Fields to couple

As the chemistry is so numerically expensive, we want two UMs: a senior UM at N216 which has most of the science, except the chemistry; and a junior UM at N96 providing the chemistry. The junior UM should be locked to senior UM by overwriting core fields, which are passed from the senior UM, while the junior UM will pass back chemistry (and maybe aerosol).

The dynamical core fields

The dynamical core fields for ENDGame are the following 3D fields

  • U
  • V
  • W
  • θvd: Virtual dry potential temperature
  • Dη/Dt
  • ρdry
  • π: Exner function
  • The mX fields:
    • m_v: mass mixing ratio of water vapour
    • m_cl: mass mixing ratio of mean liquid water content
    • m_cf: mass mixing ratio of ice water content
    • IF (l_mcr_qcf2), m_cf2: Second prognostic for cloud ice (I don't how l_mcr_qcf2 is set, it's set to FALSE in mphys_input_mod.F90)
    • IF (l_mcr_qrain), m_r: mass mixing ratio of rainwater content (it's set in namelist RUN_PRECIP in SHARED)
    • IF (l_mcr_qgraup), m_gr: mass mixing ratio of graupel (it's set in namelist RUN_PRECIP in SHARED)

and I think all the fields used require a degraded version of the Senior UM fields overwriting the Junior UM fields.

Overwriting U from Senior UM to Junior UM was pretty successful with only a few overwriting fields. Overwriting θ showed that it's important to overwrite the dynamical core fields - and this case this is more θvd than θ. The overwrite of θ was soon dropped after those plots were made, and seemed to make no difference.

Boundary conditions and surface influence

Clearly surface fields, such as sea surface temperature, will impact on the core dynamical fields. However, is this through boundary conditions or source terms to the dynamical core equations?

The equations (see section 7.8 in ENDGame formulation state that below the lowest half level, ∂u/∂η=∂v/∂η=0, i.e. free-slip (well Nigel actually argues that applying a no-slip condition instead, would probably lead to the same thing). Many of the source terms are also assumed to be `piecewise constant' below the lowest level. All this doesn't seem to leave room for boundary conditions to impact the flow, and so surface influences must come through the source terms.