Rate and strength of snow ablation on sea ice both exhibit significant hemispheric differences. Previous studies revealed that diurnal freeze–thaw cycles are prevailing the surface processes on sea ice in the Antarctic whereas strong and persistent snowmelt is characteristic in the Arctic. Our investigation aims at providing a comprehensive overview on the factors controlling the observed hemispheric contrasts in the seasonal evolution of snowmelt patterns. We use a thermodynamic snow model (SNTHERM) in combination with a microwave emission model (MEMLS) to identify the links between atmospheric forcing, snow property changes and observed microwave signals in the context of hemispheric and regional peculiarities. This provides insight into the sensitivity of microwave signals to specific thermodynamic processes during the snowmelt season. Results show that surface microwave emissivity is largely controlled by the contribution of melt, evaporation and compaction to seasonal snow thinning. This finding allows for an evaluation of the potential of satellite data to identify single stages of snowmelt and metamorphism noch over Arctic and Antarctic sea ice.