Documentation and experimentation description for the atmospheric spectral model GFDLSM V197

[ !!!Caution!!! User beware - this documentation is still under development]

---

Key documentation of model features is given by Gordon and Stern (1982) [2] , Gordon (1986 [3] , 1992 [4] ), and Gordon and Hovanec (1985) [5] , with additional details on the physics schemes provided by Miyakoda and Sirutis (1977 [12] , 1986 [6] ). Extended-range forecasting results are summarized by Miyakoda et al. (1979 [7] , 1986 [8] ), and by Stern and Miyakoda(1995) [33] .

Table of Contents

• Contact Information
• Modeling Group
• Model Representative(s)
• Experimental Implementation
• Experimental Design
• Simulation Period
• Earth Orbital Parameters
• Calendar
• Radiative Boundary Conditions
• Ocean Surface Boundary Conditions
• Orography/Land-Sea Mask
• Atmospheric Mass
• Physical Constants
• Spinup/Initialization
• Computer/Operating System
• Computational Performance
• Model Output Description
• Calculation of Standard Output Variables
• Sampling Procedures
• Interpolation Procedures
• Output Data Structure/Format/Compression
• Model Characteristics
• AMIP II Model Designation
• Model Lineage
• Model Documentation
• Numerical/Computational Properties
• Horizontal Representation
• Horizontal Resolution
• Vertical Domain
• Vertical Representation
• Vertical Resolution
• Time Integration Scheme(s)
• Smoothing/Filling
• Dynamical/Physical Properties
• Equations of State
• Diffusion
• Gravity Wave Drag
• Chemistry
• Radiation
• Convection
• Cloud Formation
• Precipitation
• Planetary Boundary Layer
• Sea Ice
• Snow Cover
• Surface Characteristics
• Surface Fluxes
• Land Surface Processes

Contact Information

Modeling Group

GFDL Experimental Prediction Group

Model Representative(s)

• Bill Stern, Tony Gordon, Joe Sirutis
• Geophysical Fluid Dynamics Laboratory/NOAA, Princeton University, P.O. Box 308, Princeton, New Jersey 08542.
• Phone: (609) 452-6545.
• (609) 987-5063.
• e-mail: ( wfs@gfdl.gov )..
• Internet World Wide Web Address: http://www.gfdl.gov/~wfs/gfdlsm/DocGFDLSM_v197.html .

Experimental Implementation

Experimental Design

• Ensemble of atmosphere only integrations for simulation period below, proposed size = 6-10.
• All ensemble members use the same prescribed observed SSTs (see Ocean Surface Boundary Conditions ).
• Ensemble members differ only through different intiial conditions used to initialize the atmosphere for 01 January 1979 (see Spinup/Initialization ).
• Experiment naming convention is as follows:
• atmNjan79 where N is defined as day in January 1979 from which the initial conditions were derived.
• Details of the Experimental Prediction Group's experimental naming convention can be found (access restricted to GFDL E-Group only) at: E-group Experiment Naming.

Simulation Period

• Start time : 00Z 1 January 1979.
• Stop time: 00Z 1 January 1997.

Earth Orbital Parameters

Values of

• Obliquity = 23.439 degrees.
• Eccentricity = 0.016718.
• Longitude of perihelion = 102.932 degrees.

Calendar

• Julian calendar used for model integration with leap years.
• vernal equinox defined as March x, where x = 20.41 - 0.0078(Y - 1987) + 0.25Y(modulo 4), Y is the year and Y(modulo 4) is the remainder after dividing Y by 4.

Radiative Boundary Conditions

• Solar constant = 1365 W/m**2.
• Solar cycles present: seasonal and diurnal, Short and long wave calculations every 2 hours.
• Carbon dioxide concentration = 348 ppm.
• Ozone concentration ???????? (recommended: zonal-average monthly climatology of Wang et al. 1995).

Ocean Surface Boundary Conditions

• Daily observed SSTs produced by MOM2 ocean model assimilation for domain from 65.5N -> 78.5S
• Reynolds monthly mean observed SSTs used in regions outside of ocean model domain (1979 -> 1992 EOF fitting analysis; 1993-> OI Aanalysis)
• Linear temporal interpolation of SSTs to precise model time.

