Geology Reference
In-Depth Information
detection of the two dyes together, it seems probable
that infiltrating waters follow bedding-parallel
fractures for distances up to c. 100 m, but also
migrate vertically down intersecting joints and
small faults. The properties of the 3-D network of
fractures are such that each dye probably spreads
through a volume shaped approximately as an
upright half-cone with its apex at the point of
injection, its axis vertical, and the vertical plane of
bisection of the apical angle lying parallel to the
strike of the limestone beds. The half-angle at the
apex of this cone is about 358. Such strong lateral
dispersion of the tracers implies that infiltrating
water derived from any single point on the surface
must mix extensively with water from other points.
However the fact that DY96 was detected strongly
at only one site indicates that some parts of the
flow network are effectively channeled parallel to
the bedding and that these channeled pathways
encounter relatively few distributary junctions. Such
pathways may be expected to show less mixing, and
seepages fed by them may therefore possess distinc-
tive chemistry and/or hydrological characteristics
compared with others that are derived from better
mixed parts of the network. Such contrasts do in
fact occur between the three seepage sites that
were monitored as part of this work.
Variations in drip discharge rates at three sites
measured at 30-minute intervals across three annual
hydrological cycles are shown in Figure 5 and show
that drip responses to the seasonal pattern of winter
rain differ dramatically at each site. The Gib04a
speleothem site (Mattey et al. 2008) is fed at a low
discharge of 0.04 to 0.06 l/d which shows no
direct relationship to the winter - summer cycle
of recharge or to individual high-rainfall events.
Dye-tracing suggests it is fed by mixed water from
a recharge zone that lies upslope and directly
above the cave chamber where the roof thickness
is 64 m (Fig. 3). The drip rate pattern is consistent
with flow through roof rock which has a capacity
for significant storage and mixing before emerging
at the drip site.
The drip monitoring site in the Dark Rift, situ-
ated at a higher level in the cave system (Fig. 2)
shows a more complex drip response that is highly
correlated with periods of heavy rainfall. The enla-
rged scale in Figure 5 shows that the drip rate has
a base flow component of a similar magnitude to
that seen at a lower level at the Gib04a site, but
this is punctuated by short-lived but intense dis-
charge events which may peak at over 8 l/d for
a few hours. At the beginning of the winter season
there is a six-hour lag between the start of a rainfall
event and the sharp rise in drip rate which
subsequently declines usually to a new, higher
level of base flow. The sharp response may be
termed quick-flow and is suggestive of a channeled
pathway. Each winter season begins with a high
degree of correlation between rain and quick-flow
events but as the wet season develops this relation-
ship becomes less clear, presumably as recharge
fills the fracture network and flow switching
begins to contribute water from other reservoirs.
Neither dye was detected at this site, but Photine
CU was found at two drips a few metres away,
suggesting that the the main recharge zone for
base flow lies directly above the cave chamber
where the roof thickness is around 38 m. Analogy
with the spread of DY96 tracer suggests that the
recharge for the quick-flow component may
lie somewhat upslope and feed the same reservoir
as the baseflow. The pattern of water chemistry at
this site suggests a source that is quite distinct
from both Gib04a and the flowstone site.
The flowstone site is located at the same level
and less than 4 m away from the Gib04a site
(Fig. 3). Here the drip response shows remarkable
annual regularity with peak discharges around mid-
January exceeding 80 l/d. The discharge rate then
steadily declines for the remaining winter season
and by the start of the following autumn the site is
close to drying up. In winter 2005/6 the flowstone
discharge ceased completely until mid-January
but during 2006/7 and 2007/8, although flows
increased erratically at the start of the winter rainfall
season, the peak discharge was delivered in each
case at the same time in mid-January. Dye-tracing
reveals that this water is fed from a recharge zone
near the summit of the Rock, and groundwater
follows bedding places down to the flowstone site.
The unusual cyclical pattern of delayed peak dis-
charge followed by an exponential decline suggests
the operation of a siphon which carries over accu-
mulated early winter rainfall into a second reservoir
that directly feeds the flowstone. During heavy
periods of rainfall the siphon can be by-passed creat-
ing irregular pulses observed before the main dis-
charge event in January. Water reaches this site
via a longer flow path (c. 120 m) and the drip
water, whose composition is also quite different
with higher levels of Ca and total alkalinity (see
below), is depositing flowstone on the cave floor.
Cave water hydrochemistry
The main features of compositional variation in
cave waters are illustrated in Figures 6 and 7. The
Mg/Ca compositions of lake and drip water are
compared on Figure 6a and show coherent trends
consistent with extensive prior calcite precipitation
(PCP) from parental waters having Ca values at
or above the upper limits of the water analyzed
(Fairchild et al. 2000). The rationale for this is that
calcites have much lower Mg/Ca than the waters
Search WWH ::
Custom Search