The temporal variation of millimeter wave backscatter signature of snowcover is studied based on a dense medium radiative transfer (DMRT) theory and a one-dimensional mass and energy balance model of snow named SNTHERM. The multilayer DMRT scattering model of snowcover developed in this work takes into account the reflection and refraction at the snow-snow interfaces. Appropriate boundary conditions, quadrature points and weights are selected for using the discrete-ordinate eigenanalysis method to solve the multilayer DMRT equations. It shows that the inclusion of reflection and refraction at the snow-snow interfaces may affect the model prediction. Cohesive spherical particles are applied to account for the clustering feature of snow grains. To model the time-varying behavior of snowcover, SNTHERM is employed to simulate the aging process of snow. SNTHERM provides pertinent snow parameters, such as grain size, density, liquid water content, and stratification with a high resolution in time and depth. These snow properties are then used in modeling the temporal radar signatures of snowcover. The comparisons of temporal model responses with time series polarimetric backscatter data at 35 and 95 GHz are presented. Good agreement is demonstrated between model and measurements in both timing and magnitude. The results indicated that the coupled DMRT and SNTHERM model can be useful in studying the electromagnetic interactions with time-varying snow microstructure.