A novel semi-analytical model is derived for the calculation of water saturation levels in the near subsurface using passive temperature measurements. The omnipresent diel signal of natural temperature variations is exploited, however the solution is general so that a cyclical temperature signal of any period could be used. The model is based on the first-principles advection-conduction-dispersion equation, which is fully general for porous media, and directly considers the saturation dependency of thermal conductivity which has been avoided in previous studies. An established empirical model for thermal conductivity of variably saturated porous media is incorporated and two solutions for saturation are derived. A spatially sequential implementation of one of these solutions is shown to predict the vertical saturation profile to within 2