using my own oceanographic netcdf files

408 views
Skip to first unread message

Lysel Garavelli

unread,
Nov 18, 2014, 3:16:02 PM11/18/14
to connectivity-mod...@googlegroups.com
Hi,

I just installed the CMS on my computer and I had some problems to run it.
I wanted to run a simulation using my own oceanographic data (which are ROMS files). Reading the CMS user's guide, I modified the nest_1.nml file according to what it was described and I put my ROMS files in /expt_example/nests. When I want to run the CMS using ./cms example, the error message is: "NetCDF: Variable not found".
I put three different output files with the same name and u, v, or w at the end of the name file. But when I am doing this, it can find the file... 
Could you describe me exactly how you are doing a simulation using your own oceanographic data? Do you have some nest_x.nml file example?

Thanks a lot,

Regards,

Lysel Garavelli

Robin Faillettaz

unread,
Nov 19, 2014, 4:32:31 AM11/19/14
to Lysel Garavelli, connectivity-mod...@googlegroups.com
Hi Lysel,


> Le 18 nov. 2014 à 21:16, Lysel Garavelli <lyselga...@gmail.com> a écrit :
>
> I wanted to run a simulation using my own oceanographic data (which are ROMS files). Reading the CMS user's guide, I modified the nest_1.nml file according to what it was described and I put my ROMS files in /expt_example/nests. When I want to run the CMS using ./cms example

This is the right way to do it. If you set the right of longitude, latitude, depth and time variables, and u, v and w names, the problem may not come from the nest file.

> the error message is: "NetCDF: Variable not found ».
> I put three different output files with the same name and u, v, or w at the end of the name file. But when I am doing this, it can find the file...
> Could you describe me exactly how you are doing a simulation using your own oceanographic data? Do you have some nest_x.nml file example?

If you are using the last version of the CMS (the one online right now) you need to gather u, v and w in a single nc file. This is probably causing the error.

I have done it with MARS3D outputs and I attached my nest_1.nml file if it can help.

Tell me if that works, or not !
Cheers,

nest_1.nml

Bex Ross

unread,
Nov 19, 2014, 4:33:11 AM11/19/14
to connectivity-mod...@googlegroups.com
Hi Lysel,

I am just a fellow user and I haven't tried using ROMS files with CMS (or any model with separate u/v/w files), but from the sounds of it, if you are labelling your files correctly with the e.g. nest_nestnumber_yyyymmddhhmmssu.nc format then the error is probably not that it can't find the file, rather it is that you have named the variable that is inside the file incorrectly in the nest_x.nml file.

Try an "ncdump -c nest_nestnumber_yyyymmddhhmmssu.nc" (or in matlab an ncdisp) to see what your ROMS files name the variables stored in them and make sure that the names in there match what you write under the fields lon_name, lat_name, depth_name, time_name, uvel_name, vvel_name, wvel_name, etc in the nest_x.nml file. As it says on p28 of the CMS v1.1 users guide you can add fields if any of the names change between your u/v/w files too.

E.g. I ran CMS with POLCOMS files which looked like this inside...(from a matlap ncdisp):

Variables:

    lon

           Size:       198x1

           Dimensions: lon

           Datatype:   single

           Attributes:

                       longname = 'Longitude'

                       units    = 'degrees_east'

                       bounds   = 'lonbnd'

    lat

           Size:       224x1

           Dimensions: lat

           Datatype:   single

           Attributes:

                       longname = 'Latitude'

                       units    = 'degrees_north'

                       bounds   = 'latbnd'

    time

           Size:       31x1

           Dimensions: time

           Datatype:   int32

           Attributes:

                       longname = 'time'

                       units    = 'seconds since 2004-01-01'

                       calendar = 'standard'

    xUx

           Size:       198x224x40x31

           Dimensions: lon,lat,z,time

           Datatype:   double

           Attributes:

                       units    = 'm/s'

                       longname = 'Eastward Velocity'

    xVx

           Size:       198x224x40x31

           Dimensions: lon,lat,z,time

           Datatype:   double

           Attributes:

                       units    = 'm/s'

                       longname = 'Northward Velocity'


So...

lon_name = "lon"

lat_name = "lat"

time_name = "time"

uvel_name = "xUx"

vvel_name = "xVx"



If you keep having a problem, post again with as much info as possible including a copy of your nest_x.nml file and an ncdump -h of an example u,v and w file (same with salinitiy and temp if you have them too) and we'll do our best to troubleshoot with you.

