Using MACRO to simulate preferential flow in structured and non-structured soils.

摘要

A model, MACRO, was evaluated for short-termed simulation of water flow and non-reactive solute (chloride or bromide) transport in three soils (Maury, Cecil, and Lakeland) with contrasting structural characteristics. Long-termed drainage flow was also simulated for Cecil.; In two-domain simulations, the model was sensitive to two parameters, effective diffusion path-length and boundary hydraulic conductivity, because they control the exchange of water and solute between the two domains. Flow was fairly uniform near the surface of the Cecil and Lakeland soils because of low clay content and weak soil structure. Bypass flow was quite pronounced within the surface 15 cm of the Maury silt loam. However, preferential flow and transport were significant below 20-cm depth of all three soils because of strong primary structure in Maury; primary secondary, and tertiary structure in Cecil; and finger development in the structureless Lakeland. Flow in the Lakeland was simulated by assuming that fingers represented interpedal pores (pores between the aggregates) and this approach worked well for the simulations. The model simulated water flow and solute transport relatively well by setting small values (1–10 mm) of effective diffusion path-length near the soil surface and large values (100 mm) of it below around 20-cm depth.; Although a constant tortuosity factor of 0.5 in micropores was used for three soils, the tortuosity factor in macropores changed with depth and soil. Relatively large values of tortuosity (the range of 1.5–2.5) were used for Maury and Cecil, whereas the smallest value, 1.0, was used in Lakeland within the surface 100 cm. Three parameters required by the model affect solute transport: (1) excluded volumetric water content, (2) dispersivity, and (3) mixing depth.; The performance of the model was evaluated graphically (plotting measured and model values with depth and time) and statistically (coefficient of residual mass, model efficiency, and correlation coefficient). Preferential flow in these three soils was relatively well simulated by the model.; In general, the parameters defined for the model for each soil related to the flow and transport characteristics of the soils. The existence of more uniform flow near the soil surface and more preferential flow below a certain depth was represented by the model with reasonable parameter values in each soil. However, although some parameters differed among the three soils, the parameter of effective diffusion path-length did not reflect known differences in the physical characteristics of the three soils. (Abstract shortened by UMI.)