From: TideMan on
On Apr 26, 4:58 pm, "Neelima " <neelima_oc...(a)rediffmail.com> wrote:
> Dear all,
>             I have a 3 hourly, monthly nc file, nww3.wind.199802.nc, converted it to  ascii in matlab. Thus the obtained ascii data is having time, latitude, longitude , u component and v component of winds. The u and v components are of (225 x 157 x 288) size, where , 225 is time, 157 is latitude and 288 is longitude.
>
> Now I want that data to be separated for 73 x 81 (latitude: -50S-30N & longitude: 30-120E), for every 3 hours. Means, I want 225 individual ascii datasets of u & v components of wind, of size 73 x 81.  Can anyone please help me , in obtaining the data?
>
> Also the time, is given as, 173676, 173679,173682,...........173748 (225 values). How to convert this number to time format like, 01-Feb-1998 03:00:00, 01-Feb-1998 06:00:00 ........ 28-Feb-1998 21:00:00?? I have used the ordinary procedures like, 'datestr' and 'julian date conversions', but it is showing some abnormal results like, 03- Nov-2577.  The starting time of the data is given as late January, 1997. ( Exact date and time are not mentioned in the website, from where I have downloaded the data).
>
> Thanku in advance.
>
> Regards,
> Neel

Well, I would just download the data I wanted into Matlab, rather than
the whole bloody lot.
How did you "convert to ascii in matlab"?
If you used netcdf, you can download a portion using this format:
nc=netcdf('nww3.wind.199802.nc','nowrite');
u=nc{varname}(:,indx,jndx)
where:
varname is the variable name, e.g., 'ugrd10'
indx is the range of longitude addresses
jndx is the range of latitude addresses.
(I'm not sure of the order for your .nc file - you can check this
using ncdump)

Alternatively, you can avoid downloading the .nc files altogether by
using GrADSDODS - Google it.