Skip to content

Ocean diagnostics

Ocean diagnostics in MOM6 are configured via the diag_table file, which controls all runtime diagnostic output. Full details of the diag_table format and semantics are available in the MOM6 docs.

In ACCESS-OM3, diag_table is treated as a derived file and should not be edited directly, as any manual changes may be overwritten when configurations are regenerated. Instead, it is generated from the user-editable configuration file diag_table_source.yaml using the Python helper script make_diag_table.py, which is available under /g/data/vk83/apps/make_diag_table/make_diag_table.py. Hence this documentation mainly focuses on the resulting diagnostic filename conventions, rather than the generation workflow.

ACCESS-OM3 diagnostic filename conventions

The diagnostic filename conventions used by ACCESS-OM3 configurations are summarised here. At a high level, diagnostic files follow the pattern: 

<file_prefix>.<model>.<dimension>[.<field>[+<vertical_coordinate>][+d2]].<frequency>.<time_cell_method>.<datestamp>.nc

Note

In practice, filenames fall into a small number of well-defined cases, described below.

Common components

  • <file_prefix>: Always access-om3,
  • <model>: Always mom6 for the ocean model,
  • <dimension>:
    • One of 2d or 3d for spatial diagnostics,
    • static for static grid data,
    • scalar for scalar diagnostics.

Field name or mode token

  • <field>
    • For standard 2d / 3d diagnostics, this is the diagnostic field name.
    • For special cases:
      • static files do not include a field name,
      • scalar files do not include a field name and instead contain multiple scalar diagnostics in a single file.

Note

The ACCESS-OM3 convention is to write one physical field per file, except in the two special cases above.

Temporal components

  • <frequency>: The output interval between records, following naming conventions given in this schema.
  • <time_cell_method>: Specifies how values are accumulated within each output interval, such as "mean", "max", "min", "snap" etc.
  • <datestamp>: Uses FMS time-string formatting, such as %4yr (4-digit year), other options can be %dy (day), %mo (month), %hr (hour), %mn (minute), %sc ("second") etc. The datestamp indicates the frequency at which new files are written. E.g. YYYY indicates yearly, YYYY-MM indicates monthly etc.

Note

static files do not include any temporal components

Optional spatial components

  • <vertical_coordinate>: Is included only for non-native vertical coordinates, such as z, rho2 etc

  • d2: Indicates half-resolution (downsampling) diagnostics. This is only supported for standard 2d / 3d spatial diagnostics and requires:

    • NIGLOBAL / Layout_X divisible by 2
    • NJGLOBAL / Layout_Y divisible by 2

    where NIGLOBAL and NJGLOBAL are the global horizontal grid sizes, and Layout_X, Layout_Y are the processor layouts. Related issue was discussed here.

Practical filename cases

In practice, with the above conventions ACCESS-OM3 uses three distinct filename classes.

1. Standard 2D / 3D diagnostics (most common)

access-om3.mom6.<2d|3d>.<field_name>[+<vertical_coordinate>][+d2].<frequency>.<time_cell_method>.<datestamp>.nc

Example: 

access-om3.mom6.3d.thkcello+rho2+d2.1mon.mean.1990.nc
Charateristics:

  • One 2d or 3d diagnostic field per file,
  • <vertical_coordinate> is included only for non-native vertical coordinates (e.g. z, rho2),
  • d2 indicates half-resolution output and is supported only for standard 2d / 3d diagnostics.

2. Static grid diagnostics

Static grid variables are written once per run and grouped into a single file. 

access-om3.mom6.static.nc
Characteristics:

  • No field name,
  • No frequency, time method, or datestamp,
  • No d2 support
  • Contains multiple static grid variables

3. Scalar (global integral) diagnostics

Global scalar diagnostics are grouped into a single file. 

access-om3.mom6.scalar.<frequency>.<time_cell_method>.<datestamp>.nc

Example:

access-om3.mom6.scalar.1day.snap.1990.nc

Characteristics:

  • No field name
  • Contains multiple scalar diagnostics (e.g. masso, soga, thetaoga, tosga)