Can't set LES-Domain bl_pbl scheme to 0...

[nluchett]

Hi all,

I am running an LES-scale real case study and I am having issues with ./real.exe when setting the LES-domain's boundary layer scheme = 0. I am initializing with HRRR 3Km data with 3 domains (3km, 1km, and 100m)... therefore I only want to run the 100m domain in LES-mode... I've set diff_opt = 2, and km_opt = 2 as is instructed for my 100 m domain... My 3km and 1km domains I have the bl_pbl_physics set to 5... However, when I set my 100m domain's bl_pbl_physics to 0, and then run ./real.exe it times out and provides no errors in the log... But when I set the 100m domains bl_pbl_physics to 5, like the 3km and 1km domains, ./real.exe runs fine and so does ./wrf.exe....

I'm using WRFV4.0.3 and here is my namelist.input (version that has 100m domain's bl_pbl_physics set to 5):

Does anyone have any ideas on why it doesn't like me setting my LES domain bl_pbl_physics = 0? Am I missing something else that needs to be done to make this domain in LES-mode?

Thanks!

Nick Luchetti


&time_control
run_days = 0,
run_hours = 47,
run_minutes = 0,
run_seconds = 0,
start_year = 2013, 2013, 2013, 2013, 2013,
start_month = 06, 06, 06, 06, 06,
start_day = 29, 29, 30, 30, 30,
start_hour = 00, 00, 12, 12, 12,
end_year = 2013, 2013, 2013, 2013, 2013,
end_month = 06, 06, 06, 07, 07,
end_day = 30, 30, 30, 01, 01,
end_hour = 23, 23, 23, 06, 06,
interval_seconds = 3600
input_from_file = .true.,.true.,.true.,.true.,.true.,
history_interval = 180, 60, 1, 1, 1,
frames_per_outfile = 100000, 100000, 200, 200, 200,
restart = .false.,
restart_interval = 10,
io_form_history = 2
io_form_restart = 2
io_form_input = 2
io_form_boundary = 2
debug_level = 100
/

&domains
time_step = 5,
time_step_fract_num = 0,
time_step_fract_den = 1,
max_dom = 3,
e_we = 302, 316, 631, 154, 133,
e_sn = 302, 316, 631, 154, 133,
e_vert = 73, 73, 73, 73, 73,
p_top_requested = 10000,
num_metgrid_levels = 51,
num_metgrid_soil_levels = 8,
dx = 3000, 1000, 100, 333.3333, 111.1111,
dy = 3000, 1000, 100, 333.3333, 111.1111,
grid_id = 1, 2, 3, 4, 5,
parent_id = 0, 1, 2, 3, 4,
i_parent_start = 1, 110, 70, 15, 15,
j_parent_start = 1, 100, 70, 25, 15,
parent_grid_ratio = 1, 3, 10, 3, 3,
parent_time_step_ratio = 1, 3, 4, 4, 10,
feedback = 0,
smooth_option = 0,
eta_levels = 1.000000000000000,
0.998508104197351, 0.996925394125426, 0.995243951206412, 0.993457754922932, 0.991560435895111, 0.989545258100294, 0.987405100523324, 0.985132438271090,
0.982719323195605, 0.980157364082011, 0.977437706471938, 0.974551012208590, 0.971487438808267, 0.968236618783778, 0.964787639068830, 0.961129020719153,
0.957248699096215, 0.953134004773270, 0.948771645441416, 0.944147689135808, 0.939247549149480, 0.934055971054791, 0.928557022310711, 0.922734084998379,
0.916569852297947, 0.910046329396935, 0.903144839604523, 0.895846036537385, 0.888129923341037, 0.879975880015997, 0.871362700030039, 0.862268637515973,
0.852671466477644, 0.842548553554091, 0.831876946021130, 0.820633476838704, 0.808794888678084, 0.796337978981484, 0.783239768212893, 0.769477693546955,
0.755029830304960, 0.739875143474624, 0.723993771632570, 0.707367345512618, 0.689979343314522, 0.671815484609511, 0.652864164351548, 0.633116928024758,
0.612568988323634, 0.591219782947057, 0.569073572061315, 0.546140072721662, 0.522435126006340, 0.497981390782009, 0.472809055857948, 0.446956559775163,
0.420471304598278, 0.393410346824067, 0.365841044891482, 0.337841638785613, 0.309501732893113, 0.280922648613755, 0.252217608278586, 0.223511706654224,
0.194941620637448, 0.166655001434949, 0.138809486098476, 0.111571255837581, 0.085113055457951, 0.059611568867944, 0.035244015442395, 0.000000000000000,
/



