easyclimate.filter.redfit¶

Red-noise spectra estimating

Functions¶

`calc_redfit`(data[, timearray, nsim, mctest, rhopre, ...])	Estimating red-noise spectra directly from unevenly spaced paleoclimatic time series.
`calc_redfit_cross`(data_x, data_y[, timearray_x, ...])	Estimating red-noise spectra directly from unevenly spaced paleoclimatic time series.

Module Contents¶

easyclimate.filter.redfit.calc_redfit(data: xarray.DataArray, timearray: numpy.array = None, nsim: int = 1000, mctest: bool = False, rhopre: float = -99.0, ofac: float = 1.0, hifac: float = 1.0, n50: int = 1, iwin: Literal['rectangular', 'welch', 'hanning', 'triangular', 'blackmanharris'] = 'rectangular')¶

Estimating red-noise spectra directly from unevenly spaced paleoclimatic time series.

Parameters¶

data: xarray.DataArray: Input time series data
timearray: numpy.array: Time series data array
nsim: int: Number of Monte-Carlo simulations (1000-2000 should be o.k. in most cases)
mctest: bool: Toggle calculation of false-alarm levels based on Monte-Carlo simulation, if set to True : perform Monte-Carlo test, if set to False : skip Monte-Carlo test (default).
rhopre: float: Prescibed value for \(\rho\); unused if < 0 (default = -99.0)
ofac: float: Oversampling factor for Lomb-Scargle Fourier transform (typical values: 2.0-4.0)
hifac: float: Max. frequency to analyze is set to hifac * <fNyq> (default = 1.0)
n50: int: Number of WOSA segments (with 50 % overlap)
iwin: {“rectangular”, “welch”, “hanning”, “triangular”, “blackmanharris”}: Window-type identifier used to suppress sidelobes in spectral analysis: ({“rectangular”, “welch”, “hanning”, “triangular”, “blackmanharris”}, optional)

Caution

Parameters ofac, hifac, n50 and window type are identical to the SPECTRUM program (see Schulz and Stattegger, 1997 for further details). Except mctest, hifac and rhopre all parameters must be specified.

Returns¶

The red-noise spectra (xarray.Dataset).

Parameters¶

data_x:xarray.DataArray: First input time series data
data_y: xarray.DataArray: Second input time series data
timearray_x: numpy.array: First time series data array
timearray_y: numpy.array: Second time series data array
x_sign: bool: Change the sign of the first time series: if True: The sign of the data is changed if False: The sign of the data is not changed (default)
y_sign: bool: Change the sign of the second time series: if True: The sign of the data is changed if False: The sign of the data is not changed (default)
nsim: int: Number of Monte Carlo simulations (1000-2000 is recommended)
mctest: bool: Estimate the significance of auto and coherency spectrum with Monte Carlo simulations if True: perform Monte Carlo simulations if False: do not perform Monte Carlo simulations
mctest_phi: bool: Estimate Monte Carlo confidence interval for the phase spectrum if True: perform Monte Carlo simulations (mctest needs to be true as well) if False: do not perform Monte Carlo simulations
rhopre_1: float: Prescribed value for \(\rho\) for the first time series, not used if \(\rho < 0\) (default = -999.0).
rhopre_2: float: Prescribed value for \(\rho\) for the second time series, not used if \(\rho< 0\) (default = -999.0).
ofac: float: Oversampling factor for Lomb-Scargle Fourier transform (typical values: 2.0-4.0).
hifac: float: Max. frequency to analyze is set to hifac * <fNyq> (default = 1.0).
n50: int: Number of WOSA segments (with 50 % overlap)
alpha: float: Significance level (Note: only 0.01, 0.05 [default], or 0.1 are allowed).
iwin: {“rectangular”, “welch”, “hanning”, “triangular”, “blackmanharris”}: Window-type identifier used to suppress sidelobes in spectral analysis: ({“rectangular”, “welch”, “hanning”, “triangular”, “blackmanharris”}, optional).

Caution

Parameters ofac, hifac, n50 and window type are identical to the SPECTRUM program (see Schulz and Stattegger, 1997 for further details). Except mctest, hifac, rhopre(1) and rhopre(2) all parameters must be specified.