4/3-km Simulations of WRF version 3.1.1 for Cases Nov 2009-Jan 2010, Lowest-Sigma Level Data

The Problems

Winds are too strong over the land. Daytime high temperatures over the Intermountain West are too cold.

Working on Solutions

To help fix these, several approaches have been tried:
1) turning off the Noah LSM
2) modifying surface friction through roughness length or vertical diffusion
3) modifying the roughness length or friction velocity based upon the variance (36-km variance WA 36-km variance, 12-km variance, WA 12-km variance) of the subgrid scale terrain.

Click here for the statistics and plots that use PBL-scheme 2-m and 10-mtemperatures and winds instead of the lowest sigma level values.


The loops in this table illustrate the attempts and how they impact the simulations.

Table 1. Standard Plots
InitializationDomainProductHourCases (control click to select multiple)
00Z Jan 7, 2010 is a really bad case for standard runs

Table 2. Geographic, Mean Statistics Plots
PeriodFieldStatisticHourStation GroupCases (control click to select multiple)
Case Name(s)
namelist.input file
namelist.input diff from std		 Description
1. UW WRF-GFS
std
namelist.input		current standard run, with Kain-Frisch
2. wrfapr08
namelist.input
diff from std		 v3.0.1.1, standard apr08 version
3. wrfv311stauffer2
namelist.input
diff from std		 v3.1.1, halved vertical mixing coefficients K, no KF, no 6th order diffusion, new Q nudging
4. wrfv311stauffer3
namelist.input
diff from std		 v3.1.1, 1/4 vertical mixing coefficients K, no KF, no 6th order diffusion, new Q nudging
5. wrfv311ustar3
namelist.input
diff from std		 v3.1.1, Increased ustar (friction velocity) by factor of 2 only over land, no 6th order diffusion, new Q nudging
6. wrfv311ustar4
namelist.input
diff from std		 v3.1.1, Increased ustar (friction velocity) by factor of 1.5 only over land, no 6th order diffusion, new Q nudging
7. wrfv311ovens1
namelist.input
diff from std		 v3.1.1, Combination of wrfv311stauffer2 wrfv311ustar4: 1/2 vertical mixing coefficients and 1.5 times increase in ustar (friction velocity) only over land, no 6th order diffusion, new Q nudging
8. wrfv311u3s2
namelist.input
diff from std		 v3.1.1, Combination of wrfv311stauffer2 wrfv311ustar3: 1/2 vertical mixing coefficients and 2 times increase in ustar (friction velocity) only over land, no 6th order diffusion, new Q nudging
9. wrfv311u3s2oldlsm1
namelist.input
diff from std		 v3.1.1, Combination of wrfv311stauffer2 wrfv311ustar3: 1/2 vertical mixing coefficients and 2 times increase in ustar (friction velocity) only over land and some modifications attempting to use older Noah LSM, no 6th order diffusion, new Q nudging
10. wrfv311oldqn
namelist.input
diff from std		 v3.1.1, Older (apr08) nudging coefficients of moisture and temperature fields
11. wrfv311modis
namelist.input
diff from std		 v3.1.1, Used new high-resolution MODIS landuse.
12. wrfv311nolsm
namelist.input
diff from std		 v3.1.1, Removed Noah LSM and used thermal diffusion scheme.
13. wrfv311u4nolsm
namelist.input
diff from std		 v3.1.1, 1.5 times increase in ustar (friction velocity) only over land (ustar4) and removed Noah LSM and used thermal diffusion scheme.
14. wrfv311sfz2
namelist.input
diff from std		 v3.1.1, 2 times increase in roughness length SFZ0 over land and water, PLUS 1.5 times increase in ustar (friction velocity) only over land (ustar4), plus removed Noah LSM and used thermal diffusion scheme.
15. wrfv311sfz3
namelist.input
diff from std		 v3.1.1, 1.5 times increase in roughness length SFZ0 over land and water, plus removed Noah LSM and used thermal diffusion scheme.
16. wrfv311sfz4
namelist.input
diff from std		 v3.1.1, 1.25 times increase in roughness length SFZ0 over land, plus removed Noah LSM and used thermal diffusion scheme.
17. wrfv311sfztervarnolsm
aka wrfv311sfztervarnolsnamelist.input
diff from std		 v3.1.1, roughness length SFZ0 modified over land using terrain variance, plus removed Noah LSM and used thermal diffusion scheme.
Linear Formula applied to variances >= 25: sfz0*min(10.0,var*8/475 + 30/19)
Formula plot
18. wrfv311sfz5
namelist.input
diff from std		 v3.1.1, roughness length SFZ0 modified over land using terrain variance, plus removed Noah LSM and used thermal diffusion scheme.
Linear Formula applied to variances >= 1: sfz0*min(8.0,var*6/475 + 32/19)
Formula plot
19. wrfv311sfz6
namelist.input
diff from std		 v3.1.1, roughness length SFZ0 modified over land using terrain variance, plus removed Noah LSM and used thermal diffusion scheme.
Linear Formula applied to variances >= 1: sfz0*min(8.0,var*8/475 + 30/19)
Formula plot
20. wrfv311sfz7
namelist.input
diff from std		 v3.1.1, roughness length SFZ0 modified over land using simple factor of 7.5 increase, plus removed Noah LSM and used thermal diffusion scheme.
Formula = sfz0 * 7.5
NOTE: this case had stability problems and missed 4 runs
21. wrfv311sfz8
namelist.input
diff from std		 v3.1.1, roughness length SFZ0 modified over land using terrain variance, plus removed Noah LSM and used thermal diffusion scheme.
Linear Formula applied to variances >= 1: sfz0*min(8.0,var*4/381.25 + 4)
Formula plot
22. wrfv311ustar5
namelist.input
diff from std		 v3.1.1, friction velocity, ustar, modified over land using terrain variance, plus removed Noah LSM and used thermal diffusion scheme.
Linear Formula applied to ustar based on variance: var >= 0, max(1,min(3,var*2/500+0.65))
Formula plot

Statistics Jan 1-Feb 8 2009 for wind speeds >= 3 knots