Drip Water Hydrology and Speleothems

Recent research has established speleothems (mineral deposits growing in caves) as one of the most important archives of past environments and climates. Here we are concerned with the most common calcareous types that grow within the unsaturated zone of karstified carbonate rocks (i.e., partly or completely lithified carbonates containing significant post-depositional cavities). The properties of these rocks significantly influence the delivery of water that supply the raw ingredients for speleothem growth, and hence our focus in this article on dripwater hydrology. We focus on the ‘standard case,' where karstification occurs by infiltration of meteoric water that has been acidified by soil-derived carbon dioxide, rather than by fluids rising from depth. The focus on dripwater offers a complementary perspective to the usual emphasis in karst hydrology on the fast transmission of water to conduits emerging at springs that are important for water supply (Ford & Williams 2007).

Carbonate sediments are converted to rocks by sets of diagenetic processes in marine, meteoric, or burial environments. Many carbonates become tightly cemented, and are sometimes dolomitized, during burial, but then are exposed to karstification by freshwater when they are uplifted. When freshwater contacts carbonate, particularly where carbon dioxide pressures are raised in soil zones, calcium carbonate dissolution occurs. Because primary cavities are small and poorly connected (and tend to be absent in pre-Mesozoic carbonates), matrix permeability is low. As a consequence, fluid movement occurs preferentially along lithological boundaries where slightly enhanced porosity is already developed (inception horizons) or fractures generated by small tectonic movements or gravitational disturbance. Significant karstification, including the development of caves, can occur over timescales of 10⁴–10⁶ years (Dreybrodt 1996). In a mature cave system, water drains efficiently underground, leading to unsaturated conditions near the surface, allowing degassing of CO₂-rich percolating water and the formation of speleothems. In these settings, drainage of infiltrating water tends to occur sub-vertically (Figure 1), but is principally guided by networks of fractures (or bedding, where it is inclined). The density of joints (master fractures) is inversely proportional to thickness of beds (Klimchouk & Ford 2000). In well-lithified carbonate rocks, what is often called matrix permeability in reality consist of sets of micro-fractures. This enables water storage to occur such that in some slow-flowing drips, mean water ages deduced from tracer studies may reach periods of years (Kluge et al. 2010) to decades (Kaufman et al. 2003). Figure 1 illustrates a range of scenarios for speleothem development beneath segments of carbonate rock with different properties, Figure 2 illustrates a range of speleothem forms, and Figure 3 exemplifies the importance of fracture flow shown by a line of stalactites.

Figure 1: Some possible flow routes that water can take through the limestone bedrock to a cave.

This figure is based on a hypothetical cave in pre-Mesozoic bedded limestone which has undergone folding and which contains paleokarstic, lithified caves. Stalagmites (A, B, C and D) are all fed by different flow pathways. Water dripping onto stalagmite A comes from the overlying limestone bed. No fissures can be seen in this limestone bed, and it is therefore likely that the stalagmite is mostly fed by diffuse flow, through either the limestone matrix or through very fine fractures. Stalagmite B is fed by a greater proportion of the fracture flow. Flow is likely to be a mixture of relatively fast fracture flow, as shown by the dotted line, as well as slow diffuse flow from the overlying strata. Water at stalagmite C has taken a more complex flow route, which also includes passing through an overlying cave which is full of sediment, labelled ‘1’. The sediment in the cave can hold water and therefore act as a water store. Stalagmite D has the most complicated flow route. As well as passing through two sediment filled caves ('1' and ‘2’), the water also passes through an active, water filled cave (‘3’). All three caves may vary in their size and the way they fill with water and subsequently drain.

© 2012 Nature Education All rights reserved.

Figure 2: Speleothems in Grand Roc Cave, Dordogne, France.

The cave floor is covered with a flowstone, where water intermittently flows from near-right to far-left (solid arrows). A columnar stalagmite a few cm wide is visible at the lower right, which is fed by a relatively low discharge drip source (dotted arrows). Next to this is a broad stalagmite, indicative of seasonal or continuous high discharge dripping, with flow from the stalagmite contributing to the flowstone deposit. Stalactites can be seen in the upper left of the picture.

© 2012 Nature Education Courtesy of D. Genty. All rights reserved.

Figure 3: Line of stalactites, each consisting of a solid upper part and a hollow (soda straw) termination.

