[Wrf-users] Global WRF with nests and time-varying lower boundary conditions

[Eric Altshuler]

Hello,

I am using WRF 3.0.1.1 to run global simulations with nests and time-varying lower boundary conditions (sst_update = 1). I have found that the namelist.input.global file provided in WRFV3/test/em_real does not specify values for interval_seconds and input_from_file. Two problems arise from the omission of these two namelist variables:

1. If interval_seconds is not specified, it defaults to 43200 seconds (12 h), and the wrflowinp_d* files only contain 12-hourly lower boundary conditions even when I set the following:

 auxinput4_inname                    = "wrflowinp_d<domain>",
 auxinput4_interval                  = 360,  360,  360,

So even when I request 6-hourly data for the lower boundary condition file, I get 12-hourly output instead. Since I have a global domain and lateral boundary conditions are unnecessary (and wrfbdy_d01 is no longer produced in version 3.0.1.1), the value of interval_seconds (which pertains to lateral boundary conditions, not lower boundary conditions) shouldn't be relevant, but it still overrides the value of auxinput4_interval for lower boundary conditions.

2. If input_from_file is not specified, it defaults to .false. and no wrfinput_d02 file is produced for domain 2. Also, no wrflowinp_d02 is created. Apparently, if wrfinput and/or wrflowinp files are desired for nests, then input_from_file has to be set to .true. in namelist.input.global.

Also, I discovered that the wrflowinp_d* files are identical regardless of whether usemonalb is .true. or .false. Can someone explain why this is the case?

Finally, I intend to run a global domain with several nests and time-varying SST and sea ice. For a global domain with CONSTANT lower boundary conditions, it is only necessary to run WPS on the initial data. However, if time-varying lower boundary conditions are desired, then WPS must be run on the data for all times throughout the simulation period (3, 6, 12-hourly or whatever). In truth, only the lower boundary data (SST, sea ice) are required for multiple times; since this is a global simulation, the full 3-D atmospheric data are only really needed at the initial time, yet, as I understand things, WPS (specifically metgrid) cannot produce a met_em* file for a particular time if the 3-D meteorological data are missing. Is there some way to run WPS so that the initial met_em file contains all the 3-D and lower boundary data (as usual) but subsequent met_em files contain only the necessary time-varying lower boundary data (SST, sea ice, vegetation fraction and albedo)?

Thanks in advance for your feedback.

Best regards,

Eric L. Altshuler
Assistant Research Scientist
Center for Ocean-Land-Atmosphere Studies
4041 Powder Mill Road, Suite 302
Calverton, MD 20705-3106
USA

E-mail: ela at cola.iges.org
Phone: (301) 902-1257
Fax: (301) 595-9793