Orography/Land-Sea Mask

• Orography obtained from a 1 x 1-degree Scripps dataset (Gates and Nelson 1975 [26] ) is interpolated to the model's Gaussian grid (see Horizontal Resolution ). The heights are transformed to spectral space, truncated at T42 resolution and then an isotropic filter to eliminate gibbs error is applied (Navarra, et al. 1994 [34] )
• The land-sea mask is based on the Scripps Orography dataset (described above) after interpolation to the model grid but before any spectral truncation.

Atmospheric Mass

Global-average value of model surface pressure ~ 979.14, approx. 0.4 hPa gain / decade.

Physical Constants

Model values of standard physical constants.(under construction).

Spinup/Initialization

• The model atmosphere is initialized for 1 January 1979 from NCEP re-analyses at 00Z selected from the first 15 days of January 1979.
• soil moisture and snow cover/depth are initialized from gcm climatologies produced from a 1979 - 1988 AMIP I rerun (i.e., exp = t42cr822).

Computer/Operating System

Model simulations run on a Cray T90 computer using a single processor in a UNICOS operating environment.

Computational Performance

Model computation time per simulated month Atmosphere only ~ 1.8 hrs. (or ~3.5 mn / day).

Model Output Description

Calculation of Standard Output Variables

• Method for calculation of percentage time that a pressure surface is below ground (recommended: procedure of Boer, 1995 Mon. Wea. Rev., 114, 885-902).
• Method for calculation of monthly mean tendencies at 17 WMO standard pressure levels:
• Temperature tendency due to total diabatic heating.
• Temperature tendencies due to short-wave and long-wave radiation.
• Temperature tendency due to moist convection.
• Temperature tendency due to dry convection.
• Temperature tendency due to large-scale/stratiform precipitation.
• Total moisture tendency due to diabatic processes.
• Method for calculation of cloud properties:
• cloud water/ice (if applicable).
• extinction coefficient (cloud optical thickness/layer depth).
• cloud emittance.
• Method for calculation of surface variables (recommended: procedure of Hess and McAvaney, 1995 Aust. Met. Mag., 44, 139-145)
• 10 m winds
• 2m specific humidity
• 2m temperature
• Method for calculation of mean sea-level pressure (recommended: ECMWF algorithm--code to be provided by PCMDI).
• Method for calculation of clear-sky radiation and cloud radiative forcing ( recommended: procedure of Potter et al., 1992 J. Geophys. Res., 97, 20507-20518).
• Method for calculation of potential vorticity, if supplied (recommended : procedure of Hoskins et al., 1985 Quart. J. R. Met. Soc., 111 , 877-946).
• Method for calculation of planetary boundary layer height, if supplied ( recommended: see Holtslag and Boville, 1993 J. Climate, 6, 1825-1842 and Vogelezang and Holtslag, 1996 Bound. Layer Met., 81 , 245-269).

Sampling Procedures

• Sampling procedure for calculation of monthly means of standard output variables (e.g., accumulation over every model time step vs accumulation of 6-hourly averages). Recommendations: See the AMIP II Guidelines for variable-dependent sampling procedures.
• Relevant references.

Interpolation Procedures

• Algorithm for interpolation of standard output variables to 17 WMO pressure surfaces (e.g., monthly averages computed in model coordinates, then interpolated to constant pressure surfaces vs monthly averages computed in model coordinates weighted by time varying mass on the 17 WMO pressure surfaces).
• Algorithm for treatment of variables on pressure surfaces below ground (if applicable). Recommended: ECMWF algorithm, if below-ground values are calculated--code to be provided by PCMDI.
• Relevant references.

Output Data Structure/Format/Compression

• Structure/format of output data (AMIP II specification: LATS structure in either NetCDF or GRIB format).
• Original and compressed word length of data (in bits per word) and description of compression algorithm.
• Relevant references.

