A Mesopore and Matrix infiltration model based on soil structure

After decades of research into the subject, preferential flow in soils still plagues those of us who hope to provide better solutions to society's natural resource management conundrums. To that end, simulation models that strike the balance between simplicity and robustness are a high priority. We present the two-domain, Mesopore and Matrix (M&M), water-infiltration module based on soil structure for the Precision Agricultural-Landscape Modeling System (PALMS), by combining laminar flow of water through interaggregate slits with water movement from slits into aggregates using Darcy's law. The M&M model is based on new assumptions using soil aggregate geometry that allow the model to be more easily parameterized for landscapes with varying soil properties. The vertical and horizontal arrangements of aggregates and the slit width (2B(Ped)(theta)) are based on cubic geometry, where the width of the cubes represents the width (w(ped)(theta,z)) of soil structural units, which depends on soil water content (theta) and depth (z). The M&M module can be parameterized so that mesopore infiltration resembles that of the Richards equation as a starting point for the preferential-flow parameterization. Using field measurements of ped size and in situ mesopore volume to calculate B-ped(B) and w(ped)(theta,z), the parameters from the Richards equation starting point can be modified to capture mesopore effects. PALMS was run using this approach with and without M&M for the 1996 frost-free growing season (March 1 to Nov. 1). While both models slightly under-predicted 1.4 m depth drainage compared with independently measured field data from the same period, PALMS with the M&M module increased estimated drainage by 32% in response to the largest 1996 storm event (similar to 100 mm in 2 days) and 11% seasonally (627 mm cumulative precipitation) over the original PALMS model using the Green and Ampt approach. For the 2008 growing season, a period of unusually high rainfall rates (820 mm annual precipitation), PALMS with M&M simulated a 57% increase in drainage response to the season's largest storm (similar to 200 mm in 2 days) and a 20% increase in seasonal drainage compared to PALMS with the Green and Ampt equations. For both 1996 and 2008 seasons, the drainage response to large storms predicted by PALMS with M&M occurred much more rapidly (hours as opposed to days) than PALMS with the Green and Ampt approach. (C) 2009 Elsevier B.V. All rights reserved.