Jökull - 01.12.1990, Page 130

Figure 8. Chemistry of precipitation in Keldnaholt
1974-1982. —Efnainnihald í úrkomu í Keldnaholti
1974-1982.
following discussion is based mainly on an unpub-
lished treatise of the analyses data up to and including
1982 (Sigurðsson, 1985 b), Figs. 7 and 8.
The concentration of each component is varying
from month to month and between years for any month
of the year. The same applies to the ratios between the
components. This will not be dealt with any further
here, but mean values for each month over a number of
years show a rather smooth curve of distribution over
the year, with a high in winter and a low in summer
for most of the components (see Fig. 8).
The winter high is related to a number of factors,
the weight of each of them not being known (wind ve-
locity, temperature of the sea, distance from the shore
and from the oceanic source etc. Matthess, 1973;
Sigurðsson, 1985a). A deductive factoral analysis of
the chemical composition is therefore at present not
possible. Despite the wide variations in the ratios
between components for every month, there is a cer-
tain correlation obviously present. Chloride is the
strongest component in sea water as well as in the pre-
cipitation. The distribution of chloride in groundwa-
ter indicates, that it is predominantly of marine origin
(Sigurðsson and Einarsson, 1988). If the ratios of the
other components to chloride in the precipitation are
known, it is possible to correct for the marine compo-
nent in the groundwater, thus restricting the composi-
tion to hydrogeological factors.
Plotted against chloride, the other components
show a wide scatter of values (points). This scatter
has as a rule a rather regular lower limit, which can
be represented by a curved line. This line is thought
to represent a basic marine factor, the higher values
being caused by industrial smoke, dissolution from
mineral dust and extreme meteorological or oceano-
graphic conditions. This theory can not be proved
(see above) except by testing. The lower-limit curve
can be approximated by a polynomial. The contribu-
tion of the higher members appears to be numerically
minimal but the errors to be great. The linear and the
quadratic members can be reasonably calculated, but
only the linear one can so far be regarded as having a
tolerable accuracy. A pure linear approximation must
disregard the highest chloride values, but represents
probably the most proper fit at present (Sigurðsson,
1985b). The correction coefficients, thus obtained,
are in good agreement with those by Gíslason et al.
(1990), which were obtained from a different set of
data and with a simple, linear correlation, and which
were also used to a similar purpose as in the present
article.
This linear ratio between chloride and the other
components can be used as a correction coefficient to
subtract the contribution of the marine factor in the
precipitation to the chemistry of the groundwater. It is
then further assumed, that the concentration of solutes
through evapotranspiration at the surface is similar
for all components see Fig. 9. In the following the
country-wide distribution of various components will
be discussed, which have been corrected in the above
described way. These distributions appear to be very
126 JÖKULL, No. 40, 1990