Improved partitioning between matrix and macropore flow: Novel bimodal lognormal functions for water retention and hydraulic conductivity in pumice and non-pumice soils

Dual-porosity models have been shown to improve models of soil-water movement and enhance the water balance of structured soils. In this study we introduce novel, continuous, closed-form, bimodal, lognormal functions for soil-water retention, theta(psi), and unsaturated hydraulic conductivity, K(psi), enhancing traditional models and significantly improving predictions, particularly for pumice soils. These functions incorporate the thresholds for (a) water pressure, (b) soil water content, and (c) unsaturated conductivity, which accurately differentiates macropores from matrix pores. Validation using 313 observation points from laboratory data shows an increase in the Nash-Sutcliffe efficiency coefficient of theta(psi) from 0.94 to 0.97, and for K(psi) from 0.83 to 0.95. The model also requires six constant, semi-empirical tortuosity parameters and provides a physically constrained approach for the hydraulic parameters that reduces non-uniqueness risks. The derived functions yield improved predictions and enable the computation of macropore soil water content and flow contributions, with potential applications for the advancement of preferential flow modelling.