atmos_flux_inversion.covariances module

Support classes for covariances.

See also

atmos_flux_inversion.correlations

class atmos_flux_inversion.covariances.CorrelationStandardDeviation(*args, **kwargs)

Bases: atmos_flux_inversion.linalg_interface.ProductLinearOperator, atmos_flux_inversion.linalg.SelfAdjointLinearOperator

Represent correlation-std product.

sqrt()

Find S such that S.T @ S == self.

atmos_flux_inversion.covariances.OBSERVATION_INTERVAL = array(1, dtype='timedelta64[h]')

The time between individual observations at one site.

More precisely, these are the units used with the correlation function used in observation_covariance_matrix().

atmos_flux_inversion.covariances.observation_covariance_matrix(variance_series, correlation_function)

Create the observation covariance matrix given parammeters.

Assumes the observations are far enough apart to be uncorrelated.

Parameters

• variance_series (pandas.Series[n_observations]) –

• correlation_function (DistanceCorrelationFunction) –

Returns

Return type

np.ndarray