plot_stlgcp: Plot spatio-temporal log-Gaussian Cox process (LGCP)...

In SepTest: Tests for First-Order Separability in Spatio-Temporal Point Processes

Plot spatio-temporal log-Gaussian Cox process (LGCP) realizations as time-sliced maps

Description

Produces a sequence of spatial raster maps displaying the evolution of a simulated or fitted spatio-temporal log-Gaussian Cox process (LGCP) over time. Each map shows the latent Gaussian random field (intensity surface) at a given time slice, with the corresponding observed point events overlaid. This visualization helps interpret temporal evolution and spatial clustering patterns in simulated or fitted LGCP models.

Usage

plot_stlgcp(data)

Arguments

data

A list containing components from a spatio-temporal LGCP model or simulation output: RF A list describing the latent Gaussian random field, with elements xcoord, ycoord, tcoord, and Z, where Z is a 3D array of intensity (or log-intensity) surfaces over space and time. st.lgcp A data frame with columns x, y, and t giving the spatial and temporal coordinates of observed events.

Details

The function plots up to 10 evenly spaced time slices from the latent intensity field and overlays the corresponding point events accumulated up to each time point. Each panel represents a spatial realization at a fixed time t_k, providing a visual summary of the dynamic spatio-temporal structure of the LGCP.

This approach follows the visualization principles used in Ghorbani et al. (2021, 2025), where LGCPs are employed to assess the behavior of separability diagnostics and kernel-based tests in complex, non-separable point process settings. The raster intensity surfaces illustrate latent heterogeneity, while the overlaid points display observed event clustering relative to the underlying field.

Time slices are selected at evenly spaced quantiles of the temporal domain, and each plot includes:

• A raster map of the latent intensity surface at time t_k;

• White contour lines showing equal-intensity regions;

• Overlaid black points representing events observed up to t_k.

The resulting plots are arranged into a single grid layout using the patchwork package for ease of comparison.

Value

A combined ggplot object displaying up to ten spatial raster maps arranged in a grid layout (by default, two rows and up to five columns). Each panel corresponds to one time slice. The function returns the combined plot object.

Author(s)

Nafiseh Vafaei [e-mail]

References

Ghorbani M., Vafaei N., Dvořák J., Myllymäki M. (2021). Testing the first-order separability hypothesis for spatio-temporal point patterns. Computational Statistics & Data Analysis, 161, 107245.

See Also

Gauss.st.F for simulating spatio-temporal Gaussian random fields; get.lambda.function and rstpoispp for generating intensity-based spatio-temporal point processes.

Examples

# Example: visualize a spatio-temporal LGCP simulation
out <- rstLGCPP(xlim = c(0,1),
                ylim = c(0,1),
                tlim = c(0,1),
                grid = c(15,15,10))
plot_stlgcp(data = out)

SepTest documentation built on Feb. 3, 2026, 5:07 p.m.