All the best,
Bex

Bex Ross

unread,
Nov 19, 2014, 4:38:42 AM11/19/14
to connectivity-mod...@googlegroups.com
Apologies:

lon_name = "Longitude"
lat_name = "Latitude"

the rest are right ;)

Robin, you should be able to run the u,v, and w files separately in CMS v1.1: there is info on this in the user guide on p18 and p28. I'm guessing you couldn't run it though getdata in this format though which might be the difference?

All the best,
Bex

On Tuesday, 18 November 2014 20:16:02 UTC, Lysel Garavelli wrote:

Robin Faillettaz

unread,
Nov 19, 2014, 5:04:57 AM11/19/14
to Bex Ross, connectivity-mod...@googlegroups.com
Le 19 nov. 2014 à 10:38, Bex Ross <rebecc...@plymouth.ac.uk> a écrit :
>
> Robin, you should be able to run the u,v, and w files separately in CMS v1.1: there is info on this in the user guide on p18 and p28. I'm guessing you couldn't run it though getdata in this format though which might be the difference?

Maybe, I don’t know then because I work with CMS v1.0 (which the only online version??).

Bex was much clearer than me… sorry about it! I try again.
From what I have done working with the CMS and local nc files, most of the errors come from the structure of these files that the CMS doesn’t understand. It’s very important to give each variable the correct dimensions and in the order it expects. In my case, I work without getdata and places my ncfiles in the nests/ directory. I also have to merge u, v and w files into a single one to get it to work. This is what my nc files look like (I don’t have w velocities):

dimensions:
Longitude = 1101 ;
Latitude = 463 ;
Depth = 60 ;
Time = 1 ;
variables:
float Longitude(Longitude) ;
Longitude:units = "degrees" ;
float Latitude(Latitude) ;
Latitude:units = "degree" ;
float Depth(Depth) ;
Depth:units = "degree" ;
double Time(Time) ;
Time:long_name = "Time" ;
Time:calendar = "standard" ;
float zu(Time, Depth, Latitude, Longitude) ;
zu:units = "cms.s-1" ;
zu:_FillValue = 0.f ;
float zv(Time, Depth, Latitude, Longitude) ;
zv:units = "cms.s-1" ;
zv:_FillValue = 0.f ;

which correspond to, in the nest files (attach below):
xaxis= "Longitude"
yaxis= "Latitude"
zaxis= "Depth"
taxis= « Time »

lon_name= "Longitude"
lat_name= "Latitude"
depth_name= "Depth"
time_name= "Time"
uvel_name= "zu"
vvel_name= "zv"

As you can see, xaxis, yaxis, zaxis and taxis AND lon_name, lat_name, depth_name and time_name are different but should all be with a single dimension in the netCDF files. But uvel and vvel (and wvel) should have 4, Time, Depth, Latitude and Longitude, ordered like this. If you get new errors you may need to reorganize them. I do it with R, just let me know if you need it.

Good luck,

————————————————

Robin Faillettaz, Ph.D candidate
---
Université Pierre et Marie Curie - Paris 6
Laboratoire d'Océanographie de Villefranche - UMR7093
Station Zoologique
181 chemin du Lazaret
06230 Villefranche sur Mer

Tel: +33 (0)4-93-76-38-10
Port: +33 (0)6-84-91-43-33
faill...@obs-vlfr.fr

Journée des Jeunes Chercheurs de l’Observatoire de Villefranche
http://biodev.obs-vlfr.fr/j2c



nest_1.nml

Lysel Garavelli

unread,
Nov 19, 2014, 11:46:16 AM11/19/14
to connectivity-mod...@googlegroups.com
Hi,

thanks to both of you for your answers. 
I think I had not understood the u, v, w nc files. For each month of my oceanographic data, I just have one nc file as Robin. All my variables are in one file.
So I changed my variables names according to the description of my nc file and it is still not working: "NetCDF: Variable not found".
I attached here my nest_1.nml and a ncdump of my nc file... I am not quite sure on which longitude or latitude names using, lon_rho or xi_rho for example... I tried all the possible names and it did not work either.
As pointed by Robin, my U-velocity and V-velocity values have 4 dimensions but not in the right order (longitude, latitude, depth, time), so that could cause my problem too... However, in the User's guide, it is mentioned that longitude and latitude can also have 2 dimensions ordered as (latitude,longitude). So I have two dimensions but again, not in the right order.
Let me know if you see any error in my nest_1.nml file... 

