ghis2s enhancements

[smahanam]

The "ghis2s" S2S software requires the following urgent enhancements:

1. Include HydroSCS Data: Add HydroSCS as an observational dataset.
2. Topographic Corrections: Introduce an option to apply lapse rate and aspect ratio corrections to meteorological forecasts.
3. Refactor the bcsd Function: The bcsd function within the S2Srun class currently runs seven sequential steps (mf_regridding, mf_biascorr, mf_temporal_disaggr, pr_regridding, pr_biascorr, pr_temporal_disaggr, and combine_forcings). These should be modularized into seven independent internal functions, enabling users to call only the specific steps required for their particular projects.
4. Relocate the S2Srun Class: The S2Srun class is the core module that develops the forecast experiment and writes the automation workflow. Currently, it is embedded inside the s2s_run.py driver; it needs to be moved out of the driver and directly into the ghis2s package.
5. Optimize File Sizes: High-resolution products (e.g., HydroSFS) generate excessively large files. File sizes must be reduced by packing data as int16 variables instead of float32, utilizing scale_factor and add_offset attributes.
6. Improve Execution Speed: Reduce LISDA and FCST wall-clock times by implementing domain decomposition.