Ask Your Question
0

High current magnitudes

asked 2019-06-21 00:52:38 +0100

yordan gravatar image

Hi all. For some time now, I have been trying to run the CROCO model in a domain covering the Caribbean Sea during the period from January 1, 2013 to December 17, 2013. However, outputs are so bad, exhibiting strong currents with errors above 400%. I have used several settings and different domains. In the last attempt, the CROCO model was run by using the bulk formulations over the region 53-99°W and 4-33 °N in a grid with horizontal resolution = 18 km. Additionally, the time step was 240 s.

I thank who can help me with this problem.

I am attaching the cppdefs.h, param.h, and .in files

C:\fakepath\croco.in C:\fakepath\cppdefs.h C:\fakepath\param.h

edit retag flag offensive close merge delete

4 answers

Sort by » oldest newest most voted
0

answered 2021-01-28 16:14:22 +0100

Hi Sabrina, in param.h, try setting NPP=1 (but keep your values of NP_XI and NP_ETA)

edit flag offensive delete link more
0

answered 2020-12-23 07:03:49 +0100

In cppdefs.h you have NEW_S_COORD undefined. This option goes with Vtransform=1 in pre-processing file crocotools_param.m (Vtranform=2 for #define NEW_S_COORD). Is it the case here? A mismatch of vertical levels between pre-processed input data (especially mass fields) and model setup is often the cause of unrealistic currents ... Patrick

edit flag offensive delete link more
0

answered 2020-12-22 20:54:56 +0100

Hi Yordan. I have been facing a similar problem. I believe that this problem is associated with regions of large bathymetry gradients because I can see in same areas strong currents as like as strong pressure gradients. Therefore, to solve this problem the strategies adopted were (1) to change the values of viscosity, thetas, thetab and N, as such as to (2) smooth the model's bathymetry and (3) modify the bottom friction. However, none of these tests changed significantly the current speeds. Please, could you suggest any idea to solve it? Thank you so much.

edit flag offensive delete link more
-1

answered 2021-01-28 14:45:58 +0100

Hi Patrick. When I compared the results of the sea surface level obtained by simulations with data collected in 3 different moorings (M) located at brazilian continental shelf, the values for the correlation coefficient (R2) were variables between 0.87 and 0.92. For mean square error (RMSE) the values were surrounded between 0.14 and 0.24. Although, when velocities were considered the values of R2 and RMSE were those presented below:

Along-shelf        
        R2    RMSE
     M1 0.27   0.35
     M2 0.69   0.14
     M3 0.87   0.12

Cross-shelf
        R2    RMSE
     M1 0.09   0.11
     M2 0.01   0.13
     M3 0.65   0.08

To improve my results I tried to change the values of viscosity, thetas, thetab and N , as such as to smooth the model's bathymetry and modify the bottom friction. However, none of these tests changed significantly the velocities. Please, could you suggest any idea that could help me? Thank you so much.

*Attached the cppdefs.h, param.h, and .in files.

C:\fakepath\croco_inter.in C:\fakepath\param.h C:\fakepath\cppdefs.h

edit flag offensive delete link more
Login/Signup to Answer

Question Tools

1 follower

Stats

Asked: 2019-06-21 00:52:38 +0100

Seen: 131 times

Last updated: Jan 28