xoa.krig.empirical_variogram

xoa.krig.empirical_variogram(da: DataArray, nbin=30, nbin0=10, nmax=1500, dist_units='m', distmax=None, errfunc=None)[source]

Compute the semi-variogram from data

Parameters:

  • da (xarray.dataArray) – Data array with lon and lat coordinates.

  • nmax (optional) – Above this number, size of the sample is reduced by a crude undersampling.

  • binned (optional) – If set to a number, data are arranged in bins to estimate variogram. If set to None, data are arranged in bins if the number of pairs of points is greater than nbindef*nbmin.

  • nbindef (optional) – Default number of bins (not used if binned is a number).

  • nbin0 (optional) – If set to a number > 1, the first bin is split into nbin0 sub-bins. If set to None, it is evaluated with min(bins[1]/nbmin, nbin).

  • nbmin (optional) – Minimal number of points in a bin.

  • dist_units (str, int, xoa.geo.distance_units) – Distance units as one of: 0|"m"|"meters"|"meter", 1|"km"|"kilometers"|"kilometer"

  • distmax (optional) – Max distance to consider.

  • errfunc (optional) – Callable function to compute “errors” like square root difference between to z values. It take two arguments and defaults to \((z1-z0)^2/2\).

Returns:

xarray.DataArray – Values as 1D array with name “semivariogram” and with “dist” as distance coordinate in km