Jökull - 01.01.2005, Blaðsíða 144

L. A. Rasmussen
Table 4: Sensitivity (m w.e.) of summer balance bs and net balance bn to +1◦C temperature change and to 10
percent increase in precipitation flux F . Model numbers (1,2,3) are identified in Table 3. – icelandic
−dbs/dT −dbn/dT dbn/dF
Glacier 1 2 1 2 3 1 2 3
1 Tungnaárjökull 0.61 1.25 0.90 1.56 0.77 0.15 0.24 0.30
2 Köldukvíslarjökull 0.62 0.80 0.82 0.93 0.50 0.15 0.19 0.28
3 Dyngjujökull 1.03 1.75 1.23 2.01 0.49 0.17 0.18 0.28
4 Brúarjökull 0.93 1.49 1.18 1.67 0.68 0.15 0.16 0.30
5 Eyjabakkajökull 0.97 1.36 1.26 1.76 0.74 0.18 0.34 0.27
and 5) the r(bw) is large, and when the difference is
large (such as between glaciers 2 and 4 or glaciers 2
and 5) r(bw) is small. An exception is that between
glaciers 3 and 4 or glaciers 3 and 5, the φ′ difference
is large and so is r(bw).
By contrast, glacier to glacier correlation of bs is
much more uniform. The standard deviation of bs is
also larger than of bw (Table 1), indicating that vari-
ations of bn are more influenced by those of bs than
by bw. Also, year to year variations of bs are spatially
more coherent than those of bw (Table 2).
RECONSTRUCTED MASS BALANCE,
1958-2003
For each glacier, coefficients of Equations (4) and (5)
obtained over the period of record were applied to
upper-air data for the 1958–2003 period. Cumulative
b∗
n
since 1958, which is formed from b∗
n
= b∗
w
+ b∗
s
,
is shown in Figure 5. The total 1958–2003 change
∫
b∗
n
shows a regular variation spatially (Figure 4):
strongly negative in the west, positive in the north,
strongly negative in the east.
Over 1973–1992, changes in cumulative bn are
consistent for four of the glaciers (Figure 6) with
changes in terminus position reported by Williams
et al. (1997). The major exception is Brúarjökull
(glacier 4), which surged 8 km in 1964. Perhaps its
strong retreat over 1973–1992, when there was little
change in cumulative bn (Figure 5), was the result of
strong ablation of the large amount of mass deposited
at low altitude during the surge. A minor exception is
Köldukvíslarjökull (glacier 2).
A two-part piecewise-constant function was fit to
each of the five b∗
n
(t) series. The best-fitting function
has one constant value over the first part of the series
and another constant value over the rest of it, the year
of shift between the two parts being found empirically
to produce the best fit overall.
Figure 5. Reconstructed mass balance
∫
b∗
n
cumu-
lated since 1958. Glaciers are identified in Table 1.
– icelandic
144 JÖKULL No. 55