Right now we assume constant thermal conductivity and constant heat capacity. For water-ice, these vary significantly between the solid and liquid phases near the freezing temperature.
The variable thermal conductivity should be more straightforward than variable heat capacity.
Perhaps now that FEniCS automatically differentiates our nonlinear form, and we don't have to derive the Gateaux derivative by hand, this won't take too much work to implement.
Right now we assume constant thermal conductivity and constant heat capacity. For water-ice, these vary significantly between the solid and liquid phases near the freezing temperature.
The variable thermal conductivity should be more straightforward than variable heat capacity.