Species distributions are an important EBV in the ‘species populations’ class. Knowing where species are is essential for understanding biodiversity patterns and informing conservation efforts. However, less than 10% of the world is well sampled, and even the longest running and well-sampled biodiversity observation networks have substantial data gaps. Information on species occurrences is often sparse and heavily spatially and taxonomically biased, necessitating the need for species distribution models (SDMs) to fill these data gaps and provide a better, less biased idea of where species are. SDM outputs be used as key base layers for a wide variety of purposes including: creating maps for sampling prioritization, quantifying the impact of environmental stressors on species, mapping habitat suitability for at-risk species, mapping biodiversity hotspots across the landscape, identifying the locations of conservation priorities and protected area expansion, identifying sampling gaps and the needed locations of future sampling, and calculating a range of biodiversity indicators including the Species Habitat Index (SHI), the Species Protection Index (SPI)
Authors: Sarah Valentin, Guillaume Larocque, François Rousseu
Review status: Under development
The MaxEnt pipeline builds a species distribution model using occurrence records from the Global Biodiversity Information Facility (GBIF) and environmental raster layers from the GEO BON STAC catalog. The pipeline retrieves GBIF observations for the selected taxon or taxa, cleans occurrence coordinates, removes highly collinear environmental predictors, generates background points, and fits a MaxEnt model using the ENMeval R package (Kass et al. 2021). MaxEnt is a presence-background modeling approach, meaning it compares known species presences with background environmental conditions across the study area. The MaxEnt SDM is run by 1) partitioning occurrence and background points into subsets for training and evaluation, 2) building the model with different algorithmic settings (model tuning), and 3) evaluating their performance (see package vignette). Lastly, the pipeline computes the 95% confidence interval using bootstrapping and cross validation techniques.
The pipeline evaluates different MaxEnt settings, including feature classes and regularization multipliers, and selects a tuned model based on model performance. It produces a habitat suitability prediction raster, cleaned occurrence records, selected environmental predictors, a GBIF download DOI, and a raster summarizing variability among model runs.
SDMs predict where species are likely to occur based on a suite of environmental variables that are associated with known occurrences (Peterson, 2001; Elith and Leathwick, 2009). The MaxEnt pipeline can be used to estimate the potential distribution or relative habitat suitability of one or more species within a selected study area. Outputs can support conservation planning, sampling prioritization, identification of biodiversity hotspots, protected area planning, risk assessment for species of conservation concern, and environmental impact assessments.
The results can also be used as inputs to other biodiversity analyses and indicators, such as identifying areas where species are likely to occur, comparing predicted habitat suitability across regions, or highlighting areas where additional occurrence sampling may be needed. Because the pipeline retrieves both GBIF observations and environmental predictor layers, it provides a reproducible workflow for generating species distribution maps from public biodiversity and environmental data.
MaxEnt uses presence-background data, not confirmed absence data. Predictions should be interpreted as relative habitat suitability or relative occurrence potential, not confirmed species presence or absence.
GBIF records may contain spatial, taxonomic, and temporal biases. The pipeline applies coordinate-cleaning steps, but users should still inspect the cleaned presences and interpret results cautiously, especially for poorly sampled taxa or regions.
Model quality depends strongly on the number and quality of occurrence records. Very small numbers of cleaned presences may produce unreliable predictions.
The choice of environmental predictors, background sampling method, feature classes, regularization multipliers, and partitioning method can affect model outputs. Users should treat the model as sensitive to these settings, especially for final analyses.
Environmental predictors must be ecologically relevant to the species being modeled. Including many correlated or irrelevant predictors can reduce interpretability and increase overfitting risk.
The pipeline estimates suitability based on the predictor layers supplied by the user. It does not directly account for dispersal limits, biotic interactions, land-use barriers, species detectability, or future environmental change unless those factors are represented in the input data.
Larger study areas, finer spatial resolutions, more environmental predictors, and more model runs increase computation time and memory use.
A GBIF API key is required to run this pipeline and can be added into the runner.env file.
Before running the pipeline, choose the taxon or taxa you want to model and make sure the names match the GBIF taxonomic backbone. Species names can be checked on the GBIF website.
