Premium
A parameter‐efficient hydrologic infiltration model
Author(s) -
Smith R. E.,
Parlange J.Y.
Publication year - 1978
Publication title -
water resources research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.863
H-Index - 217
eISSN - 1944-7973
pISSN - 0043-1397
DOI - 10.1029/wr014i003p00533
Subject(s) - ponding , sorptivity , infiltrometer , infiltration (hvac) , saturation (graph theory) , hydraulic conductivity , soil science , soil water , hydrology (agriculture) , mathematics , environmental science , geotechnical engineering , geology , thermodynamics , physics , drainage , porosity , ecology , combinatorics , biology
By adopting two extreme assumptions concerning the behavior of unsaturated soil hydraulic conductivity K near saturation, we derived a two‐branched model for ponding time and infiltration rate decay for arbitrary rainfall rates. One assumption was that K varies slowly near saturation and leads to an expression for ponding time and infiltration decay. For initially ponded conditions, ponding time is zero, and with rainfall rate r → ∞, the familiar Green and Ampt (1911) expression results. The other, rather opposite assumption was that K varies rapidly, e.g., exponentially, near saturation. This model also holds for both rainfall and ponded surface conditions, and for ponded conditions the expression is identical to that of Parlange (1971). Each model uses only two parameters, saturated soil conductivity K s and a parameter that is roughly related to sorptivity and responds nearly linearly to variations in initial saturation. Both parameters are physically related to measurable soil properties. Methods are presented to estimate parameters of either model from infiltrometer tests. The two models are compared with a precise numerical solution of the unsaturated soil water diffusion equations for three soils that represent a range of soil behaviors near saturation. Our results show that either assumption would be an excellent model for most hydrologic purposes, and the relative goodness of fit of each model is generally consistent with the appropriate behavior of K (θ → θ S ).