Subsurface Hydrograph Analysis by Convolution

Abstract
Subsurface flow from an 0.8-acre watershed in the Southern Coastal Plain at Tifton, GA, is intercepted and gaged independently of any surface flows. The hydrographs of seven flow events following significant rainfall were analyzed by the nonlinear technique of two-stage convolution. The convolutional model is based on two conceptual components. A characteristic function represents the three-dimensional storage capability of the soil profile and thus could be visualized as a map of the flow potential of the drainage area. A state function expresses the movement, or routing, of the characteristic function to the outlet. This routing is the first-stage convolution and produces a unit response function. Second-stage convolution of the response functions with effective rainfall produces the outflow hydrograph. The consistent results should allow identification of the parameters, and thus the two quasiphysical model components, with physical features of the drainage area.