Thanks a lot for your help!

Cheers,

Lysel
nest_1.nml
ncdump roms file.rtf

Robin Faillettaz

unread,
Nov 19, 2014, 12:28:46 PM11/19/14
to Lysel Garavelli, connectivity-mod...@googlegroups.com
Hi,

Indeed, you were using the « long names » of the variables, which are used to describe them. Your nest file should be more like below.

&nest_input
xaxis= "xi_rho"
yaxis= "eta_rho"
zaxis= "s_rho"
taxis= "time"
xstart=-66
xend=-63.8
ystart=17.3
yend=19.2
zstart=0
zend=500.00
zaxis_positive_direction= "down"
tstart_yy=2007
tstart_mm=01
tstart_dd=01
tend_yy=2007
tend_mm=1
tend_dd=30
time_step=3600
time_units= "seconds"
jdate_yy=2007
jdate_mm=1
jdate_dd=1
lon_name= "lon_rho"
lat_name= "lat_rho"
depth_name= "zeta"
time_name= "scrum_time"
uvel_name= "u"
vvel_name= "v"
wvel_name= "w"
wvel_positive_direction = ""
velocity_conversion_factor=1
temp_name= "temp"
saln_name= "salt"
fill_value=0
$end

You can try, but I’m pretty sure it will return errors, as at least your longitude and latitude are in two dimensions while they should have a single one, and you have 31 time steps in a single file.
So what you need to do is first :
- split each monthly file into a daily file.I guess you can do it with nco functions but I have never done it
If it still doesn’t work:
- Extract the a single line of your longitude variable and save it into a new variable
- Extract a single column of your latitude variable and save it into a new variable
- Extract the time value if it’s not already done after splitting per day
- Use these newly created variables as lon, lat and time variable

And then it will work !

Additionally (but that’s not causing the error), check your time step which is one hour while I understood that you have daily data. So it would be more 86400s.

All the best,


> Le 19 nov. 2014 à 17:46, Lysel Garavelli <lyselga...@gmail.com> a écrit :
>
> Hi,
>
> thanks to both of you for your answers.
> I think I had not understood the u, v, w nc files. For each month of my oceanographic data, I just have one nc file as Robin. All my variables are in one file.
> So I changed my variables names according to the description of my nc file and it is still not working: "NetCDF: Variable not found".
> I attached here my nest_1.nml and a ncdump of my nc file... I am not quite sure on which longitude or latitude names using, lon_rho or xi_rho for example... I tried all the possible names and it did not work either.
> As pointed by Robin, my U-velocity and V-velocity values have 4 dimensions but not in the right order (longitude, latitude, depth, time), so that could cause my problem too... However, in the User's guide, it is mentioned that longitude and latitude can also have 2 dimensions ordered as (latitude,longitude). So I have two dimensions but again, not in the right order.
> Let me know if you see any error in my nest_1.nml file...
>
> Thanks a lot for your help!
>
> Cheers,
>
> Lysel

Rebecca Ross

unread,
Nov 19, 2014, 12:35:42 PM11/19/14
to Lysel Garavelli, connectivity-mod...@googlegroups.com

Hi Lysel,

 

I’m afraid I have no quick fix here, your ROMS file looks a lot more complex than the netcdf files I am used to working with: I don’t blame you for getting tied in a knot! Definitely don’t bother adding the u/v/w to the nest filename then if they are all in one file, though you will need to make multiple symbolic links to each file as you have more than one snapshot in there (31 days I assume, therefore you will need 31 links each named according to the day they are pointing to in the file as mentioned in the user guide).On the subject of dimensions, the order of these shouldn’t be an issue as they are defined in the nest_x.nml file. Xi_rho/eta_rho are definitely the dimension names (like x/y in the e.g nest file) and lon_rho/lat_rho will be the long and lat names you need I recon. S_rho is never described in there but I assume it is the z dimension? It doesn’t look like you have depth in there though?

 

BIG follow-up question, now I think of it: Is this a sigma grid model (terrain following)? I suspect it is with a bit of googling and all the mention of S-coordinates. If it is you will have problems anyway as CMS assumes it is working with z-level data - if it assumes an S level is a depth level it will end up allowing particles (larvae?) to flow along the surface of that grid layer which, in actuality, can have massively varying depth due to the terrain underneath it so I would be wary if that is the case (I had that problem with POLCOMS model outputs and had to have someone help me to write a matlab script to interpolate each file to z-level).