The position of these speleothems is guided by a fracture in the overlying Carboniferous limestone (Crag Cave, Ireland). Hydrological conditions at this site have been studied by Tooth & Fairchild (2003) and Baldini et al. (2006).

© 2012 Nature Education Courtesy of Ian Fairchild. All rights reserved.

The only method of determining the actual nature of water movement to an individual speleothem is through empirical observations of discharge within the caves. Early studies used manual or tipping-bucket automated gauges (e.g., Gunn 1974) which enabled the development of the first conceptual models of cave dripwater hydrology. Smart & Friederich (1987), and later Baker et al. (1997) classified waters by their mean drip rate and drip variability. Most stalagmite dripwaters were characterised by a discharge of <10^-5 L s^-1 and a CV (coefficient of variation, i.e., relative standard deviation) of between 10 and 200%. Higher CVs were associated with dripwaters that displayed a significant seasonal variation in flow regime, with an apparent increase in the proportion of seepage waters routed via preferential flow as CV increases. Figure 4 presents a summary of speleothem forming water fluxes after Baker et al. (1997), showing a lack of relationship between flow variability and mean discharge, and significant inter-annual variability in water flux (mean and variability) between years.

Figure 4: Relationship between discharge and the intra-annual variability of discharge compiled in published studies on Carboniferous limestone cave sites in England.

The dataset includes sites where drip waters are both unsaturated and saturated with respect to calcite. Intra-annual variability of stalagmite forming drip waters at Lower Cave, England, measured in 1991–1992 and 1994–1995 are labelled and linked by arrows.

© 2012 Nature Education After Baker et al. 1997. All rights reserved.

Significant improvements to our understanding of water movement to speleothems have come from the development of instruments to continuously measure drip rates in caves. Genty & Deflandre (1998) monitored drip-waters below a stalactite in the Père Nöel Cave, S. Belgium, for more than five years. Each year there was a pronounced period of low-flow in summer coinciding with the development of a soilwater deficit. They also identified periods of water flushing from the soil or epikarst, indicating piston, or displacement flow, with sequential water movement through linked water stores. Additional high frequency, low amplitude variations in drip-water discharge were inversely correlated with air pressure, suggesting two-phase flow. Baker & Brunsdon (2003) continuously monitored six drip sites for up to four years at Stump Cross Caverns, UK, and also observed two-phase flow and non-linear drip rate behaviour. Figure 5 presents a drip logger data from multiple sites at Cathedral Cave, Wellington responding to a single recharge event. Non-linear behaviour of water movement is widely observed across these studies at drip rates of ~1 drip/minute or higher, and has also been observed at the much greater discharge observed in karst springs (Labat et al. 2000), suggesting that this behaviour was possible at a wide range of water movement rates. One might hypothesize that non-linear behaviour in speleothem-forming drip water discharge is to be expected where water movement is affected by secondary porosity.

Figure 5: Four different cave dripwater responses to a rain event in July 2010 at Wellington, New South Wales, Australia.

The top graph is the amount of rainfall recorded at Wellington by the Bureau of Meteorology. The lower four graphs show different dripwater responses (the number of drips in a 15 minute period) ranked in order of mean discharge. From top to bottom: a delayed, ‘overflow’ response requiring a ground water store to be filled followed by a subsequent fracture-low response; an immediate, fracture-flow dominated response; a damped and lagged response; drip with no response, small changes in drip rate are responding to atmospheric pressure changes.

© 2012 Nature Education All rights reserved.

Recently, quantitative hydrological models of water movement to speleothems have been developed (Fairchild et al. 2006, Baker et al. 2010, Baker & Bradley 2010, Bradley et al. 2010). To date, these have taken the form of simple lumped parameter models, where the hydrology is represented by one or more interconnected reservoirs (Figure 6). The reservoirs represent stored water, which may be within dissolutionally widened fractures or overlying caves, which may be water-, air-, or sediment-filled. Typically, water is routed directly between reservoirs to represent fractures. Where high matrix permeability or dense networks of micro-fractures occurs, this is typically modelled with a proportion of flow with a lagged and damped response of the recharge. These simple hydrological models can capture some of the basic characteristics observed by automated drip loggers such as non-linear flow responses to water input, but more complex models are still needed to successfully model dripwater chemistry (Fairchild et al. 2006).

