Skip to contents

\(\sqrt{sill/L}\sum_l z_l(\langle x, u_l\rangle)\) over L bands with spectral line processes \(z_l(t) = \sqrt{2/M}\sum_m \cos(r_{lm} t + \phi_{lm})\), frequencies from the radial spectral law of the covariance (Matheron 1973; Mantoglou and Wilson 1982; Emery and Lantuejoul 2006). 2-D: directions \(\pi(l + U)/L\), models exponential, gaussian, matern, geometric anisotropy; 3-D: Fibonacci directions under a uniform random rotation (Shoemake 1992 quaternion), gaussian and exponential. Philox streams: 0 offset, then 3l + 1 radii, 3l + 2 phases, 3l + 3 extra normals; identical to the Python arm.

Usage

TurningBands(
  coords,
  cov_model = "exponential",
  sill = 1,
  range_ = 1,
  nu = 0.5,
  n_bands = 64L,
  n_waves = 50L,
  angle = 0,
  ratio = 1,
  seed = 1L
)

Arguments

coords

Two- or three-column matrix of locations.

cov_model

"exponential", "gaussian" or "matern" (2-D).

sill

Variance.

range_

Range.

nu

Matern smoothness.

n_bands

Number of bands.

n_waves

Waves per band.

angle

Anisotropy angle (2-D).

ratio

Anisotropy ratio (2-D).

seed

Philox seed.

Value

List with field and directions.

References

Matheron, G. (1973). The intrinsic random functions and their applications. Advances in Applied Probability 5, 439-468.

Mantoglou, A. and Wilson, J. L. (1982). The turning bands method for simulation of random fields using line generation by a spectral method. Water Resources Research 18, 1379-1394.

Emery, X. and Lantuejoul, C. (2006). TBSIM: a computer program for conditional simulation of three-dimensional Gaussian random fields via the turning bands method. Computers and Geosciences 32, 1615-1628.

Shoemake, K. (1992). Uniform random rotations. Graphics Gems III, 124-132.

Examples

TurningBands(rbind(c(0, 0), c(1, .5)), "gaussian", n_bands = 4, n_waves = 3, seed = 2)$field
#> [1] 0.8813583 0.4004213