[pyngl-talk] mapping regional data with mpLimitMode corners

Tue Jun 26 15:36:09 MDT 2018

Hi Mary,

The combination below eliminated the map edge problems I was having when mapping the regional stereographic COAMPS grid. The mp resources used are analogous to your NCL native example 6. Originally, I had mpCenterLatF = 80. And although this is the center lat of the grid, the relativeCenterLat is 90 for the stereographic projection. I believe this is the correct logic.

    mpres.mpLimitMode                 = "Corners"
    mpres.mpProjection                = "Stereographic"
    mpres.mpLeftCornerLatF            = lat[0,0]
    mpres.mpLeftCornerLonF            = lon[0,0]
    mpres.mpRightCornerLatF           = lat[ny-1,nx-1]
    mpres.mpRightCornerLonF           = lon[ny-1,nx-1]

    mpres.mpRelativeCenterLon         = True
    mpres.mpCenterLonF                = 358.
    mpres.mpRelativeCenterLat         = True
    mpres.mpCenterLatF                = 90.     ###  Original value of 80.
    mpres.tfDoNDCOverlay              = "NDCViewport"

Thank you,
Matt

From: Mary Haley [mailto:haley at ucar.edu]
Sent: Monday, June 25, 2018 11:05 AM
To: Fearon, Dr. Matthew, Contractor, Code 7533
Cc: pyngl-talk at ucar.edu
Subject: Re: [pyngl-talk] mapping regional data with mpLimitMode corners

Hi Matt,

Sorry for the delay. I was creating a new NCL example to help explain things.

Plotting data in its native projection can be tricky, especially if you have a rotated pole or if you are not sure of the exact parameters for the map projection.

I think you already know this, but these resources are going to be critical to getting the map projection correct:

    mpres.mpProjection                = "Stereographic"
    mpres.mpLeftCornerLatF            = lat[0,0]
    mpres.mpLeftCornerLonF            = lon[0,0]
    mpres.mpRightCornerLatF           = lat[ny-1,nx-1]
    mpres.mpRightCornerLonF           = lon[ny-1,nx-1]

#    mpres.mpCenterRotF                = -1
    mpres.mpRelativeCenterLon         = True
    mpres.mpCenterLonF                = 358.
    mpres.mpRelativeCenterLat         = True
    mpres.mpCenterLatF                = 80.

It's likely going to be the center lat/lon settings that are the issue, and not the four corners of your data, since the way you are setting the corners looks correct to me.

Did the data come from a file that had the map projection parameters defined? If so, could you include the output from an ncdump -h of that file so I can see if there's anything that might have been left out?

I don't know if you saw these examples, but they might help (even though they are NCL-based, the same principles apply):

http://www.ncl.ucar.edu/Applications/native.shtml#ex1
http://www.ncl.ucar.edu/Applications/native.shtml#ex6

Example native_6.ncl is the one I just created.  This was an example where we didn't know the correct center longitude, so I had to create a do loop to try different ones out. I also drew the lat/lon lines associated with the data (since I had them), so I could see if they were straight or not. If the lines are straight, then this is a good sign that you have the correct map projection resources set.

Since you have the lat/lon arrays, you can also add the grid for debugging purposes. You can use Ngl.add_polyline for this. I've attached an example of adding lat/lon grid lines using dummy data.

Hope this helps,

--Mary

On Wed, Jun 20, 2018 at 2:14 PM, Fearon, Dr. Matthew, Contractor, Code 7533 via pyngl-talk <pyngl-talk at ucar.edu<mailto:pyngl-talk at ucar.edu>> wrote:
Hello,

I’m using the following Ngl resources to map regional raster data with a stereographic projection onto a basemap. Things work fairly well, except I am able to see edges of the underlying basemap. Please see figure one attached. When I include mpCenterRotF= -1, the basemap edge visibility is reduced, but still partially shown along the NW and SE edges. Please see figure two. I would like the data overlay to fill the window fully.  I’ve tried adjusting the lat corners by 0.5 to 1 degree, eg, mpRightCornerLatF = lat[ny-1,nx-1]-0.5 and mpLeftCornerLatF = lat[0,0]+0.5, to figure two but doing so seems to reveal the underlying basemap more – perhaps because rotation of -1 is active after I contract my lat corners. Seems like this should be a simple fix, but here I struggle. Perhaps I should be mapping regional data differently? Any help or suggestions would be greatly appreciated.

Thank you,
Matt

    cnres.sfXArray                   = lon
    cnres.sfYArray                   = lat

    mpres.mpFillOn                   = True
    mpres.mpLandFillColor            = "Tan1"
    mpres.mpOceanFillColor           = "SkyBlue"
    mpres.mpInlandWaterFillColor     = "SkyBlue"
    mpres.mpGeophysicalLineThicknessF= 2.5
    mpres.mpLimitMode                = "Corners"  #"LatLon"

    mpres.mpProjection                = "Stereographic"
    mpres.mpLeftCornerLatF            = lat[0,0]
    mpres.mpLeftCornerLonF            = lon[0,0]
    mpres.mpRightCornerLatF           = lat[ny-1,nx-1]
    mpres.mpRightCornerLonF           = lon[ny-1,nx-1]

#    mpres.mpCenterRotF                = -1
    mpres.mpRelativeCenterLon         = True
    mpres.mpCenterLonF                = 358.
    mpres.mpRelativeCenterLat         = True
    mpres.mpCenterLatF                = 80.
    mpres.tfDoNDCOverlay              = "DataTransform" #NDCViewport"

    mpres.mpGridLineThicknessF       = 0.5
    mpres.mpGridLineDashPattern      = 2
    mpres.mpGridSpacingF             = 1

    mpres.mpDataBaseVersion          = "MediumRes"
    map_plot                         = Ngl.map(wks,mpres)

    cnres.cnFillOn                   = True
    cnres.cnLinesOn                  = False
    cnres.cnLineLabelsOn             = False
    cmap                             = Ngl.read_colormap_file("MPL_PuOr")
    cnres.cnFillPalette              = cmap[::-1,:]
    cnres.cnLevelSelectionMode       = "ExplicitLevels"
    cnres.cnLevels                   = [-10,-8,-6,-4,-2,2,4,6,8,10]

    cnres.lbAutoManage               = False
    cnres.pmLabelBarWidthF           = 0.02
    cnres.lbLabelFontHeightF         = 0.009
    contour_plot                     = Ngl.contour(wks,cvar,cnres)

    Ngl.overlay(map_plot,contour_plot)

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

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.ucar.edu/pipermail/pyngl-talk/attachments/20180626/13dbe38f/attachment-0001.html>