# plotting TIE-GCM netcdf output
**URL:** https://discourse.paraview.org/t/plotting-tie-gcm-netcdf-output/3946
**Category:** ParaView Support
**Created:** [April 2, 2020, 1:37pm UTC](https://discourse.paraview.org/t/plotting-tie-gcm-netcdf-output/3946 "2020-04-02T13:37:52Z")
**Posts on this page:** 5
**Page:** 1
### Author:  [@rainbowsend](https://discourse.paraview.org/u/rainbowsend)
#### Post date: [April 2, 2020, 1:37pm UTC](https://discourse.paraview.org/t/plotting-tie-gcm-netcdf-output/3946/1 "2020-04-02T13:37:52Z")
Hi all,
I am new to Para View, and this forum. I have already done the first quarter of the tutorial and tried to find some answers in the Internet but I was not very successful. I am using Para View 5.8.0 on Manjaro Kyria 19.0.2.
I want to plot the output of the NCAR [TIE-GCM](https://www.hao.ucar.edu/modeling/tgcm/tie.php) (Thermosphere-Ionosphere-Electrodynamics General Circulation Model).
The output is saved in netCDF files following the CF-Conventions (CF-1.0). The fields are given on spherical coordinates: latitude, longitude and pressure scale height as vertical component. When choosing 5 degree resolution the netcdf file has the following entries (extract):
```
dimensions:
time = UNLIMITED; // (4 currently)
lon = 72;
lat = 36;
ilev = 29;
variables:
double time(time=4);
:long_name = "time";
:units = "minutes since 2004-3-20 0:0:0";
:initial_year = 2004; // int
:initial_day = 80; // int
:initial_mtime = 80, 0, 0; // int
double lon(lon=72);
:long_name = "geographic longitude (-west, +east)";
:units = "degrees_east";
:standard_name = "longitude";
double lat(lat=36);
:long_name = "geographic latitude (-south, +north)";
:units = "degrees_north";
:standard_name = "latitude";
double ilev(ilev=29);
:long_name = "interface levels";
:short_name = "ln(p0/p)";
:units = "";
:positive = "up";
:standard_name = "atmosphere_ln_pressure_coordinate";
:formula_terms = "p0: p0 lev: ilev";
:formula = "p(k) = p0 * exp(-ilev(k))";
:computed_standard_name = "air_pressure";
float NE(time=4, ilev=29, lat=36, lon=72);
:long_name = "ELECTRON DENSITY";
:units = "cm-3";
:missing_value = 1.0E36; // double
float ZG(time=4, ilev=29, lat=36, lon=72);
:long_name = "Geometric Height ZG";
:units = "cm";
:missing_value = 1.0E36; // double
```
I have successfully opened a corresponding dataset using the ‘NetCDF Reader’, but I have some questions
1. **Representation**
If I choose ‘outline’ the boundary is a semicircle. I would expect a sphere
2. **Vertical Coordinate**
I would like to use the geometric height (field ZG ) as vertical coordinate instead of the pressure scale height. There is the filter ‘Warp by Vector’ which probably can do this. But ZG is a 4D scalar field and not a vector field. Of course I can compute the cartesian coordinates from ZG, latitude and longitude, and include them in the netcdf file, but Para View can not open 5D variables (lat,lon,time,ilev,dim vector). I get the error message:
I have tried ‘Warp by Scalar’ but all points are then shifted a long the the same axis.
So is there a way to use the geometric coordinates using Para View?
1. **Contour**
After applying the ‘Cell Data to Point Data Filter’ it is possible to use the contour filter.
But I get the following error message:
Are there other ways to solve these problems, e.g., converting the netcdf file to a different format?
Thanks a lot for your help!
---
### Author:  [@mwestphal](https://discourse.paraview.org/u/mwestphal)
#### Post date: [April 6, 2020, 8:08am UTC](https://discourse.paraview.org/t/plotting-tie-gcm-netcdf-output/3946/2 "2020-04-06T08:08:26Z")
please share your file.
---
### Author:  [@rainbowsend](https://discourse.paraview.org/u/rainbowsend)
#### Post date: [April 6, 2020, 8:12am UTC](https://discourse.paraview.org/t/plotting-tie-gcm-netcdf-output/3946/3 "2020-04-06T08:12:40Z")
Sorry, I forget to include the file. You can download it here:
[https://uni-bonn.sciebo.de/s/WBBtavjODIDnKoe](https://uni-bonn.sciebo.de/s/WBBtavjODIDnKoe)
---
### Author:  [@mwestphal](https://discourse.paraview.org/u/mwestphal)
#### Post date: [April 6, 2020, 8:35am UTC](https://discourse.paraview.org/t/plotting-tie-gcm-netcdf-output/3946/4 "2020-04-06T08:35:40Z")
> If I choose ‘outline’ the boundary is a semicircle. I would expect a sphere
This is how curvilinear grid outline is represented in ParaView.
> When changing the representation to ‘points’, only a slice of all internal points is shown. Why are not all points plotted?
Only the point on the “surface” are reoresented, it is kinda breaking up for curvilinear grid though.
> But ZG is a 4D scalar field and not a vector field.
Use calculator filter to create a 3D vector.
> After applying the ‘Cell Data to Point Data Filter’ it is possible to use the contour filter. But I get the following error message:
Looks like a bug with curvilinear grid and contour filter. The results still seems correct though.
In any case, you can convert your dataset to unstructured grid by using a AppendDataset filter.
---
### Author:  [@rainbowsend](https://discourse.paraview.org/u/rainbowsend)
#### Post date: [April 6, 2020, 10:09am UTC](https://discourse.paraview.org/t/plotting-tie-gcm-netcdf-output/3946/5 "2020-04-06T10:09:26Z")
Thanks for your help!
I explain my pipeline here in more detail for other potential users:
1. Load TIE-GCM history file using the ‘NetCDF Reader’.
2. Apply the `Append Datasets` Filter to use unstructured grid representation
3. Apply the `Cell Data to Point Data` Filter
4. Apply the `Calculator` Filter to compute a height vector
Result Array Name: height
`ZG/100000*norm(coords)`
divide by 100000 to get result in km
5. Apply the `Warp by Vector` Filter
In Properties select ‘height’ in the Vectors drop-down list
The result looks as follows:
