Wrong coordinates and broken values in SARAH cutouts

[euronion]

The coordinates in SARAH cutouts are messed up, i.e. mirrored, see the image below.

> albedo
<xarray.DataArray 'albedo' (y: 160, x: 270)>
array([[0.04827 , 0.048171, 0.048043, ..., 0.103032, 0.104033, 0.104248],
       [0.047978, 0.047966, 0.047898, ..., 0.104392, 0.109037, 0.11042 ],
       [0.047583, 0.04764 , 0.047696, ..., 0.105399, 0.109951, 0.111676],
       ...,
       [0.03567 , 0.03552 , 0.035488, ..., 0.353771, 0.35935 , 0.36295 ],
       [0.035606, 0.035525, 0.035507, ..., 0.356749, 0.357823, 0.359936],
       [0.035569, 0.0355  , 0.035465, ..., 0.360018, 0.361162, 0.359274]],
      dtype=float32)
Coordinates:
    time     datetime64[ns] 2013-06-01T12:00:00
  * x        (x) float64 -12.0 -11.8 -11.6 -11.4 -11.2 ... 41.2 41.4 41.6 41.8
  * y        (y) float64 33.0 33.2 33.4 33.6 33.8 ... 64.0 64.2 64.4 64.6 64.8
    lon      (x) float64 -12.0 -11.8 -11.6 -11.4 -11.2 ... 41.2 41.4 41.6 41.8
    lat      (y) float64 33.0 33.2 33.4 33.6 33.8 ... 64.0 64.2 64.4 64.6 64.8
Attributes:
    units:      (0 - 1)
    long_name:  Albedo

The cutout

> cutout
<Cutout europe-2013-sarah x=-12.00-41.80 y=33.00-64.80 time=2013-01-01-2013-12-31 prepared_features=['temperature', 'influx'] is_view=False>

with the messed up coordinates

> cutout.data
<xarray.Dataset>
Dimensions:         (time: 8760, x: 270, y: 160)
Coordinates:
  * time            (time) datetime64[ns] 2013-01-01 ... 2013-12-31T23:00:00
  * x               (x) float64 -12.0 -11.8 -11.6 -11.4 ... 41.2 41.4 41.6 41.8
  * y               (y) float64 33.0 33.2 33.4 33.6 33.8 ... 64.2 64.4 64.6 64.8
    lon             (x) float64 -12.0 -11.8 -11.6 -11.4 ... 41.2 41.4 41.6 41.8
    lat             (y) float64 33.0 33.2 33.4 33.6 33.8 ... 64.2 64.4 64.6 64.8
Data variables:
    temperature     (time, y, x) float32 ...
    influx_toa      (time, y, x) float32 ...
    albedo          (time, y, x) float32 ...
    influx_direct   (time, y, x) float32 ...
    influx_diffuse  (time, y, x) float32 ...
Attributes:
    module:               sarah
    prepared_features:    ['temperature', 'influx']
    creation_parameters:  {'resolution': 0.2, 'x': slice(-12.0, 42.0, None), ...

Maybe an Australian PoV?

[euronion]

Addendum:
albedo is only choosen as an example. temperature and influx_toa are affected similarly. influx_diffuse and influx_direct have a different problem: All values are simply nan. (see below)

[euronion]

Update:
I recreated the cutout again to exclude problems during creation or from transfer / faulty input files.
Problems regarding coordinates persist.
The problem with influx_diffuse and influx_direct is a bit different from what I described earlier:
The values for both variables for January until including November are zero or nan. Only December has usable data in these variables.

[FabianHofmann]

The problem lays not directly in the sarah.py script but in the era5 script. The albedo and other influx variables are retrieved from era5 even though when the module is set to sarah (which is kind of weird, but I think the intention is to have a complete dataset...)

ERA5 datasets have reversed y axis. In the normal case this is fixed in the standard prepare_cutout function when the merging the empty cutout (coordinates only) with the feature datasets. In the sarah.py script the axis is not double-checked and overwritten here.

An easy fix is to make a double check and correction in the era5.py script, by adding

    ds = maybe_swap_spatial_dims(ds)

into the _rename_and_clean_coords function. This ensures the correct alignment of the datasets.
maybe_swap_spatial_dims is defined in gis.py