[ncl-talk] using gc_inout over the Central Arctic

[Dave Allured - NOAA Affiliate]

Another approach would be to break up your polygon into three separate
wedge shaped polygons, each of which is carefully designed to exclude the
pole and a tiny region around it.  Then create three separate grid masks
with gc_inout, and merge them for your final data mask.  This would avoid
the need to rotate or reproject the data.

I suggest making each wedge so that the pole is excluded not only from the
polygon itself, but also from the convex hull of the polygon.  I would also
suggest making the wedges overlap each other slightly along their common
east/west boundaries, to eliminate the rare possibility of missing grid
points that sit exactly on a boundary line.

If your data grid actually includes the pole itself, then manually set
those mask values(s) to True to handle the special case.

--Dave


On Thu, Dec 1, 2016 at 8:54 AM, Rick Brownrigg <brownrig at ucar.edu> wrote:

> Hi Rym,
>
> I don't definitively know a solution, but I wonder if you might be able to
> get what you need by either i) rotating/translating your data somewhere off
> pole to do the in/out calculation, then rotating back,  or ii) performing
> the calculation in a map projection, and again, deprojecting the results
> back ?
>
> Rick
>
> On Thu, Dec 1, 2016 at 7:10 AM, Rym Msadek <msadek at cerfacs.fr> wrote:
>
>> Hi,
>>
>> I am trying to use the gc_inout function to define a polygon over the
>> Arctic region and use that as a mask to average an oceanic field like
>> potential temperature.
>> I have an issue when the longitudes and latitudes of the polygon cover a
>> region that crosses the North pole. gcinout gives me a contour that is not
>> closed and I can’t figure out how to overcome this issue.
>>
>> Here are the lon lat that define my polygon, over the Eurasian basin of
>> the Arctic.
>> ---------
>> latC=(/81,81.5,82,82.7,82.4,81.6,80,78.8,77,78,79,81,82,84,8
>> 8,88.5,89,86,85,85,84,83,82,82,81/)
>> lonC=(/27,36,48,60,77,94,107,114,121,134,140,141,145,149,132
>> ,149,-79,-55,-43,-22,-20,-17,-9,5,27/)
>>
>> in_obs=gc_inout(nav_lono,nav_lato,lonC,latC)
>>
>> mask_eurasia = where(in_obs.eq.True,1,0)
>> plot= gsn_csm_contour_map_polar(wks,mask_eurasia,res)
>> --------
>> I’m attaching the resulting plot that I get.
>> Note that I have used this same gc_inout function to define other
>> polygons in the Arctic and it works fine when the region was not over the
>> central Arctic.
>>
>> Any help would be appreciated.
>>
>> Thanks,
>>
>> Rym
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/ncl-talk/attachments/20161201/ff71cae1/attachment.html