Jökull - 01.01.2011, Qupperneq 5

Daily temperature in Iceland 1949–2010
Table 1. Average number of stations within 50 (100) km of group-2 and group-3 stations. – Fjöldi veðurstöðva
innan 50 (100) km frá stöðvum í flokki 2 og flokki 3.
Statistics Min 25% Median Mean 75% Max
Nb. of G-1 st. near G-2 st. 0 (1) 1.4 (3.9) 2.6 (5.4) 2.6 (5.8) 3.5 (7.2) 6.4 (11.4)
Nb. of G-1 st. near G-3 st. 0.7 (0.8) 2.5 (4.5) 3.2 (6.1) 3.4 (6.2) 4.0 (7.6) 6.8 (15.5)
Nb. of G-1 and G-2 st. near G-3 st. 2.9 (7.9) 5.8 (14.0) 7.5 (17.7) 10.4 (18.5) 15.2 (22.0) 27.2 (38.2)
Table 2. Station elevation statistics (in m a.s.l.). – Hæð veðurstöðva (í m y.s.).
Statistics/Period 1950 1960 1970 1980 1990 2000 2010
25% (Group-1) 14 12 13 14 15 16 14
Median (Group-1) 23 20 27 27 25 25 25
Mean (Group-1) 48 50 75 76 73 71 60
75% (Group-1) 44 39 80 79 83 52 50
Max (Group-1) 384 450 672 641 641 641 384
25% (Group-2) NA NA NA NA NA 10 16
Median (Group-2) NA NA NA NA NA 48 52
Mean (Group-2) NA NA NA NA NA 230 219
75% (Group-2) NA NA NA NA NA 550 450
Max (Group-2) NA NA NA NA NA 949 949
25% (Group-3) NA NA NA NA NA 40 44
Median (Group-3) NA NA NA NA NA 267 226
Mean (Group-3) NA NA NA NA NA 252 231
75% (Group-3) NA NA NA NA NA 397 370
Max (Group-3) NA NA NA NA NA 600 600
temperature fields should not be interpreted as repre-
senting the actual 2-m temperature over glaciers on
warm days when melting of snow or ice takes place.
Rather, the gridded temperatures may be interpreted
as an estimate of the 2-m air temperature over a hy-
pothetical ice-free surface at the location and altitude
in question. It could be argued that the tempera-
ture estimates obtained here for the glaciers should be
masked out from the gridded temperature fields be-
cause they do not represent an unbiased estimate of
the in-situ 2-m temperature that would be measured
by a weather station located on the glacier surface as
explained above. Rather than doing this, we provide a
mask on the same grid identifying the glacier-covered
areas of Iceland so that the users can mask out tem-
peratures over glaciers if they so choose. The tem-
perature estimated here for the glaciers based only on
measurements outside the glaciers are as mentioned
above useful for various modelling purposes and are
therefore a valuable part of the data set. We also note
that other gridded temperature data sets for Iceland
(e.g. the data set of Björnsson et al., 2007) are also
based on measurements from outside the glaciers and
will therefore be similar to our data set in this regard.
Gridding method
The method used to construct the gridded temperature
fields is similar to the one used by Dodson and Marks
(1997), the so-called linear lapse rate adjustment. It
combines a spatial interpolation technique with an el-
evation correction based on a spatially constant lapse-
rate. This method has been used at IMO over the
past 10 years to produce 3-hourly temperature maps in
near real time for the purpose of weather monitoring
(unpublished work). First, as temperature is known to
be influenced by elevation (Daly, 2006), elevation dif-
ferences between stations are corrected by estimating
JÖKULL No. 61, 2011 5