mesmer.core.computation.gaspari_cohn#

mesmer.core.computation.gaspari_cohn(r)#

smooth, exponentially decaying Gaspari-Cohn correlation function

Parameters:

r (xr.DataArray, np.ndarray) – Values for which to calculate the value of the Gaspari-Cohn correlation function (e.g. normalised geographical distances)

Returns:

out (xr.DataArray, , np.ndarray) – Gaspari-Cohn correlation function

Notes

  • Smooth exponentially decaying correlation function which mimics a Gaussian distribution but vanishes at r = 2, i.e., 2 x the localisation radius (L)

  • based on Gaspari-Cohn 1999 [1] (as taken from Carrassi et al., 2018 [2])

  • r = d / L, with d = geographical distance in km, L = localisation radius in km