Re: [ferret_users] ascii grid format with header, read into ferret, convert to netCDF

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

Hi Matthew,
I haven't tried your commands, but what you've done look fine, except 
that you'll need a

     /SKIP=6

on the FILE command to skip over the header lines.

To assign a date to the data coming in from each of these files, you 
need the concept called "timestamping".  Essentially you define a time 
axis, and then add 0*time to the variable at that time step, to give it 
a time.  This is described here,

http://ferret.pmel.noaa.gov/Ferret/documentation/users-guide/commands-reference/SAVE#timestamp

-Ansley

On 5/3/2013 3:25 PM, Matthew Cooper wrote:
> Hi all,
>
> I am working with ascii grid format data, meaning the first 6 rows of 
> the file is header info, then the grid of data represents the model 
> output at each grid cell (example attached):
>
>  ncols                  347
>  nrows                 979
>  xllcorner       588970.54128361004
>  yllcorner       4864497.1359059000
>  cellsize        100.00000
>  NODATA_value   -9999.0000
>
> 0.79   0.79   0.79   0.81   0.81   0.82 ....1,ncols
> 0.79   0.79   0.79   0.81   0.81   0.82   ....
> 0.79   0.79   0.79   0.81   0.81   0.82   ....
> 0.79   0.79   0.79   0.81   0.81   0.82   ....
> .
> .
> .
> nrows,1
>
> So, each value (0.79 for example) represents the precipitation at that 
> 100x100 m grid cell.
>
> I have gone over the ferret documentation 2.5: Ascii data and there 
> doesn't seem to be information about reading data in this format. For 
> example I have tried:
>
> yes? define axis/x=1:347:1 longax
> yes? define axis/y=1:979:1 latax
> yes? define grid/x=longax/y=latax gridd
> yes? file/var=ppt/grid=gridd ppt_test.asc
> yes? sh data
>      currently SET data sets:
>     1> ./ppt_test.asc  (default)
>  name     title                             I J         K         L   
>       M         N
>  PPT      PPT                              1:347 1:2       ...       
> ...       ...       ...
>
>
> This isn't right.
>
> I have one of these files for each timestep of the model run. I'd like 
> to read these files into ferret, then stack all of the files into a 
> netcdf file so I can arrange model runs by year, where each netcdf 
> file contains 365 ascii files representing each day of the model run.
>
> Does ferret have the ability to read and store the header info? Would 
> it be better if I changed the write command in my code to leave out 
> the header info so it just gives a grid of output?
>
> Thank you in advance! To be clear, my primary question is how to 
> properly read in the data in the format I described (example file 
> attached). If possible, help on converting the batch of files into a 
> netcdf would also be great.
>
> Matt
> --
> Matt Cooper
> M.S. candidate, Water Resources Science
> College of Earth, Ocean and Atmospheric Sciences
> Oregon State University