6.4. Ice thermodynamics

For modelling ice thermodynamics, five different options are available, which can be chosen in the run-specs header by the parameter CALCMOD:

• -1: ISOT: isothermal method, constant temperature and age.

• 0: COLD: cold-ice method, resetting of temperatures above pressure melting.

• 1: POLY: polythermal method, separate domains for cold and temperate ice.

• 2: ENTC: conventional enthalpy method.

• 3: ENTM: melting-CTS enthalpy method.

Options 1 and 3 are the most sophisticated solvers and thus recommended for most real-world problems. The different methods are discussed by Greve and Blatter [31].

A temperate ice surface is not allowed. To avoid this situation, the surface temperature is always capped at a maximum value of \(-0.001^\circ\mathrm{C}\). Therefore, any vertical ice profile has an upper cold layer, which may or may not be underlain by a temperate layer (see also Fig. 6.2).

As explained in Section “Spatial grid”, for the polythermal method (POLY) separate vertical domains are employed for the cold-ice layer (\(\zeta_\mathrm{c}\) domain) and the temperate-ice layer (\(\zeta_\mathrm{t}\) domain). In all other cases, the \(\zeta_\mathrm{c}\) domain is used for the entire ice column from the base to the surface, the \(\zeta_\mathrm{t}\) domain is redundant, and the parameter KTMAX should be set to 2.

If the polythermal method is chosen (CALCMOD = 1), the parameter CTS_MELTING_FREEZING determines whether melting and freezing conditions are distinguished (1), or melting conditions are always assumed (2). In principle, the former is physically more adequate. However, in practice, freezing conditions occur only for a very small fraction of the areas with a temperate layer, and their treatment can cause numerical issues due to the discontinuities of the temperature gradient and the water content at the CTS (cold-temperate transition surface). Therefore, setting CTS_MELTING_FREEZING to 2 is safer and usually sufficient.