add missing manual file

docs/modules/multi-class-ice.md

@@ -0,0 +1,13 @@
+#Theory
+The column physics package of the Los Alamos sea ice model (CICE), the Icepack (v1.3.4) has been couple with SCHISM. Besides the zero-layer thermodynamics, two more sophisticated thermodynamic formulations, the Bitz and Lipscomb (1999; BL99) thermodynamics formulation for constant salinity profiles, and the mushy layer thermodynamics formulation for evolving salinity (Turner et al., 2013), are also implemented. At the sub-grid scale, thin ice and thick ice coexist, and therefore an ice thickness distribution (ITD; Lipscomb, 2001; Bitz et al., 2001) has been implemented in order to describe the unresolved spatial heterogeneity of the thickness field. The ITD offers a prognostic statistical description of the sea ice thickness, which it divides into multiple categories, along with the ice area fraction corresponding to each category – a more detailed approach than the single fraction used in the previous implementation. More tracers and more ice processes are added in this coupled model by Icepack, including multiple melt pond parameterizations (Hunke et al., 2013) and a mechanical redistribution parameterization (Lipscomb et al., 2007) that responds to sea ice convergence by piling up thin sea ice and therefore mimicking ridging and rafting events. The interaction between the shortwave radiation and the sea ice in Icepack is addressed using two formulations: the Community Climate System Model (CCSM3) formulation, which relates the surface albedo to the surface sea ice temperature, and the delta–Eddington formulation (Briegleb et al., 2007), which relates the albedo to inherent optical properties of sea ice and snow. The dynamic solver is not included in Icepack and we used two approaches: (1) the classic elastic–viscous–plastic method (EVP; Hunke and Dukowicz, 1997) and (2) the modified elastic–viscous–plastic method (mEVP; Kimmritz et al., 2015). Both methods are inherited from the previous single-class ice and snow formulation (Zhang et al., 2023).
+
+#Usage
+1. Compile with USE_MICE and USE_EVAP on;
+2. Provide snow flux data in sflux/sflux_prc* files (named as srate), the format is the same as rainfall data (prate);
+3. The main parameter input for this module is mice.nml and namelist.icepack (you can find a sample in sample_input/);
+4. For the parameters in the mice.nml, ice_advection=6 (hybrid TVD-upwind) is recommended. For the parameters in namelist.icepack, you can find them all in the [Icepack manual](/~https://github.com/CICE-Consortium/Icepack/wiki/Icepack-Release-Table).
+5. For the output, all variables for iof_mice in param.nml are enabled under new scribe IO, but only the variables in namelist.icepack are enabled under OLDIO; this needs to be updated in newer versions.
+6. Optional: ihot_mice in mice.nml. ihot_mice =0 is the cold start; ihot_mice = 1, the initial file (hotstart.nc) should contain ice concentration, thickness and velocity and from e.g. HYCOM, scripts have been attached; ihot_mice = 2, it is a restart function, the ice variables are stored in outputs/hotstart*, and will be restarted by calling restart_icepack in icedrv_io.f90, but not stable for too many variables, like floe size distribution (tr_fsd).
+
+
+#References