|Toride, Nobuo - SAGA UNIV., JAPAN|
|Leij, Feike - U.C. RIVERSIDE|
Submitted to: Soil Science Society of America Journal
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: August 5, 1997
Publication Date: N/A
Interpretive Summary: The stream tube model was used for several cases of chemical transport from the surface of soils with variable water velocity (v) and chemical adsorption (Kd). The chemical is dispersed in both the vertical and the horizontal direction as result of the variability. Chemicals such as fertilizers can be applied to the soil surface, a so-called boundary value problem (BVP), or they can be mixed in the top layer of the soil, an initial value problem (IVP). The concentration averaged over the entire field versus depth exhibits more spreading for the BVP than the IVP. Results of the stream tube model are compared with those of the regular transport equation (i.e., the CDE). In the latter case the parameters are assumed constant in the horizontal direction of the field while they are made to depend on depth by trying to account for the soil variability. This approach is too simplistic compared to the stream tube model if different types of transport problems are studied. On the other hand, the results of the stream tube model compare favorably with those obtained from a detailed numerical simulation, which is considered a benchmark, of chemical transport and water flow in a heterogeneous soil.
Technical Abstract: The use of the stream tube model developed in the first part of this study is illustrated for several examples with a stochastic pore-water velocity, v, and distribution coefficient, Kd. Increased vertical solute spreading due to stochastic local-scale parameters is accompanied by increased horizontal variations of the field-scale mean concentration. Solute application at the surface is modeled as a boundary value problem (BVP) and an initial value problem (IVP). The field-averaged concentration versus depth exhibits more spreading for the BVP than the IVP. The use of a random v instead of Ks is preferable for small variations in water content. Results of the stream tube model are compared with those of a one-dimensional macroscopic CDE with effective parameters (i.e., depth-dependent constants). The stream tube model and the macroscopic CDE will give different results if the effective parameters are used to model other transport scenarios. Finally, the stream tube model was fitted to the concentrations obtained from a detailed numerical simulation of flow and transport in a heterogeneous field. The (simple) stream tube model provides a sensible description of the field-averaged concentration and variance.