Geoscience Reference
In-Depth Information
way of calculating the volume of runoff production from a particular rainfall (Horton, 1933), he also had
a hydrological laboratory in his wooded back garden in Voorheesville, New York State (Horton, 1936)
where he would surely not have observed infiltration excess overland flow very often. Beven (2004a),
using data collected by Horton on infiltration capacities and local rainfall statistics, has suggested that
overland flow might have occurred on the LaGrange Brook catchment only once in every 2-5 years. There
is also some experimental evidence that suggests that some of the overland flow collected on runoff plots
at the site might have been generated by a return flow mechanism (Beven, 2004a). Horton was an excellent
scientist who published papers on a wide variety of hydrological and meteorological phenomena (see
Box 1.1). His perceptual model of infiltration was different from the idea that infiltration is controlled by
the hydraulic gradients within the soil profile. He thought that conditions at the surface were much more
important and recognised seasonal effects due to cultivation, the redistribution of particles by rainsplash
blocking larger pores, and the irregularity of the surface in allowing air to escape. His perceptual model
surely involved a much wider range of processes than the model that now bears his name (see, for example,
the work of Horton from 1936 and his process descriptions from 1942; see also the summary of some of
his archived papers by Beven, 2004b, 2004c).
In the same period as Horton, however, Charles R. Hursh was working in the Coweeta watersheds in
Georgia in the United States. These Southern Appalachian catchments are forested with soils that are
deeply weathered and have generally high infiltration capacities. Surface runoff is restricted mainly to the
channels, so here the storm runoff production must be controlled by subsurface responses (Figure 1.4d).
Hursh published a number of articles dealing with subsurface responses to rainfall (see, for example,
Hursh and Brater, 1941). A later director of the Coweeta laboratory, John Hewlett, was also influential
in getting the importance of subsurface stormflow more widely recognised in the 1960s (Hewlett and
Hibbert, 1967; Hewlett, 1974).
Independently in the 1960s, studies within the Tennessee Valley Authority (which at that time served
as one of the major hydrological agencies in the eastern United States) were revealing that it was very
difficult to predict runoff production in many catchments under the assumption that infiltration excess
surface runoff was produced everywhere on the hillslopes. The information on infiltration capacities of
the soils and rainfall rates could not support such a model. Betson (1964) suggested that it would be
more usual that only part of a catchment would produce runoff in any particular storm and that, since
infiltration capacities tend to decrease with increasing soil moisture and the downslope flow of water
on hillslopes tends to result in wetter soils at the base of hillslopes, the area of surface runoff would
tend to start close to the channel and expand upslope. This
partial area model
(Figure 1.4b) allowed for
a generalisation of the Horton conceptualisation. It is now realised that the variation in overland flow
velocities and the heterogeneities of soil characteristics and infiltration rates are important in controlling
partial area responses. If runoff generated on one part of a slope flows onto an area of higher infiltration
capacity further downslope it will infiltrate (the “run-on” process). If the high intensity rainfall producing
the overland flow is of short duration, then it is also possible that the water will infiltrate before it reaches
the nearest rill or stream channel. Bergkamp (1998), for example, estimated that for some plot scale
experiments with artificial rainfalls at an intensity of 70 mm/h, the average travel distance for overland
flow was of the order of 1 m!
Another form of partial area response was revealed by studies in a different environment by Dunne
and Black (1970) working in Vermont. They observed surface runoff production, but on soils with
high surface infiltration capacities. The surface runoff resulted from a saturation excess mechanism
(Figure 1.4c), a type of response that had been previously suggested by Cappus (1960), working in
France (and published in French).
These four major conceptualisations are all subsets of the more general perceptual model outlined
previously. We now know that infiltration excess, saturation excess or purely subsurface responses might
all occur in the same catchment at different times or different places due to different antecedent conditions,
soil characteristics or rainfall intensities. In addition, an infiltration excess mechanism might take place
