- 3D Transient Data - AGU request for additional diagnostics

Summary

The original single-level transient data highlighted the large variety of simulated tropical transient activity in the APE models.  However the data were insufficient to diagnose the vertical structure of dominant propagating features, necessary to catalogue and understand differences between the models and to compare with observed tropical structures, including the Madden-Julian Oscillation.

Additional 6-hourly multi-level data are now requested from a 1-year rerun of the APE control experiment.
Global data are requested, so that transient statistics can also be produced for the extratropical storm-tracks.

We request that you rerun a spun-up 360-day or 365-day period of the APE control experiment using the same model version as submitted to the main APE database, even if not an exact (bit-level) reproduction of the original experiment.  If you can contribute only with a modified model version, please contact the AGU group.  Data should be made available for download by AGU (ftp/http, or on DVD if data volumes prove too large for online transfer).

Data can be submitted to AGU from the beginning of March 2007, with a deadline of 31 May 2007.
Please contact the AGU group now if you can contribute to this additional data request.

Diagnostic Specification

Additional single-level and multi-level variables are requested, with a time resolution of 6-hours over a 360-day period.

Data should be submitted on the global horizontal grid.  If data processing and storage space make this impossible, tropical data (30N-30S) may be acceptable for most variables - please contact us.

Several single-level variables are requested in a single file MODEL-TAG_TR00_control.nc, similar to the original TR data request.

Each multi-level variable is requested in a separate file MODEL-TAG_TRnn_control.nc.   We recommend:

• multi-level dynamical variables (TR11 to TR17 below) are supplied on the 17 standard isobaric levels,
  using dimension metadata from the ML file
  
  
• multi-level tendency/forcing variables (TR18 to TR27 below) are supplied on model levels,
  using dimension metadata from the PF file

The (meta)data specification for the new variables follows the original TR transient data, with the following NetCDF attributes.

file_id \ diag_id \ variable   \  ape_name                          \   units
-----------------------------------------------------------------------------
_TR00_  \  TR01  \  tr_tppn    \  precipitation_flux                \  kg m-2 s-1
_TR00_  \  TR02  \  tr_lw_toa  \  toa_outgoing_longwave_flux        \  W m-2
_TR00_  \  TR06^ \  tr_mslp    \  air_pressure_at_sea_level         \  Pa
_TR00_  \  TR07  \  tr_slh     \  surface_upward_latent_heat_flux   \  W m-2
_TR00_  \  TR08  \  tr_ssh     \  surface_upward_sensible_heat_flux \  W m-2

file_id \ diag_id \ variable   \  ape_name                          \   units
-----------------------------------------------------------------------------
_TR11_  \  TR11^ \  tr_u       \  eastward_wind 	            \  m s-1
_TR12_  \  TR12^ \  tr_v       \  northward_wind	            \  m s-1
_TR13_  \  TR13^ \  tr_t       \  air_temperature	            \  K
_TR14_  \  TR14^ \  tr_om      \  omega 		            \  Pa s-1
_TR15_  \  TR15^ \  tr_z       \  geopotential_height	            \  m
_TR16_  \  TR16^ \  tr_q       \  specific_humidity	            \  1
_TR17_  \  TR17^ \  tr_rh      \  relative_humidity	            \  1

file_id \ diag_id \ variable     \  ape_name                                     \  units
-----------------------------------------------------------------------------------------
_TR18_  \  TR18* \  tr_t_pftotal \ Temperature tendency - Parameterization total \  K s-1
_TR19_  \  TR19  \  tr_t_sw      \ Temperature tendency - Short wave             \  K s-1
_TR20_  \  TR20  \  tr_t_lw      \ Temperature tendency - Long wave              \  K s-1
_TR21_  \  TR21  \  tr_t_turb    \ Temperature tendency - Turbulence             \  K s-1
_TR22_  \  TR22  \  tr_t_conv    \ Temperature tendency - Convection             \  K s-1
_TR23_  \  TR23  \  tr_t_cld     \ Temperature tendency - Cloud	                 \  K s-1
_TR24_  \  TR24* \  tr_q_pftotal \ Absolute humidity tendency - Parameterization total \  kg kg-1 s-1
_TR25_  \  TR25  \  tr_q_turb    \ Absolute humidity tendency - Turbulence       \  kg kg-1 s-1
_TR26_  \  TR26  \  tr_q_conv    \ Absolute humidity tendency - Convection       \  kg kg-1 s-1
_TR27_  \  TR27  \  tr_q_cld     \ Absolute humidity tendency - Cloud            \  kg kg-1 s-1

-----------------------------------------------------------------------------------------
compare original data:
_TR_	\  TR01  \  tr_tppn    \  precipitation_flux	      \  kg m-2 s-1
etc
-----------------------------------------------------------------------------------------

(^) denotes 6-hourly samples of instantaneous values - for all dynamical variables.
All other variables (fluxes and parameterization forcing) are averages over all timesteps in each 6-hour period.  The averaging period (centred on or preceding the nominal time) should be defined by the metadata, either the time_bounds or the cell_methods attribute.

(*) Total temperature and absolute humidity tendencies refer to total parameterization tendency (i.e. omitting the dynamical tendency).

Data / File Sizes

Assuming 32-bit float representation in NetCDF, a T42L30 model with a 128x64 grid produces the following approximate sizes for 360-days at 6-hour resolution.  Data for the tropics only (30S-30N) reduces file sizes by 2/3.

                               global grid       tropical grid    #variables
----------------------------------------------------------------------------
1 single-level variable =      47.2 MByte         15.8 MByte          5
1     17-level variable =       803 MByte        267.5 MByte          7
1     30-level variable =      1415 MByte        471.6 MByte         10
----------------------------------------------------------------------------
           total size   =      20.0 GByte          6.7 GByte