Skip to content

Description

Jump to bottom
Rick Pernak edited this page Feb 1, 2021 · 7 revisions

RRTMG_SW is a radiative transfer model that utilizes the correlated-k approach to calculate shortwave fluxes and heating rates efficiently and accurately for application to GCMs.

Clear sky comparison of the latest version of RRTMG_SW (using 2 streams) relative to RRTM_SW (using 16 streams with DISORT).

Key Features

Key features of RRTMG_SW are:

  • Absorption coefficient data for the k-distributions are obtained directly from the line-by-line radiative transfer model, LBLRTM, which has been extensively validated against observations, principally at the ARM SGP site. Data are consistent with those used in RRTM_SW_v2.5, which is described on its Wiki.
  • Fluxes and heating rates can be calculated over fourteen contiguous bands in the shortwave (820-50000 cm-1, or 0.2-12.20 microns). The individual band ranges (in wavenumbers, cm-1) are:
Band v1 v2
1 2600 3250
2 3250 4000
3 4000 4650
4 4650 5150
5 5150 6150
6 6150 7700
7 7700 8050
8 8050 12850
9 12850 16000
10 16000 22650
11 22650 29000
12 29000 38000
13 38000 50000
14 820 2600

The last band is coded out of sequence to preserve spectral continuity with the longwave bands.

  • Modeled sources of extinction are water vapor, carbon dioxide, ozone, methane, oxygen, nitrogen, aerosols, and Rayleigh scattering
  • A two-stream algorithm is used to perform scattering calculations; Reference: Oreopoulos and Barker (1999).
  • Uses reduced set of g-intervals (112) for integration over extinction in each band relative to full set of g-intervals used in RRTM_SW (224).
  • Includes McICA (Monte-Carlo Independant Column Approximation) capability to represent sub-grid cloud variability with random, maximum-random, maximum, exponential, and exponential-random options for cloud overlap; the exponential and exponential-random methods allow specification of the required decorrelation length as a constant or as a variable that varies as a function of latitude and day of the year; References: Barker et al. (2003); Pincus et al., JGR, (2003); Oreopoulos et al., ACP, (2012)
  • Coding has been reformatted to use many FORTRAN90 features.
  • Model able to run either as a column model or as a callable subroutine.
  • Clear sky agreement with RRTM_SW is generally within 3 W m-2 for flux and for heating rate within 0.1 K day-1 in the troposphere and 0.3 K day-1 in the stratosphere
  • Water clouds:
    • The optical properties of water clouds are calculated for each spectral band from the Hu and Stamnes parameterization. The optical depth, single-scattering albedo, and asymmetry parameter are parameterized as a function of cloud equivalent radius and liquid water path.
    • Reference: Hu, Y. X., and K. Stamnes, An accurate parameterization of the radiative properties of water clouds suitable for use in climate models. J. Climate, Vol. 6, 728-742, 1993.
  • Ice clouds:
    • The optical properties of ice clouds are calculated for each spectral band from the Fu parameterization, which assumes the ice crystals are hexagonal and randomly-oriented in space. The optical depth, single-scattering albedo, and asymmetry parameter are parameterized as a function of the generalized effective size of the ice crystals and the ice water content. (It is important to note that the generalized effective size is not equivalent to the effective mean size.)
    • Reference: Fu, Q., An accurate parameterization of the solar radiative properties of cirrus clouds for climate models. J. Climate, 9, 2058-2082, 1996.
  • Absorption coefficients and other initialization data can be optionally input through a netCDF data file. This feature was developed and provided by Patrick Hofmann and Robert Pincus of NOAA.
  • Several options for treating solar variability are available based on the NRLSSI2 solar source model, which allows the use of a fixed solar constant (1360.85 W m-2), a time-varying total solar irradiance based on the mean solar cycle, or a varying total solar irradiance defined by specific time-varying input indices for facular brightening and sunspot dimming. The constant total solar irradiance (1368.22 W m-2) based on the Kurucz model in all versions of RRTMG_SW prior to v4.0 also remains available.

RRTMG_SW vs. RRTM_SW

Key differences between RRTMG_SW and RRTM_SW are:

  • RRTMG_SW uses a two-stream algorithm for multiple scattering, and RRTM_SW uses DISORT; Reference: Oreopoulos and Barker (1999).
  • RRTMG_SW uses reduced set of g-intervals (112) for integration over extinction in each band relative to full set of g-intervals used in RRTM_SW (224).
  • RRTMG_SW includes McICA (Monte-Carlo Independant Column Approximation) capability to represent sub-grid cloud variability with random, maximum-random, maximum, exponential, and exponential-random options for cloud overlap; References: Barker et al. (2003), Pincus et al., JGR, (2003); Oreopoulos et al., ACP, (2012); RRTM_SW does not include McICA and does not have cloud fraction capability; it is limited to clear or fully overcast cloud conditions.
  • RRTMG_SW coding has been reformatted to use many FORTRAN90 features.
  • RRTMG_SW can be used as a callable subroutine and adapted for use within global or regional models.
  • RRTMG_SW can optionally read the required input data either from a netCDF file or from the original RRTM_SW source data statements.
  • RRTMG_SW can treat the solar source function and solar variability based on the NRLSSI2 solar model. RRTM_SW uses the Kurucz solar source function with a fixed solar constant of 1368.22 W m-2.
Clone this wiki locally