 

I have probably created more problems there than I have solved. Sorry! Keep emailing if you are still attempting to work with these files, the more people paying attention to your questions the more likely you’ll find someone who can help more definitively!

 

All the best,

Bex

 

P.S. I see Robin has already messaged while I have been writing this… and has many of the same points - good spot on the timestep Robin! Like he says you should be able to get this working though, but do think on the sigma issue, it maybe that it is moot to keep working with these files as they currently stand if they will not have the correct effect on your simulations! Also I don’t think zeta is depth having only the dimensions lat, long and time, I think it is just the surface of the ocean i.e. sea surface height?

--
You received this message because you are subscribed to the Google Groups "connectivity-modeling-system-club" group.
To unsubscribe from this group and stop receiving emails from it, send an email to connectivity-modeling-...@googlegroups.com.
To post to this group, send email to connectivity-mod...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.




This email and any files with it are confidential and intended solely for the use of the recipient to whom it is addressed. If you are not the intended recipient then copying, distribution or other use of the information contained is strictly prohibited and you should not rely on it. If you have received this email in error please let the sender know immediately and delete it from your system(s). Internet emails are not necessarily secure. While we take every care, Plymouth University accepts no responsibility for viruses and it is your responsibility to scan emails and their attachments. Plymouth University does not accept responsibility for any changes made after it was sent. Nothing in this email or its attachments constitutes an order for goods or services unless accompanied by an official order form.

Nandini Ramesh

unread,
Feb 18, 2015, 7:20:53 PM2/18/15
to connectivity-mod...@googlegroups.com, lyselga...@gmail.com
I am also trying to use the CMS with my own velocity files (i.e., without the use of "getdata"), and receive the same error message. Was this issue resolved at all?

I have attached my nest_1.nml and ncdump -h of my file. 
I am trying to use monthly velocity data and have each snapshot in a separate file (with both u and v in the same file).
Any help would be appreciated. Thank you very much!

Cheers, 
Nandini

To unsubscribe from this group and stop receiving emails from it, send an email to connectivity-modeling-system-club+unsubscribe@googlegroups.com.
To post to this group, send email to connectivity-modeling-system-cl...@googlegroups.com.


For more options, visit https://groups.google.com/d/optout.

nest_nml.txt
ncdump_myfile.txt

Robin Faillettaz

unread,
Feb 19, 2015, 3:14:01 AM2/19/15
to Nandini Ramesh, connectivity-mod...@googlegroups.com, lyselga...@gmail.com
Hi !

It seems that you may not be using the correct variables names. Your dimensions and variables have the same names (LAT126_135, LON286_300, etc) and thus should be the one to use in the nest file, like this:

input
xaxis= "LON286_300"
yaxis= "LAT126_135"
zaxis= "LEV1_1"
taxis= "TIME"
xstart=285
xend=300
ystart=35
yend=45
zstart=0
zend=10
zaxis_positive_direction= "down"
tstart_yy=2009
tstart_mm=01
tstart_dd=15
tend_yy=2009
tend_mm=06
tend_dd=15
time_step=2635200
time_units= "days"
jdate_yy=0001
jdate_mm=01
jdate_dd=01
lon_name= "LON286_300"
lat_name= "LAT126_135"
depth_name= "LEV1_1"
depth_conversion_factor=1
time_name= "TIME"
uvel_name= "UO"
vvel_name= "UO"
velocity_conversion_factor=1
fill_value=9.96921E36
$end


Try this and tell us if it works !
Good luck,



> Le 19 févr. 2015 à 01:20, Nandini Ramesh <the.n...@gmail.com> a écrit :
>
> I am also trying to use the CMS with my own velocity files (i.e., without the use of "getdata"), and receive the same error message. Was this issue resolved at all?
>
> I have attached my nest_1.nml and ncdump -h of my file.
> I am trying to use monthly velocity data and have each snapshot in a separate file (with both u and v in the same file).
> Any help would be appreciated. Thank you very much!



