Re: How to use 2D lon/lat (curvilinear) with gsn_contour_csm_map_ce and gsnZonalMean

Mario,

More information provided by Dennis Shea: the idea behind a zonal mean
plot is that you are averaging all values across the longitude
dimension for each latitude value. The assumption, then, is that your
longitude coordinate values are constant across a particular latitude.
This is what 1D coordinate arrays provide.

With 2D lat/lon arrays, you can't make this assumption.

--Mary

On Wed, 15 Jul 2009, Mary Haley wrote:

> Hi Mario,
>
> You posted your question to the PyNGL email list. This is more
> appropriate for the ncl-talk email list.
>
> I'm afraid the zonal means plot only works if you have 1D coordinate
> arrays. The error message
> is misleading because the script was written before 2D lat/lon arrays
> were allowed.
>
> I'm in the middle of a workshop right now and can't look at this.
>
> You might try calculating the zonal mean yourself (using dim_avg) and
> then generating an
> XY plot of the data.
>
> I'll try to look at this later this week.
>
> --Mary
>
>
> On Jul 14, 2009, at 3:21 AM, Mario Krapp wrote:
>
>> Hi,
>>
>> I want to contour the data of a curvilinear grid using
>> gsn_contour_csm_map_ce and add a zonal mean as in
>> http://www.ncl.ucar.edu/Applications/Scripts/dev_1.ncl
>> I use version 5.0.0.
>>
>> The lon, lat values are 2D arrays so I use
>> res_at_sfXArray = lon
>> res_at_sfYArray = lat
>>
>>
>> Here is the part I changed according to the script mentioned above:
>> ---cut---
>> fn = "tho_6m_mean.nc" ; define filename
>> in = addfile(fn,"r") ; open netcdf file
>> ts = in->var2(0,0,:,:) ; select variable
>> to ave
>> lat = in->lat
>> lon = in->lon
>>
>> ;************************************************
>> ; plot parameters
>> ;************************************************
>> wks = gsn_open_wks("ps","stddev") ; Opens a ps file
>> gsn_define_colormap(wks,"BlWhRe") ; choose colormap
>>
>> plot = new(2,graphic) ; create graphical
>> array
>> res = True ; plot mods desired
>> ;************************************************
>> ; original data
>> ;************************************************
>>
>> res_at_gsnZonalMean = True ; add zonal plot
>> res_at_mpOutlineOn = False
>> res_at_sfXArray = lon
>> res_at_sfYArray = lat
>>
>> res_at_vpWidthF = 0.7
>>
>> plot(0) = gsn_csm_contour_map_ce(wks,ts(:,:),res)
>> ---cut---
>>
>> The message I receive is:
>> (0) gsn_csm_contour_map_ce: Warning: The resource gsnZonalMean
>> can only be set to True
>> fatal:Plus: Dimension size, for dimension number 0, of operands
>> does not match, can't continue
>>
>> So I uncomment
>> ;res_at_gsnZonalMean = True ; add zonal plot
>>
>> now I receive:
>> warning:ScalarFieldSetValues: 2d coordinate array sfXArray has an
>> incorrect dimension size: defaulting sfXArray
>> warning:ScalarFieldSetValues: 2d coordinate array sfYArray has an
>> incorrect dimension size: defaulting sfYArray
>>
>> I don't know what this means and made a 'ncl_filedump':
>> ---cut---
>> Variable: f (file variable)
>>
>> filename: tho_6m_mean
>> path: tho_6m_mean.nc
>> file global attributes:
>> CDI : Climate Data Interface version 1.3.2
>> Conventions : CF-1.0
>> dimensions:
>> x = 362
>> y = 192
>> nv = 4
>> lev = 1
>> time = 1 // unlimited
>> variables:
>> float lon ( y, x )
>> long_name : longitude
>> units : degrees
>> standard_name : longitude
>> bounds : lon_bounds
>>
>> float lon_bounds ( y, x, nv )
>>
>> float lat ( y, x )
>> long_name : latitude
>> units : degrees
>> standard_name : latitude
>> bounds : lat_bounds
>>
>> float lat_bounds ( y, x, nv )
>>
>> double lev ( lev )
>> long_name : generic
>> units : level
>>
>> double time ( time )
>> units : day as %Y%m%d.%f
>> calendar : proleptic_gregorian
>>
>> float var2 ( time, lev, y, x )
>> coordinates : lon lat
>> _FillValue : -9e+33
>> ---cut---
>>
>> If I plot the data just using gsn_contour_map everything works
>> fine, but gsnZonalMean is not supported.
>>
>> Thanks and Regards
>>
>> Mario Krapp
>>
>>
>> --
>> Mario Krapp
>> Max Planck Institute for Meteorology
>>
>> Bundesstraße 53
>> 20146 Hamburg
>> Germany
>> tel. +49 40 41173 168
>> email mario.krapp_at_zmaw.de
>> www http://www.mpimet.mpg.de
>> _______________________________________________
>> pyngl-talk mailing list
>> List instructions, subscriber options, unsubscribe:
>> http://mailman.ucar.edu/mailman/listinfo/pyngl-talk
>
> _______________________________________________
> pyngl-talk mailing list
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/pyngl-talk
>

_______________________________________________
pyngl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/pyngl-talk
Received on Wed Jul 15 2009 - 22:24:03 MDT