Metadata diference between CMIP5 output and Questionnaire

[charlotte]

A. Charlotte Report on mis-match between the forcing metadata required in the CMIP5 netCDF headers and the forcing info captured in the conformance section of the quesitonnaire.

[charlotte]

I made a  wiki page for documenting the changes to the list of attributes
from  CMOR user guide version 2.0 Jan 19 2010
to  CMIP5 Model Output Requirements April 28 2010
via Requirements for CMIP5 Output, March 19 2010

New attributes in the CMIP5 Model Output Requirements document are

• branch time: time in parent experiment when this simulation started
• product = "output": indicates that the data you are writing is model ouput
• parent_experiment_id: indicates which experiment this simulation branched from
• cell_measures: string that appears in the cell_measures column of the  CMIP5 Requested Output tables
• flag_values: defined with values taken from the "flag values" columns of the  CMIP5 Requested Output tables
• flag_meanings: defined with values taken from the "flag meanings" columns of the  CMIP5 Requested Output tables
• grid_mapping: name of the variable containing the grid_mapping information

Attributes that are not in the CMIP5 Model Output Requirements document are

• cmor_version
• grid_mapping_name: names of various mapping parameters

[charlotte]

Forcing

We can add the DRS forcings as a drop down list on the simulation page.

• GHG includes only CO2 and methane
• N/A for control type runs with no forcing

A list of the forcing controlled vocabulary is on ticket #671

Model output

Model output can now be submitted in native format: even if this is not cartesian latlon.

Parent_experiment_id

we caputre this implicity in the questionnaire because all simulations are associated with a related simulaiton and all simulations are associated with experiments.

[charlotte]

branch time in questionnaire and branch time in CMOR is different

[charlotte]

instructions for realisation in CMIP5 documentation do not make sense.

[charlotte]

I think the forcing labels refer to the individual forcins specified for the 7.3 historical??? experiments. I've asked Karl Taylor for clarification.

[charlotte]

We are simplifying the way we ask for forcing inputs so that in effect we will only require a text description.  This will compliment rather than duplicate what is captured in the netcdf headers.