This paper describes a methodology for mathematical modeling, computing, and research of two mutually connected problems: 2D non-stationary temperature distribution in logs stored for a long time in an open warehouse at periodically changing ambient air temperature in winter and change in the icing degrees of the logs during this time. Mathematical descriptions of the periodically changing ambient air temperature and of three types of relative icing degree of the logs that result under the influence of that temperature have been presented. These descriptions are introduced in coupled 2D non-linear mathematical models of the heat distribution in logs during their freezing and defrosting. The paper presents solutions of the models with explicit form of the finite-difference method. Results from a simulative investigation of the 2D non-stationary temperature distribution, average mass temperature, and three types of icing degree of beech logs with industrial dimensions (diameter of 0.4 m and length of 0.8 m), moisture content of 0.6 kg·kg⁻¹, and initial temperature of 0°C during their 5 days and nights alternating freezing and defrosting in an open warehouse at sinusoidal change of the ambient air temperature with various initial values below −5°C and different amplitudes are graphically presented and analyzed.