Simulate joint species distribution (presence/absence) data

Presence y_{i,s} ~ Bernoulli(psi_{i,s}) with logit psi = X beta_s, the per-species occupancy coefficients drawn from Gaussian community hyperpriors. Matches the jsdm() family, which observes presence directly (no detection process), so the response is an N x n_species presence matrix.

Usage

simulate_jsdm(
  N = 100,
  n_species = 10,
  beta_comm_mean = c(0, 0.5),
  beta_comm_sd = c(0.5, 0.3),
  seed = NULL
)

Arguments

N

Number of sites (default 100).

n_species

Number of species (default 10).

beta_comm_mean

Community-mean occupancy coefficients, length 1 + n_occ_covs (intercept first).

beta_comm_sd

Between-species SD of each occupancy coefficient (same length as beta_comm_mean).

seed

Random seed.

Value

A list with y (an N x n_species 0/1 presence matrix), data, and truth (per-species coefficients and the community hyperparameters).

Examples

sim <- simulate_jsdm(N = 60, n_species = 5, seed = 1)
dim(sim$y)   # 60 sites x 5 species presence matrix