Estimation of water extractability and hydraulic conductivity in tropical mollisols, ultisols, and andisols

Abstract

Simple methods to estimate water extractability and hydraulic conductivity were tested in nine Mollisols, Ultisols and Andisols. A neutron hydroprobe was used to monitor soil water content over depth and time. A method to estimate soil water extractability developed by Ritchie was adapted. The applicability of the calibrated method was tested with independent soil data using the CERES model, a crop simulation model developed by a multidisciplinary team of scientists from the Grassland, Soil, and Water Research Laboratory in Temple, Texas. Using the estimated values of plant extractable water as inputs, the CERES model simulated 72% of the soil water content data points ranged from 0.08 to 0.50 m^3/m^3 with deviations of less than 0.03 m^3/m^3 from the observed values. The method requires inputs of bulk density, organic matter content, and 1.5 MPa water content for Andisols, and bulk density, organic matter content, and sand and silt content for other soils. Simple methods to estimate hydraulic conductivity developed by Libardi et ala (1980), Chong et a1. (1981), and Sisson et ala (1980) were compared. The methods require only soil water content measurements. Besides having one assumption associated with each method, all methods operate on the assumption of unit hydraulic gradient during redistribution and an exponential relationship between hydraulic conductivity and water content. The results showed that the method which assumes a power function between water content and time consistently gave higher estimates of saturated hydraulic conductivity than methods which assume a logarithmic or exponential functions. Using means and variances of the estimated saturated hydraulic conductivity for estimating spatial variability of soil water flux showed that the choice of method only affected the estimation of spatial variability of soil water flux at early time after cessation of ponding. At longer times, the differences among the estimated saturated hydraulic conductivity by each method did not greatly influence the estimated spatial variability of the soil water flux.