Jökull - 01.01.2020, Blaðsíða 29

Hannesdóttir et al.
can now be extended back in time and compared with
changes in terminus position derived from the glacier
outline inventory. The retreat rates were highest dur-
ing the most recent time periods (2000–2010, 2010–
2014 and 2014–2019, see Table 3), and the maximum
rates for most ice caps and glaciers observed in 2010–
2014. The rapid retreat during the 1930s and 1940s,
as documented in the terminus variations database of
the Iceland Glaciological society, is not well repre-
sented in the outline inventory, due to poor temporal
resolution during the first half of the 20th century. The
rate of area decrease may during this time have been
similar as in the first two decades of the 21st century.
The frontal measurements are an important source of
information about glacier variations and complement
other data about glacier change.
Multi-temporal glacier outline inventories are im-
portant for large-scale (i.e. regional) studies of geode-
tic mass balance and glacier variations, where the
outlines are needed as input data but require tedious
work of digitization, clear aims and detailed work-
flow (e.g. Brun et al., 2017; Dussaillant et al., 2019).
Areal changes can also easily be analysed and com-
pared globally on the basis of multi-temporal glacier-
outline inventories. The total area of glaciers in Ice-
land has decreased by 18% since ∼1890. The four
largest ice caps (Vatnajökull, Langjökull, Hofsjök-
ull and Mýrdalsjökull) have lost 12–30% since the
end of the 19th century, whereas the intermediate-size
glaciers have decreased by 35–80%. For compari-
son, most glaciers worldwide outside Antarctica and
Greenland have experienced an area loss of about 30–
60% since their mapped maximum LIA extent (Paul
and Bolch, 2019), although time periods, climatic
regimes, glacier characteristics and sample sizes dif-
fer globally.
An increasing number of terminal lakes that are
formed as the glaciers retreat, enhance melting of
ice and increase glacier retreat, and they have caused
rapid changes in the proglacial area of many glaciers
in Iceland in the past two decades (Guðmundsson et
al., 2020). The development of the terminus lakes
will in the future be monitored as part of the moni-
toring of glacier variations in Iceland and their extent
will be submitted to GLIMS as part of the next verion
of the Icelandic glacier outline database. Also, indi-
vidual flow basins will be delineated and submitted
to GLIMS, thus the area changes of individual outlet
glaciers can then be extracted.
One possible explanation of the widely different
extents of Drangajökull ice cap according to different
historical sources and field investigations described in
a previous section, is that perennial snow may have
covered large areas of the plateau near the glacier for
several decades during the cold climate of the 18th
and 19th centuries. Such areas should be included
within the glacier outline according the GLIMS def-
inition (Raup and Khalsa, 2010), but this definition
is not easy to apply for LIA glacier extents when the
location of the glacier margin is partly based on geo-
morphological evidence such as moraines. Large ar-
eas on the highland to the southeast from Drangajök-
ull may have been covered by perennial snow during
parts of the 18th and 19th centuries and these areas are
not included within our LIA maximum outline. Bryn-
jólfsson et al. (2014) describe that negligible glacial
geomorphological imprints in specific areas at eleva-
tions 500–600 m around Drangajökull, suggest a thin
and not very dynamic glacier ice.
In this paper, we have described the new Icelandic
glacier inventory that has been sent to GLIMS. The
glacier outline database will be updated with more
detailed information based on DEMs and orthoim-
ages that are being created from the historical aerial
images of NLSI from the 1940s and 1960s, as well
as images from declassified Hexagon KH9 satellites
from 1977–1980. Work is ongoing to refine the max-
imum LIA glacier extent in some areas, where de-
tailed studies of the forefields has not yet been under-
taken, for example on Tröllaskagi and Þrándarjökull
and Hofsjökull eystri. Complications regarding the
maximum LIA extent of glaciers on Tröllaskagi and
in some other areas, where glaciogenic landforms are
influenced by permafrost (rock glaciers and ice-cored
moraines), need to be more thoroughly considered
(e.g. Wangensteen et al., 2006; Lilleøren et al., 2013;
Tanarro et al., 2019).
The digital outlines provide a baseline for future
monitoring of glacier changes and a reference against
which changes can be compared. Since the outline
26 JÖKULL No. 70, 2020