Skip to main content
ARS Home » Research » Publications at this Location » Publication #105707
Title: RETCML: INCORPORATING MAXIMUM-LIKELIHOOD ESTIMATION PRINCIPLES IN THE RETC SOIL HYDRAULIC PARAMETER ESTIMATION CODE

Author
item HOLLENBECK, K - TECH UNIV LYNGBY DENMARK
item SIMUNEK, J - UC RIVERSIDE, CA
item Van Genuchten, Martinus

Submitted to: Computers and Geosciences
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/1/1999
Publication Date: N/A
Citation: N/A

Interpretive Summary: The RETC (RETention Curve) code is a widely used computer program developed at the U. S. Salinity Laboratory for estimating parameters of the retention curve and hydraulic conductivity functions of unsaturated soils. While the retention curve characterizes the energy status of the soil water, the unsaturated hydraulic conductivity function describes the ability of the porous medium to conduct water. The retention and hydraulic conductivity functions constitute properties needed for simulating flow and solute transport in the unsaturated zone. Computer models using analytical descriptions of soil hydraulic properties are now routinely used in both research and management to predict the movement of water and chemicals in the vadose zone between the soil surface and the ground water table. Interest in the vadose zone has dramatically increased in recent years because of growing evidence that the quality of the subsurface environment is being adversely affected by industrial, municipal, and agricultural activities. In this paper we present an improved version of RETC, called RETCML, for which the uncertainty analysis is based on maximum-likelihood theory for the special case of weighted least-squares estimators.

Technical Abstract: RETC is a public domain computer code for estimating parameters of the water retention curve and hydraulic conductivity functions of unsaturated soils. RETC was developed at the U. S. Salinity Laboratory and is now used world-wide with thousands of copies distributed. Evaluation of the final estimation results in the code has been improved to yield a new version, RETCML, based on maximum-likelihood theory for the special case of weighte least-squares estimators. This paper first explains the theory of maximum- likelihood and introduces the principles of model adequacy and parameter uncertainty on a formal basis. Next, this paper presents a user guide for the code. RETCML is also free and has been programmed to be almost fully compatible with the original RETC input files, thus making it easy to re- analyze previous data. The output of RETCML includes a thorough evaluation of estimation results.