Re: ccsm to vis5d

Hi Michael,

These functions were written many years ago by somebody who is
now gone. We don't even build these into the standard NCL binaries
because people haven't been using them.

However, that doesn't mean we won't try to answer questions about them!

Looking at the interface for v5d_write, it looks like the first two arguments
must be scalar integers, and the 3rd argument must be a 3D float array of
any dimensionality.

I think you can keep using v5d_write to write 3D arrays to the same file, before
you call v5d_close.

I've attached a sample test script that I found.

If this still doesn't work, let me know if you are getting an error message,
and exactly what is happening.

--Mary

On Mar 23, 2012, at 9:38 AM, Michael Notaro wrote:

> The script below gives an error in the v5d_write line about
> the dimensions being wrong. I have never used these v5d
> functions before and so I am doing something wrong. I
> am trying to write 3 variables to a file and there are no such
> examples to follow. To make things easier, I converted all
> 3 variables to 3D and then combined them all into a variable
> called data and then tried to write data to a v5d file with no
> success. I am not really sure what I am doing wrong here.
> So the problem is related to my confusion with these v5d ncl
> function and I don't think the v5d user group will be able to
> help me with that since they likely don't even use ncl.
> Any advice would be appreciated here. Sorry I am so confused
> with this.
>
>
>
>
>
>
> a=addfile("test1.nc","r")
> b=addfile("test2.nc","r")
> psl=a->PSL(0,:,:)
> cloud=b->CLOUD(0,:,:,:)
> pcp=(b->PRECL(0,:,:))+(b->PRECC(0,:,:))
> pcp=pcp*60.*60.*24.*1000. ; mm/d
> lev=doubletofloat(a->lev)
> lev=lev(::-1)
> copy_VarCoords(a->lev,lev)
> lat=doubletofloat(a->lat)
> lon=doubletofloat(a->lon)
> lat=lat(::-1)
> lon=lon(::-1)
> copy_VarCoords(a->lat,lat)
> copy_VarCoords(a->lon,lon)
>
> psl=psl(::-1,:)
> pcp=pcp(::-1,:)
> cloud=cloud(:,::-1,:)
>
> psl!0="lat"
> psl!1="lon"
> psl&lat=lat
> psl&lon=lon
> pcp!0="lat"
> pcp!1="lon"
> pcp&lat=lat
> pcp&lon=lon
> cloud!0="lev"
> cloud!1="lat"
> cloud!2="lon"
> cloud&lev=lev
> cloud&lat=lat
> cloud&lon=lon
>
> numtimes = 1
> numvars = 3
> nr = 96
> nc = 144
> nl = (/26,26,26/)
>
> varnames=(/"psl","pcp","cloud"/)
>
> timestamp = (/1/)
> datestamp = (/1/)
> compress = 1
> northlat = 90.0
> latinc = 1.9
> westlon = 0.0
> loninc = 2.5
> projection = 1
> proj_args = (/northlat, westlon, latinc, loninc/)
> vert = 3
> vert_args=lev
>
> data=new((/96,144,3*26/),float)
> do i=0,25
> data(:,:,i)=psl
> end do
> do i=26,51
> data(:,:,i)=pcp
> end do
> data(:,:,52:77)=cloud(lat|:,lon|:,lev|:)
>
> v5d_create("monsoon.v5d", numtimes, numvars, nr, nc, nl, varnames, timestamp, datestamp, compress, projection, proj_args, vert, vert_args)
> v5d_write(numtimes, numvars, data)
> v5d_close()
>
>
>
>
>
> On Mar 23, 2012, at 10:29 AM, Mary Haley wrote:
>
>> Sorry, I misunderstood your question then. I thought the issue was with
>> writing a vis5d file. Can you explain what you mean not being sure how to
>> handle vertical coordinates?
>>
>> --Mary
>>
>> On Mar 23, 2012, at 8:21 AM, Michael Notaro wrote:
>>
>>> Thanks.
>>> Hopefully someone will email an answer, although I doubt it since
>>> it is more of an NCL question than a vis5d question.
>>>
>>>
>>>
>>> On Mar 23, 2012, at 7:56 AM, Mary Haley wrote:
>>>
>>>> Hi Michael,
>>>>
>>>> I'm afraid we don't have any vis5d experience in our group. I've CCed the vis5dplus group on this question and your other one.
>>>>
>>>> If you want to join that group:
>>>>
>>>> http://mailman.ucar.edu/mailman/admin/vis5dplus
>>>>
>>>> --Mary
>>>>
>>>> On Mar 21, 2012, at 5:20 PM, Michael Notaro wrote:
>>>>
>>>>> I am trying to convert sea-level pressure, precipitation, and cloud
>>>>> cover at all 26 levels on 1.9x2.5 degree grid from CCSM3.5 to
>>>>> a vi5d file for a single time step. Could you someone take a look
>>>>> at the code below and offer some suggestions? I am particularly
>>>>> not sure how to handle the vertical coordinates or multiple variables.
>>>>> Sorry, I have no experience with vis5d up to now so this is my first
>>>>> try to convert the data into that format.
>>>>>
>>>>> Michael
>>>>>
>>>>>
>>>>>
>>>>> a=addfile("test1.nc","r")
>>>>> b=addfile("test2.nc","r")
>>>>> psl=a->PSL(0,:,:)
>>>>> cloud=b->CLOUD(0,:,:,:)
>>>>> pcp=(b->PRECL(0,:,:))+(b->PRECC(0,:,:))
>>>>> pcp=pcp*60.*60.*24.*1000. ; mm/d
>>>>> copy_VarCoords(psl,pcp)
>>>>> lat=doubletofloat(a->lat)
>>>>> lon=doubletofloat(a->lon)
>>>>> copy_VarCoords(a->lat,lat)
>>>>> copy_VarCoords(a->lon,lon)
>>>>>
>>>>> numtimes = 1
>>>>> numvars = 3
>>>>> nr = 96
>>>>> nc = 144
>>>>> nl = 26
>>>>>
>>>>> timestamp = (/1/)
>>>>> datestamp = (/1/)
>>>>> compress = 1
>>>>> northlat = 90.0
>>>>> latinc = 1.9
>>>>> westlon = 0.0
>>>>> loninc = 2.5
>>>>> projection = 1
>>>>> proj_args = (/northlat, westlon, latinc, loninc/)
>>>>> bottomhgt = 0.0
>>>>> hgtinc = 1.0
>>>>> vert = 3
>>>>> vert_args = (/bottomhgt, hgtinc/)
>>>>> v5d_create("UVdata.v5d", numtimes, numvars, nr, nc, nl, (/"psl","pcp","cloud"/), timestamp, datestamp, compress, proj, proj_args, vert, vert_args)
>>>>> v5d_write(numtimes, numvars, (/"psl","pcp","cloud"/))
>>>>> v5d_close()
>>>>>
>>>>> _______________________________________________
>>>>> ncl-talk mailing list
>>>>> List instructions, subscriber options, unsubscribe:
>>>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>>>
>>>> _______________________________________________
>>>> ncl-talk mailing list
>>>> List instructions, subscriber options, unsubscribe:
>>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>>
>>
>

_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk