Abstract:Time domain reflectometry(TDR) is increasingly used to estimate soil water content in field and laboratory tests. Errors arising from estimation of soil water content in field are analyzed and their sources are summarized as adhesiveness of expansive soils and gaps between tube and expansive soils. Based on these,three types of laboratory test are designed to quantitatively analyze the corresponding sources of the errors. It is shown that with high adhesiveness of expansive soils,the amount of bound water is large and dielectric constant of soil is small. The measured values of volumetric water content by TDR are smaller than those by oven dry,but there is a corresponding relationship between volumetric water content measured by TDR and that calculated from mass water content. Thus a linear regression equation is proposed to represent the calibration curve. Gaps generated from shrinkage of expansive soils can be divided into gaps around the tube and radial gaps. Errors dramatically increase as gaps around the tube exist. An air gap around the tube about 1 mm will result in an underestimation of 40% of soil water content,and the errors increase with the increase of soil water content. However,there is no obvious influence caused by radial gaps,which only results in errors less than 5%. The reason is that the measurement sensitivity of TDR probe rapidly decreases with the increase of measuring distance. Finally,the data from field monitoring are applied to examine the test results,and measures to reduce errors are put forward.