Representing snow equi-temperature metamorphism (ETM) is key to model the evolution and properties of the snow cover. Recently, a phase-field model describing mean curvature flow evolution on 3-D microstructures was proposed (Bretin et al., 2019). In the present work, this model is used to simulate snow ETM at the pore scale, considering the only process of moving interfaces by sublimation-deposition driven by curvatures. We take 3-D micro-tomographic images of snow as input in the model and obtain a time series of simulated microstructures as output. Relating the numerical time, as defined in the model, to the real physical time involves the condensation coefficient, a poorly-constrained parameter in literature. A calibration was performed by fitting simulations to experimental data through the evolution of specific surface area (SSA) of snow under ETM at -2°C. A value of the condensation coefficient was obtained: (9.8 ± 0.7) 10-4 and was used in all the following simulations. We then show that the calibrated model enables to well reproduce an independent time series of ETM at -2°C in terms of SSA, covariance length, and mean curvature distribution. Finally, the calibrated model was used to investigate the effect of ETM on microstructure and effective transport properties (thermal conductivity, vapor diffusion, permeability), for four different samples. As an interesting preliminary result, simulations show an enhancement of the structural anisotropy of snow in the case of initially anisotropic microstructures such as depth hoar. Results highlight the potential of such micro-scale models for the development of snow property predictions for large-scale snowpack models.