When rainfall falls on the land surface, in many environments the first thing it hits will be a vegetation canopy. This has the effect of retaining some of the rain on leaf surfaces as interception and redistributing the rest down to the ground surface as through/all and stemflow. At the ground surface, therefore, the rate of supply of water will no longer be spatially uniform. There may be local concentrations at the base of stems or trunks; there may be other areas that receive lower intensity inputs.
This distribution of intensities falls onto a surface that will have a variable infiltration capacity due to variations in soil properties and initial moisture status. In particular, the properties at the soil surface will be very important in controlling how much of the rainfall can infiltrate into the soil. However, in many environments much of the rainfall input will infiltrate into the soil unless the soil is already completely saturated. It will do so by one of two routes, by direct infiltration into the soil matrix and by infiltration into pathways due to larger structural voids (cracks, root channels, animal or insect burrows, etc). The latter may be very important since the faster flow velocities associated with structural porosity can mean that the water can move rapidly into the soil, bypassing parts of the soil matrix as a preferential flow (see, e.g., Beven and Germann, 1982). Some of this water will be absorbed into the soil matrix at depth, some may move rapidly downslope within the preferential flow pathways, and some may be moved rapidly vertically to the local water table. Rates of movement will depend on the input rainfall intensity, the permeability and initial moisture content of the soil matrix, the structural characteristics of the soil, and the depth to the water table. It is well known from soil thin sections, however, that water movement in structural pathways can lead to transport of clay particles that are deposited in thin layers called cutans, which can then restrict the infiltration of water into the soil matrix and prolong the preferential flow. It is also known that preferential flow can be important in the transport of contaminants, such as pesticides and herbicides, in runoff since such contaminants are often sorbed to fine particles.
Within the soil matrix, water movement into soil that is not fully saturated will take place primarily vertically as a wetting front. The propagation of the wetting front will again depend on the antecedent wetness of the soil, the input intensity, and the matrix hydraulic characteristics. Flow within a continuous soil matrix can usually be described by the Richards equation, which is based on the unsaturated form of Darcy's law. The Richards equation is difficult to solve for general situations of practical hydrological interest, but there are now very many approximate numerical solution codes available (see Chapter 28). All such solution algorithms will require the specification of the soil hydraulic characteristics, which will depend on the texture and organic matter content of the matrix.
Based on many thousands of measurements, empirical relationships have been developed between the easily measured texture characteristics and the much more difficult to measure soil hydraulic properties (see, e.g., Rawls and Brakensiek,
1989). These empirical regressions are often useful but must be used with care. The estimates obtained in this way are subject to significant estimation error and are also only as good as the original data on which they are based. In this case the measurements were usually based on small soil samples and did not include any effects of the structural porosity. In any case, preferential flows may not be well described by the Richards equation, and it has proven very difficult to develop a comprehensive description of preferential flow with parameters that can easily be estimated in applications.
A useful simple analogy for the movement of water in both matrix and preferential flow pathways is that of the wetting front as a kinematic shock or locally pistonlike front, in which the rate of propagation of the front is given as c = I/Ad where c is the velocity of the front, / is the local input intensity, and Ad is an effective change in moisture content across the front. This is an approximation, applicable only where the input rate / does not exceed the infiltration capacity of the soil at any depth, but it is then readily seen that the wave speed c increases with the input intensity and decreases with the change in moisture content. Thus, if A8 is small, either in a preferential flow pathway (ignoring losses due to sorption into the matrix) or because the soil matrix is already wet, the front may move quickly into the soil. For a low input intensity /, infiltrating into a dry soil with an effectively high Ad, the speed of the wetting front may be very much lower. With a variation of effective intensities at the ground surface, and a variety of effective local AO values in different flow pathways, there will be a distribution of wetting front velocities in the soil.
Some of the infiltrating water will be retained in the soil and later evaporated or transpired back to the atmosphere (see Chapter 26), but some of the wetting resulting from a storm rainfall may reach an existing water table, or will induce saturation at the base of the soil profile, or a perched zone of saturation above a horizon of lower permeability in the soil profile. The wetting or recharge will induce a response in the saturated zone that will ultimately produce some subsurface runoff.
Was this article helpful?