Latteral boundary condition problem

[layeghi2001]

Hello everybody
I run the ROMS model for PERSIAN Gulf and Oman Sea for 10 years and its results are in good acceptance with sattelite images and other model results. I used GRADIENT LBC for east boundary. but when I extend the domain to Indian Ocean, the model blow_up. it is a sample of SST contour. please help me to find proper LBC.

[kate]

For this sort of domain, I always use the RadNud option, providing values from a larger scale model at the open boundaries. If you are modeling recent years, you can download the HYCOM solutions and extract boundary files from them.

[layeghi2001]

Thanks for your suggestion. I saw the HYCOM website. it has analysis and reanalysis data. each reanalysis data file for 3-hour interval is about 2.5 Gb. how can I get the necessary data from these files?

[kate]

It depends on your tool of choice. I use Python scripts for the download and interpolation. Others use Matlab or whatever.

[layeghi2001]

Is there any matlab script for downloading and preparing boundary data from HYCOM website in ROMS data bank?

[layeghi2001]

I changed the LBC to RadNud, but this error apeared:
READ_PHYPAR - could not find input file: ocean_nud.nc

so, can you help me to make this file?

[kate]

Have you downloaded the Matlab tools from the myroms.org site? Look through there.

[johnluick]

Here is a link to some somewhat dated Matlab functions - two are for downloading HYCOM reanalysis, one is for preparing a boundary file. Beware that I wrote them in 2013 and I would probably do the boundary file function differently if I did it now (I'm including it as you may find it a useful starting point). It is possible that HYCOM has moved the data to a new machine (url). I ended up not using HYCOM data, but not because of any particular problem with HYCOM. The download required a script because it was to be used in batch mode daily (ie as part of a cron job). You probably don't require that at this stage so could combine the two download functions into a single script or function.
The link: https://1drv.ms/u/s!AvnPG7Gh4JsAgvZQYq04-vFWY-QEPA
Please let me know if and when you download it.
Best of luck.

[layeghi2001]

thanks for your great help. I used the scripts, but it needs a file named 'latlontopo.mat'. how can I find or produce it?

[johnluick]

My apologies. The link at the end of this contains LatLonTopo.mat (as well as the matlab functions). It is simply a mat-file version of:
http://tds.hycom.org/thredds/dodsC/data ... epth_2d.nc
...which is what HYCOM uses for seafloor topography.

I cannot remember why I created a mat-file when I could have just used the netcdf. You could adapt the Matlab to go straight to the netcdf file. In fact I think I created subset versions of LatLonTopo.mat for different projects and I also can't remember why I did that. Maybe it seemed like a good idea at the time.

-- John

https://1drv.ms/u/s!AvnPG7Gh4JsAgvdeORb6Bq_P0gZ5ZA

Again, please let me know how it goes, in case others need this in future.

[yadusharma]

Hi John,

I want to try giving daily boundary values of tracers (to try Nudging) from HYCOM in my model domain for inter-annual simulations. I would be grateful if you can share the matlab code (which you mentioned here)for handling the HYCOM data.

Thanks in advance.

[johnluick]

Yadu,
I'm not sure when I'll get in to the office where the code you requested is located. Until then you might try following some more recent notes I made to myself after downloading HYCOM data (obviously your times and lats and lons will be different, and I only downloaded the surface data (level 1):
Notes:
http://tds.hycom.org/thredds/GLBa0.08/e ... -2009-uvel
http://tds.hycom.org/thredds/GLBa120.08 ... -2009-vvel
Choose Netcdf Subset
SAM: Select u, -31N -38N 123E 141E, select vertical level 1, select “Add Lat/Lon”, select a month of data (subsetter chokes if too much data); likewise for v.
Run p_combine_HYCOM.m to convert the nc files into a single pdf.
The script p_combine_HYCOM.m (nc_varget.m will order your data matrices as [t,lat,lon], which is why I still use nc_varget rather than ncread+permute):
%Combine monthly HYCOM netcdf files into a single matfile
clear
%*************************************************************
fnu(1,:)='C:\mystuff\MODELS\HYCOM\DATA\SAM\uvel\200908.nc';
fnu(2,:)='C:\mystuff\MODELS\HYCOM\DATA\SAM\uvel\200909.nc';
fnu(3,:)='C:\mystuff\MODELS\HYCOM\DATA\SAM\uvel\200910.nc';
fnu(4,:)='C:\mystuff\MODELS\HYCOM\DATA\SAM\uvel\200911.nc';
fnu(5,:)='C:\mystuff\MODELS\HYCOM\DATA\SAM\uvel\200912.nc';
fnv(1,:)='C:\mystuff\MODELS\HYCOM\DATA\SAM\vvel\200908.nc';
fnv(2,:)='C:\mystuff\MODELS\HYCOM\DATA\SAM\vvel\200909.nc';
fnv(3,:)='C:\mystuff\MODELS\HYCOM\DATA\SAM\vvel\200910.nc';
fnv(4,:)='C:\mystuff\MODELS\HYCOM\DATA\SAM\vvel\200911.nc';
fnv(5,:)='C:\mystuff\MODELS\HYCOM\DATA\SAM\vvel\200912.nc';
fnOUT='C:\mystuff\MODELS\HYCOM\DATA\SAM\HYCOM.mat';
%*************************************************************

[Nf,~]=size(fnu);
t=nc_varget(fnu(1,:),'MT');
t=HYCOM2dnum(t);
dat.Lat=nc_varget(fnu(1,:),'Latitude');
dat.Lon=nc_varget(fnv(1,:),'Longitude');
u=nc_varget(fnu(1,:),'u');
u=squeeze(u(:,1,:,:));
v=nc_varget(fnv(1,:),'v');
v=squeeze(v(:,1,:,:));

for k=2:Nf
t1=nc_varget(fnu(k,:),'MT');
t=vertcat(t,HYCOM2dnum(t1(:)));
u4=nc_varget(fnu(k,:),'u');
u3=squeeze(u4(:,1,:,:));
u=vertcat(u,u3);
v4=nc_varget(fnv(k,:),'v');
v3=squeeze(v4(:,1,:,:));
v=vertcat(v,v3);
end

dat.t=t;
dat.u=u;
dat.v=v;

save(fnOUT,'dat')