sp_scalewise: Extract scale-wise spatial processes

In spCF: Coarse-to-Fine Spatial and Spatio-Temporal Modeling

Extract scale-wise spatial processes

Description

Evaluate mean and standard deviation of the (multiscale) spatial process for bandwidth values within a pre-specified range. For a spatio-temporal fit from cf_dglm, the process can additionally be averaged over a user-specified time range, returning the temporally averaged spatial process at each (sample / prediction) location.

Usage

sp_scalewise(mod, bw_range = c(0, Inf), time_range = c(-Inf, Inf))

Arguments

mod	Output object from the cf_lm, cf_glm or cf_dglm function.
bw_range	Range of bandwidth values of the synthesized spatial processes, treated as the half-open interval [min, max). For example, bw_range = c(10, 20) synthesizes scales with bandwidth b such that 10 <= b < 20. The half-open convention lets contiguous ranges (e.g. c(0, 10) and c(10, Inf)) partition the scales without double-counting a scale whose bandwidth equals the shared endpoint. The default c(0, Inf) synthesizes all scales.
time_range	Range of time points over which the spatio-temporal process is averaged. Only used when mod is a cf_dglm fit (which carries a time index for every row of Z). For example, time_range = c(5, 10) averages the process over time points 5 to 10 at each location. The default c(-Inf, Inf) averages over all time points. Ignored (with a warning if set to a non-default value) for purely spatial fits.

Value

A list with the following elements:

pred

Means and standard deviations of the spatial process at the sample sites. For a cf_dglm fit, one row per (unique) location with the temporally averaged process, together with its coordinates and the number of averaged time points.

pred0

The same at the prediction sites. NULL when mod was fitted without prediction sites, and also (with a warning) when no prediction site falls inside time_range.

Author(s)

Daisuke Murakami

See Also

cf_lm, cf_glm, cf_dglm

spCF documentation built on Oct. 5, 2026, 5:07 p.m.