See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/366274338

A curated list of R packages for ecological niche modelling

Article in Ecological Modelling · February 2023

DOI: 10.1016/j.ecolmodel.2022.110242

CITATIONS READS

2 993

4 authors:

Neftalí Sillero João Carlos Campos

University of Porto CIBIO Research Center in Biodiversity and Genetic Resources
180 PUBLICATIONS 3,444 CITATIONS 48 PUBLICATIONS 1,283 CITATIONS

SEE PROFILE SEE PROFILE

Salvador Arenas-Castro A. Márcia Barbosa

University of Cordoba (Spain) Biogeography Consulting & Training
86 PUBLICATIONS 432 CITATIONS 153 PUBLICATIONS 3,722 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

PhD Project: Eco-physiology and niche models to assess the vulnerability to climate change View project

ECOPOTENTIAL View project

All content following this page was uploaded by A. Márcia Barbosa on 19 December 2022.

The user has requested enhancement of the downloaded file.

 Ecological Modelling 476 (2023) 110242

Contents lists available at ScienceDirect

Ecological Modelling
journal homepage: www.elsevier.com/locate/ecolmodel

A curated list of R packages for ecological niche modelling

Neftalí Sillero a, *, João Carlos Campos a, Salvador Arenas-Castro b, A.Márcia Barbosa a
a
CICGE – Centro de Investigação em Ciências Geo-Espaciais, Faculty of Sciences, University of Porto, Alameda do Monte da Virgem, Vila Nova de Gaia 4430-146,
Portugal
b
Area of Ecology – Department of Botany, Ecology and Plant Physiology, Faculty of Sciences (University of Córdoba). Campus de Rabanales, Córdoba 14014, Spain

A R T I C L E I N F O A B S T R A C T

Editor: Dr Juan Blanco The R language provides most applications (packages) currently available for ecological niche modelling. In the
last few years, these packages have increased substantially. There are now numerous packages for computing
Keywords: ecological niche models (ENMs) with different algorithms, comparing species niches and niche models, creating
Species distribution models virtual species, getting and preparing species and related environmental data, filtering species records, and
R language
evaluating models, amongst other tasks. This profusion of package options may be daunting for both beginners
R packages
and more experienced R users. Here we provide an organised and annotated list of available packages for each
Biodiversity modelling
Biogeography task. We present a concise history of the creation of these packages, we briefly describe what each package does,
Spatial analyses and we discuss whether the current set of R packages covers all the necessities for modelling species’ ecological
niches. This list will be accessible and updated over time on a dedicated website.

1. Introduction operating systems such as Microsoft Windows. ModEco integrates

several modelling tools within a GIS framework: algorithms (Support
Ecological niche models (ENMs), which are closely related to species Vector Machines, Bioclim, Domain, Generalised Linear Model,
distribution models (SDM), are empirical or mathematical estimations of Maximum Likelihood Classification, Artificial Neural Network, Rough
a species’ ecological niche (Sillero et al., 2021). ENMs are statistical Set, Maximum Entropy, Classification Trees, Ensemble Model), data
methods or theoretically derived response surfaces that relate physio­ management and visualisation, feature analysis, model training and
logical or distribution data to environmental variables, to describe prediction, and accuracy assessment. OpenModeller includes tools for
and/or predict species distributions (Franklin, 2010; Peterson et al., sampling points, modelling, evaluating and projecting models into
2011; Sillero, 2011; Guisan et al., 2017; Sillero et al., 2021). ENMs can different environmental scenarios. Several algorithms are available as
be calculated with statistical software (e.g. SPSS, Statistica), geograph­ plugins: Bioclim, Domain, Mahalanobis Distance, GARP, Maxent, ENFA
ical information systems (e.g. ArcGIS, QGIS, DIVA-GIS), or with dedi­ and Support Vector Machines.
cated graphical user interface (GUI) software (e.g. OpenModeller, Maxent is the java software that runs the Maximum Entropy algo­
ModEco, Maxent, Biomapper, Wallace, ShinyBiomod). Traditional sta­ rithm (Phillips et al., 2017). Biomapper is the software that runs ENFA
tistical software with a GUI can run an ENM (for example, GLM or GAM) (Hirzel et al., 2002). Wallace (Kass et al., 2018) is a modular Shiny
if the species and environmental data are compiled as a table. Some platform that guides the user through a complete modelling process,
modelling algorithms purely designed for ENMs, such as Maximum making parameter choices explicit, from data acquisition to plotting
Entropy (Phillips et al., 2006, 2017) or Ecological Niche Factor Analysis model predictions on an interactive map. ShinyBiomod (Ondo et al.,
(ENFA; Hirzel et al., 2002), are not available in traditional statistical 2019) is another user-friendly Shiny application that uses the R
programs but in dedicated ones. modelling package biomod2 (Thuiller et al., 2009, 2012). Some appli­
Some dedicated programs offer modelling solutions with all stages of cations are built as plugins for other programs. For example, SDMtool­
the modelling process. ModEco (Guo and Liu, 2010) and OpenModeller box is a Python-based toolbox for ArcGIS that includes functions for
(Enrique et al., 2009) are stand-alone programs available for some modelling pre and post-processing tasks (Brown, 2014; Brown et al.,

