Influencing factors and pedo-transfer functions of topsoil saturated hydraulic conductivity in the coastal farmlands of reclaimed tidal flats

Determining the factors that influence saturated hydraulic conductivity of topsoils and developing appropriate pedo-transfer functions are important prerequisites for modeling field soil water and salt movement and the amendment of soil salinization in salt-laden coastal farmlands. In order to achieve these goals, soil saturated hydraulic conductivity and a number of associated soil physico-chemical properties were measured in a typical salt-laden costal farmland. The characterization of saturated hydraulic conductivity of the soil profile was discussed and principal component analysis conducted on the soil basic properties with potential impacts on saturated hydraulic conductivity. The pedo-transfer functions suitable for indirect estimation of saturated hydraulic conductivity in the study area were established. Results indicated that soil saturated hydraulic conductivity generally exhibited a decreasing trend from surface soil to subsurface soil, and an increasing trend from subsurface soil to substratum of the soil profile. The 20?40 cm soil layer had the least saturated hydraulic conductivity, with a range of 2.75?6.73 cm·d^-1. It was classified as the lowest water permeability due to the presence of soil compaction and plow pan. The profile characterization of bulk density was generally contrary to that of saturated hydraulic conductivity, and this was not surprising as bulk density represented the total state of porosity which was closely related to soil permeability. Results of correlation analysis showed that the contents of sand particles and organic matter had significant positive correlation with saturated hydraulic conductivity. Also soil bulk density, soil salinity and sodium adsorption ratio (SAR) had adverse impacts on saturated hydraulic conductivity. In addition to soil physical attributes (such as bulk density, soil porosity and soil texture), soil chemical factors (including fertility and salinity) also considerably influenced saturated hydraulic conductivity. The correlation between saturated hydraulic conductivity and soil basic properties provided the basis for further principal component analysis. The factors influencing saturated hydraulic conductivity were grouped into four principal components - soil water retention capacity component, state of soil salinity and alkalinity, soil fertility characteristics and soil texture. With a total of 78.17% of the variance explained, the explainable total variances of the four principal components were 28.91%, 21.02%, 17.59% and 10.65%, respectively. This also indicated that some important basic soil properties were not taken into account in the use of the datasets in this study. The most commonly used Vereecken pedo-transfer function exhibited satisfactory fitting and estimation performance. The prediction capability improved when soil salinity was added to soil data input for Vereecken pedo-transfer function. The modified Vereecken_1 pedo-transfer function (in which soil salinity was directly used) was the most appropriate method with the best prediction accuracy and suitability. The established Vereecken_1 pedo-transfer function in this study can be used to estimate saturated hydraulic conductivity from easily measurable or readily available soil properties such as soil sand/clay content, soil bulk density, soil salinity and soil organic matter content. The research findings provided support for indirect estimation of saturated hydraulic conductivity and numerical modeling of soil water flow and salt transport in salt-laden coastal farmlands.