SimCE: Circulant embedding simulation on a regular spatial grid

In GeoModels: Procedures for Gaussian and Non Gaussian Geostatistical (Large) Data Analysis

Circulant embedding simulation on a regular spatial grid

Description

Internal subroutine used by GeoSimapprox to simulate a Gaussian random field on an exact, regular two-dimensional grid by circulant embedding.

Usage

SimCE(M, N, x, y, z = NULL, corrmodel, param,
      mean.val = 0, max.ext = 1L)

Arguments

M	Integer; number of coordinates in x.
N	Integer; number of coordinates in y.
x	A finite, strictly increasing, equally spaced numeric vector of length M.
y	A finite, strictly increasing, equally spaced numeric vector of length N.
z	Must be NULL. Three-dimensional CE grids are not implemented.
corrmodel	String; the name of a spatial correlation model.
param	A list of correlation parameters. sill is the marginal variance and nugget is the nugget proportion.
mean.val	A finite scalar or a vector with one value per grid point.
max.ext	Positive integer; maximum number of successive doubling attempts for the embedding.

Details

The simulated coordinates are exactly expand.grid(x, y). The method uses Euclidean distance. If the embedding still has substantial negative eigenvalues at the maximum extension, they are truncated to zero and a warning is issued.

Value

A list containing X, the simulated values, grid.points, the exact grid coordinates, k0, the selected extension attempt, approx.flag, and the minimum embedding eigenvalue.

Author(s)

Moreno Bevilacqua, [e-mail], Víctor Morales Oñate, [e-mail], Christian Caamaño-Carrillo, [e-mail]

See Also

GeoSimapprox

GeoModels documentation built on Sept. 23, 2026, 5:07 p.m.