Framework and initial code created by Brian Leung, with code adapted for package by Dat Nguyen and Brian Leung.
THIS PACKAGE IS IN ITS EARLY STAGES AND A WORK IN PROGRESS. USE AT YOUR OWN RISK.
This R package adapts the code used for the Sightings, Survey and Bias-adjustment Kernel (S2BaK) framework within Leung et al., 2019, a flexible, efficient and integrative bias-adjustment model to fit species distribution models (SDMs) for multiple species. The framework allows the user to apply any SDM approach to the framework, such as Generalized Additive Models (GAMs) or MaxEnt (Phillips et al., 2006).
An explanation of S2BaK will be added to this README. Please refer to Leung et al., 2019 in the meanwhile.
The s2bak package, currently hosted on GitHub, can be installed using devtools R package:
# install.packages("devtools")
library(devtools)
install_github("https://github.com/LeungEcoLab/s2bak")
library(s2bak)Five data objects are primarily required to fit the S2BaK model. Making the data structures more flexible is a current goal:
- An environment data.frame with each row corresponding to a site for species sightings, including spatial predictors of bias.
- An environment data.frame with each row corresponding to a site for surveys, including spatial predictors of bias.
- A two-column data.frame of species sightings linked to the row-index of the environment data.
- A two-column data.frame of species presences from surveys linked to the row-index of the environment data. Note that with surveys that true absences are assumed if the species is not present at a site.
- A species-trait data.frame with each row being a species, and each trait being a column.
Simulated sample data can be generated using the function s2bakSim, which is used for demonstration and to illustrate the data structure.
The functions s2bak.S2, s2bak.SO and s2bak.S2BaK fit SDMs for multiple species. The package currently supports fitting SDM model functions that support formulae (e.g., mgcv::gam and glm). Supporting the most popular SDM methods is a goal, such as MaxEnt.
Given that the framework is intended to be used with multiple species, the package has integrated parallelization using the doParallel package. Within the functions s2bak.S2, s2bak.SO and s2bak.S2BaK functions, which are all used to fit SDMs, the user can specify the number of cores using the ncores parameter.
Leung, B., Hudgins, E. J., Potapova, Anna, & Ruiz-Jaen, Maria. (2019). A new baseline for countrywide α-diversity and species distributions: illustration using >6000 plant species in Panama. Ecological Applications.
Microsoft Corporation and Steve Weston (2022). doParallel: Foreach Parallel Adaptor for the 'parallel' Package. R package version 1.0.17. https://CRAN.R-project.org/package=doParallel
Phillips, S. J., Anderson, R. P., & Schapire, R. E. (2006). Maximum entropy modeling of species geographic distributions. Ecological modelling, 190(3-4), 231-259.
Wood, S.N. (2011) Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. Journal of the Royal Statistical Society (B) 73(1):3-36