————————————————

Robin Faillettaz, Ph.D candidate in Marine Ecology
---
Université Pierre et Marie Curie - Paris 6
Laboratoire d'Océanographie de Villefranche - UMR7093
Station Zoologique
181 chemin du Lazaret
06230 Villefranche sur Mer

Tel: +33 (0)4-93-76-38-04

Nandini Ramesh

unread,
Feb 19, 2015, 9:39:35 AM2/19/15
to connectivity-mod...@googlegroups.com, the.n...@gmail.com, lyselga...@gmail.com
Thanks Robin, 

I tried this out, but still get the same error message. 

Do you know whether there's some way I can look into where the error is coming from, since the error message is not very informative?

Cheers,
Nandini

bexeross

unread,
Feb 19, 2015, 11:37:01 AM2/19/15
to connectivity-mod...@googlegroups.com, the.n...@gmail.com, lyselga...@gmail.com
Hi Nandini,

The error message (if it is the same as Lysel) is "NetCDF:variable not found" which is actually pretty reassuring if you ask me - it means it can read your file but one of the variables you have typed in the nest file does not match up with what's inside, so whatever you are doing stick with looking at variable names to try and fix this. 

I don't have my files right here to cross reference with, so don't give this much credence over Robin's advice who probably did have his to look at, but in case you were going down the right vein the first time with 

lon_name= "Longitude"
lat_name= "Latitude"
depth_name= "Depth"

I note that your ncdump names do not start with capital letters...

LON286_300:standard_name = "longitude" ;
LAT126_135:standard_name = "latitude" ;
LEV1_1:standard_name = "depth" ;

So that might be something to try changing too.

All the best,
Bex

Nandini Ramesh

unread,
Feb 19, 2015, 4:22:16 PM2/19/15
to bexeross, lyselga...@gmail.com, connectivity-mod...@googlegroups.com

Thank you, Bex.

Unfortunately, that didn't work either - I still get the same error message. I'm sure that the problem is not with the data itself as I have been able to open the file and plot the velocities.

Do either of you have examples of input that worked, that I could compare with? I was also unable to run the example experiment that came with the code using getdata. Initially I had assumed that this was because the data was no longer served in the same location (as the error displayed said that it was unable to access the file), but perhaps something else is wrong.

Thanks,
Nandini

Ana Vaz

unread,
Feb 20, 2015, 3:17:05 PM2/20/15
to connectivity-mod...@googlegroups.com, bexe...@gmail.com, lyselga...@gmail.com
Hello all,

I have some comments/questions:

1) You need to include agrid=.false. on the nest.nml file for using local files not on agrids (look at the nest_1.nml on the example folder from the downloaded code). Did you both use it? I haven't seen it on your files

2) your files need to be named as explained on the userguide (nest_nestnumber_yearmonthdayhourminsecvariable.nc, so you will have one nest_....u.nc, nest_...v.nc, etc), is that the case?

3) as Bex pointed out, for now the depths must be in zlevels (0,10,20, etc), so you will need to pos-process your files in case they are not to interpolate the s-coordinate into z. Alternatively, if you are interested in 2D transport on one of the layers, you can extract that layer, create nest files with only that layer, and use it as a fixed depth (say 10 m). I have done it before to follow transport in the mixed layer using HYCOM hybrid ones. 

Bellow I copied the userguide description of using local files on b or c grids:

I hope this helps, any questions write again! It is great to have a community of CMS users to brainstorm.

As of CMS v1.1, users can also directly provide their hydrographic data in the nests folder (so not in raw, and not using getdata). Advantage of this method is that the hydrographic files in any grid (A, B or C) can be used. If you are using native B or C grids, set the agrid *variable in nest_x.nml to *.false., and if using an A grid, set it to .true.

Since on C grids, the zonal and meridional components of the velocity are given at different locations, if using this method you will need to provide at least two nest files for each time step, one for u and one for v, named nest1yyyymmddhhmmssu.nc and nest1yyyymmddhhmmssv.nc

The date format yyyymmddhhmmss is the same date string as also used by getdata. 

You can also add vertical (w), temperature (t), and salinity (s) files as needed. So for each time step for each oceanographic files are available, you will need at least two files (and up to 5 different files). Each file will store only one field (u, v, w, t or s).

For each of your variables, you can define different variable names for longitude, latitude and depth. The names of the variables dimensions can be set in the nest_1.nml file as lon_nameUlat_nameUdep_nameUlon_nameV, lat_nameVdep_nameVlon_nameWlon_nameT, etc. 

The names of the variables themselves can also be set in nest_1.nml as uvelname, vvelname etc. If not using agrid, CMS will use all these values in nest_1.nml to determine how to read the fields.


