Initialising em_tropical_cyclone for the southern hemisphere

[pegwan]

I am attempting to run an idealised tropical cyclone simulation for the southern hemisphere, however I am running into some trouble when initialising the vortex.

In order to change the starting conditions, I am modifying 'module_initialize_tropical_cyclone.F' (user settings shown below).

Flipping the sign of the Coriolis parameter results in the intended Coriolis force for the southern hemisphere, however the actual winds of the initial vortex remain spinning counter-clockwise:

!-----------------------------------------------------------------------
! USER SETTINGS

! Parameters for analytic vortex:
! Reference: Rotunno and Emanuel, 1987, JAS, p. 549

r0 = 600000.0 ! outer radius (m)
rmax = 135000.0 ! approximate radius of max winds (m)
vmax = 15.0 ! approximate value of max wind speed (m/s)
zdd = 20000.0 ! depth of vortex (m)


! other settings:

fcor = -5.0e-5 ! Coriolis parameter (1/s)
sst = 30.0 ! sea-surface temperature (Celsius)

!-----------------------------------------------------------------------

For context, below is an ncview visualisation of horizontal-axis winds using the above settings:


Unfortunately, flipping the sign of vmax seemingly has no effect, with the vortex appearing identical to the image above.

Flipping the signs of r0 and rmax results in the complete absence of a vortex.

What actions can I take to create the desired simulation? Is there something here I have missed?

 

[kwerner]

Hi,
I would like to apologize for the long delay in response. Can you let me know if you are still experiencing this issue? Thanks!

 

[pegwan]

Yes I am! Some assistance would be greatly appreciated.

 

[kwerner]

Have you also tried changing the center lat/lon so that the domain is located in the southern hemisphere? Near the top of module_initialize_tropical_cyclone.F, these are the default settings:

  CALL nl_set_cen_lat(1,20.)
  CALL nl_set_cen_lon(1,-105.)

What happens if you change those, as well?

Just to note: I'm sure you are, but make sure you are recompiling the code each time you make a change to the module_* file. Because you aren't modifying anything with the registry, you don't need to clean or reconfigure the code each time. Just simply recompile.

 

[pegwan]

Hi,

I finally got around to testing this out. Unfortunately, changing the central latitude to –20° using the call in module_initialize_tropical_cyclone.F (leaving all vortex parameters unmodified, as if I were conducting a northern-hemisphere run) did not seem to change the actual value of XLAT that was recorded in the wrfoutput files (20.35°). The cyclone consequently continued to spin counter-clockwise.

Do you have any idea why this might be the case? Where might this oddly specific value of XLAT be coming from? What else could I modify to change it?

 

[weiwang]

Hi, Can you please give the following modification a try - all changes are in module_initialize_tropical_cyclone.F (assuming you're using a recent version of the code, the line numbers should be roughly correct) (< lines are for changed code, and > are original):

215c215
< CALL nl_set_cen_lat(1,-20.)
---
> CALL nl_set_cen_lat(1,20.)
248c248
< grid%f(i,j) = -fcor
---
> grid%f(i,j) = fcor
879c879
< v(k) = -(vref(i1,k)+( vref(i2,k)- vref(i1,k))*frac )*cos(angle)
---
> v(k) = (vref(i1,k)+( vref(i2,k)- vref(i1,k))*frac )*cos(angle)
922c922
< u(k) = (vref(i1,k)+( vref(i2,k)- vref(i1,k))*frac )*sin(angle)
---
> u(k) = -(vref(i1,k)+( vref(i2,k)- vref(i1,k))*frac )*sin(angle)