Merge pull request #14 from m2lines/add-yog-rebase

Implement YOG scheme in code

bld/namelist_files/namelist_defaults_cam.xml

@@ -1413,7 +1413,7 @@
 <deep_scheme phys='spcam_m2005'  >SPCAM    </deep_scheme>
 
 <yog_scheme                      >off </yog_scheme>
-<nn_weights                      >NONE</nn_weights>
+<yog_nn_weights                  >NONE</yog_nn_weights>
 <SAM_sounding                    >NONE</SAM_sounding>
 
 <shallow_scheme pbl="none"         >NONE     </shallow_scheme>

bld/namelist_files/namelist_definition.xml

@@ -3277,7 +3277,7 @@ Default: true
 </entry>
 
 <!-- YOG parameters -->
-<entry id="nn_weights" type="char*132" input_pathname="abs" category="conv"
+<entry id="yog_nn_weights" type="char*132" input_pathname="abs" category="conv"
        group="yog_params_nl" valid_values="" >
 Absolute path to the neural net weights used for the YOG deep convection scheme.
 </entry>

src/control/runtime_opts.F90

@@ -52,6 +52,7 @@ subroutine read_namelist(nlfilename, single_column, scmlat, scmlon)
    use rk_stratiform,       only: rk_stratiform_readnl
    use unicon_cam,          only: unicon_cam_readnl
    use zm_conv_intr,        only: zm_conv_readnl
+   use yog_intr,            only: yog_readnl
    use hk_conv,             only: hkconv_readnl
    use uwshcu,              only: uwshcu_readnl
    use pkg_cld_sediment,    only: cld_sediment_readnl
@@ -144,6 +145,7 @@ subroutine read_namelist(nlfilename, single_column, scmlat, scmlon)
    call cldfrc2m_readnl(nlfilename)
    call unicon_cam_readnl(nlfilename)
    call zm_conv_readnl(nlfilename)
+   call yog_readnl(nlfilename)
    call rk_stratiform_readnl(nlfilename)
    call hkconv_readnl(nlfilename)
    call uwshcu_readnl(nlfilename)

src/physics/cam/SAM_consts.F90

@@ -0,0 +1,129 @@
+module SAM_consts_mod
+  !! Physical constants from the SAM Model required for variable conversions
+
+  use netcdf
+
+  implicit none
+  public
+
+  ! From params.f90 in SAM
+  ! These are constants used in equations as part of the parameterisation.
+  ! Since the net is trained with them they should be left as defined here, rather than
+  ! forced to match the external model
+
+  ! Physical Constants:
+
+  != unit (m / s^2)
+  real(8), parameter :: ggr = 9.81
+      !! Gravity acceleration, m/s2
+
+  != unit J / kg :: lcond, lsub, lfus
+  real(8), parameter :: lfus = 0.3336e+06
+      !! Latent heat of fusion
+  real(8), parameter :: lcond = 2.5104e+06
+      !! Latent heat of condensation
+  real(8), parameter :: lsub = 2.8440e+06
+      !! Latent heat of sublimation, J/kg
+
+  != unit (J / kg) / K :: cp
+  real(8), parameter :: cp = 1004.
+      !! Specific heat of air
+
+  ! unit K :: fac_cond, fac_fus, fac_sub
+  real(8), parameter :: fac_cond = lcond/cp
+      !!
+  real(8), parameter :: fac_fus = lfus/cp
+      !!
+  real(8), parameter :: fac_sub = lsub/cp
+      !!
+
+  ! Temperatures limits for various hydrometeors
+  != unit K :: tprmin
+  real(8), parameter :: tprmin = 268.16
+      !! Minimum temperature for rain
+
+  != unit K :: tprmax
+  real(8), parameter :: tprmax = 283.16
+      !! Maximum temperature for snow+graupel, K
+
+  != unit K :: tbgmin
+  real(8), parameter :: tbgmin = 253.16
+      !! Minimum temperature for cloud water.
+
+  != unit K :: tbgmin
+  real(8), parameter :: tbgmax = 273.16
+      !! Maximum temperature for cloud ice, K
+
+  != unit K :: tbgmin
+  real(8), parameter :: tgrmin = 223.16
+      !! Maximum temperature for snow, K
+
+  != unit K :: tbgmin
+  real(8), parameter :: tgrmax = 283.16
+      !! Maximum temperature for graupel, K
+
+  ! Misc. microphysics variables
+  ! != unit 1 / K :: a_pr
+  real(8), parameter :: a_pr = 1./(tprmax-tprmin)
+      !! Misc. microphysics variables
+
+  != unit 1 / K :: a_bg
+  real(8), parameter :: a_bg = 1./(tbgmax-tbgmin)
+      !! Misc. microphysics variables
+
+  real(8), parameter :: an = 1./(tbgmax-tbgmin)
+  real(8), parameter :: bn = tbgmin * an
+  real(8), parameter :: ap = 1./(tprmax-tprmin)
+  real(8), parameter :: bp = tprmin * ap
+  real(8), parameter :: fac1 = fac_cond+(1+bp)*fac_fus
+  real(8), parameter :: fac2 = fac_fus*ap
+  real(8), parameter :: ag = 1./(tgrmax-tgrmin)
+
+  ! ---------------------------
+  ! SAM Grid Variables
+  ! ---------------------------
+  integer, parameter :: input_ver_dim = 30
+      !! Set to 48 in setparm.f90 of SAM. Same as nz_gl, but trained on a run with 30.
+
+  ! Outputs from NN are supplied at lowest 30 half-model levels for sedimentation fluxes,
+  ! and at 29 levels for fluxes (as flux at bottom boundary is zero).
+  integer, parameter :: nrf = 30
+      !! number of vertical levels the NN uses
+  integer, parameter :: nrfq = nrf - 1
+      !! number of vertical levels the NN uses when boundary condition is set to 0
+
+  real(8), parameter :: dt_sam = 30.0
+      !! SAM timestep in seconds
+
+!---------------------------------------------------------------------
+! Functions and Subroutines
+
+contains
+
+  != unit 1 :: omegan
+  real(8) function omegan(tabs)
+      != unit K :: tabs
+      real(8), intent(in) :: tabs
+          !! Absolute temperature
+
+      omegan = max(0., min(1., (tabs-tbgmin)*a_bg))
+
+      return
+
+  end function omegan
+
+
+  subroutine check(err_status)
+  !! Check error status after netcdf call and print message for
+  !! error codes.
+
+  integer, intent(in) :: err_status
+    !! error status from nf90 function
+
+  if(err_status /= nf90_noerr) then
+     write(*, *) trim(nf90_strerror(err_status))
+  end if
+
+  end subroutine check
+
+end module SAM_consts_mod