CMS uses the tstartyy, tstartmm, and tstart_dd in nest_1.nml to determine when to start the run, so it is important to also provide these (see earlier in the section).

In agrid-is-false-mode, CMS can also handle netcdf files with multiple snapshots in a file. Easiest way to get CMS to work with files that have for instance five snapshots, is to create five symbolic links to the same file, following the normal CMS calendar convention. Do this for every file. CMS will then automatically iterate through the individual snapshots in a file. It is important to note, however, that the tsart values (see above) point to the first snapshot of a file, and the tend values point to the last snapshot.

Currently, work is underway to generalize the grids that CMS can handle even further. As of yet, the grid needs to be “orthogonal”, meaning that the values for longitude have to be independent of latitude and vice versa. A common Mercator grid will work in CMS but more complicated tri-polar grids do not yet work. Use getdata if you want to work with these.

Nandini Ramesh

unread,
Feb 22, 2015, 6:51:03 PM2/22/15
to Ana Vaz, connectivity-mod...@googlegroups.com, bexeross, lyselga...@gmail.com
Hi all, 

Thanks for the suggestions. I am still not able to get my test run to work. 

I tried all of the following:

1. Changing the files between having a single file for u & v (e.g., a file named nest_1_yyyymmddhhmmss.nc containing a snapshot of both u and v) and having separate files for u and v (using files named as before, except with "u" or "v" at the end of the name). I still got the "NetCDF: Variable not found" error message.

2. I am using a single z-layer (centered on ~5 m) and am only interested in the 2D transport within this layer. I tried setting zstart and zend both equal to this value. I also tried setting zstart = 0 and zend = 10, as these are the boundaries of this particular depth layer in the original netCDF file. I still got the same error.

3. I tried changing the dimension names as pointed out by Bex. This also resulted in the same error.

4. I added a line with "agrid = .true." to the end of the file. The example nest_1.nml that came with the download did not have this line, but that shown in the user guide did. When I did this, it displayed a new error message:
PGFIO-F-239/namelist read/unit=2/entity name is not member of group
File name = input_mytest/nest_1.nml formatted, sequential access, record = 33
In source file getnestinfo.f90, at line number 84
Record 33 in the nest_1.nml file corresponds to the line with agrid = .true. so presumably the code does not expect that line.

Do any of you have examples of files that worked? I was not able to get the example file from the downloaded code to work (using getdata) either, so there's possibly something wrong with the installation. If I could take a look at the nest_1.nml file and the thredds address for an example that worked, that would be extremely helpful for me to compare. 

Thank you very much, 
Nandini


--
You received this message because you are subscribed to a topic in the Google Groups "connectivity-modeling-system-club" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/connectivity-modeling-system-club/XHCjqLvIMeU/unsubscribe.
To unsubscribe from this group and all its topics, send an email to connectivity-modeling-...@googlegroups.com.
To post to this group, send email to connectivity-mod...@googlegroups.com.

Rebecca Ross

unread,
Feb 23, 2015, 5:18:56 AM2/23/15
to Nandini Ramesh, Ana Vaz, connectivity-mod...@googlegroups.com, bexeross, lyselga...@gmail.com

OOPS, meant to send this to the group…

 

Hey Nadini,

 

Firstly you need to be clear on which version of the cms you are using. If you downloaded the zip file you have version 1.0 and will need this userguide. The latest  version of CMS can be checked out by SVN with instructions on the source tab, amounting to version 1.1 so you would need the latest user guide. There is a difference in the nest file namelist between the two versions so make sure you are following the right instruction manual!

 

There should be no problem using a single file containing U and V (I am doing that), or with separate files with u and v labels, as long as you follow the instructions for that particular set up.

 

