Ks Estimates Using Macroscopic Capillary Length Estimated from Soil Hydraulic Shape Coefficients and Haverkamp Infiltration Model

Abstract

In this study, we developed a new general approach to estimate the macroscopic capillary length (lambda(c)) using different hydraulic function models and related shape parameters, along with the Haverkamp infiltration model constant. We next applied this new approach to the van Genuchten model coupled with a Burdine condition (vGB) to estimate lambda(c). Then, we applied the new lambda(c) computation to three different methods for estimating saturated hydraulic conductivity (K-s), and analyzed two sets of constant infiltration data: 1) an analytically generated Beerkan-type dataset and 2) constant head and Beerkan-type infiltration tests performed at the Ambark & ouml;pr & uuml; Experimental Station of Blacksea Agricultural Research Institute in Samsun, Turkey. Our new approach provided accurate K-s estimates when applied to the analytical Beerkan infiltration data. The highest error was observed for a silt soil, with 30 % error for one formulation versus <15 % for the others. For synthetic coarse-textured soils such loamy sand and sandy loam, the error was <10 %. For the field data and Beerkan-type experiments, the new approach gave consistent estimates of K-s regardless of analytical interpretation. However, ANOVA analysis revealed that K-s varied between different infiltration test types, with constant head infiltrometry with 5 cm of applied water head having greater K-s values than the Beerkan tests (p < 0.05). Estimated K-s values also differed between land use types (p < 0.01), with a maize field having significantly greater K-s compared to a soybean field. Overall, we conclude that the proposed approach represents an efficient and appropriate method for characterizing point-scale saturated hydraulic conductivity, so long as experimental artifacts such as ring insertion deep and preferential flows are considered.