Problem with wind interpolation/rotation around 0°

[efvik]

I am getting some strange/artificial wind around 0 degrees, which I guess is introduced at some point during interpolation and rotation (pyfimex?). Below I have compared the data from different sources. What could be the issue here?

For the wind direction, I have changed the range from [0,360) to [-180,180) for visualization purposes (to see that it is mostly continuous).

Script used:

import dnora as dn
from dnora.grid.mask import Edges
import numpy as np

# Bathymetry
grid = dn.grid.GEBCO(lon=(11.331293, 15.7666452),
                       lat=(77.295424, 78.3997957),
                       name="Svalbard")

# Set spacing and boundary points
grid.set_spacing(dm=250)

grid.import_topo()    
grid.mesh_grid()

# Set the boundaries
grid.set_boundary_points(dn.grid.mask.Edges(edges=["W", "N", "S"], step=10))

# Model
model = dn.modelrun.NORA3(grid, start_time="2019-01-01T01:00", end_time="2019-01-05T01:00")

# Boundary Spectra
model.import_spectra(dn.read.spectra.metno.NORA3(),dn.pick.NearestGridPoint())

# Import Forcing
model.import_wind()
model.import_ice()

exporter = dn.export.SWAN(model)
exporter.export_grid()
exporter.export_spectra()
exporter.export_wind()
exporter.export_ice()
executer = dn.executer.SWAN(model)
executer.write_input_file()
executer.run_model(nproc=8)

[bjorkqvi]

When using the branch v2_prerelase the fimex files seems ok. Also the ascii files that are provided to SWAN seems to have no issues with the magnitude and direction.

We are still trying to figure out if something is happening in the SWAN interpolation or what is going on.