Re: Offset terrain in WRF Polar Stereographic

Hi Don,

I think I see what the issue is, and Dave Brown is looking into it.

Your map actually is supposed to be a polar stereographic projection,
and hence the data should be projected into a circular space.

However, for some reason, when you use tfDoNDCOverlay = True, it is
trying to project the data into a rectangular space.

The work-around for now is to set the sfXArray/sfYArray resources to
the lon2d/lat2d values read off the file.

I'll let you know if we can get the other method to work.

--Mary

On Feb 17, 2010, at 11:29 AM, Don Morton wrote:

> Thank you, Mary.
>
> Yes, that's the approach I've been using (you're actually the one
> who pointed me to it several years ago - you taught me everything I
> know!!), and that "skew" has always been an issue, and people
> comment on it, BUT, it puts the data in the right place. You can
> see we even use your "Mary Haley Shove Those Darned Gridpoints
> Where They Belong" method in our daily runs - the right hand side
> is skewed just a little.
>
> http://weather.arsc.edu/ForecastProducts/AKTrial2/webImages/html/
> forecast-d02.php
>
> So, I guess I'm wondering if this is something that the WRF NCL
> routines should be expected to handle. I don't know if it matters
> or not, but (for wrfhelp's benefit) here are the namelist.wps
> entries used to create the domain. If anyone wants to look at it,
> I happen to have an 81km (and 27km) resolution wrfout file of the
> same region, so those would be smaller files (one and two orders of
> magnitude, respectively).
>
>
>
> &geogrid
> parent_id = 1,
> parent_grid_ratio = 1,
> i_parent_start = 1,
> j_parent_start = 1,
> e_we = 675,
> e_sn = 675,
> geog_data_res = '2m','2m',
> dx = 9000,
> dy = 9000,
> map_proj = 'polar',
> ref_lat = 64.85,
> ref_lon = -147.85,
> truelat1 = 64.85,
> stand_lon = -147.85,
> geog_data_path = '/datadir/morton/WRF/geog'
> /
>
>
>
>
> On Tue, Feb 16, 2010 at 2:37 PM, Mary Haley <haley@ucar.edu> wrote:
> Don,
>
> I converted your wrf_xxxx to use gsn_xxxx functions, so that I
> could use the XLAT and XLONG values on the file to do the
> transformation, instead of depending on the map projection to be
> exactly right.
>
> I did use the same map parameters that I believe wrf_xxxx is trying
> to use.
>
> As you can see, the contours now fall within the borders much
> better, but it also looks like the rotation is skewed a little. I
> think maybe either the wrf_map_overlays function isn't doing the
> right thing, or else the map parameters on your file are not
> correct. My guess is that mpCenterLonF or mpCenterLatF may be
> incorrect, or one of the Corners resources is incorrect.
>
> Perhaps wrfhelp has some information here...
>
> --Mary
>
>
> On Feb 16, 2010, at 12:04 PM, Don Morton wrote:
>
>> Hi,
>>
>> I thought this problem had been resolved ages ago, but maybe only
>> for Lambert projections.
>>
>> I'm using NCL 5.1.1 on WRF v3.1.1 netCDF output with polar
>> stereographic projection, and I'm running into the same old
>> terrain-offset issues that some of us have mentioned over the past
>> several years.
>>
>> Am I correct in my expectations that the WRF NCL mapping routines
>> "should" handle this these days? Or, am I possibly missing
>> something key? I can go back and handle this the old way, using
>> the standard NCL plotting functions I once used, with the
>> following sorts of resource settings
>>
>> res@gsnAddCyclic = False ; Regional data,
>> not cyclic
>> res@tfDoNDCOverlay = False ; Set True for
>> native projection
>>
>> ; Specify map boundaries
>> res@mpLimitMode = "Corners" ; Use the corners
>> limit mode
>>
>>
>> and not using all those convenient WRF functions, but now I'm
>> wondering...
>>
>>
>> My resulting image is at
>>
>> <http://weather.arsc.edu/Miscellaneous/NCLWoes/Hgt.png>
>>
>> With a simplified script available at
>>
>> <http://weather.arsc.edu/Miscellaneous/NCLWoes/Hgt.ncl>
>>
>> and the wrfout file (870 Mbytes) available at
>>
>> <http://weather.arsc.edu/Miscellaneous/NCLWoes/
>> wrfout_d01_2009-12-20_00:00:00>
>>
>>
>> --
>> Arctic Region Supercomputing Center
>> http://www.arsc.edu/~morton/
>> _______________________________________________
>> ncl-talk mailing list
>> List instructions, subscriber options, unsubscribe:
>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>
>
>
>
>
> --
> Arctic Region Supercomputing Center
> http://www.arsc.edu/~morton/

_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Thu Feb 18 08:11:21 2010