The agrid is .false. notation personally I didn’t use because I wanted my files to be interpolated to A-grid, as I am working in the complicated tripolar region of HYCOM, but if you don’t want the additional interpolation then you do need this as Ana mentions but only if you are using the LATEST version of cms (SVN), otherwise you cannot include a line about the agrid and you cannot avoid the interpolation. My working Hycom raw ncdump and nest file are attached above. If you want to try with some Hycom files but can’t get the opendap download working for some reason you can still easily download a couple of files from the Hycom website. I tend to use a bash script to batch download/use the opendap service in cms, but there are many other ways to download of which the easiest for you to just try some may be the NCSS (e.g. here). Incidentally you are right that the urls have changed from when they were written in the user manual – they migrated a bunch of servers and also now have a secondary reanalysis dataset. If you were using the latest manual where it says filename= "http://tds.hycom.org/thredds/dodsC/glb_analysis" it looks like you now need to look up which experiment you need to download from ( you can just follow the links to the data you want from the Hycom data server above and when you get to a page listing the access options choose the one that says OPENDAP to replace the nest file url with e.g. if you want data from 2014 from the global 1/12 degree analysis you will end up here and will use this url http://tds.hycom.org/thredds/dodsC/GLBa0.08/expt_91.1/2014. Hopefully that will allow you to get the example working. The HYCOM google forum is very useful for help if you think the urls have changed again or there is an issue there.

 

And just double checking – the nestfile you sent us at the start of this conversation began with the word “nput” rather than “input”, just wanted to make sure you had corrected that – I suspect you would get a different error if that was the problem but just in case! It is all about being absolutely precise with the nestfile wording.

 

That’s all I have for you at the moment. If you are still having trouble can you try sending us your monthly snapshot file (might need to be via dropbox or similar – these files aren’t small!) and we’ll give it a go too.

--

You received this message because you are subscribed to the Google Groups "connectivity-modeling-system-club" group.

To unsubscribe from this group and stop receiving emails from it, send an email to connectivity-modeling-...@googlegroups.com.


To post to this group, send email to connectivity-mod...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

ncdump_rawfileHYCOM.txt
nest_1HYCOM.txt

Nandini Ramesh

unread,
Mar 2, 2015, 2:04:56 PM3/2/15
to Rebecca Ross, Ana Vaz, connectivity-mod...@googlegroups.com, bexeross, lyselga...@gmail.com
Hi Rebecca, 

Thank you very much, it turns out I was using the v1.1 manual with the v1.0 zip file, so I tried to re-install the code.

Now, when I try to install the CMS, I receive the following error: /usr/bin/ld: cannot find -limf

I am using a gfortran compiler and made the changes according to the manual in the cms.f90 file to remove the mpi-related lines.  I also changed the paths in the arch/mpi_compiler file to match those of the computer I am using.

I've attached the mpi_compiler file. Has anyone had similar problems?

I'm not sure what the -limf flag does (I'm still quite new to fortran).

Thank you, 
Nandini


mpi_compiler

Fábio Matos

unread,
May 26, 2015, 1:15:59 PM5/26/15
to connectivity-mod...@googlegroups.com, lyselga...@gmail.com, rebecc...@plymouth.ac.uk, bexe...@gmail.com, anaca...@gmail.com
Hi,

I'm trying to install the CMS v1.1b for the first time and I don't have any background in Fortran. My machine is running the Bio-Linux 8.0.5 and after installed the compilers requested and the NetCDF I get the same error that Nandini:

/usr/bin/ld: cannot find -limf
collect2: error: ld returned 1 exit status
make: *** [cms] Error 1

How can I fix it?

Thank you.

Regards,
Fábio

To unsubscribe from this group and all its topics, send an email to connectivity-modeling-system-club+unsubscribe@googlegroups.com.
To post to this group, send email to connectivity-modeling-system-cl...@googlegroups.com.


For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "connectivity-modeling-system-club" group.

To unsubscribe from this group and stop receiving emails from it, send an email to connectivity-modeling-system-club+unsubscribe@googlegroups.com.
To post to this group, send email to connectivity-modeling-system-cl...@googlegroups.com.


For more options, visit https://groups.google.com/d/optout.

Nandini Ramesh

unread,
May 26, 2015, 1:40:34 PM5/26/15
to Fábio Matos, connectivity-mod...@googlegroups.com, Lysel Garavelli, mberc plymouth, Rebecca Ross, Ana Vaz
Hi Fabio,

Sorry, I just realized that I hadn't sent my email with the fix to the full group.

I was able to compile the code after removing the -limf flag and adding -lnetcdff. It turns out that gfortran needed this extra flag. Also, I had to add /usr/local/lib to my path, as it was not on the default ubuntu path.

I wasn't using the same OS as you, so the solution may not be exactly the same - changing -limf to -lnetcdff and making sure that all the relevant directories are in your $PATH variable, similar to what is discussed in this thread: http://askubuntu.com/questions/402353/how-to-add-home-username-bin-to-path

Cheers,
Nandini

To unsubscribe from this group and all its topics, send an email to connectivity-modeling-...@googlegroups.com.
To post to this group, send email to connectivity-mod...@googlegroups.com.


