quantReclass {modEvA}

Reclassify continuous values based on quantiles

Description

This function takes the continuous predictions of a model of suitability (e.g. the continuous Bioclim envelope model, computed by the bioclim function of the dismo package or the envelope function of the predicts package), and reclassifies them according to their quantiles.

Usage

quantReclass(pred, by = 0.01, na.rm = TRUE)

Arguments

pred

a 'numeric' vector or a 'SpatRaster' map of predicted suitability values.

by

numeric value indicating which quantiles to compute, e.g. 0.01 for percentiles (the default), 0.1 for deciles, etc.

na.rm

logical value indicating whether NA values should be ignored when computing the quantiles. Defaults to TRUE.

Details

This function was created by Formoso-Freire et al. (accepted) to reclassify continuous Bioclim predictions into ranked suitability values, rescaling them into relative suitability. Modern implementations of Bioclim compute a percentile distribution of the values of each environmental variable at species presence localities. Then, the closer to the 50th percentile (the median), the more suitable a location is according to that variable (Hijmans et al. 2020; Hijmans 2023). However, the more variables are included in the model, the less suitable any location becomes, because it is less likely to be close to the median for all variables. The proposed rescaling procedure removes the dependence of Bioclim predictions on the number of variables included, and it has shown to provide more realistic predictions (Formoso-Freire et al., accepted).

Value

This function returns an object of the same class as 'pred' with the reclassified values.

Author(s)

A. Marcia Barbosa, Victoria Formoso-Freire, Andres Baselga, Carola Gomez-Rodriguez

References

Formoso-Freire V., Barbosa A.M., Baselga A., Gomez-Rodriguez C. (accepted) Predicting the spatio-temporal pattern of range expansion under lack of equilibrium with climate. Biological Conservation

Hijmans R.J., Phillips S., Leathwick J. & Elith J. (2020). dismo: Species distribution modelling (1.3.5). https://CRAN.R-project.org/package=dismo

Hijmans R.J. (2023). predicts: Spatial Prediction Tools. R package version 0.1-11. https://CRAN.R-project.org/package=predicts

See Also

getThreshold, bioclim in package dismo, envelope in package predicts

Examples

# simulate some sample data:
set.seed(2023)
bioclim_pred <- runif(n = 10, min = 0, max = 1)
bioclim_pred

quantReclass(pred = bioclim_pred, by = 0.1)