[ncl-talk] dimension changed in ncl definition different ctl file's description

[David Brown]

Have you read the applications page on handling GrADS data
(http://www.ncl.ucar.edu/Applications/grads.shtml)? The Fortran
records in the file contain only a single lat/lon slice of the data.
You will need to read multiple records in a loop to get all the U
data.

If you look at the sample script on the web page, you will see that
you need to preallocate the variable to the size of the variable you
are expecting and then read the data in a loop. I don't know for sure
how GrADS binary data files are organized, and I am not sure how the
time dimension is being handled: but based on the fact that each
variable has a 0 in its specification I am guessing that only one time
step is actually in the file even though it gives the time dimension
as 21. Assuming that all the U data comes first you could read it
using something like this:

nlat = 180
nlon = 360
nlev = 29

U = new((/nlev,nlat,nlon/), float)
do i = 0, nlev -1
   U(i,:,:) = fbinrecread(fili,i, (/nlat,nlon/))
end do

There is a useful function, fbinnumrec
(http://www.ncl.ucar.edu/Document/Functions/Built-in/fbinnumrec.shtml),
 that will tell you the number of records in the file. It can be
helpful to know this number when trying to figure out layout of the
data in the file.
 -dave


On Fri, Mar 13, 2015 at 4:02 AM, dyjbean soybean <dyjbean at gmail.com> wrote:
> hi, i met a strange problem between grads and ncl.
>
> the below is the ctl file for the data,
> +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>
> dset E:\lnw\postvar2010010100
> options sequential big_endian
> options sequential
> title post output from grapes
> undef -9.99E+33
> xdef  360 linear    0.0000    1.0000
> ydef  180 linear  -89.5000    1.0000
> zdef   29 levels
>  1000.000000
>  962.5000000
>  925.0000000
>  887.5000000
>  850.0000000
>  800.0000000
>  750.0000000
>  700.0000000
>  650.0000000
>  600.0000000
>  550.0000000
>  500.0000000
>  450.0000000
>  400.0000000
>  350.0000000
>  300.0000000
>  275.0000000
>  250.0000000
>  225.0000000
>  200.0000000
>  175.0000000
>  150.0000000
>  125.0000000
>  100.0000000
>  70.00000000
>  50.00000000
>  30.00000000
>  20.00000000
>  10.00000000
> tdef   21 linear 00z01JAN2010   360mn
> vars 46
> u 29 0 u_wind
> v 29 0 v_wind
> t 29 0 temperature
> h 29 0 geopotential height
> q2 29 0 specific humidity
> q3 29 0 specific humidity
> q4 29 0 specific humidity
> q5 29 0 specific humidity
> q6 29 0 specific humidity
> q7 29 0 specific humidity
> ozone 29 0 cloud fraction from microphy
> w 29 0 vertical wind
> ps 0 0 surface pressure
> psl 0 0 sea level pressure
> rainc 0 0 precipitation
> rainnc 0 0 precipitation
> ts 0 0 surface temperature
> glw 0 0 surface downward lw
> gsw 0 0 surface net sw
> glwu 0 0 surface upward lw
> tglwu 0 0 toa upward lw
> gclw 0 0 surface clear sky downward lw
> gclwu 0 0 surface clear sky upward lw
> tgclwu 0 0 toa clear sky upward lw
> gswu 0 0 surface upward sw
> tgsw 0 0 toa net sw
> tgswu 0 0 toa upward sw
> gcsw 0 0 surface clear sky net sw
> gcswu 0 0 surface clear sky upward sw
> tgcsw 0 0 toa clear sky net sw
> tgcswu 0 0 toa clear sky upward sw
> tcc 0 0 total cloud cover
> hcc 0 0 high cloud cover
> mcc 0 0 middle cloud cover
> lcc 0 0 low cloud cover
> tvw 0 0 column integrated qv
> tcw 0 0 column integrated qc
> tiw 0 0 column integrated qi
> hfx 0 0 sueface heat flux
> qfx 0 0 sueface vapour flux
> zs 0 0 terrain
> q2m 0 0 specific humidity at 2m
> t2m 0 0 temperature at 2m
> u10m 0 0 u wind at 10m
> v10m 0 0 v wind at 10m
> pblh 0 0 pbl height
> endvars
>
> +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>
> according to the above ctl file, we can find some variables 4D and others
> 1D,
> so i gave the following code  in ncl for test:
>
> ++++++++++++++++++++++++++++++++++++++++
> fili="postvar2010010100"
> nlat=180
> mlon=360
> nlev=29
> ntim=21
>
> setfileoption("bin","ReadByteOrder","BigEndian")
> uwind=fbinrecread(fili,0,(/mlon,nlat,nlev,ntim/),"float")
>
> ++++++++++++++++++++++++++++++++++++++++
>
> but there come the following warning,
> +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
> warning:fbinrecread: size specified is less than record size, some data will
> not be read
> +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>
> i changed its dimension description with "-1" instead of
> (/mlon,nlat,nlev,ntim/),as follows,
>
> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
> uwind1=fbinrecread(fili,0,-1,"float")
> print(dimsizes(uwind1))
>
> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>
> then, come to the following result,
>
> +++++++++++++++++++++++++++++++++
>  print(dimsizes(uu))
> (0)     64800
> +++++++++++++++++++++++++++++++++
>
> so i use the following code to read the first variable,
>
> ++++++++++++++++++++++++++++++++++++++++++++++++++++
> uwind2=fbinrecread(fili,0,(/mlon,nlat/),"float")
> ++++++++++++++++++++++++++++++++++++++++++++++++++++
>
> there is no warning appeared.
>
>
> the above result shows that the first variable has the dimension "mlon*nlat"
> not "mlon*nlat*nlev*ntim",which is so wierd.
> and i cannot distinguish which level and time the uwind2 represented.
>
> can somebody give me some advice about this difference between ctl file's
> description and ncl realization?
>
> if i should add two cycle, like following,
>
> ++++++++++++++++++++++++++++++++++++++++++++++++++++++
> do i=0,ntim-1
>   do j=0,nlev-1
>      uwind=fbinrecread(fili,0,(/mlon,nlat/),"float")
>      vwind=fbinrecread(fili,1,(/mlon,nlat/),"float")
>      ttemp=fbinrecread(fili,2,(/mlon,nlat/),"float")
>
>      ....................
>      u10m=fbinrecread(fili,43,(/mlon,nlat/),"float")
>      v10m=fbinrecread(fili,44,(/mlon,nlat/),"float")
>      pblh=fbinrecread(fili,45,(/mlon,nlat/),"float")
>   end do
> end do
>
> +++++++++++++++++++++++++++++++++++++++++++++++++++++
>
>
> any help will be appreciated,thanks
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> _______________________________________________
> ncl-talk mailing list
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>