Handle exceptionally rare linear algebra error

[Zeitsperre]

Pull Request Checklist:

• This PR addresses an already opened issue (for bug fixes / features)
  • This PR fixes #xyz
• Tests for the changes have been added (for bug fixes / features)
  • (If applicable) Documentation has been added / updated (for bug fixes / features)
• CHANGELOG.rst has been updated (with summary of main changes)
  • Link to issue (:issue:number) and pull request (:pull:number) has been added

What kind of change does this PR introduce?

• Adds handling for a rare but mathematically possible error that has happened at least once in xclim

Does this PR introduce a breaking change?

No.

Other information:

  _______________________ test_loess_smoothing_nan[False] ________________________
  [gw2] darwin -- Python 3.11.9 /Users/runner/work/xclim/xclim/.tox/py311-coverage-extras/bin/python
  
  use_dask = False
  
      @pytest.mark.slow
      @pytest.mark.parametrize("use_dask", [True, False])
      def test_loess_smoothing_nan(use_dask):
          # create data with one axis full of nan
          data = np.random.randn(2, 2, 10)
          data[0, 0] = [np.nan] * 10
          da = xr.DataArray(
              data,
              dims=["scenario", "model", "time"],
              coords={"time": pd.date_range("2000-01-01", periods=10, freq="YS")},
          ).chunk({"time": -1})
      
          out = loess_smoothing(da)
      
          assert out.dims == da.dims
          # check that the output is all nan on the axis with nan in the input
  >       assert np.isnan(out.values[0, 0]).all()
  
  tests/test_sdba/test_loess.py:87: 
  _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
  .tox/py311-coverage-extras/lib/python3.11/site-packages/xarray/core/dataarray.py:814: in values
      return self.variable.values
  .tox/py311-coverage-extras/lib/python3.11/site-packages/xarray/core/variable.py:566: in values
      return _as_array_or_item(self._data)
  .tox/py311-coverage-extras/lib/python3.11/site-packages/xarray/core/variable.py:363: in _as_array_or_item
      data = np.asarray(data)
  .tox/py311-coverage-extras/lib/python3.11/site-packages/dask/array/core.py:1709: in __array__
      x = self.compute()
  .tox/py311-coverage-extras/lib/python3.11/site-packages/dask/base.py:372: in compute
      (result,) = compute(self, traverse=False, **kwargs)
  .tox/py311-coverage-extras/lib/python3.11/site-packages/dask/base.py:660: in compute
      results = schedule(dsk, keys, **kwargs)
  .tox/py311-coverage-extras/lib/python3.11/site-packages/numpy/lib/_function_base_impl.py:2397: in __call__
      return self._call_as_normal(*args, **kwargs)
  .tox/py311-coverage-extras/lib/python3.11/site-packages/numpy/lib/_function_base_impl.py:2390: in _call_as_normal
      return self._vectorize_call(func=func, args=vargs)
  .tox/py311-coverage-extras/lib/python3.11/site-packages/numpy/lib/_function_base_impl.py:2471: in _vectorize_call
      res = self._vectorize_call_with_signature(func, args)
  .tox/py311-coverage-extras/lib/python3.11/site-packages/numpy/lib/_function_base_impl.py:2511: in _vectorize_call_with_signature
      results = func(*(arg[index] for arg in args))
  _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
  
  >   raise np.linalg.LinAlgError(
          "Matrix is singular to machine precision.")
  E   numpy.linalg.LinAlgError: Matrix is singular to machine precision.
  
  .tox/py311-coverage-extras/lib/python3.11/site-packages/numba/np/linalg.py:899: LinAlgError

[SarahG-579462]

lol. LGTM