The study of how cave drip-water discharge responds to recharge events is fundamental to evaluating the potential of actively forming speleothems as high-resolution climate archives. Most previous research has focused on caves of the Northern Hemisphere middle latitudes, where recharge is strongly seasonal. Few studies have explored drip-water behaviour from regions where the expected seasonal rainfall pattern is significantly perturbed on an irregular basis by changing regional atmospheric circulation patterns. Here, we report the results of a 4-year study of cave drip-water-climate relationships from two caves in eastern Australia. The discharge of 10 drip sites located beneath bedrock thicknesses of 12, 22 and 45 m was monitored either continuously (using automated infrared sensors) or at discrete approximately monthly intervals and compared with local rainfall and water balance data. The study period traversed two major droughts, including the severe 2002-2003 El Niño. Drips at 12 and 22 m depths responded almost simultaneously to individual recharge events, although the time lag between individual events varied according to the volume of recharge and pre-event storage. Overall, a steady decline in discharge is evident through the moisture-deficit period, with increased flows through phases of positive water balance. Speleothems growing at these and similar shallow-chamber sites have potential for reconstructing palaeo-rainfall trends at high-resolution, although the highly variable nature of year-to-year recharge would make it difficult to obtain data on a calendrical time-scale. Drips at 45 m depth did not respond consistently to individual recharge events and displayed hydrological behaviour markedly dissimilar to one another and to the near-surface drip sites, indicating great complexity in karst architecture and the absence of fissure flow. Although speleothems at this depth may well preserve information on longer-term rainfall trends, their potential to encode a palaeo-rainfall variability signal at interannual resolution is poor.
Hydrological Processes Vol. 21, Issue 13, p. 1737-1748