Jökull - 01.12.1972, Síða 59

taken to be 10%. According to measurements
on glaciers in Norway this was a reasonable
choice (Liestj/l 1969, pers. comm.). In calculat-
ing volumes of releasecl meltwater (q) from
the surface layer regard was paid to the de-
crease of the melting area during the observa-
tion period.
As discussed before the computed daily values
of the ice phase ablation are not believed to
be quite reliable due to imperfect circum-
stances for estimation of the turbulent encrgy
fluxes at the glacier surface. Short time values
of ablation measured directly from stakes are
not reliable either (see Fig. 3). The free water
content was not measurecl and measurements
of the density variations were quite rough. But
by summing up the q values for the whole
period (July 1 to August 5) one obtains a value
of 2.06- 10° m3 which corresponcls fairly well
with the change in ice and snow storage AS =
2.16 • 10cm3 ± 10%. This storage AS was deter-
mined by the glaciological method and includes
the liquid water storage in the surface layer
above the summer surface (Björnsson 1971, p.
20). Due to this agreement for the whole mid-
sumrner period it is assumed that the calculat-
ed q values in Fig. 8 represent the time varia-
tions in the storage averaged over shorter time
intervals like a week or so.
Fig. 8 shows a marked change in the re-
sponse of the hydrological system to the in-
put of meltwater and rain. After July 15 the
response time is short and there is a high
correlation between daily values of input and
output. It is characteristic for this period that
the melting was continuous. Melting had then
lastecl for long time and presumably most of
the meltwater passages were opened and only
small amounts of meltwater becarne stored in
the glacier. Linear system analysis would be
meaningful for this period. A quantitative
synthesis of the water balance components fails
however. This might he due to run-off of de-
layed meltwater and an underestimate of rain
volumes. The latter is obviously the case dur-
ing the peak in the last week of July.
In the first fourteen days of July the system
hehaves quite differently. Up to that time
periods of melting had been frequently abrupt-
ed by freezing and the generated meltwater
percolated deep into the glacier or drained
Fig. 8. Water budget components in 1968 for
the drainage basin (7.5 km2) above the limni-
graph. About 23% of the basin is covered with
the glacier. Daily variation of run-off in the
Bægisá-river, and volumes of measured rain
and computed meltwater.
Mynd 8. Dagsgildi i vatnsbúskap sumarið 1968.
Rennsli Bœgisár, mœlt regn ofan rennslisrnæli-
stöðvar og reiknuð leysing á Bœgisárjökli.
out. In the last week of June tliere had been
frost and a part of the precipitation on July 1
fell as snow when the air temperature fell for
some hours below 0 °C during the night time.
The same was the case on July 11 (see Björns-
son 1971, fig. 6). In both cases an isothermal
state of 0 °C was rapidly restored in the glacier
hut the input of rain was abruptly damped
out. The “soil moisture deficiency” was less
than the volume of infiltrated water. In the
whole first half of July tliere is no clear cor-
relation between input and output. At the same
time there is an obvious excess in run-off which
must be explained as a delayed rain and melt-
water from a previous melting period. A longer
record of the water budget components, by
preference for the whole hydrological year
would have been iteeded to determine the delay
of the liquid water storage in the glacier.
The results from Bægisárjökull illustrate that
water balance measurements should be used
with care in estimating glacier mass balance.
The same concerns the use of any linear system
analysis (e.g. multiple linear regression) in for-
casting glacier run-off for the summer as a
whole.
Much theoretical work remains to be done
JÖKULL 22. ÁR 57