measurement_vector

Workspace.measurement_vector: Vector

The measurment vector for, e.g., a sensor.

This must often be the same size as measurement_sensor.

The notation in ARTS, for the purpose of OEM(), is that

\[\vec{y} = \mathbf{F}\left(\vec{x}\right) + \vec{y}_\epsilon\left(\vec{x}\right) + \epsilon\]

where \(\mathbf{F}\) is the forward model function of the physics of the simulation space, \(\vec{x}\) is the model_state_vector, \(\vec{y}_\epsilon\) is the measurement_vector_error, and \(\epsilon\) are any additional errors, such as random noise.

Throughout ARTS, measurement_vector have different contextual meanings. These are:

1. \(\vec{y}\) - i.e., measured data.

2. \(\vec{y} - \epsilon\) - e.g., the best fit to measured data, measurement_vector_fitted.

3. \(\mathbf{F}\left(\vec{x}\right)\) - i.e., the physical model of the measurement.

Input to workspace methods

OEM() measurement_vector_fittedFromMeasurement()

Modified by workspace method

measurement_vectorConditionalAddError()

Output from workspace methods

measurement_vectorFromOperatorPath() measurement_vectorFromSensor()

Related workspace variables

measurement_averaging_kernel measurement_gain_matrix measurement_inversion_agenda measurement_jacobian measurement_jacobian_error

measurement_sensor measurement_vector_error measurement_vector_error_covariance_matrix measurement_vector_fitted