…tter initialization of surface for SCM

… in new variables in gmtb_scm_input (using new subroutines); have not propagated new variables outside of the input subroutine yet though

… for surface variables (set physics vars directly from input, rather than using the state DDT as a middleman; read in all new variables; separate setting of physics variables between physics_associate and physics_set, the former is used for variables that the SCM forcing may change outside of physics and use pointers, while the latter merely sets the initial values of variables

…ithin gmtb_scm_input.F90

…or checking for missing data; add code to handle missing data (abort or not) in physics_set (only got through scalar variables as of this commit)

…d start code from FV3GFS_io.F90

…CCPP

…ug in gmtb_scm_input related specified surface fluxes; fix associating surface_skin_temperature with forcing temperature for specified surface flux cases; update plot_config to use data from supported suites

…CPP scheme init stage)

…y have incorrect logic)

…le pointer; update GFS_typedefs.[F90/meta] for latest ccpp/physics; add metadata for newly required physical constant; fix logic for testing for missing data for stc/slc/smc