Jökull - 01.12.1969, Qupperneq 56

decades when J and E is available. The best
values of c obtained in this manner seem to
be:
TABLE 6
Jan Feb Mar Apr May Jun
-0.9 - 1.2 - 1.5 - 1.7 - 1.7 - 1.1
Jul Aug Sep Oct Nov Dec
-0.3 0.3 1.2 1.4 0.5 -0.2
We have thus determined all tlie necessary
constants and coefficients in Eq. 2. This equa-
tion can then be solved by the graph in Fig. 6.
Here the monthly ice forecast may be read
from the sloping lines on the right hand side,
knowing J and E and the season. As an
example we have the JM-temperature 0.8 and
the value E = 10.5 mb for January and we
want the forecast for February. The JM-temp-
erature 0.8 is followecl upwards from the scale
in the lower left until the beginning of Febru-
ary. From there we follow a horizontal line to
the right till we find the value E = 10.5 mb.
There we read the ice forecast 0.1 for Febru-
ary, i.e. some 3 days of ice incidence.
This method has been tested on 352 months
in the last decades, giving the following re-
sults:
TABLE 7
Estimated Actual ice incidence
ice incidence 0 .1—.5 .6-1.0
0 252 15 3
.1—.5 38 21 2
.6-1.0 3 3 15
This table is not directly comparable with
the forecasts made at the end of November
(Table 4) since in that comparison we had
much more data available, generally in icefree
years, yielding a great many forecasts of ice
incidence 0, and most of those months were
also actually icefree. Considering this the
monthly forecasts are probably better. Of 20
cases with heavy ice, 0.6 to 1.0, 15 forecasts are
correct (75%) whereas only 11 were correct in
the former case. Of 39 cases witli light ice,
1—5 tenths of a month, 21 forecasts are success-
ful in Table 7 (54%) whereas only 38% were
correct in Table 4. However, we liave now 6
cases with entirely wrong forecasts, consider-
able ice being forecast and no ice experienced,
or vice versa. Of the forecasts in Table 4 only
3 were entirely wrong.
But even if this inclusion of the west wind
north of Northwest-Iceland improves tlie fore-
casts, it is evident that without taking the JM-
temperature into account in the monthly fore-
casts they would be of very little value. We have
as an example checked the westwind in all
months preceding months with ice incidence
1.0. In four cases the west wincl component
was above normal, but in four cases it was be-
loio normal for the season.
ACKNOWLEDGEMENT
This stucly was initiated by a request of a
special Ice Committee of the Althing, the Ice-
landic Parliament.
REFERENCES
Biidel, J. 1950: Atlas der Eisverhaltnisse des
Nord-Atlantischen Ozeans, Hamburg.
Meteorologisk Aarbog, 1920—1968. Copenhag-
en.
Nansen, F. 1924: Blandt Sel og Bjbrn. Oslo.
Norsk Meteorologisk Árbok, 1920—1968. Oslo.
Slefánsson, U. 1962: North Icelandic Waters.
Reykjavík.
52 JÖKULL 19. AR