sim_MSMocc: Simulate data from the multi-species, misclassification count...

In StrattonCh/occupancy: Fit and Analyze Occupancy Models

Description

This function simulates data from he multi-species, misclassification count occupancy model first developed by Wright et al. (2020).

Usage

sim_MSMocc(
  K = 2,
  M = 20,
  M.conf = 2,
  max_j = 10,
  beta_psi = matrix(c(0, 1, 0, 1.2), ncol = K),
  beta_lambda = matrix(c(0, 1, 0, 1.2), ncol = K),
  theta = matrix(c(0.8, 0.2, 0.2, 0.8), byrow = T, ncol = K),
  seed = NULL,
  rand_visits = TRUE
)

Arguments

K	number of species
M	number of sites
M.conf	number of sites that have confirmed observations
max_j	maximum number of visits to each site. If rand_visits = FALSE, this value is the number of visits to each site. If rand_visits = TRUE, each site is visited a random (sample(2:max_j, size = 1)) number of times.
beta_psi	vector of regression coefficients used to generate psi
beta_lambda	vector of regression coefficients used to generate lambda
theta	confusion matrix for automatic call classifications. Rows of theta should sum to one, diagonal elements are correct identification probabilities, first row corresponds to probabilities for calls truly from species 1.
seed	optional seed for reproducibility
rand_visits	logical; should each site be visited a random number of times? See details.

Details

This function simulates data from the single season, multiple species occupancy model to estimate occupancy and relative activity for multiple species simultaneously. If rand_visits = TRUE, each site is visited a random number of times between two and max_j. Covariates are drawn from the uniform(0, 1.5) distribution so that the effect of the direction of each regression coefficient is intuitive. Note that if covariates are not desired, beta_psi and beta_lambda can be set to intercepts that generate the desired derived probabilities.

Value

object of class list containing the following elements:

• beta_psi vector of regression coefficients used to generate psi

• beta_lambda vector of regression coefficients used to generate lambda

• psi_cov matrix of site level covariates

• lambda_cov array of detection level covariates; each slice represents a single covariate

• psi vector of derived site level occupancy probabilities

• lambda matrix of derived visit level expected count detection parameter

• z matrix of latent occupancy states for each site/species

• Y array of observed Bernoulli responses for site/visit/species

• n_visits vector of number of visits to each site

• c.conf how calls confirmed to species K were classified for confirmed visits

• c.sum.a and c.sum.b unconfirmed calls classified TO species K from sites with some confirmations ("a") or no confirmations ("b")

• sites.conf site locations with confirmations

• sites.unconf.a and sites.unconf.b site locations without confirmations ("a" indecates some observations at these sites were confirmed, "b" sites had no confirmations)

• data a data frame containing all information necessary to fit the model

Examples

n.species <- 4
sim <- sim_MSMocc(K = n.species, M = 100, M.conf = 10, max_j = 20,
                  beta_psi = matrix(c(0, 1,
                                      0, 1.1,
                                      0, 1.2,
                                      0, 1.3),
                                    ncol = n.species),
                  beta_lambda = matrix(c(0, 1, 1,
                                         0, 1, 1,
                                         0, 1, 1.2,
                                         0, 1, 1.4),
                                       ncol = n.species),
                  theta = matrix(c(0.7, 0.1, 0.1, 0.1,
                            0.1, 0.7, 0.1, 0.1,
                            0.1, 0.1, 0.7, 0.1,
                            0.1, 0.1, 0.1, 0.7), nrow = n.species),
                  rand_visits = TRUE)
sim$Y

StrattonCh/occupancy documentation built on Feb. 17, 2021, 6:36 a.m.