Multiple Linear Regression of Complex Permittivity of a Till at Frequency Range from 200 MHz to 400 MHz

Abstract

The complex permittivities of 40 compacted natural clayey till specimens at various water content, density and pore fluid salinity as independent variables are analyzed. A multiple linear regression model is constructed and the statistical analyses for the relative permittivity (e') and dielectric loss (e '')-regression equations in the frequency range from 200 to 400MHz are performed. It is concluded that the multiple linear regression model of the complex permittivity is highly significant. The coefficient of determination (R 2) and standard error (SE) of the linear regression equations vary with frequency. At 200MHz for e ' -regression and 300MHz for e '-regression, all three independent soil variables are significant at a probability of 0.95. The complex permittivity regression model with two variables at 250MHz for e 'and 350 for e '-regression equations. Upon further study on the effect of soil mineralogy to the complex permittivity, the results obtained in this study may serves as the principle of the characterization of subsurface and detection of soil and groundwater contamination.