Select a study area using the bounding box and CRS input. The CRS and spatial resolution determine the scale of the analysis, so choose a CRS appropriate for the region and make sure the spatial resolution is in the units of that CRS.
Choose environmental predictor layers from the STAC catalog that are ecologically relevant to the species being modeled. For example, climate, vegetation, elevation, land cover, or habitat-related predictors may be appropriate depending on the species. Avoid including many predictors that represent the same underlying environmental gradient.
The BON in a Box pipeline allows you to run an SDM for a specific region and species (or multiple species) of interest. The pipeline has the following inputs:
random: background points are randomly sampled throughout the whole study extent.weighted_raster: background points are sampled in proportion to the number of observations of a target group in an observation density raster. unweighted_raster: background points are sampled only in cells where there are observations from a target group. inclusion_buffer: background points are sampled within a buffer around observations. thickening: background points are sampled in proportion the local density of observations by sampling in a buffer around each observation. Block: partitions the bounding box into four equally sized quadrants and assigns groups by quadrantCheckerboard 1: Generates checkerboard from the study area and assigns groups based on what square the points fall inCheckerboard 2: Similar to checkerboard 1 but performs this separately for occurrence and background pointsJackknife: Does not partition the background points into testing and training (uses them all), performs leave one out cross validation (recommended for small datasets only)Random k-fold: Does not partition the background points into testing and training, partitions groups randomly into a user specified (K) number of bins, and runs the model k times, with each bin used once as testing.This step retrieves occurrence records for the target species from GBIF and downloads environmental predictor layers from the GEO BON STAC catalog. Environmental layers may include climatic, topographic, and land-cover variables used to characterize habitat suitability.
This step cleans the GBIF occurence records by multiple empirical tests. Environmental predictors are also screened for collinearity, and highly correlated variables are removed to reduce model overfitting.
This step creates a set of pseudo-absences (background points) using one of the five available methods: random, inclusion buffer, weighted raster, unweighted raster, thickening.
This step extracts environmental predictor values at species occurrence and background locations and combines them into a modelling dataset suitable for species distribution modelling.
This step runs the SDM on the clean data using the MaxEnt algorithm using the ENMeval R package (Kass et al. 2021). The MaxEnt SDM is run by 1) partitioning occurrence and background points into subsets for training and evaluation, 2) building the model with different algorithmic settings (model tuning), and 3) evaluating their performance (see package vignette). This step produces spatial predictions of habitat suitability across the study area.
This step computes the 95% confidence interval using bootstrapping and cross validation techniques.
The pipeline creates the following outputs:
Baston D (2025). exactextractr: Fast Extraction from Raster Datasets using Polygons. doi:10.32614/CRAN.package.exactextractr
Elith, J., & Leathwick, J. R. (2009). Species Distribution Models: Ecological Explanation and Prediction Across Space and Time. Annual Review of Ecology, Evolution, and Systematics, 40(Volume 40, 2009), 677–697. https://doi.org/10.1146/annurev.ecolsys.110308.120159
Kass JM, Muscarella R, Galante PJ, Bohl CL, Pinilla-Buitrago GE, Boria RA, Soley-Guardia M, Anderson RP (2021). “ENMeval 2.0: Redesigned for customizable and reproducible modeling of species’ niches and distributions.” Methods in Ecology and Evolution, 12(9), 1602-1608. https://doi.org/10.1111/2041-210X.13628.
Peterson, A. T. (2001). Predicting Species’ Geographic Distributions Based on Ecological Niche Modeling. The Condor, 103(3), 599–605. https://doi.org/10.1093/condor/103.3.599
Phillips, S. J., Dudík, M., Elith, J., Graham, C. H., Lehmann, A., Leathwick, J., & Ferrier, S. (2009). Sample selection bias and presence‐only distribution models: implications for background and pseudo‐absence data. Ecological applications, 19(1), 181-197. https://doi.org/10.1890/07-2153.1
Vollering, J., Halvorsen, R., Auestad, I., & Rydgren, K. (2019). Bunching up the background betters bias in species distribution models. Ecography, 42(10), 1717-1727. https://doi.org/10.1111/ecog.04503
Author: Michael D. Catchen
Review status: Under development
This document describes the methodology behind the BON in a Box pipeline for using Boosted Regression Trees (BRTs) for species distribution modeling.
