Simulate a multi-season integrated occupancy data set

Draws from the dyn_int_occu() model: a dynamic (multi-season) occupancy process (psi1, colonization gamma, extinction eps) observed by S detection sources with per-source detection probability.

Usage

simulate_dyn_int_occu(
  N = 200,
  T_seasons = 4,
  S = 2,
  J = 3,
  psi1 = 0.5,
  gamma = 0.3,
  eps = 0.2,
  p = c(0.4, 0.6),
  field = NULL,
  trend = NULL,
  source_seasons = NULL,
  seed = NULL
)

Arguments

N

Number of sites (default 200).

T_seasons

Number of seasons (default 4).

S

Number of detection sources (default 2).

J

Visits per source (scalar or length-S; default 3).

psi1, gamma, eps

Season-1 occupancy, colonization, extinction (defaults 0.5 / 0.3 / 0.2).

p

Per-source detection probabilities (length-S; default 0.4 / 0.6).

field

Optional per-site shared areal field (length N) added to the first-season occupancy logit – the shared field of the multi-season integrated spatial model (stIntPGOcc). Default NULL (no field).

trend

Optional per-site varying-coefficient (SVC) areal field (length N); with a covariate w ~ N(0,1) stored in data, the first-season logit gains w * trend on top of field – the svcTIntPGOcc surface. Default NULL. When set, data carries a cell node index (1..N) and the covariate w for the bar spatial(~ 1 + w || cell, graph).

source_seasons

Optional length-S list; source_seasons[[s]] is the integer vector of seasons source s observes (partial season overlap). The seasons a source does not cover are set to NA in its array – the staggered survey where sources rarely share the full season grid. Default NULL (every source observes every season).

seed

Optional random seed.

Value

A list with y (a length-S list of [N x J x T] arrays), data, sources, and truth.

Examples

sim <- simulate_dyn_int_occu(N = 40, T_seasons = 3, seed = 1)
lapply(sim$y, dim)