Model Characteristics

AMIP II Model Designation

Institution_Acronym, Model/Version_Name, (Horizontal_Resolution Vertical_Resolution), Simulation_Year

Example: NCAR CCM3 (T42 L18) 1997

Model Lineage

• Predecessor model(s) from which the AMIP II model is descended, and the major changes made in the former in order to arrive at the latter.
• Designation of most similar model documented for AMIP I (see summary documentation at World Wide Web address
  
  
  
  
  http://www-pcmdi.llnl.gov/pcmdi/modldoc/amip/amip.html )
  
  Example: NCAR CCM2 (T42 L18) 1992

Model Documentation

Bibliography of key documents describing model characteristics.

Numerical/Computational Properties

Horizontal Representation

• Formulation of horizontal variation of model variables (e.g., second-order finite differences on a C grid, spectral basis functions with transformation to Gaussian grid, etc.).
• Variable-dependent differences in formulation, if present (e.g., spectral dynamical variables vs semi-Lagrangian grid-point water vapor and chemistry).

Horizontal Resolution

• Spectral triangular 42 (T42), using a transform grid roughly equivalent to 2.8125 x 2.8125 degrees latitude- longitude (Gorden and Stern 1982 [2] )

Vertical Domain

• Model top in units of hPa.
• Pressure of lowest atmospheric level (in hPa) when surface pressure is 1000 hPa.

Vertical Representation

• Vertical coordinates (e.g. sigma, modified sigma, hybrid, etc.).
• Vertical differencing scheme, conservation constraints if any.
• Relevant references.

Vertical Resolution

• Total number of vertical levels.
• Number of levels below 800 hPa and above 200 hPa for a surface pressure of 1000 hPa.

Time Integration Scheme(s)

• Type of scheme used for time integration.
• Explicit vs (semi-)implicit scheme.
• Corrector steps, method of damping computational mode, etc.
• Time filter, if any.
• Time step length for dynamics vs physics, frequency of recalculating radiative fluxes.
• Relevant references.

Smoothing/Filling

• Method of
• smoothing orography and/or other fields.
• filling spurious negative moisture values.
• correcting nonconserved quantities (atmospheric mass, etc.).
• Relevant references.

Dynamical/Physical Properties

Equations of State

• Variables used to define model history of state.
• Prediction equations and framework (e.g., primitive equations in a spectral Eulerian framework for dynamics, in a grid-point semi-Lagrangian framework for water vapor and chemistry).
• Relevant references.

Diffusion

• Representation of horizontal diffusion:
• variables to which applied and variable-dependent differences in application (e.g., diffusion of moisture treated by semi-Lagrangian formulation).
• linear vs nonlinear.
• numerical order (e.g., second-order or del-squared).
• variation in application according to height, horizontal resolution.
• application on constant sigma surfaces, pressure surfaces, etc.
• Representation of vertical diffusion above the surface layer:
• variables to which applied and variable-dependent differences in application.
• dependence on stability and/or height.
• local vs nonlocal.
• linear K-theory or another approach (e.g., prognostic turbulence kinetic energy scheme).
• Relevant references.

Gravity Wave Drag

• Description of gravity wave drag parameterization.
• Relevant references.

Chemistry

• Enumeration of radiatively active gases (CO2, ozone, water vapor, other trace gases) and aerosols.
• Description of main properties (diagnostic vs prognostic ozone, concentrations, temporal dependence, distributions--globally uniform, zonal mean, 3-D).
• Algorithm for chemistry transport, if present (e.g., grid-point semi-Lagrangian transport of water vapor, and on option, chemical tracers).
• Dataset references.

Radiation

• Treatment of clear-sky shortwave radiation, including number and boundaries (expressed in microns) of spectral bands.
• Treatment of clear-sky longwave radiation, including number and boundaries (expressed as wavenumbers in m-1) of spectral bands.
• Treatment of interactions with clouds, including dependence of shortwave/longwave optical properties on (prescribed, diagnostic, or prognostic) cloud liquid water; cloud vertical overlap assumptions for radiation calculations (e.g., full vs random overlap).
• Treatment of interactions with aerosols, if present.
• Relevant references.

Convection

• Treatment of penetrative (deep) convection.
• Treatment of shallow convection.
• Relevant references.