Summary
This pipeline builds a model to predict the distribution of a species (a type of essential biodiversity variable), by using occurrence data from the Global Biodiversity Information Facility (GBIF), and environmental predictors from an arbitrary STAC Catalogue.
In particular, this pipeline uses a specific model called a Boosted Regression
Tree (BRT), a machine-learning model which tends to work well with spatial data. The
details of how a BRT works are in the description of the key script in the
pipeline, fitBRT.jl.
Inputs:
Outputs
See an example pipeline output here
[!IMPORTANT] Using BRTs to fit a species distribution model requires absence data. For the majority of species where no absence data is available, there are various methods to generate pseudoabsences (PAs) based on heuristics about species occurrence. However, the performance characteristics of an SDM fit using PAs can be widely variable depending on the method and parameters used to generate PAs. This means the results of BRT should be explicitly considered as a function of how PAs were generated, and sensitivity analysis to different PAs is highly encouraged.
Pipeline Steps
flowchart LR
a{input species} --> b[Load GBIF Occurrences]
c{input bounding box} --> b
d{input layers} --> e[Load Layers from STAC]
c --> e
b --> f[Clean presences]
f --> g[Generate Pseudoabsences]
c --> g
g --> h[Fit BRT]
e --> h
h --> i(predicted sdm)
h --> j(uncertainty map)
c --> k[create water mask]
h --> l[model fit statistics]
h --> m[diagnostic plots]
k --> hAuthors: François Rousseu, Guillaume Blanchet, Dominique Gravel
Review status: Under development
Methods: The species distribution modeling method provided in the package ewlgcpSDM (Effort-Weighted Log-Gaussian Cox Process) is based on spatial point processes and presence-only observations. It implements the method proposed by Simpson et al. (2016) to estimate log-Gaussian Cox processes using INLA (Rue et al. 2009) and the SPDE approach (Lindgren et al. 2009). The model relies on a discrete grid (the mesh) of arbitrary resolution to approximate the spatial component of the model. The method proposed in ewlgcpSDM contains three key aspects for species distribution modeling, namely:
The current version of the pipeline does not make use of the spatial component yet as some more work is needed to allow the adjustments necessary for the spatial component to work properly. The current version of the pipeline thus corresponds to an effort-weighted inhomogeneous Poisson point process.
BON in a Box pipeline: The pipeline is used to run an SDM for a set of species in a specific region and using a set of environmental predictors. Some inputs are yet to be added to the list of inputs required by the user. Currently, the pipeline mostly reuses the same inputs as the MaxEnt pipeline, namely:
The pipeline creates the following outputs:
See an example pipeline output here
Citations: Lindgren, F., Rue, H., and Lindström, J. 2011. An explicit link between Gaussian fields and Gaussian Markov random fields: the stochastic partial differential equation approach. Journal of the Royal Statistical Society Series B: Statistical Methodology, 73(4): 423-498.
Phillips, S. J., Dudík, M., Elith, J., Graham, C. H., Lehmann, A., Leathwick, J. and Ferrier, S. 2009. Sample selection bias and presence-only distribution models: implications for background and pseudo-absence data. Ecological Applications, 19(1): 181-197, https://doi.org/10.1890/07-2153.1
Rue, H., Martino, S. and Chopin, N. 2009. Approximate Bayesian Inference for Latent Gaussian models by using Integrated Nested Laplace Approximations, Journal of the Royal Statistical Society Series B: Statistical Methodology, 71(2): 319–392, https://doi.org/10.1111/j.1467-9868.2008.00700.x
Simpson, D., Illian, J. B., Lindgren, F., Sørbye, S. H. and H. Rue. 2016. Going off grid: computationally efficient inference for log-Gaussian Cox processes, Biometrika 103(1): 49–70, https://doi.org/10.1093/biomet/asv064