Re: new user

[Ansley Manke <ansley@xxxxxxxxxxxxx>]

Hi Steve,
This is similar to what's in the FAQ, "How can I plot a time series for a
variable which exists in multiple files?" at
http://ferret.pmel.noaa.gov/Ferret/FAQ/data_management/multi_dataset.html

First, you need to specify which time step to write out in the SAVE/APPEND
commands, i.e.
   save/file=Qscat_concat.nc/append U_T, V_T
needs to be
   save/file=Qscat_concat.nc/append U_T[L=2], V_T[L=2]

I can't tell what might be happening with lat/lon.  Specifying lat and lon in the
grid problem might do the trick (use range and delta from the lat/lon in
your data)

  define axis/t="1-jan-2001:00":"4-jan-2001:18":6/unit=hours t96hours
  define axis/X=0:359:deltax  xax            ( use delta x from your data, e.g.2 )
  define axis/Y=-89.5:89.5:deltay  yax     ( use delta y from your data)
  define grid/t=t96hours/x=xax/y=yax gg

  let U_T=t[g=gg]*0 + U[gx=gg,gy=gg]
  let V_T=t[g=gg]*0 + V[gx=gg,gy=gg]



Steve Cocks wrote:

> Hello,
>
> I am a new user of Ferret, and in a middle of a number of problems.    I
> ran a script to concantenate 16 files into one; each file is a
> particular time step for a given year, julian date and hour.  Each file
> carries within it the variables time (unit="yyyydddhh", range is 1),
> lat(units="degree-north", range is 1-171), lon(units="degrees-east",
> range is 1-360), u(units=""(m/s)^2) and v(units=""(m/s)^2").  The goal I
> had was to run a script to concatenate these files into 1 file with all
> of the above variables but with a time range of 16, or l=1:16.  The
> script I used is included below:
> ***********************************************************
> define axis/t="1-jan-2001:00":"4-jan-2001:18":6/unit=hours t96hours
> define grid/t=t96hours gg
>
> let U_T=t[g=gg]*0 + U
> let V_T=t[g=gg]*0 + V
>
> set variable/title="U"/units="m^2/s^2" U_T
> set variable/title="V"/units="m^2/s^2" V_T
>
> use "/rlp/scocks/Scatterometer/2001/00/gcv_QSCAT_200100100.nc"
> save/file=Qscat_concat.nc/append U_T, V_T
> .
> .    ETC ETC ETC, boy the list get's long
> .
> use "/rlp/scocks/Scatterometer/2001/18/gcv_QSCAT_200100418.nc"
> save/file=Qscat_concat.nc/append U_T, V_T
> quit
> ************************************************************
> This script did indeed write a concat file with l=1:16 although when
> each source file was read, the note
>
> "*** NOTE: Units on axis "time" are not recognized: yyyydddhh
>  *** NOTE: They will not be convertible:
> save/file=Qscat_concat.nc/append U_T, V_T
>  LISTing to file Qscat_concat.nc
>  **TMAP ERR: Requested data range is outside of data set limits
>              Expected: U_T[I= 1 : 360 ] Found: U_T[I= 111 : 271 ]"
>
> appeared indicating an error.  When I tried to look at my data, I found
> that my Lon and Lat variables had been restricted to a much smaller lat
> and lon range.  Also, my other Go scripts all of sudden didn't recognize
> the Lon (and probably Lat as well) variable.  The comparisons of ncdump
> -c on the source file and the new concatentated file showed big
> differences in variable ranges (excluding the change in time as was
> expected).  My questions are:
>
> 1)  What happened, and how can I concatenate the files without losing
> the lat and lon data, yet have the variables with l=1:16?
> 2)  Is there a way to define string variables, say YYYY, DDD, HH, and
> then substitute into a loop so I won't have to write down every single
> file into the program?
>
> Would greatly appreciate some insight into this matter.  Thanks in
> advance.
>
> Scocks
> Texas A&M University