Cloud Formation

• Treatment of stratiform cloud (e.g., prognostic vs diagnostic, cloud types considered, cloud fraction determination, height-dependence)
• Treatment of sub-gridscale convective cloud
• Relevant references.

Precipitation

• Criterion for large-scale precipitation formation (including whether method is diagnostic or prognostic).
• Criterion for convective precipitation formation (including whether method is diagnostic or prognostic).
• Treatment of evaporation of falling large-scale and/or convective precipitation.
• Sub-gridscale distribution of precipitation over land, if applicable.
• Relevant references.

Planetary Boundary Layer

• Treatment of PBL (e.g., K-theory, nonlocal diffusion, prognostic turbulence kinetic energy (TKE) scheme, etc.).
• Determination of PBL top (e.g., prognostic scheme, stability-dependent algorithm, assumed sigma level, etc.).
• Relevant references.

Sea Ice

The prescription of the model's sea ice extent and concentration should be described in the experimental implementation of Ocean Surface Boundary Conditions . Here, provide relevant references and describe the method of determining:

• sea ice depth and surface temperature
• effects of snow on sea ice (e.g., effects on thermal properties, roughness, etc.), if applicable.

Snow Cover

• Criterion for formation, accumulation, and melting of prognostic snow, or reference to data set used for prescribed snow cover/depth. Value of snow density used for conversion of snowmass to snow depth (water equivalent).
• Algorithm for (sub-gridscale) fractional snow cover, if applicable.
• Contribution of sublimation to surface evaporative flux, effects of snow on surface roughness or thermal properties, if any.
• Relevant references.

Surface Characteristics

• Distinguished surface types (i.e., ocean, sea ice, continent, bare and vegetated ground, continental ice, etc.).
• Vegetation types, if distinguished.
• Roughness lengths for distinguished surface types (if spatially variable, provide dataset reference). Dependence of roughness on ocean surface wind stress, snow cover, etc.
• Specification of surface albedo for distinguished surfaces. Dependence of albedo on snow cover, soil moisture, solar zenith angle/wavelength, etc.
• Specification of longwave emissivity for distinguished surfaces.
• Relevant references.

Surface Fluxes

For distinguished surface types:

• Treatment of surface radiative fluxes.
• Treatment of turbulent eddy fluxes of surface momentum, heat, and moisture (e.g., Monin-Obukhov similarity theory implemented by the approach of...). Dependence of surface evaporation on vegetation, soil moisture, etc.
• Dependence of turbulent drag coefficients on stability, roughness, etc. Provision of different magnitudes of neutral drag coefficients for heat, moisture, momentum.
• Relevant references.

Land Surface Processes

• Method of computing soil heating, including number and depth of layers. Dependence of soil heat capacity/conductance on snow cover.
• Method of computing soil moisture, including sources and sinks, number and depth of layers, prescribed soil parameters, etc.
• Treatment of vegetation, including canopy moisture storage/evaporation, evapotranspiration mechanisms, etc.
• Relevant references.

Further Instructions for Modeling Group Representatives (after this document becomes final)

Obtaining a Template

You may obtain a copy of this documentation template in Hypertext Markup Language by saving it as a source file using a Web browser or by Internet ftp from address

sprite.llnl.gov/pub/phillips/amipdoc/doctemplate.html --under construction

Alternatively, you may obtain a plain text version of this template from ftp address

sprite.llnl.gov/pub/phillips/amipdoc/doctemplate.txt --under construction

or by requesting an e-mail copy from Tom Phillips ( phillips@pcmdi.llnl.gov )..

Note that the description of model characteristics only need include differences from the most similar model already documented for AMIP I--see the model summary documentation at Internet World Wide Web address

http://www-pcmdi.llnl.gov/pcmdi/modldoc/amip/amip.html .

Returning the Completed Template

Whenever possible, please return a completed copy (Unix-compressed if necessary) of this documentation template by electronic mail to Tom Phillips ( phillips@pcmdi.llnl.gov ). Otherwise, please fax a printed copy to Tom at +1-510-422-7675 or send it to him by surface mail at PCMDI, L-264, Lawrence Livermore National Laboratory, P.O. Box 808, Livermore, California (USA) 94551.