An eddy stay the initial locaiton on the beta plane in Two-layer model

[shifei90]

Dear, Edward

I run an example with an initial bump in the reduced gravity model and two-layer model on the beta plane, respectively. An eddy propagrate normally westward in the reduced gravity model. However, an eddy keep still and gradually decay in the two layer model. Why does it not propagrate westward. Is there anything wrong this example?

Some codes are as follows:

def bump(X, Y): iniH = 0. + 100np.exp(-((6e5-X)2 + (5e5-Y)2)/(21e5**2))

def coriolis(X,Y): r = Y omega = 7.292115e-5 # (1/s) (Groten, 2004) f0 = 2omeganp.sin(15np.pi/180) # 3.77e-05 beta = 2omeganp.cos(np.pi15/180)/6371e3 #2e-11 f = f0 + beta*r

def beta_plane_bump_n_layer(): print("Running n-layer beta-plane bump example") xlen = 2.0e6 ylen = 1.5e6 nx = 400 ny = 300 layers = 2 # also set in aronnax.conf file

aronnax.Grid(nx, ny, layers, dx, dy, x0=0, y0=0)

grid = aro.Grid(nx, ny, layers, xlen / nx, ylen / ny, x0=120,y0=15)

with working_directory(p.join(self_path)):
    drv.simulate(work_dir='.', config_path='aronnax.conf',
                 initHfile=[lambda X, Y: 100. + bump(X, Y),  800.],  
                 fUfile=[coriolis],
                 fVfile=[coriolis],                      
                 nx=nx, ny=ny, dx=xlen/nx, dy=ylen/ny,
                 exe='aronnax_external_solver',
                 dt=600, nTimeSteps=4640)

Thank you.

Fei

[edoddridge]

Hi Fei,

This looks ok to me. It may be that the bottom drag is very strong, so the 2-layer simulation is dominated by the viscous stress, and the eddy decays rapidly. Are you running each simulation for an equivalent amount of model time? You will likely need a shorter time step for the 2-layer simulation, so it will need more time steps.

[shifei90]

Thank you, Edward

You are right.

[edoddridge]

You're welcome. I'll close this issue since it seems like you've solved it.