Re: [ferret_users] Getting LAT and LON to be dimensions of a Variable

[Ansley Manke <ansley.b.manke@xxxxxxxx>]

Hi,
IF the latitudes are the same at all i, and the longitudes are the same at all j,  you can define coordinate axes from one line of the lon,lat grids:

   yes? DEFINE AXIS/X/units=degrees_east lon_axis = lon[j=1]
   yes? DEFINE AXIS/Y/units=degrees_north lat_axis = lat[i=1]

Now there are several ways to think of a "monthly axis".  Is this model data? What calendar did they use? Some use a 360-day calendar with 30-day months.  Others may use a Gregorian calendar, or one with leap years, but not making other corrections.  Here's one way to do it, with the standard Proleptic Gregorian calendar and equal-length months (we could set /units= or /t0= but I've let Ferret set those using its defaults).

   yes? DEFINE AXIS/T=1-jan-1999:31-dec-2008/edges/npoints=120 time_axis
   yes? list t[gt=time_axis]

Then you can define a sst variable on these axes, and Ferret plots will be labeled with longitude, latitude and time.

   yes? LET/TITLE="Sea Surface Temperature"/UNITS="degC" sst2  = sst[gx=lon_axis@asn,gy=lat_axis@asn,gt=time_axis@asn]

Ansley

On 8/7/2014 5:20 PM, Molulaqhooa Maoyi wrote:

Thank you Ansley for the reply

Yes the Lat and Lon are the same for i and j respectively.

Here is the ncdump I did on my netcdf file:

The time is monthly data from 1999 to 2008

netcdf ROMS_SST{

dimensions:

M = 194 ;

L = 230 ;

TDIM = 120 ;

variables:

double lat(M, L) ;

lat:long_name = "latitude" ;

lat:units = "degN" ;

double lon(M, L) ;

lon:long_name = "longitude" ;

lon:units = "degE" ;

double SST(TDIM, M, L) ;

SST:long_name = "Sea Surface Temperature" ;

SST:units = "degC" ;

}

On Fri, Aug 8, 2014 at 1:51 AM, Ansley Manke <ansley.b.manke@xxxxxxxx> wrote:

Hi Molulaqhoo,
We have plans to handle data on "curvilienar grid" such as this transparently, using attributes in the dataset. This is not yet implemented. For now you need to do the 3-argument FILL command, with one exception.  Some grids  are defined this way but are in fact rectilinear.  If the latitudes are the same at all values of i in the grid, and the longitudes are the same at all values of j in the grid, then it'd be a matter of defining 1-D coordinate axes; we can help you do that.

The axis that is being marked as a K axis doesn't have attributes that tell Ferret to mark it as a time axis.  Is there a coordinate variable for the dimension of length 120 in the file?  The Unix command ncdump would help us understand what is in the file:

> ncdump -h filename.nc

You could initialize the file in Ferret with the /ORDER=   qualifier:

yes? use/order=xyt  filename.nc

and then if the data should have a formatted  time axis, you could use the DEFINE AXIS command, setting up a T axis of length 120, with the correct units and time origin, then put the data onto that axis with a command like:

yes? let new_sst = sst[gt=time_axis@asn]

-Ansley

On 8/7/2014 3:05 PM, Molulaqhooa Maoyi wrote:

Dear all

I am new into ferret and I have two questions:

I am battling to get the SST data to include LAT and LON dimension instead of them  being separate. I would like to do a fill on the SST data without having to specify LAT and LON e.g fill SST[k=1], LON,LAT

 name     title                                  I              J         K         L         M         N

 LAT      latitude                             1:230     1:194     ...       ...       ...       ...

 LON      longitude                          1:230     1:194     ...       ...       ...       ...

 SST      Sea Surface Temperature  1:230     1:194     1:120     ...       ...       .

 

My second question is how to move a dimension from K to L on the SST variable.

Regards 

Molulaqhooa

--

"Try and Fail than to Fail to Try" - Israel Maoyi

--

"Try and Fail than to Fail to Try" - Israel Maoyi