WRF Forecast Testing

Extensive testing and evaluation of various physics options have been performed to ensure these schemes 
can realistically describe physical and dynamical process in the atmosphere. The tested schemes pull 
heavily from the new contributions from users, but the suite also includes older options to allow 
comparisons with packages whose behavior is more thoroughly known. 

The WRF ARW tests are conducted for the summer period from June 1 to June 28, 2015, and for the winter 
period from January 1 to January 28 2016.
  
- Domain Configuration:

CONUS domain with 15-km resolution, as shown in figure below.



Grid Structure:
401 x 261 horizontal grid points
40 vertical levels with model top at 30 hPa
Lambert-Conformal map projection

- Initial and boundary conditions:
NCEP 0.25 degree GFS analyses at 6-h intervals are used to provide the initial and boundary conditions 
for each of the testing runs.  

Each single case is initialized daily at 0000 UTC and run for 48-hour forecast. For both 
summer and winter, 28 cases are conducted for testing and evaluation.

- Physics options 

Various combinations of physics options are tested using WRFV4.1. A baseline test is conducted using 
WRFV4.0.1 for comparison. 

Table 1 Configurations for tests conducted (unless specified, all tests are based on WRFV4.1, in which 
hybrid vertical coordinate is the default option)



- Verification
All WRF results are initialized with and compared to 0.25 degree GFS surface and upper 
air analysis.  Domain averaged BIAS and RMSE are calculated for conventional fields.
The results of the STD suite are also verified against stations observations and soundings 
archived in GDAS historical dataset.