1. Shortwave Radiation Options in the WRF Model
An oh-so fascinating study of the Dudhia, Goddard and RRTMG shortwave schemes

2. Radiation in the WRF Current Schemes:
All single column, 1-D schemes – each column treated independently
Good approximation if vertical depth is much less than horizontal scale
Radiation schemes resolve atmospheric heating from:
Radiative flux divergence
Surface downward longwave and shortwave radiation [for ground heat]
Shortwave radiation:
Includes wavelengths of solar spectrum
Accounts for absorption, reflection and scattering in atmosphere and on surfaces
Upward flux dependent on albedo
In atmosphere, determined by vapor/cloud content, as well as carbon dioxide, ozone and trace gas concentrations

3. Dudhia Scheme ra_sw_physics = 1
Based on Dudhia 1989, from MM5
Uses look-up tables for clouds from Stephens 1978
Version 3 has option to account for terrain slope and shadowing effects on the surface solar flux
Simple downward integration of solar flux, which accounts for:
Clear air scattering
Water vapor absorption [Lacis and Hansen, 1974]
Cloud albedo and absorption

4. Goddard Scheme ra_sw_physics = 2
Based on Chou and Suarez 1994
Includes 11 spectral bands
Different climatological profiles available for numerous ozone options
Considers both diffuse and direct solar radiation in 2- stream approach, accounts for scattering and reflection

5. RRTMG Scheme ra_sw_physics = 4
Uses MCICA [Monte Carlo Independent Column Approximation] method of random cloud overlap – statistical method to resolve sub-grid scale cloud variability
Finer resolution runs usually associated with WRF model means that clouds will most likely take up the entire grid space [binary clouds], in which case MCICA will not work.

6. Temperature ----- Meeting Notes (11/29/10 10:20) -----
No significant differences, perhaps some reducting in warm layer near surface in schemes 1 and 2 around sunrise-ish.

7. Relative Humidity ----- Meeting Notes (11/29/10 10:20) -----
Some drying at mid levels in Option 4 near the end of the run.

8. Zonal Winds ----- Meeting Notes (11/29/10 10:27) -----
Odd - weak winds thoughout the plot, especially aloft.

9. Meridional Winds ----- Meeting Notes (11/29/10 10:27) -----
Very similar.

10. Vertical Winds ----- Meeting Notes (11/29/10 10:27) -----
Note diurnal vertical motions.

11. Top of Atmosphere Radiation Longwave Radiation Upward
----- Meeting Notes (11/29/10 10:27) -----
Units: Watts/(m^2)

12. Top of Atmosphere Radiation Longwave Radiation Upward Differences

13. Surface Radiation Longwave

14. Surface Radiation Longwave Differences

15. Surface Radiation Shortwave
----- Meeting Notes (11/29/10 10:37) -----
Goddard scheme initializes almost 200 w/(m^2) below other 2 schemes, and overshoots later in the model run.

16. Surface Radiation Shortwave Differences

17. Surface Radiation Longwave Radiation Upward

18. Surface Radiation Longwave Radiation Upward Differences

19. Surface Radiation Longwave Radiation Downward

20. Surface Radiation Longwave Radiation Downward Differences

21. Surface Heat Flux Ground Heat

22. Surface Heat Flux Ground Heat Differences

23. Surface Heat Flux Sensible Heat

24. Surface Heat Flux Sensible Heat Differences

25. Surface Heat Flux Latent Heat

26. Surface Heat Flux Latent Heat Differences

27. Significant Variations and Conclusions
Goddard Scheme (ra_sw_physics=2) initialized differently and gave the most extreme values
Most variations were insignificant, other than mid-level drying in RRTMG scheme.
Much larger flux differences arise if clouds are sparse or absent during peak diurnal heating
Surface fluxes
Clear sky conditions – algorithmic differences in handling gaseous absorption/emission of longwave radiation and extinction of shortwave radiation
Differences in initial concentrations of trace gases
Differences in allowable cloud fractions

28. Resources
“Assessment of Radiation Options in the Advances Research WRF Weather Forecast Model”, Iacono and Nehrkorn
“A Description of the Advanced Research WRF Version 3”, Skamarock et al.