Discussion
Figures 10 and 11 illustrate the implications of the model when extrapolated to longer slopes and more extreme rainfalls. Using the same parameter as in the field data, runoff coefficients are estimated as a function of storm rainfall for slopes of 5 to 50m in figure 12, where they are compared with SCS curve number relationships. Over the storm sizes seen in the field data, the greatest relative divergences between the two approaches are found for the large number of storms of less than 30 mm and with less than 10% runoff, for which the SCS method consistently underpredicts the small volumes of runoff.
The importance of the proposed alternative runoff estimate is not, however, seen to lie in the quality of fit to individual data sets, all of which show wide variations that can be contained within either form of expression. What seems much more important and useful is that the proposed expression explicitly includes scaling for both rainfall intensity and slope length, providing a model with much greater possibilities for transference across scales and between sites and climates. Experimentation within the model environment also shows that the parameters in equation (4) also responds rationally to changes in infiltration parameters and their spatial variability, to gradient and to micro-topography expressed through the potential for locally divergent flow.
The potential to apply a model at different spatial scales within a catchment is of value, not only in support of field experiments but also to distribute runoff and sediment transport within a field area or within a landscape evolution model.