Yongping Chen, Univ Connecticut, Dept Civil & Environm Engn, Storrs, CT 06269, Paolo Castiglione, Montana State University, Montana State University, 818 Leon Johnson Hall, Bozeman, MT 59715, Jon Wraith, PO Box 173120, Montana State University, Montana State University, LRES Department, Bozeman, MT 59717-3120, and Dani Or, EPFL Ecole Polytechnique Federale, EPFL Ecole Polytechnique Federale, GR B1 399 Station 2, Lausanne, 1015, Switzerland.
Many non-destructive techniques for measuring the volumetric water content, including TDR, capacitance, GPR, and remote sensing, rely on measurement of the soil dielectric permittivity. As a result of the temperature dependence of soil permittivity, measured water content may exhibit a marked temperature effect. A clear understanding of the underlying phenomena is desirable from fundamental and practical standpoints. The measured temperature dependence of soil permittivity may be expected to differ substantially among different regions of the electromagnetic frequency spectrum. For example, the decrease in permittivity of water with temperature, and the concurrent increase in electrical conductivity which enhances interfacial polarization phenomena such as the Maxwell-Wagner effect, result in a crossover frequency below which the soil bulk permittivity increases with temperature, and above which it decreases with increasing temperature. An additional phenomenon, observed primarily in finer textured soils, is the enhanced mobility of bound water with increasing temperature, which results in an increase in the measured bulk soil permittivity. We review the different processes responsible for the temperature dependence of soil permittivity, and indicate potential temperature effects at different frequencies. Our theoretical considerations are supported by a series of measurements of soil permittivity spectra, carried out with a Vector Network Analyzer for different soil textures, water contents, and temperatures. We also measured the apparent dielectric constant using TDR. Our results confirm that TDR measurements are only slightly affected by temperature variation, and that different-textured soils exhibit different temperature effects. On the other hand, we show that sensors operating at lower frequencies are likely to be strongly influenced by temperature variation, and that these effects may be only partially compensated for.