Figure 6: Conceptual diagrams of numeric models of karst water movement to stalagmites.

(a) Two-layer model of Fairchild et al. (2006). (b) Single reservoir with overflow feed model of Baker et al. (2010). (c) Single reservoir model with underflow feed of Baker & Bradley (2010) and (d) lumped parameter model of Bradley et al. (2010). See original articles for explanations of the symbols used.

© 2012 Nature Education All rights reserved.

In many parts of the world, water supply to speleothems over time is non-linear due to recharge occurring only when the precipitation>evapotranspiration threshold is passed, although localized recharge through macropores is always possible. Coupled with the complexity of subsurface flow routes, such as overflow behaviour from storage reservoirs, water movement to speleothems can be expected to display non-linear behaviour. This could be on a seasonal basis, where dripping may even cease, or on the longer-term. Where a stalagmite is fed by a substantial amount of long-lived stored water, candlestick-shaped stalagmites may result: this stored water means that these stalagmites may have a damped influence from the surface atmosphere, soil, and epikarst. For speleothems fed by both reservoir and fracture components, a long growth duration might be expected if fed by a significantly large reservoir compared to water outlets, but with decreasing reservoir size, non-linear water supply is increasingly likely, ultimately leading to seasonal or episodic growth and more complex growth shapes.

The effects of hydrological change are seen in the morphology and in the physical and chemical microstructure of speleothems. Figure 7 illustrates a set of stalagmites from Refugio Cave, Spain, a cave less than 15 m below the surface within fractured dolomitic marbles (Mudarra et al. 2009). The density of stalactites suggests that the drip water hydrology is dominated by matrix or abundant micro-fractures rather than fracture flow. Storage is limited, however, since flow of many drips ceases in the summer. On the left of the field of view are a series of ‘candlestick' shaped stalagmites. They show a common pattern of variation in diameter, with wider portions reflecting higher drip discharge (Kaufmann 2003), and in cross-section have growth hiatuses associated with narrower portions. This illustrates the clear role that hydrology makes to maintaining the growth and controlling the diameter of the stalagmites. Figure 8 illustrates successive annual fluorescence laminae in a stalagmite from Uamh an Tartair, Scotland, indicative of enhanced flushing of organic matter from the overlying peaty soils into caves during the autumn season. This cave is less than 20 m from the surface and formed in Cambro-Ordovician dolomite (Fuller et al. 2008). Matrix porosity would be negligible and drip water hydrology dominated by fracture flow. Stalagmites at this site have been shown to contain thousands of years of continuous fluorescent annual laminae, which requires the presence of both a fracture flow component to rapidly transport discrete fluorescent laminae, and sufficient storage to maintain drip rate (Proctor et al. 2000). Finally, Figure 9 illustrates chemical data illustrating a pattern of covarying increases in magnesium in strontium in a cave (Clamouse, southern France), which displays a characteristic response of prior calcite precipitation in response to seasonal and longer-term droughts (McMillan et al. 2005). The gradual nature of the changes reflects the seepage flow from abundant secondary matrix porosity that has been generated by partial dedolomitization of the host Mesozoic carbonates.

Figure 7: Set of stalagmites with diameter varying with height.

This illustrates a common forcing by varying drip rates over time, higher drip rates leading to a widening of speleothem growth (Kaufmann 2003). Refugio cave, Southeast Spain.

© 2012 Nature Education Courtesy of Ian Fairchild. All rights reserved.

Figure 8: Photograph (0.6 mm high) of optically-stimulated fluorescence of a sectioned stalagmite.

Photograph (0.6 mm high) of optically-stimulated fluorescence of a sectioned stalagmite showing up to 12 luminescent growth zones corresponding to autumnal flush of soil organic matter into the cave. Tartair cave, stalagmite SU96-7 as in Proctor et al. (2000).

© 2012 Nature Education Courtesy of Ian Fairchild. All rights reserved.

Figure 9: Traverse across several annual layers in a stalagmite from Clamouse Cave, southern France.

Illustrates seasonal enrichments in Sr and Mg interpreted to correspond to dry summers (replotted from data in McMillan et al. 2005). Arrows point to successive winters.

© 2012 Nature Education All rights reserved.