pylawr.transform.spatial#

pylawr.transform.spatial.BeamExpansion()

Filter to apply/remove the beam expansion effect to/from a reflectivity field.

pylawr.transform.spatial.Interpolator([...])

This class can be used to fill holes within a radar field with given remapping algorithm.

pylawr.transform.spatial.BeamExpansion#

class pylawr.transform.spatial.BeamExpansion[source]#

Bases: Transformer

Filter to apply/remove the beam expansion effect to/from a reflectivity field. See radartutorial.eu for reference.

This filter basically just does

\[input \cdot R ^ 2\]

in the forward direction and

\[\frac{input}{R ^ 2}\]

in the inverse direction.

_abc_impl = <_abc._abc_data object>#
static _determine_inv(array)[source]#

Method to determine the transform direction if no inverse is given.

Parameters

array (xarray.DataArray) – The linear reflectivity array to operate on.

Returns

inverse – The direction of the transformation with apply if False or removed if True.

Return type

bool

transform(array, grid=None, inverse=None, *args, **kwargs)[source]#

Apply (or remove if inverse=True) the beam expansion effect from a given array.

Tag the array with TAG_BEAM_EXPANSION_CORR or TAG_BEAM_EXPANSION_UNCORR accordingly.

Parameters

  • array (xarray.DataArray) – The linear reflectivity array to operate on.

  • grid (child of pylawr.grid.base.BaseGrid or None) – This grid is used to determine the distance to the radar. If no grid is given it will be determined from the array.

  • inverse (bool or None, optional) – Invert the transformation? Defaults to None, which means to try to determine the direction automatically by checking the array’s tags. If that fails, inverse defaults to False.

Returns

transformed_array – The array with the beam expansion effect applied/removed

Return type

xarray.DataArray

Raises

  • AttributeError – An AttributeError is raised if no grid was specified and no grid is set for the array.

  • TypeError – A TypeError is raised if the given grid is not a valid grid.

  • ValueError – A ValueError is raised if the given grid does not have the same shape as the given array.

pylawr.transform.spatial.Interpolator#

class pylawr.transform.spatial.Interpolator(threshold=0.75, algorithm=None, polar=True, cov_thres=4, zero_field=(11, 11), zero_thres=0.34)[source]#

Bases: Transformer

This class can be used to fill holes within a radar field with given remapping algorithm. For the interpolation temporary UnstructuredGrid is used.

Parameters

  • threshold (float) – If more values are nan than field size * threshold, then no interpolation is possible. Default is 0.75.

  • algorithm (child of pylawr.remap.base.BaseRemap or None) – This initialized algorithm is used for hole filling. If this is None, NearestNeighbor with one nearest neighbor is used as algorithm. Default is None.

  • polar (boolean, optional) – If the given reflectivity is in polar coordinates or in other coordinates. If polar coordinates are used, the padding for the masks is calculated with polar padding, else a reflect padding is used.

  • cov_thres (float, optional) – If this interpolation covariance (uncertainty) above this threshold ( default \(= 4~ ext{dbz}^{2}), then the value is not interpolated. If given interpolation algorithm has no uncertainty estimation, this step is skipped. To use no thresholding this values can be set to :py:class:\).

  • zero_field (tuple(int)) – The size of the local receptive field to search for valid values, which have rain. If the number of rain values is below a given threshold, the missing values are set to zero. Default is (11, 11).

  • zero_thres (float, optional) – If the number of rain values within a local receptive field is below this threshold, the missing values are set to zero. Default is 0.34.

Warning

The algorithm is refitted every time such that old fits are overwritten.

_abc_impl = <_abc._abc_data object>#
_get_source_mask(reflectivity)[source]#

Get a mask for all source points based on given reflectivity. Two different conditions are necessary such that a value is declared as source value:

  • the value is not nan

  • the value has no surrounding nan value or is below 5 dBZ

Parameters

reflectivity (xarray.DataArray) – This reflectivity array is used to create the source mask.

Returns

source_mask – All source values are True, all other are False.

Return type

numpy.ndarray (bool)

static _get_target_mask(reflectivity)[source]#

Mask all values, which should be interpolated. Currently, only numpy.nan are set to True, while all other are set to False.

Parameters

reflectivity (xarray.DataArray) – This reflectivity array is used to create the target mask. All nan- values are used as target.

Returns

target_mask – The inferred boolean target mask with the same shape as the given reflectivity.

Return type

numpy.ndarray (bool)

_get_valid_mask(reflectivity)[source]#

Get a mask for all valid values. A value is declared as valid value, if the number of rain values (> 5 dBZ) within a local receptive field is larger than a set threshold.

Parameters

reflectivity (xarray.DataArray) – This reflectivity array is used to create the valid mask.

Returns

valid_mask – All valid values are True, all other are False.

Return type

numpy.ndarray (bool)

_prefill_noninterp_vals(reflectivity)[source]#

All missing and not valid values within a given array are set to -32.5 dBZ.

Parameters

reflectivity (xarray.DataArray) – This reflectivity array is used to fill non valid values

Returns

filled_values – In this array all missing and not valid values are replaced with -32.5 dBZ.

Return type

xarray.DataArray

_repl_uncertain_interpolation(refl_array, repl_value=nan)[source]#

This method replaces values in a given interpolation array, where the interpolation covariance are above a set threshold.

property algorithm#
transform(refl_array, grid=None, *args, **kwargs)[source]#

Interpolate holes within given array.

Parameters

  • refl_array (xarray.DataArray) – This array is used as input array. This array is handled as single reflectivity array such that only one time step at a time can be processed.

  • grid (child of pylawr.grid.base.BaseGrid or None) – This grid is used for the interpolation. If this grid is None, the grid from given array is used. Default is None.

Returns

filled_array – In this interpolated array all nan-values are replaced by their interpolated value. The interpolated array will be returned as logarithmic reflectivity in dBZ.

Return type

xarray.DataArray

Warning

For interpolation of missing values a grid is needed.