Academic Journals Database
Disseminating quality controlled scientific knowledge

Evaluating models for predicting hydraulic characteristics of layered soils

ADD TO MY LIST
 
Author(s): S. S. W. Mavimbela | L. D. van Rensburg

Journal: Hydrology and Earth System Sciences Discussions
ISSN 1812-2108

Volume: 9;
Issue: 1;
Start page: 301;
Date: 2012;
VIEW PDF   PDF DOWNLOAD PDF   Download PDF Original page

ABSTRACT
Soil water characteristic curve (SWCC) and unsaturated hydraulic conductivity (K-coefficient) are critical hydraulic properties governing soil water activity on layered soils. Sustainable soil water conservation would not be possible without accurate knowledge of these hydraulic properties. Infield rainwater harvesting (IRWH) is one conservation technique adopted to improve the soil water regime of a number of clay soils found in the semi arid areas of Free State province of South Africa. Given that SWCC is much easier to measure, most soil water studies rely on SWCC information to predict in-situ K-coefficients. This work validated this practice on the Tukulu, Sepane and Swartland layered soil profiles. The measured SWCC was first described using Brooks and Corey (1964), van Genuchten (1980) and Kasugi (1996) parametric models. The conductivity functions of these models were then required to fit in-situ based K-coefficients derived from instantaneous profile method (IPM). The same K-coefficient was also fitted by HYDRUS 1-D using optimised SWCC parameters. Although all parametric models fitted the measured SWCC fairly well their corresponding conductivity functions could not do the same when fitting the in-situ based K-coefficients. Overestimates of more than 2 orders of magnitude especially at low soil water content (SWC) were observed. This phenomenon was pronounced among the upper horizons that overlaid a clayey horizon. However, optimized α and n parameters using HYDRUS 1-D showed remarkable agreement between fitted and in-situ K-coefficient with root sum of squares error (RMSE) recording values not exceeding unity. During this exercise the Brooks and Corey was replaced by modified van Genuchten model (Vogel and Cislerova, 1988) since it failed to produce unique inverse solutions. The models performance appeared to be soil specific with van Genuchten-Mualem (1980) performing fairly well on the Orthic and neucutanic horizons while its modified form fitted very well the prismatic and pedo-cutanic horizons. The lognormal distribution model of Kasugi (1996) showed an extraordinary good fit among the Swartland profile horizons especially the saprolite rock layer. It was therefore concluded that in-situ KL-coefficient estimates from SWCC parameters could be acceptable if only rough estimates were required. Optimization of parameters for in-situ conditions especially for HYDRUS 1-D carried much prospects in characterising the hydraulic properties of most of the layered soils earmarked for IRWH in the province.
Save time & money - Smart Internet Solutions      Why do you need a reservation system?