Discontinuity in thermal conductivity
The thermal conductivity for snow exhibits a discontinuity at theta[ICE]==0.55, how it is now implemented now in Snowpack.cc:
else if (Edata.theta[ICE] > 0.55 || Edata.theta[ICE] < min_ice_content) { Keff = Edata.theta[AIR] * Constants::conductivity_air + Edata.theta[ICE] * Constants::conductivity_ice + Edata.theta[WATER] * Constants::conductivity_water + Edata.theta[SOIL] * Edata.soil[SOIL_K]; } else { Keff = SnLaws::compSnowThermalConductivity(Edata, dvdz, !alpine3d); //do not show the warning for Alpine3D }
The reason is that this transitional regime is not well studied, so the validity of parametrizations for snow is limited to typical snow densities. However, a discontinuity is not so nice, so probably we should ""stich"" both regimes together, e.g., using logistic or linear functions to provide a smooth overlap in both regimes.