Changes between Initial Version and Version 1 of PDAFomi_obs_op_gatheronly


Ignore:
Timestamp:
Jun 11, 2025, 4:07:58 PM (6 days ago)
Author:
lnerger
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • PDAFomi_obs_op_gatheronly

    v1 v1  
     1= PDAFomi_obs_op_gatheronly =
     2
     3This page documents the routine `PDAFomi_obs_op_gatheronly` of PDAF-OMI.
     4
     5This routine provides the application of observation operator for the case that model variables are observed in a coupled model system only in another model component. Thus, `dim_obs_p=0` for the calling model component.  Accordingly, this observation operator only performs the gather operation to obtain the full observations. An example for using this observation operator is in the case of a coupled atmosphere-ocean model in which the atmosphere component model runs with different processes than the ocean components. If one assimilates observations of the ocean with strongly-coupled data assimilation into the atmosphere in this system, the atmosphere would call `PDAFomi_obs_op_gatheronly`, while the ocean would call, e.g., an observation operator involving interpolation.
     6
     7The routine is usually called in `obs_op_OBSTYPE` in an [wiki:OMI_observation_modules_PDAF3 OMI observation module].
     8
     9|| See the [wiki:OMI_observation_operators_PDAF3 page describing PDAF-OMI observation operators] for a general overview.||
     10
     11The interface is:
     12{{{
     13  SUBROUTINE PDAFomi_obs_op_gatheronly(thisobs, state_p, obs_f_all)
     14
     15    TYPE(obs_f), INTENT(inout) :: thisobs  ! Data type with full observation
     16    REAL, INTENT(in)    :: state_p(:)      ! Process-local model state provided by PDAF
     17    REAL, INTENT(inout) :: obs_f_all(:)    ! Full observed state for all observation types (array provided by PDAF)
     18}}}
     19
     20'''Note:'''
     21 * This observation operator requires that the filter processes for all coupled model components are in the same MPI communicator, so that a joint multi-component state vector exists.