Xarray masking at certain location

Question

I have a xarray data set and want to transform the values of my variable in a specific region within the dataset to Nan values. Using dataset.where I only get the specific region as the output but I want to keep the values at the locations around the mask in the array. So basically I want to keep my dataset but replace the values of my variable at certain coordinates with NaN.

[The dataset is called Lae21 and the lon/lat_mask are sliced coordinates from the dataset]

`print(Lae21)
mask = Lae21.where((Lae21.x!=lon_mask) & (Lae21.y!=lat_mask)) 
print(mask)`

    <xarray.Dataset>
Dimensions:  (y: 1614, x: 7682)
Coordinates:
  * y        (y) float64 5.261e+06 5.261e+06 5.261e+06 ... 5.256e+06 5.256e+06
  * x        (x) float64 4.389e+05 4.389e+05 4.389e+05 ... 4.619e+05 4.619e+05
Data variables:
    NDVI     (y, x) float32 ...
Attributes:
    transform:            (3.0, 0.0, 438857.72, 0.0, -3.0, 5260962.7)
    crs:                  +init=epsg:32632
    res:                  (3.0, 3.0)
    is_tiled:             0
    nodatavals:           (nan,)
    scales:               (1.0,)
    offsets:              (0.0,)
    AREA_OR_POINT:        Area
    TIFFTAG_XRESOLUTION:  1
    TIFFTAG_YRESOLUTION:  1
<xarray.DataArray 'NDVI' (y: 3, x: 3)>
array([[nan, nan, nan],
       [nan, nan, nan],
       [nan, nan, nan]], dtype=float32)
Coordinates:
  * y        (y) float64 5.259e+06 5.259e+06 5.259e+06
  * x        (x) float64 4.521e+05 4.521e+05 4.521e+05

So the resulting array should be all the values from the original dataset Lae21 and NaN values at the specific location of the mask. (Looking at examples of the .where function I thought that should be the output...).

Thanks a lot for your answers!

Answer 1

I would use .loc method for this:

Lae21.NDVI.loc[{"x": lat_mask, "y": lon_mask}] = np.NaN

See this specific chapter of xarray indexing guide for more details.