Fix science issues in make_bcs (GEOSsurface_GridComp/Utils/Raster/)

[gmao-rreichle]

#763 identifed the following issues with the generation of the EASE grid tile space:

• The Pfafstetter (catchment) ID assigned to a tile is that of the last contributing raster grid cell and not that of the dominant catchment within the EASE grid cell. An analysis by @zyj8881357 showed that approximately 36% of EASEv2_M09 tiles and 80% of EASEv2_M36 tiles do not represent the dominant catchment.
• The min/max lat/lon values for each tile in the catchment.def file seem to be associated with the EASE grid cell and not the desired lat/lon boundaries of the tile.
• It remains unclear if the adjustment of the land tile elevation properly accounts for lake and landice surfaces. This requires further examination.