Calculate abundance distribution model uncertainty

This function calculates the uncertainty of an abundance distribution model by performing a bootstrap procedure. It refits the model multiple times on resampled data and then calculates the standard deviation of the predictions across all iterations.

Usage

adm_uncertainty(
  models,
  training_data,
  response,
  pred,
  iteration = 50,
  n_cores = 1
)

Arguments

models

A model object from `fit_abund_*` or `tune_abund_*` functions.

training_data

A data.frame or tibble with abundance data and predictors.

response

character. Column name of the response variable.

pred

A SpatRaster object with the environmental layers for projection.

iteration

numeric. The number of bootstrap iterations. Default 50.

n_cores

numeric. The number of cores to use for parallel processing. Default 1.

...

Additional arguments passed to refitting functions or adm_predict (e.g., x, y, rasters, sample_size for CNN; custom_architecture for DNN/CNN; invert_transform, transform_negative for spatial prediction).

Value

A SpatRaster object with a single layer representing the model uncertainty, calculated as the standard deviation of the bootstrap predictions.

See also

adm_predict

Examples

if (FALSE) { # \dontrun{
require(terra)
require(dplyr)

# Load data
data("cretusa_data")
cretusa_predictors <- system.file("external/cretusa_predictors.tif", package = "adm")
cretusa_predictors <- terra::rast(cretusa_predictors)

species_data <- adm_extract(
  data = cretusa_data,
  x = "x",
  y = "y",
  env_layer = cretusa_predictors
)

# Fit model
mraf <- fit_abund_raf(
  data = species_data,
  response = "ind_ha",
  predictors = c("PC1", "PC2", "PC3"),
  partition = ".part"
)

# Calculate uncertainty
unc <- adm_uncertainty(
  models = mraf,
  training_data = species_data,
  response = "ind_ha",
  pred = cretusa_predictors[[c("PC1", "PC2", "PC3")]],
  iteration = 10,
  n_cores = 2
)

plot(unc)
} # }