For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "connectivity-modeling-system-club" group.

To unsubscribe from this group and stop receiving emails from it, send an email to connectivity-modeling-...@googlegroups.com.
To post to this group, send email to connectivity-mod...@googlegroups.com.


For more options, visit https://groups.google.com/d/optout.




This email and any files with it are confidential and intended solely for the use of the recipient to whom it is addressed. If you are not the intended recipient then copying, distribution or other use of the information contained is strictly prohibited and you should not rely on it. If you have received this email in error please let the sender know immediately and delete it from your system(s). Internet emails are not necessarily secure. While we take every care, Plymouth University accepts no responsibility for viruses and it is your responsibility to scan emails and their attachments. Plymouth University does not accept responsibility for any changes made after it was sent. Nothing in this email or its attachments constitutes an order for goods or services unless accompanied by an official order form.

--
You received this message because you are subscribed to a topic in the Google Groups "connectivity-modeling-system-club" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/connectivity-modeling-system-club/XHCjqLvIMeU/unsubscribe.
To unsubscribe from this group and all its topics, send an email to connectivity-modeling-...@googlegroups.com.
To post to this group, send email to connectivity-mod...@googlegroups.com.

Bharti D.K

unread,
Mar 27, 2018, 10:44:49 AM3/27/18
to connectivity-modeling-system-club
Hi all,

I am coming across the same error that a few people have reported here. I am trying to run CMS by placing manually downloaded HYCOM netcdf files in the nests folder. In my case, I have followed the steps below to create my input files -

1. downloaded daily HYCOM analysis output (GLBu0.08 - expt_91.2) for the Indian Ocean from 1 May 2016 to 31 May 2016
2. changed fill value to the CMS recognized one - 1.2676506e30
3. split u and v variables
4. split the above u and v files further to have day-wise netCDF files - renamed these to make them CMS compatible

I get the following error (lines 3 and 4) on running cms using the input files created above -

Not using vertical velocity data
Not using temperature data
Error in the netCDF file: expt_5/nests/nest_1_20160501000000u.nc
NetCDF: Variable not found

I checked and the variable names match between nest_1.nml and the .nc files. I've read this thread and section 4.4 in the user guide a couple of times and I'm still not able to understand what might be the problem.

I am attaching my nest_1.nml file and ncdump of nest_1_20160501000000u.nc here. I'll be very grateful if someone can help.

Thanks,
Bharti

To unsubscribe from this group and all its topics, send an email to connectivity-modeling-system-club+unsubscribe@googlegroups.com.
To post to this group, send email to connectivity-modeling-system-cl...@googlegroups.com.


For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "connectivity-modeling-system-club" group.

To unsubscribe from this group and stop receiving emails from it, send an email to connectivity-modeling-system-club+unsubscribe@googlegroups.com.
To post to this group, send email to connectivity-modeling-system-cl...@googlegroups.com.


For more options, visit https://groups.google.com/d/optout.




This email and any files with it are confidential and intended solely for the use of the recipient to whom it is addressed. If you are not the intended recipient then copying, distribution or other use of the information contained is strictly prohibited and you should not rely on it. If you have received this email in error please let the sender know immediately and delete it from your system(s). Internet emails are not necessarily secure. While we take every care, Plymouth University accepts no responsibility for viruses and it is your responsibility to scan emails and their attachments. Plymouth University does not accept responsibility for any changes made after it was sent. Nothing in this email or its attachments constitutes an order for goods or services unless accompanied by an official order form.

--
You received this message because you are subscribed to a topic in the Google Groups "connectivity-modeling-system-club" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/connectivity-modeling-system-club/XHCjqLvIMeU/unsubscribe.
To unsubscribe from this group and all its topics, send an email to connectivity-modeling-system-club+unsubscribe@googlegroups.com.
To post to this group, send email to connectivity-modeling-system-cl...@googlegroups.com.
nc_dump
nest_1.nml

salwood82

unread,
Aug 30, 2018, 8:51:36 AM8/30/18
to connectivity-modeling-system-club
Did anyone who posted problems with the error "NetCDF: Variable not found" on here find and solve their issue?
Cheers,
Sally


On Tuesday, 18 November 2014 20:16:02 UTC, Lysel Garavelli wrote:
Reply all
Reply to author
Forward
0 new messages