Re: [ferret_users] seeing some strange output in curvilinear grids

[Kankan Sarkar <kankan.math@xxxxxxxxx>]

On 2/24/14, Russ Fiedler <russell.fiedler@xxxxxxxx> wrote:
> Hi,
>
> No, Ghansham's command is ok. You can do a 3 variable shade/fill with
> the last 2 arguments being 2D x/y coordinates.

Yes, Ghansham's command is wright. I checked once again. Filling the
map with three variables gives lat & lon with units in the axes. While
filling with one variable gives index in the axis.

>
> Actually this looks like a bug(?) with how the end points are being
> handled when the x/y coordinates contain missing values.
>
> If you make a plot like
>
> shade/j=400:800/i=400:800/lev=50 img_wv,longitude_wv,latitude_wv
>
> It looks great.
>
> Now home in near an edge where missing values for lon/lat start
>
>   shade/j=1180:1184/i=1175:1181
>
> is ok but getting distorted
>
> Now extend the range of "i' by 1
>
>   shade/j=1180:1184/i=1175:1182   ! Whoa
>
> Listing shows we get missing values entering at j=1184,i=1182
>
> yes? list/j=1180:1184/i=1181:1182
>               DATA SET: ./test.h5
>               X: 1179.5 to 1181.5
>               Y: 1178.5 to 1183.5
>               TIME: 18-FEB-2014 04:00
>   Column  1: IMG_WV is Water Vapor Count
>   Column  2: LONGITUDE_WV is longitude (degrees_east)
>   Column  3: LATITUDE_WV is latitude (degrees_north)
>                IMG_WV  LONGITU LATITUDE_WV
>   ---- J:1180 Y:   1179
> 1180   / 1181:  930.0   152.6  -43.80
> 1181   / 1182:  931.0   153.5  -43.90
>   ---- J:1181 Y:   1180
> 1180   / 1181:  933.0   153.5  -44.01
> 1181   / 1182:  950.0   154.6  -44.13
>   ---- J:1182 Y:   1181
> 1180   / 1181:  948.0   154.6  -44.25
> 1181   / 1182:  993.0   156.1  -44.41
>   ---- J:1183 Y:   1182
> 1180   / 1181:  976.0   156.0  -44.53
> 1181   / 1182:  995.0   159.2  -44.87
>   ---- J:1184 Y:   1183
> 1180   / 1181:  995.0   159.5  -45.02
> 1181   / 1182:   ....    ....    ....
>
>
>
> It looks like the polygons that are being created are somehow being
> extrapolated to the borders of the plot.  We can change the behaviour by
> specifying hlim and/or vlim
>
> shade/j=1180:1184/i=1175:1182/vlim=-45.5:-43.0
>
>
>
> I was able to get something that looks kind of ok if you fill the
> missing lons/lats but looks a bit odd at the edges but it might be
> correct for all I know.
>
> let lon=longitude_wv[x=@fnr,y=@fnr]
> let lat=latitude_wv[x=@fnr,y=@fnr]
>
> shade img_wv,lon,lat  ! Success
>
>
> Cheers,
> Russ
>
>
>
>
>
>
> On 24/02/14 15:50, Kankan Sarkar wrote:
>> Dear Ghansham,
>>
>> I think, your plot is not proper because of, you have given three
>> variable to fill together. You can make fill plot for one variable at
>> a time. You may try the following command -
>>
>> fill/PALETTE=greyscale  IMG_WV
>>
>> Regards,
>> Kankan
>>
>>
>>
>> On Fri, Feb 21, 2014 at 7:22 PM, Ghansham Sangar
>> <ghanshamsangar@xxxxxxxxx <mailto:ghanshamsangar@xxxxxxxxx>> wrote:
>>
>>     Hi All
>>
>>     Kindly see this one.
>>     I am trying to plot a curvilinear dataset
>>     This is an hdf5 product that follows Netcdf-4 classic data model
>>     and CF conventions.
>>     I am able to open it.
>>     But there are a few issues.
>>     Kindly check attached files.
>>
>>     regards
>>     ghansham
>>
>>
>>
>>
>> --
>> Kankan Sarkar
>> Senior Research Fellow
>> Physical Oceanography Division
>> CSIR - National Institute of Oceanography
>> Dona Paula - 403 004, Goa, India
>
>


-- 
Kankan Sarkar
Senior Research Fellow
Physical Oceanography Division
CSIR - National Institute of Oceanography
Dona Paula - 403 004, Goa, India