Projecting WRF output onto the WGS-84 Ellipsoid

[jordanA]

I know that both WRF and WPS both assume a spherical earth model with a radius of 6370km. I'm looking to display the data onto the WGS-84 Ellipsoid model. While I'm primarily concerned with the continental US I will eventually want to do this for the entire globe.

I've found information, most notably here: http://computing.io/wp/2016/01/wrf-lati ... nd-datums/
That suggests that WPS and WRF may just be treating the WGS-84 information in the land-use data as spherical earth information and this suggest using the XLAT and XLONG output directly from WRF and putting it onto the WGS-84 Ellipsoid. This doesn't sound 100% correct however, and certainly wind information would need to be transformed slightly differently as well well some other information, in particular I would expect the geopotential height may need to be adjusted for the change in surface shape.

I know that to do transforms between Ellipsoids I'd need to use one of the Helmert, Molodensky-Badekas, or Molodensky methods. Have any of you gone down this road before and if so how did you go about transforming the data?

[winash12]

I am not sure why you wish to complicate matters more than what is already a complicated problem.




In order to project the output onto WGS-84 ellipsoid I think you would need to write your own code.


This site can give more details - gis.stackexchange.com

[winash12]

This should tell you more what those terms mean. It is a peer reviewed website

http://gis.stackexchange.com/questions/ ... -longitude

[jordanA]

I simulate aircraft on the WGS-84 Ellipsoid, corresponding Geoid(s) and so on. We use the WGS84 as that is pretty much the standard especially with GPS navigation becoming more and more popular.

We've been looking for some nice realistic weather data. In the past we've used RUC/RAP236 grib files directly but this just gives us some static wind/temp/etc for the course of the simulation and the individual that did that didn't adjust from the spherical earth in those files.

Since then we've really beefed up all our loxodrome, and geodetic calculations and all the rest so our plans were to use WRF to fill in a bunch of data to get realistic wind/pressure/temp changes for whatever area we'd be simulating, which is especially handy for simulations that last several hours/days/weeks.

thanks for pointing to 'gis.stackexchange.com' I'll do some further searching over there.

[jordanA]

As far as the rotated grid goes, If I'm using ARW instead of MMM aren't the grids only rotated if I use the Regular Lat/lon or Cylindrical Equidistant projections with adjusted pole_lat and pole_lon values?

So if I used Lambert, Mercator, or Polar Sterographic projections it wouldn't be an issue?

[winash12]

Jordan - No. As far as I understand it at least. You can confirm that with wrfhelp.

[jordanA]

Heard back from wrfhelp. Only when pole_lat and pole_lon are specified is the grid rotated.