Verdi shapefile error

[YoonB]

Dear CMAS Forum,

hello I’m using VERDI-2.0 Beta version to analyze my CMAQ results.
I added an external shapefile to calculate the areal interpolation but when I try to plot the areal interpolation, an error occurs disclaiming
“Problem with areal interpolation calculations. Check if polygons intersect grid cells.”

Therefore I checked the shapefile with another gis file viewer software and it seems that the shapefile itself doesn’t have any issues. There were no intersection between administrative area polygons in the gis file itself. However when I apply the same file on verdi, an erroneous layer comes out with unidentifiable lines across the area.

I would like to know if its sort of projection error in verdi and how it can be fixed

The following pictures are

1. The error screen when I try to plot the areal interpolation
2. The erroneous areal interpolation plot
3. The same shapefile displayed on a different viewer software showing differences with picture 2.

[dschwede]

Thank you for your question. Could you please retry the plot with the VERDI 2.1 which was released in September?

[YoonB]

Thank you for your reply
However the same problem occurs in VERDI 2.1 as well

I’d like to share the shapefile to check for errors but it seems like the GIS format files are unable to upload in the forum…

[lizadams]

If you would like to upload your shapefiles and the output file that you are using to create the areal interpolation to the following google drive, we can take a look.

README-CMAS_User_Forum

[YoonB]

Thank you

I uploaded the shapefile and the output which includes the whole shapefile area in the drive.

[lizadams]

Hi YoonB,
There are two issues.

1. the TFLAG value and the SDATE were not the same in your egts3d file.
   I used nco to edit the TFLAG to make it 2019361, 0 to agree with the SDATE.
   I am uploading the modified egts3d file for you to use.

2. I retried using a different shapefile of South Korea, and was able to partially get it to work with the new egts3d file, but it didn’t fully create the areal interpolation plot, just seemed to hang. I then extracted one of the polygons to see if that would work, and it did.

I obtained the shapefile from this location.
https://maps.princeton.edu/catalog/stanford-dk009rq9138

I had tried to check the validity of your shapefile using QGIS, and it didn’t detect anything wrong.
However, perhaps you can use QGIS to simplify the vertices in the shapefile. https://docs.qgis.org/2.14/en/docs/training_manual/processing/generalize.html However, this dk009rq9138.shp from the above website contains fewer administrative levels.

This is the output that I see when creating the areal interpolation plot:
verdi -f $cwd/egts3d_l.2019.12.base.ncra.new.2.nc
^[[O^[[I^[[O^[[I^[[OCalculated target 2 of 17: Chungcheongnam-do size 62x41, 2542.0 in 218188ms, 0.011650503235741653squares/ms
Calculated target 3 of 17: Daegu size 11x15, 165.0 in 19ms, 8.68421052631579squares/ms
Calculated target 4 of 17: Daejeon size 8x10, 80.0 in 7ms, 11.428571428571429squares/ms
Calculated target 5 of 17: Gangwon-do size 69x58, 4002.0 in 65847ms, 0.06077725636703267squares/ms
Calculated target 6 of 17: Gwangju size 10x6, 60.0 in 5ms, 12.0squares/ms
Calculated target 7 of 17: Gyeonggi-do size 38x50, 1900.0 in 47306ms, 0.04016403838836511squares/ms
Calculated target 8 of 17: Gyeongsangbuk-do size 90x77, 6930.0 in 135997ms, 0.05095700640455304squares/ms
Calculated target 9 of 17: Gyeongsangnam-do size 49x48, 2352.0 in 383855ms, 0.006127313699183286squares/ms
Calculated target 10 of 17: Incheon size 98x97, 9506.0 in 549238ms, 0.017307615277894102squares/ms
Calculated target 11 of 17: Jeju size 25x17, 425.0 in 7560ms, 0.056216931216931214squares/ms
Calculated target 12 of 17: Jeollabuk-do size 54x32, 1728.0 in 63736ms, 0.027111836324839964squares/ms
Nov 15, 2021 11:02:36 PM org.geotools.referencing.factory.DeferredAuthorityFactory disposeBackingStore

INFO: Disposing class org.geotools.referencing.factory.epsg.UnnamedExtension backing store

Nov 15, 2021 11:02:36 PM org.geotools.referencing.factory.DeferredAuthorityFactory disposeBackingStore

INFO: Disposing class org.geotools.referencing.factory.epsg.EsriExtension backing store

I wonder if the large number of small islands makes the area interpolation calculation take a long time to calculate the area.
I simplified the shapefile to only contain 6 regions and got it to work. But again, it was very slow.

verdi -f $cwd/egts3d_l.2019.12.base.ncra.new.2.nc

Calculated target 0 of 6: Gyeongsangnam-do size 49x48, 2352.0 in 396134ms, 0.005937384824327122squares/ms
Calculated target 1 of 6: Gyeongsangbuk-do size 90x77, 6930.0 in 134850ms, 0.05139043381535039squares/ms
Calculated target 2 of 6: Daegu size 11x15, 165.0 in 19ms, 8.68421052631579squares/ms
Calculated target 3 of 6: Chungcheongbuk-do size 40x45, 1800.0 in 666ms, 2.7027027027027026squares/ms
Calculated target 4 of 6: Ulsan size 13x12, 156.0 in 1331ms, 0.11720510894064613squares/ms
Calculated target 5 of 6: Gangwon-do size 69x58, 4002.0 in 63061ms, 0.06346236184012305squares/ms
TargetCalculator calculated 6 areas in 596065ms
The data produced was as follows:

Modified egts3d and shapefiles available here skorea_shapefile_egts3d.tar.gz - Google Drive