The solutional growth of karst features involves a simple mass transfer, in which the mass removed from the walls of a void equals the mass removed in solution by flowing water. Mass removed = volume rock density, and mass in solution = discharge solute concentration. Therefore (e.g., in a solution conduit) the rate of volume increase = discharge gain in dissolved load time / rock density. Density is essentially con-stant, so conduit size depends only on the cumulative values of discharge, dissolution rate, and time. All three are essential, and all are equally important.
Discharge in a conduit depends on catchment area and water balance; and the distribu-tion of water among all solution conduits depends on hydraulic variables and conduit geometry. Dissolution rate varies with rock type, undersaturation, and solution kinetics, the last of which can be determined by laboratory and field measurements. Together, they provide a tool for quantifying the local geomorphic history.
These relationships seem simple, but applying them quantitatively is complex. This requires a finely divided 2- or 3-dimensional grid in which each segment varies in dis-charge and dissolution rate within each of many small time increments. Computer modelers use this approach to simulate conduit growh; but the results depend on the specific boundary conditions of the model.
It is more challenging to use this concept intuitively to solve real field problems, where the variables are only partly understood. In this case, one must show that the water source, dissolution rate, and available time are all great enough to account for the ob-served solution features. All three variables are closely linked by a web of interactive processes, all of which can be expressed quantitatively. Whether the goal is to under-stand what is already known, or to predict the unknown, this approach provides a solid basis for interpreting karst systems.