RetrievalAddErrorPolyFit

Workspace.RetrievalAddErrorPolyFit(self, jacobian_targets: pyarts.arts.JacobianTargets | None = None, covariance_matrix_diagonal_blocks: pyarts.arts.JacobianTargetsDiagonalCovarianceMatrixMap | None = None, measurement_sensor: pyarts.arts.ArrayOfSensorObsel | None = None, t: pyarts.arts.Vector | None = None, sensor_elem: pyarts.arts.Index | None = None, polyorder: pyarts.arts.Index | None = None, matrix: pyarts.arts.BlockMatrix | None = None, inverse: pyarts.arts.BlockMatrix | None = None) → None

Set a measurement error to polynomial fit.

This is a generic error that is simply added to measurement_vector as if

\[y = y_0 + \epsilon(p_0, p_1, ..., p_n),\]

where y represents measurement_vector and y0 is the measurement

Order 0 means constant: y := y0 + a Order 1 means linear: y := y0 + a + b * t and so on. The derivatives that are added to the model_state_vector are those with regards to a, b, etc..

Note

The rule for the sensor_elem GIN is a bit complex. Generally, methods such as measurement_sensorAddSimple() will simply add a single unique frequency grid to all the different SensorObsel that they add to the measurement_sensor. The GIN sensor_elem is 0 for the first unique frequency grid, 1 for the second, and so on. See ArrayOfSensorObsel member methods in python for help identifying and manipulating how many unique frequency grids are available in measurement_sensor.

This method wraps jacobian_targetsAddErrorPolyFit() together with adding the covariance matrices, to the covariance_matrix_diagonal_blocks, which are required to perform OEM().

The input covariance matrices must fit the size of the later computed model state represented by the jacobian_targets. The covariance matrix inverse

Author(s): Richard Larsson

Parameters:

  • jacobian_targets (JacobianTargets, optional) – A list of targets for the Jacobian Matrix calculations. See jacobian_targets, defaults to self.jacobian_targets [INOUT]

  • covariance_matrix_diagonal_blocks (JacobianTargetsDiagonalCovarianceMatrixMap, optional) – A helper map for setting the covariance matrix. See covariance_matrix_diagonal_blocks, defaults to self.covariance_matrix_diagonal_blocks [INOUT]

  • measurement_sensor (ArrayOfSensorObsel, optional) – A list of sensor elements. See measurement_sensor, defaults to self.measurement_sensor [IN]

  • t (Vector) – The grid of the perturbation. [IN]

  • sensor_elem (Index) – The sensor element whose frequency grid to use. [IN]

  • polyorder (Index, optional) – The order of the polynomial fit. Defaults to 0 [IN]

  • matrix (BlockMatrix) – The covariance diagonal block matrix. [IN]

  • inverse (BlockMatrix, optional) – The inverse covariance diagonal block matrix. Defaults to pyarts.arts.BlockMatrix() [IN]