Skip to contents

Plot site occupancy and detection rates

plot_sites.Rd

Plots species-level occupancy and detection rates combining site predictors (if included via make_data()) and site random effects (iota).

Usage

plot_sites(
  fit,
  species = NULL,
  sites = NULL,
  map = TRUE,
  intercepts = TRUE,
  back_transform = TRUE,
  include_predictors = TRUE,
  restricted = TRUE,
  ndraws = NULL,
  seed = NULL,
  ...
)

Arguments

fit

A fitted model object from fit_model().

species

character. Vector of species to plot. If NULL (default), all species are plotted. Must be one of attr(occARU_data, "species").

sites

character. Vector of sites to plot. If NULL (default), all sites are plotted. Must be one of attr(occARU_data, "sites").

map

logical. If TRUE (default), plot site effects summarised with posterior medians on a map using UTM coordinates. Requires site coordinates to have been supplied to make_data(). If FALSE, or if no coordinates are present, site effects are plotted with ggdist::stat_pointinterval().

intercepts

logical. If TRUE (default), species-level baseline log detection rates are added to the site effects. If FALSE, only the site deviations are plotted on the log scale.

back_transform

logical. Only used when intercepts = TRUE. If TRUE (default), log detection rates are back-transformed to the natural scale via exp(). If FALSE, values are left on the log scale.

include_predictors

logical. If TRUE (default), includes predictors in the site effects, if included. If FALSE, only plots the random effects.

restricted

logical. If TRUE (default), when include_predictors is FALSE, plots random site effects with orthogonal projection, i.e., \(\boldsymbol{\iota}(\boldsymbol{I} - \boldsymbol{P_{X_2}})\), where \(\boldsymbol{I} - \boldsymbol{P_{X_2}}\) is the orthogonal complement of the column space of the site design matrix. If FALSE, plots random effects without orthogonal projection, i.e., \(\boldsymbol{\iota}\) only. Has no effect when include_predictors is TRUE as the linear predictor is unaffected by orthogonal projection.

ndraws

Positive integerish. Number of draws to use for plotting, passed to tidybayes::spread_rvars(). Default: NULL (uses all draws).

seed

Positive numeric. Seed to use when subsampling draws when ndraws is not NULL. Default: random integer.

...

Additional arguments passed to ggplot2::geom_point() when map = TRUE and ggdist::stat_pointinterval() when map = FALSE.

Value

A ggplot object with occARU-specific attributes attached:

plot_data

The tibble used to produce the plot.

Details

If map = TRUE and site coordinates are present in the fitted object, effects are displayed as points sized by the magnitude of the detection rate on a map; otherwise site effects are plotted as point-intervals. When latent = TRUE was set in fit_model(), sites with median posterior occupancy of 0 are shown as red crosses. When latent = FALSE, detection rates are weighted by occupancy probability (inv_logit(logit_psi[s, i])).

See also

fit_model(), plot_intercepts(), plot_coefficients(), plot_surveys(), plot_correlations(), plot_partitions(), plot_realised()