2Dhill_problem with pressure perturbation

Anything related to the idealized simulations in WRF.

2Dhill_problem with pressure perturbation

Postby MaVy » Sat Jul 05, 2014 2:32 pm

Hi all,
I am having some trouble to understand how WRF computes the pressure perturbation.
I ran a basic simulation using the "2D flow over hill" idealized case, I changed the terrain function in the fortran code in order to obtain a 3D circular mountain (I did also some other small modifications to the default input sounding (which is "no moisture" anyway)). I used open lateral boundary conditions (open_xs, open_xe, open_ys, open_ye) and a sponge layer at half-height of domain, also, I switched off every type of ground-atmosphere interaction (PBL, heat fluxes etc..).
At the end of this message you can find the namelist that I used.
The aim is to visualize only mountain waves and the related perturbations on the background state.

The results that I get are pretty good for all the wave perturbations (w, u, potential temperature) except for pressure. In fact, the pressure perturbation becomes, in time, positive in everywhere within the domain.
If I run, for example, a simulation of 24 hours and I take a horizontal cross-section of the pressure pert. at k=1, t=12 I cannot see: positive pressure pert. along the windward side of the mountain and negative pressure pert. along the lee side (as in the first hours of simulation (for t=1:10) ), but only positive pressure even if the dipole (increased-decreased pressure along the two sides of the mountain) is still present.
This simulation is very basic, and there isn't anything in the domain that can be responsible of a changing in pressure, if no waves. Therefore... what is going on in the domain?
More specifically my question is: when I visualize the perturbation of pressure p this value is the sum of a perturbation due to waves pw (the dipole is still present) plus something else puknw. What is this something else? and... am I right to worry about this?

p=pw+puknw

In the end: I am sure this problem is related to the boundary conditions used. If I convert open BC in periodic BC I haven't this problem any more (positive pressure on windward side and negative pressure on lee side for all the time of simulation), so.. it looks like a problem of mass conservation... but could it be??
Furthermore, I tried to run the default WRF simulation for a 2Dhill idealized case (with periodic BC in y and open BC in x) and my simulation using these same boundary conditions, but in both simulations (my and default) the pressure becomes negative (this time!) in everywhere within the domain, in time... ).

If someone who knows very well how WRF works, can explain me what is going on in the domain during the pressure perturbation calculation... I'll be grateful .

Thank you very much,



Namelist employed:

&time_control
run_days = 0,
run_hours = 24,
run_minutes = 00,
run_seconds = 0,
start_year = 0001,
start_month = 01,
start_day = 01,
start_hour = 00,
start_minute = 00,
start_second = 00,
end_year = 0001,
end_month = 01,
end_day = 01,
end_hour = 24,
end_minute = 00,
end_second = 00,
history_interval = 60,
frames_per_outfile = 1000,
restart = .false.,
restart_interval = 0,
io_form_history = 2
io_form_restart = 2
io_form_input = 2
io_form_boundary = 2
debug_level = 0
/

&domains
time_step = 20,
time_step_fract_num = 0,
time_step_fract_den = 1,
max_dom = 1,
s_we = 1,
e_we = 100,
s_sn = 1,
e_sn = 100,
s_vert = 1,
e_vert = 200,
dx = 2000,
dy = 2000,
ztop = 20000,
hypsometric_opt = 1,
numtiles = 8,
/

&physics
mp_physics = 0,
ra_lw_physics = 0,
ra_sw_physics = 0,
radt = 0,
sf_sfclay_physics = 0,
sf_surface_physics = 0,
bl_pbl_physics = 0,
bldt = 0,
cu_physics = 0,
cudt = 0,
isfflx = 0,
/

&fdda
/

&dynamics
rk_ord = 3,
diff_opt = 0,
km_opt = 1,
damp_opt = 0,
zdamp = 10000.,
dampcoef = 0.003,
khdif = 00,
kvdif = 0,
smdiv = 0.1,
emdiv = 0.01,
epssm = 0.1,
time_step_sound = 6,
h_mom_adv_order = 5,
v_mom_adv_order = 3,
h_sca_adv_order = 5,
v_sca_adv_order = 3,
non_hydrostatic = .true.,
pert_coriolis = .false.,
/

&bdy_control
periodic_x = .false.,
symmetric_xs = .false.,
symmetric_xe = .false.,
open_xs = .true.,
open_xe = .true.,
periodic_y = .false.,
symmetric_ys = .false.,
symmetric_ye = .false.,
open_ys = .true.,
open_ye = .true.,
/

&grib2
/

&namelist_quilt
nio_tasks_per_group = 0,
nio_groups = 1,
/
MaVy
 
Posts: 2
Joined: Sat Jul 05, 2014 12:44 pm

Re: 2Dhill_problem with pressure perturbation

Postby MaVy » Wed Jul 23, 2014 2:10 pm

To whom it may interest:

I wrote to Wrfhelp, explaining my problem. The experts replied me in this way:

"The problem is with the open boundary conditions which have no mass constraint and can drift when perturbations reach them. Basically these boundary conditions don't have a memory of the initial inflow and outflow mass fluxes. The problem can only be delayed by moving the boundaries further from the area of interest. Nesting could be done to help this if running 3d cases. Or, as was mentioned, using periodic conditions won't have a mass drift, so a combination of nesting and periodic outer boundary conditions may work best. Other solutions may involve code changes to relax the boundaries towards the initial state.

wrfhelp "
MaVy
 
Posts: 2
Joined: Sat Jul 05, 2014 12:44 pm


Return to Idealized Simulations

Who is online

Users browsing this forum: No registered users and 2 guests

cron