Abstract:

The 1969-77 data confirm that groundwater temperature is significantly higher than air temperature at mean catchment altitude but provide only partial support for an explanation in terms of soil temperature and insulation of drainage from cold air ponding over the Plain. Higher pH of output than input streams is attributed mainly to percolation water chemistry. Water chemistry of two contrasted input streams suggests non-karst rock weathering has an important effect on allogenic input streams. An inverse relationship between carbonate hardness and output discharge is found again and attributed mainly to faster transit through the limestone at high flows. Summer has a steeper regression than winter due to precipitation and high flows depressing carbon dioxide and carbonate concentrations more in that season than in winter. Picknett graphs show how solutional capacity varies through the hydrologic system, with aggressive input streams, mainly saturated percolation water, and rarely saturated output springs because of the allogenic component in the last. The total carbonate load of Cave Creek is directly related to discharge, with little seasonal difference so the annual regression is chosen for later calculation. When the carbonate load duration curve and frequency classes for Cave Creek are compared with those for other karsts, it falls into an intermediate class in which neither very high nor low flows dominate the pattern. This is attributed to a combination of a large allogenic input with a complex routing pattern. Consideration of most input stream solute concentration on one occasion indicates such close dependence on catchment geology that doubt is cast on the smallness of the 1965-9 allocation of carbonate contribution from non-karst rock weathering to the allogenic input. This is explained by new CSIRO rainfall chemistry figures from the Yass R. catchment which are smaller than those used before and by elimination of a previous error in calculation. This time subtraction of atmospheric salts is done on a daily basis with a decaying hyperbolic function. Correction of Cave Creek output for allogenic stream input follows the method adopted in 1965-9 but on a firmer basis, with the assumption of approximately equal water yeild per unit area from the non-karst and karst parts of the catchment being more factually supported than before. It remains a substantial correction. The correction for subjacent karst input to Cave Creek is also improved by putting the calculation in part on a seasonal basis; it remains small. The exposed solute load output shows the same seasonal pattern as was determined earlier, with a winter/spring maximum, and it again evinced much variation from year to year. So did annual rates. The mean annual loss of 29 B was slightly greater than for 1965-9. If this difference is real and not an experimental error, the reduced allowance for atmospheric salts and greater annual rainfall in the second period could explain the increase. This erosion rate of 29 B from an annual runoff of about 400mm places this karst where it would be expected in the world pattern of similar determinations in terms of both runoff and its proximity to the soil covered/bare karst dichotomy of Atkinson and Smith (1976). Combined with the other work at Cooleman Plain on erosion at specific kinds of site, an estimate of the spatial distribution of the limestone solution is presented. It agrees well with the similar attempt for Mendip by Atkinson and Smith (1976), when allowance is made for certain differences in method and context. The main conclusions are the great role of solution in the superficial zone and the unimportance of the contribution from caves. Conflict between this process study and the geomorphic history of Cooleman Plain remains and once again an explanation is sought in long persistence of a Tertiary ironstone cover inhibiting surface solution.