra.
The west-east canyon seen on Figure 2a extended
into the southern part of the outlet glacier where it
flows out of the caldera. The explorers in 1919 (Jóns-
son, 2008) were of the opinion that the bottom of the
canyon marked the glacier bed. However, radio-echo
soundings have revealed that ice thickness where most
of the canyon was located in 1919 reaches several
hundred meters (Björnsson et al., 2000; Pálsson et
al., 2005). Thus, although having a floor of tephra,
the canyon was cut into both the tephra cover and
the underlying glacier, with ice underneath the tephra-
covered bottom. The most likely explanation for the
formation of the canyon is supraglacial drainage of
meltwater in the first 1–2 km of the path of meltwater
from the vents towards east, down towards the outlet
glacier. Such supraglacial flow of meltwater was ob-
served in the Gjálp eruption in 1996 (Gudmundsson
et al., 2004) and did also occur in the first days of the
Eyjafjallajökull eruption in 2010 (Magnússon et al.,
2012; Oddsson et al., 2016). Although the canyon
in Katla and its surroundings have similarities with
the Gjálp ice canyon, the thickness of tephra on this
photo appears an order of magnitude greater than the
maximum seen at Gjálp (1–2 m). A series of scaled
comparisons permits a tephra thickness estimate as
follows: A small stream, probably a few meters wide,
can be seen flowing along the bottom of the canyon
on the western (left hand) side. According to an esti-
mate made by the explorers in September 1919, who
descended to the bottom of the canyon, the stream
had a discharge of 0.2–0.3 m3/s (Jónsson, 2008). The
tephra in the walls of the canyon has a thickness that
is several times the width of the stream at the bottom.
Further rough scaling can be obtained from the ob-
servations made in 1919, where the canyon is said to
have been 1–2 km long and at most 200 m wide at
the top (Jónsson, 2008). The photograph (Figure 2a)
is taken approximately 1 km south of the main leg of
the canyon, which on the photo trends perpendicular
to the line of sight. The dimensions indicated by the
photo and the observations from 1919, suggest that
the canyon was 50–70 m deep relative to the area im-
mediately to its north. The absence of any visible ice
in the upper half of the canyon wall suggests that the
tephra pile was not less than one third of the total wall
height in the western part of the canyon, and perhaps
as much as half the height. This indicates tephra thick-
ness at this site in the range 15–35 m, a few hundred
meters away from the vents. Thus, we consider that
25±10 m is a conservative estimate of the maximum
tephra thickness outside of the craters.
The 1919 photos and the descriptions of the explor-
ers (Sveinsson, 1919 p. 54–59), together with photos
of the plume taken from known locations around the
village of Vík (Larsen and Högnadóttir, this issue),
24 JÖKULL No. 71, 2021