Species Distribution Modeling

An analysis of species distribution using ensemble methods in R.

About

In this short article, I demonstrate how to build ensemble model of rf, maxent, and brt in R using sdm package version 1.2.46 (Naimi and Araújo 2016).

Loading the libraries

library(sdm)

Species occurrence records

Here we will use the occurrence records inbuilt in the package to boost reproducibility.

species <- vect(system.file('external/species.shp', package = 'sdm'))

preds <- rast(list.files(system.file("external", package = 'sdm'), pattern = '.asc', full.names = TRUE))

Building the sdmData object.

d <- sdmData(Occurrence ~., train = species, predictors = preds)

Building the sdmModels object.

m <- sdm(Occurrence ~., data = d, methods = c('rf', 'maxent', 'brt'), replications = 'cv', cv.folds = 4, n = 3)

ens <- ensemble(x = m,
                newdata = preds,
                filename = ".tif",
                setting = list(method = "weighted", stat = "TSS", opt = 2),
                overwrite = TRUE)
plot(ens)

Cool, this is the plot of all the models “combined”.

References

Naimi, Babak, and Miguel B. Araújo. 2016. “Sdm: A Reproducible and Extensible r Platform for Species Distribution Modelling.” Ecography. https://doi.org/10.1111/ecog.01881.