&physics
!physics_suite = 'CONUS'
mp_physics = 8, 8, 8, 8, 8,
cu_physics = 0, 0, 0, 0, 0,
ra_lw_physics = 4, 4, 4, 4, 4,
ra_sw_physics = 4, 4, 4, 4, 4,
bl_pbl_physics = 5, 5, 5, 5, 5,
sf_sfclay_physics = 5, 5, 5, 5, 5,
sf_surface_physics = 5, 5, 5, 5, 5,
radt = 15, 15, 15, 15, 15,
bldt = 0, 0, 0, 0, 0,
cudt = 0, 0, 0, 0, 0,
icloud = 1,
isfflx = 1,
num_land_cat = 24,
sf_urban_physics = 0, 0, 0, 0, 0,
cu_rad_feedback = .true.,
do_radar_ref = 1,
/

&fdda
/

&dynamics
hybrid_opt = 2,
w_damping = 0,
diff_opt = 1, 1, 2, 2, 2,
km_opt = 4, 4, 2, 2, 2,
diff_6th_opt = 0, 0, 2, 2, 2,
diff_6th_factor = 0.12, 0.12, 0.12, 0.12, 0.12,
base_temp = 290.
damp_opt = 3,
zdamp = 5000., 5000., 5000., 5000., 5000.,
dampcoef = 0.2, 0.2, 1.2, 1.2, 1.2
khdif = 0, 0, 0, 0, 0,
kvdif = 0, 0, 0, 0, 0,
non_hydrostatic = .true., .true., .true., .true., .true.,
mix_isotropic = 0, 0, 1,
moist_adv_opt = 1, 1, 1, 1, 1,
scalar_adv_opt = 1, 1, 1, 1, 1,
gwd_opt = 1,
h_mom_adv_order = 5,
v_mom_adv_order = 3,
h_sca_adv_order = 5,
v_sca_adv_order = 3,
epssm = 0.1, 0.1, 0.5, 0.1,
/

&bdy_control
spec_bdy_width = 5,
spec_zone = 1,
relax_zone = 4,
specified = .true.,.false.,.false.,.false.,.false.,
nested = .false., .true., .true., .true., .true.,
/

&grib2
/

&namelist_quilt
nio_tasks_per_group = 0,
nio_groups = 1,
/

 

[kwerner]

Hi,
I'm not sure if this explains the reason for the failure with and without using PBL for d03, but I notice that you have a 10:1 grid resolution ratio between domains 2 and 3, which is probably going to give unreasonable results either way. We typically recommend odd numbered ratios, but since you have feedback turned off, in theory, having this as an even number is probably okay. However, we never recommend having a ratio any larger than about 7:1. Anything larger than that becomes too large of a jump, causing interpolation problems.

If you need to start with a resolution of 3km, and need to go down to a 100m domain, then you can try putting an additional domain between the 1km and 100m domains. Sometimes if the sizes of your domains are too different, though, this can cause problems with computing (the smaller domain won't be able to run with a large number of processors, while the large domain needs more than the small domain is capable of using). The size of your domains look okay at this point, though, so you may want to try.

I would recommend playing around with that first, then you can try to run the 100m domain with no PBL and see if that makes a difference.

 

[nluchett]

Hi Kwerner,

Thank you for the response and information.... I'm trying to mimic the NCAR operational WRF-Fire domain set up in which they have 2 nested domains (1km and 111 m) from a 3km HRRR initialization... I suppose then they must be doing a ~1:9 ratio from the 1km down to 111m? I will definitely try and adjust this, or consider a 4th domain in between 1km and 111m (however, as you noted, I've had issues with not being able to use as many cores in the past when adding too many domains of different sizes...).

I'm considering re-downloading the source code and starting from scratch. I'm really not sure why the pbl = 0 isn't working... I've done it before with older versions of WRF no problem...

Thanks again for the recommendations!

Nick