GSSHA provides two basic methods and four different options for simulating infiltration. The two basic methods are the Richards’ equation, described in Chapter 11, and the Green and Ampt method (1911), as described below. Selection of Richards’ equation over one of the Green and Ampt methods actually results in the model operating quite differently, most notably during long-term simulations (Chapter 10).
There are three Green and Ampt methods; basic Green and Ampt (1911), Green and Ampt with Redistribution (GAR) (Ogden and Saghafian 1997), and multi-layer Green and Ampt (Downer and Ogden in preparation). The multi-layer Green and Ampt model is not currently supported by WMS, and the user is referred to the GSSHA User’s Manual for more information on this option.
In the Job Control dialog, one of four choices can be specified:
- No infiltration.
- Green & Ampt.
- Infiltration redistribution (Green & Ampt with redistribution).
- Richard’s infiltration.
If no infiltration method is specified, then no parameters need be defined. If Green & Ampt is chosen, then grid parameters of soil hydraulic conductivity, porosity, wetting front suction head (capillary head), and initial moisture content must be defined. If the infiltration with redistribution option is chosen, the pore distribution index and residual saturation must also be defined. Selection of Richards’ equation requires the user to specify both global parameters and distributed grid cell parameters, as described in Chapter 11.
Green & Ampt Infiltration
In the Green and Ampt model of infiltration, water on an overland flow grid cell resulting from precipitation, overland flow, or other sources is assumed to enter the soil as a sharp wetting front. The soil behind the front is assumed to be saturated. The soil ahead of the front is assumed to be at some uniform initial moisture. The wetting front is drawn into the soil because of gravity and soil capillary pressure. As the front moves into the soil column, the effect of the soil capillary head is reduced and infiltration slows, approaching the value of the saturated hydraulic conductivity.
Four soil property parameters are required for each cell:
- Saturated hydraulic conductivity (cm/hr).
- Wetting front suction head or capillary head (cm).
- Porosity - fraction of voids in the soil (m3/m3).
- Initial moisture content - initial fraction of water in the soil (m3/m3).
The first three parameters may be assigned based on an index map of soil textures. As the land use may also affect these parameters, it is typical to create a composite land use/soil texture index map to assign these parameters. In the absence of measured field data, the parameters may be estimated based on the soil textural classification. Tables of parameters can be found in Rawls, Brakensek, and Saxton (1982) and are summarized in the GSSHA User’s Manual (Downer and Ogden in preparation). Assignment of these parameters based on soil textural classification typically requires that one, some, or all of these parameters be adjusted during calibration.
The initial moisture content must be less than or equal to the porosity and should be greater than the residual water content. Estimation of the initial soil moisture is based on antecedent conditions in the watershed. The Green and Ampt method is highly sensitive to the value of initial moisture. Accurate initial moisture estimates are required. Initial moisture estimates may be determined from field measurements, satellite measurements, or may be provided by the GSSHA model when long-term simulations are conducted (Chapter 10). Senarath et al. (2000) demonstrate the dangers of using initial moisture as a calibration parameter in single event calibrations.
When GAR (Green & Ampt with Redistribution) is selected for modeling infiltration, soil pore water is redistributed during periods of no or low-intensity rainfall, and infiltration capacity recovers for the next burst of storm intensity. The technique used for storm hiatus and post-hiatus stages is based on the method by Smith, Corradine, and Melone (1993) with minor modifications (Ogden and Saghafian 1997). In this model, the Green & Ampt equation is used for post-hiatus stage, so the four Green & Ampt parameters (described above) must be specified. In addition, the pore size distribution index (dimensionless) and residual saturation (dimensionless) are required. Without field measurements, values for these parameters may also be estimated based on soil textural classifications and assigned from literature values, e.g. Rawls, Brakensek, and Saxton (1982) and summarized in the GSSHA User’s Manual (Downer and Ogden in preparation).
When Richards’ equation is specified for simulation of infiltration, the movement of water in the soil is explicitly solved using Richards’ equation (Richards 1931). Detailed soil moisture profiles are calculated in each overland grid cell with infiltration calculated as a by-product of the solution. Richards’ equation requires many of the same parameters described above, in addition to global parameters related to the solution techniques used in the model. Richards’ equation and the proper assignment of parameters are described in detail in Chapter 11.