Add check_energy related history output variables to check_energy_diagnostics

[jimmielin]

In CAM check_energy.F90 the following history variables are written out:

    call addfld('TEINP',  horiz_only,  'A', 'J/m2', 'Total energy of physics input')
    call addfld('TEOUT',  horiz_only,  'A', 'J/m2', 'Total energy of physics output')
    call addfld('TEFIX',  horiz_only,  'A', 'J/m2', 'Total energy after fixer')
    call addfld('EFIX',   horiz_only,  'A', 'W/m2', 'Effective sensible heat flux due to energy fixer')
    call addfld('DTCORE', (/ 'lev' /), 'A', 'K/s' , 'T tendency due to dynamical core')
    call addfld('DQCORE', (/ 'lev' /), 'A', 'kg/kg/s' , 'Water vapor tendency due to dynamical core')

The following variables are missing:

• TEINP is state%te_ini(:,dyn_te_idx), and should be written out with the standard name vertically_integrated_total_energy_using_dycore_energy_formula_at_start_of_physics_timestep
• TEFIX is state%te_cur(:,dyn_te_idx). It is currently written out with the standard name vertically_integrated_total_energy_using_dycore_energy_formula but the field name is te_cur_dyn. Should be renamed to TEFIX?
• EFIX is eshflx from check_energy_fix_run. It has the standard name net_sensible_heat_flux_through_top_and_bottom_of_atmosphere_column and should be written out right after a call to check_energy_fix_run (before it is zeroed out by the next scheme) -- this might need splitting the diagnostic into a check_energy_fix_diagnostics so it is called at the right time.
• DTCORE is tendency_of_air_temperature_due_to_dynamics, it is computed at the end of physpkg after check_energy_fix_run -- could go with the same EFIX into the same diagnostics module

The following have been implemented:

• TEOUT

To capture the appropriate quantities at the right point in time there might be separate schemes needed for several of them.

• TEINP is ready after check_energy_chng_timestep_init
• TEFIX is ready after the energy fixer / tendencies applied and check_energy_chng computes the total energy ?
• EFIX is ready after check_energy_fix_run but before the next tendencies are zeroed out for check_energy_zero_fluxes + chng
• DTCORE should be a separate scheme, check when/where it is called in CAM

[jimmielin]

Splitting out DTCORE etc. into #168