Approaches for Estimating Soil Water Retention Curves at Various Bulk Densities With the Extended Van Genuchten Model

Abstract

Soil bulk density (ρb) variations influence soil hydraulic properties, such as the water retention curve (WRC), but they are usually ignored in soil water simulation models. We extend the van Genuchten WRC model parameters to account for ρb variations using a series of empirical expressions. WRC measurements made on eight soils with various ρb, and textures are used to calibrate these ρb‐related empirical equations. Accordingly, two approaches are developed to estimate WRCs of soils at various ρb. Another eight soils with a wide range of ρb and textures are used to evaluate the accuracy of the new approaches. Approach 1 estimates WRCs for each soil at various ρb using a WRC measurement made at a reference ρb and the soil texture fractions. This approach gives reasonable WRC estimates for the eight validation soils, with an average root‐mean‐square error (RMSE) of 0.025 m3/m3 and an average determination coefficient (R2) of 0.94. For Approach 2, a WRC measurement made at a reference ρb and one additional water content‐matric potential value measured at a different ρb value are used, which produces WRC estimates with an average RMSE of 0.017 m3/m3 and an average R2 of 0.97. The methodology used in Approach 2 is also applied to the Brooks and Corey WRC model to obtain accurate and precise WRC estimates. The proposed approaches have the potential to be incorporated into simulation models for estimating soil hydraulic properties that are affected by transient and variable ρb.

Soil water retention and hydraulic conductivity dynamics following tillage

“…However, results indicated large differences in θ v for different ρ b at −50 kPa. Even at −1500 kPa differences were still evident, contradicting the results of Ahuja et al (1998); as well as modeling assumptions that at low Ψ, the water retention for different ρ b is similar (Tian et al, 2018a). In-situ water retention measurements for a sandy soil (Zhang et al, 2017, Table 1) showed markedly similar changes in water retention, with water contents at −33 kPa ranging from 0.15-0.20 cm 3 cm −3 for a ρ b of˜1 g cm −3 and 0.28 cm 3 cm −3 for a ρ b of 1.33 g cm −3 .…”

“…Unfortunately no data at −1500 kPa were available for higher bulk densities. Using fitted θ r rather than θ v at −1500 kPa, an alternative suggested by Tian et al (2018a), did not improve Approach 1 as it resulted in θ r > θ v at −1500 kPa. This unreasonable result is likely due to the bimodal nature of the water retention curves (data not shown).…”

“…where θ res and θ sat are residual and saturated θ v , respectively, and α and n are empirical shape parameters. Both of the Tian et al (2018a) approaches require an initial value for θ res , θ sat , α, and n, at a specific (preferably lowest) ρ b . In this study, these parameters were determined using water retention measurements immediately following tillage.…”

“…where subscripts 1 and 0 refer to new and initial values, respectively, and ω is an empirical parameter set to 3.97 (Tian et al, 2018a). In Approach 1, n at a new ρ b value is determined based on soil texture, using:…”

“…Literature review on temporal changes in bulk density following tillage. Values were reported for smaller increments, averages are shown in this table for the sake of clarity Tian et al (2018a). recently introduced two approaches to incorporate changes in ρ b into the van Genuchten model The van Genuchten (1980).…”