Geology Reference
In-Depth Information
from which they form such that the distribution
coefficient (K) is small:
derived. Soil air has lower pCO 2 values (no more
than 10 21.4 ( 4% by volume) according to
Fig. 6a) and indeed variability in soil pCO 2 is
expected
K Mg ¼ (Mg=CaÞ water =(Mg=Ca) calcite ¼ 0:02
(at 17 8C; Huang & Fairchild 2001)
from
its
highly
varied
thickness
and
moisture content.
Figure 6b illustrates Sr data which also illustrate
the PCP effects. These data are more scattered at
each site which is probably at least partly analytical
because of the very low Sr concentrations (typically
0.1 ppm) encountered. There is a less marked
increase in Sr/Ca as Ca falls compared with
Figure 6a. This is due to a relatively high value for
K Sr (a value of 0.3 was used to construct the lines
in Figure 6b; Huang & Fairchild 2001).
Figure 7 shows monthly variations of cave water
pH, total alkalinity, Ca, Mg/Ca, and d 13 C DIC ,in
relation to drip discharge and cave air CO 2 levels
from late 2004 to late 2008/early 2009. An
obvious feature of the data is the strong seasonal
pattern of variation where drip water collected in
winter months has lower pH, highest total alkalinity,
highest Ca, lowest Mg/Ca, and lowest d 13 C DIC .
Another striking feature is the coherent pattern of
variation displayed by water collected at drip sites
which are each significantly different from each
other in terms of flow paths (Fig. 3) and discharge
patterns (Fig. 4). The flow stone drip water has a
distinctive composition having higher Ca and total
alkalinity levels yet shows the same seasonal
patterns of variation; the lake water reservoir also
displays similar, but more attenuated seasonal vari-
ations in hydrochemistry. This commonality in sea-
sonal behaviour suggests a common control that
must be located within the cave, because it affects
waters with such diverse hydrology and chemistry.
(1)
Hence as calcite is precipitated along the flowline,
Ca changes much more than Mg and drip water
compositions evolve following curves on Figure 6
such as the two PCP model lines illustrated. The
highest-Ca waters along a trendline represent
waters that are least modified and their pCO 2
value constrains the minimum value encountered
by the waters along their flow route. The highest-Ca
points are those that feed the flowstone, which has a
recharge zone at a higher altitude and is delivered
to the drip site along bedding planes via a siphon.
The Ca concentrations reach over 140 mg/l Ca,
corresponding to equilibrium at a pCO 2 of .10 21
( 7% CO 2 by volume), and since the cave air has
a much lower pCO 2 values, degassing leads to a
high degree of supersaturation and vigorous precipi-
tation of calcite as flowstone. Lake water samples
collected monthly over the monitoring period are
generally more constant in composition relative to
drip water in terms of Ca (around 80 mg/l) and
with initial Mg/Ca ratios similar to drip water
from the Gib04a site (Fig. 6a). The Gib04a drip
water shows a large seasonal decrease in Ca to
below 40 mg/l accompanied by a shift in Mg/Ca
to .1500 as a result of seasonal calcite precipitation
during the summer months (Fig. 7).
Drip waters from the Dark Rift are offset from
the other data samples. Although this could reflect
a source with a lower ratio of dolomite to calcite,
it is notable that the lowest Mg/Ca ratios are very
similar for the Dark Rift waters and the flowstone
waters and both may represent the original compo-
sition of bedrock being dissolved. These minimum
1000*Mg/Ca values are around 400, corresponding
to a molar Mg/Ca ratio of around 1/3 suggesting
dissolution of 2/3 dolomite and 1/3 limestone.
Note that the proportion of limestone bedrock
would be expected to be much lower than this
because of faster calcite dissolution (Fairchild
et al. 2000). The reason for the lower Ca values of
the Dark Rift could be that its water acquires its dis-
solved CO 2 from soil air rather than the more
CO 2 -rich source from which Gib04a is evidently
Cave air CO 2 and ventilation regimes
CO 2 mixing ratios in cave air sampled monthly from
NSM at the GIb04a sites and from the lake area are
shown on the lowest graph in Figure 7. As for the
temperature and humidity variations discussed
above, the distribution of cave air CO 2 also shows
complex patterns of spatial as well as strong tem-
poral variation which are clearly related to regular
diurnal and seasonal cycles modified by local
synoptic scale weather conditions. Overall, cave
air CO 2 mixing ratios show very regular seasonal
variations with highest concentrations in winter
Fig. 7. (Continued) Relationships between rainfall amount, drip discharge, dripwater composition and cave air pCO 2
measured between 2004 - 2009. Grey bands mark the period between mid-November and mid-April when cave air
pCO 2 is highest. From top to bottom: daily rainfall; drip discharge rates from Figure 5 plotted together on a log
scale; monthly dripwater pH, total alkalinity, Ca concentration, 1000*Mg/Ca and d 13 C DIC . Lower plot is the
concentration of CO 2 in deep cave air, measured on monthly spot samples at the Gib04a (black circles) and Lake
(grey circles) sites along with data (unpublished) obtained by continuous logging at 2 hr intervals measured at the
Gib04a site. Meteorological data # The Met Office, UK.
Search WWH ::




Custom Search