Jökull - 01.01.2019, Side 118

Remote sensing of snow patches on Tröllaskagi Peninsula, N-Iceland
represent a general wind pattern in the study area.
The prevailing wind for each year, starting in 1994
until 2017, is determined using wind speeds higher
than 10 m/s, air temperatures lower than 2◦C. Analy-
sis is restricted to data from September to May, when
snow drift is most likely to occur. For example, from
September 2016 until May 2017 the main wind direc-
tion was from the east. Northerly wind on the other
hand, bring most precipitation to the northern part of
the peninsula, this could be a reason for low winter
precipitation in 2016 and 2017 and a reason for the
limited snow patch cover in these years. The area of
snow patches in the study areas is relatively large in
the period 1999–2002. The prevailing wind direction
at that time was from the north, which may be ex-
pected to have caused elevated winter precipitation.
To show that not only temperature and precipi-
tation have an influence on the distribution and size
of snow patches, we’ve calculated the thaw degree
day (temperature >0◦C) for each weather station and
study area. Here we present the results from the
weather station in Ólafsfjörður. The number of days
between subsequent satellite images was normalized
and plotted together with the area difference relative
to the previous year (Table A4). For each tempera-
ture data set the mean daily temperature (Tmean) was
calculated and days with positive Tmean used to cal-
culate the thaw-degree-days: the sum of the positive
mean daily temperatures (
∑
Tmean). The
∑
Tmean
divided by the total number days resulted in the rela-
tive amount of thaw-degree-days. Furthermore, thaw-
degree-days were calculated with lapse rate consider-
ation. For each temperature data set, the mean daily
temperature (Tmean) and a temperature lapse rate (-
0.65◦C per 100 m height difference), which considers
the height difference between the height of the me-
teorological station and the mean height of the PSP
of the respective area, were calculated. The temper-
ature lapse rate is added to Tmean, leading to the
mean daily temperature at the mean height of the PSP,
TPSP
mean, followed by separating positive TPSP
mean. The
thaw-degree-days are the sum of the positive mean
daily temperatures at the mean PSP height (
∑
TPSP
mean).∑
TPSP
mean divided by the total number of days re-
sults in the relative amount of thaw-degree-days at the
mean PSP height. Many snow patches have the same
size or are even larger with thaw-degree-days up to
50% relative to the total number of days (e.g. Sakka
from 2001 to 2013).
In each study region avalanche channels were
identified and calculations of snow patch areas show,
that it is important to exclude them from the over-
all snow patch distribution. For example, the snow
patches in avalanche channels are up to 75% of
the snow patch extent in the Kerling area (period
1984–2017). In Almenningar and Úlfsdalir the snow
patches in avalanche channels range between 25–50%
of the total amount in the time period from 2014–
2016. In Brimnesdalur the avalanche snow patches
range between 3 and 10% of the total amount. In the
Kerling area the snow patches in avalanche channels
are different. The smallest areas are between 17 and
25% and the largest is about 75%, due to general very
little snow extent. In Búrfellsdalur the snow patches
in avalanche channels range between 5 and 13% and
in the Sakka region the snow patches in avalanche
channels are always less than 1%. The avalanche
snow patches remain over the summer due to the to-
pographic situation and are not representative for re-
gional permafrost occurrence.
During the field work a good overview of the
study areas was obtained. Field observations in sum-
mer 2016 and 2017 showed a 10–25 cm thick ice layer
at the base of the PSPs in the Kerling, Brimnesdalur
and Sakka. In Sakka an up to 25 cm thick ice layer was
measured in early August 2016. The ice at the bot-
tom of this PSP was in solid state, completely frozen
to its substratum, and the surrounding ground was
completely frozen 10–30 cm below its surface. How-
ever, the snow patch melted completely at the end of
September 2016, probably as a result of warm sum-
mer temperature and low winter precipitation which
caused less snow accumulation in the area than usu-
ally. Whether the ground remained frozen through-
out the year is unknown. Such ice layers at the base
of PSPs might be a result from refreezing of per-
colated meltwater or a result of winter rainfall and
are mentioned as typical within PSPs in other stud-
ies (e.g. Furrer and Fitze, 1970; Furrer, 1955; Allen,
1998; Kawashima et al., 1993). Frozen ground be-
JÖKULL No. 69, 2019 117