RSC change from inner to outer merging to get more data
Closed this issue · 5 comments
In level2/compute_radial_surface_current:
level2['RadialSurfaceCurrent'] = xr.concat( [level1.RadialSurfaceVelocity.sel(Antenna='Fore') - dswasv_f, level1.RadialSurfaceVelocity.sel(Antenna='Aft') - dswasv_a], 'Antenna', join='inner')
join could certainly be changed to "outer" at this stage. To test how it affects the rest of the chain
This was likely originally joined by 'outer', however I have found problems with coordinates downstream - e.g., lat/lon coordinates being padded by NaNs after joining. Currently joining by 'inner' solves these issues - however I agree that we should aim to have the full datasets present in the joined version. We will need to solve the creation and assignment of the lat/lon coordinates to the datasets in order to achieve this.
Currently these coords are taken from the LatImage / LonImage arrays, which are specific to a look direction. We will therefore need to join/merge the LatImage / LonImage arrays into one big one to cover the 'outer' joined dataset
Linked with #185
Believed to be closed with pull request #192.
Test to see that the RSC data being passed to the L2 processing is indeed 'outer' in size before closing.
(accidentally closed)