double counting NEMO nesting for velocity?

[mazpml]

when I did this (as in cookbok):

forcing_variables = [
('u', 'element'),
('v', 'element'),
('temp', 'node'),
('salinity', 'node'),
]

print('Adding 3D forcing variables:')
for fvcom_var, position in forcing_variables:
nest_manager.add_forcing_data(
nemo_interpolator,
fvcom_var,
horizontal_position=position,
)
data = nest_manager.get_forcing_data(fvcom_var)
print(f' {fvcom_var:8s} at {position:7s}: shape={data.shape}')
print('3D NEMO forcing data added successfully')

it gave logfile:
Adding 3D forcing variables:
Interpolating and rotating NEMO vozocrtx/vomecrty.
Interpolating NEMO vozocrtx to FVCOM grid for date: 2011-01-01 12:00:00.
Interpolating NEMO vozocrtx to FVCOM grid for date: 2011-01-02 12:00:00.
Interpolating NEMO vomecrty to FVCOM grid for date: 2011-01-01 12:00:00.
Interpolating NEMO vomecrty to FVCOM grid for date: 2011-01-02 12:00:00.
u at element: shape=(2, 24, 1013)
Interpolating and rotating NEMO vozocrtx/vomecrty.
Interpolating NEMO vozocrtx to FVCOM grid for date: 2011-01-01 12:00:00.
Interpolating NEMO vozocrtx to FVCOM grid for date: 2011-01-02 12:00:00.
Interpolating NEMO vomecrty to FVCOM grid for date: 2011-01-01 12:00:00.
Interpolating NEMO vomecrty to FVCOM grid for date: 2011-01-02 12:00:00.
v at element: shape=(2, 24, 1013)
Interpolating NEMO votemper to FVCOM grid.
Interpolating NEMO votemper to FVCOM grid for date: 2011-01-01 12:00:00.
Interpolating NEMO votemper to FVCOM grid for date: 2011-01-02 12:00:00.
temp at node : shape=(2, 24, 757)
Interpolating NEMO vosaline to FVCOM grid.
Interpolating NEMO vosaline to FVCOM grid for date: 2011-01-01 12:00:00.
Interpolating NEMO vosaline to FVCOM grid for date: 2011-01-02 12:00:00.
salinity at node : shape=(2, 24, 757)
3D NEMO forcing data added successfully

It seems it interpolated both vozocrtx and vomecrty when writing u, also when writing v?

[mazpml]

u and v are actually required to be called for rotation. But there is some suggestion it can be more efficient