Abbreviations: ENM, Ecological niche model; GUI, Graphical user interfacE; ENFA, Ecological Niche Factor Analysis; SDM, Species distribution model; GIS,
Geographical Information System; FOSS, Free Open Source Software; CRAN, Comprehensive R Archive Network.
* Corresponding author.
E-mail address: [email protected] (N. Sillero).

https://doi.org/10.1016/j.ecolmodel.2022.110242
Received 3 November 2022; Received in revised form 18 November 2022; Accepted 30 November 2022
0304-3800/© 2022 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
N. Sillero et al. Ecological Modelling 476 (2023) 110242

2017; http://www.sdmtoolbox.org/), or Maxent Model, a QGIS plugin 3.1. A brief history of the development of R packages for ecological niche
to preprocess input data in Maxent (https://geoinnova.org/plugin/ma models
xent/). Maxent, Wallace, SDMToolBox and ShinyBiomod are updated
regularly. To the best of our knowledge, the first R package with functions for
ENMs can be also calculated with programming languages, such as ENMs was PresenceAbsence (Freeman and Moisen, 2008), released in
Python and R. Python is a high-level, interpreted, general-purpose 2007. It does not compute models but provides functions for evaluating
programming language (Van Rossum and Drake, 1995). It supports presence-absence analysis, either ENMs or diagnostic tests. The package
multiple programming paradigms, including structured, calculates the optimal threshold for translating a probability surface into
object-orientated and functional programming. R is a programming presence-absence maps, such as percentage correctly classified, sensi­
language for statistical computing with graphical capacities (R Core tivity, specificity, and kappa. It also computes the Receiver Operator
Team, 2022). R is widely used amongst data scientists and statisticians Characteristic (ROC) curve and the area under the curve (AUC). How­
to analyse data and develop statistical software (Kabacoff, 2015). Both ever, the first R package developed specifically for the field of ENMs was
Python and R are free and open-source software (FOSS), allowing users BIOMOD, released in 2009 (Thuiller et al., 2003, 2009). A second
to create packages that augment or complement existing functions. version, biomod2, replaced it in 2012 (Thuiller et al., 2021). Both ver­
R is one of the most used FOSS, thanks to its ease of use, versatility, sions provided several ensemble forecasting methods for ENM algo­
and object-orientated framework. Indeed, users have created thousands rithms: generalised linear models, generalised additive models,
of packages. On 28/10/2022, CRAN, the main repository of R packages, multivariate adaptive regression splines (MARS), classification tree
hosted 18,761 packages (https://cran.r-project.org/web/packages/), a analysis, mixture discriminant analysis, artificial neural networks,
number that continues to increase over the years. The field of biodi­ generalised boosted models, random forests, and maximum entropy
versity modelling is not an exception. Many packages can be directly or (through java, a method used posteriorly by other R packages). biomod2
indirectly used for modelling species’ ecological niches and all related also included a rectilinear envelope algorithm similar to BIOCLIM
pre- and post-processing tasks. (Booth et al., 2014).
Here we provide an organised list of R packages for ecological niche The dismo package, released in 2010 (Hijmans et al., 2021), provides
models. We divided the list according to the four main stages of the tools for the complete modelling process, from downloading and pre­
modelling process (Sillero et al., 2021): (1) data collection and prepa­ paring data to evaluating the models. This package implements
ration; (2) model calculation; (3) model evaluation and validation; and presence-only modelling algorithms (Bioclim, Domain, Mahalanobis
(4) model application. We start our review with a brief description of the distance) and a presence-absence algorithm (boosted regression trees). It
development of R packages for ENMs over time, followed by the can also run Maxent by calling Java.
organised list. Finally, we discuss the current state of the universe of R maxlike (Royle et al., 2012), a package released in 2011, implements
packages for ENMs. a maximum likelihood algorithm as an alternative method to the
Maximum Entropy (Maxent) algorithm. The maxlike algorithm
2. Methods endeavoured to estimate the probability of occurrence from
presence-only data (Royle et al., 2012), instead of habitat suitability (but
We searched Google Scholar and Web of Science for R packages for see Osorio-Olvera et al. 2019).
ENMs, using the words ‘ecological niche model package’, ‘species dis­ In 2012, besides biomod2 (Thuiller et al., 2021), several packages
tribution package’, and ‘species distribution model package’. Therefore, were released for obtaining species records (rgbif, Chamberlain et al.,
we searched for packages whose main aim was to provide functions for 2022), calculating distribution and abundanDM; Latimer et al., 2006),
one or more of the four steps of the modelling process. We verified that analysing model uncertainties (usdm; Naimi et al., 2014), and modelling
all packages found in these searches were directly related to ENMs. We species dispersal (MigClim; Engler et al., 2012). usdm includes methods
checked whether each package was available through CRAN and if the from spatial statistics to analyse positional uncertainty (Naimi et al.,
package had a development site on GitHub or another similar platform. 2014); it also implemented the Variance Inflation Factor (VIF) to analyse
We considered as release year the first year of the package publication in collinearity between variables (Naimi et al., 2014).
CRAN, or GitHub if the package is not on CRAN. After obtaining a list of Several packages released in 2014 implemented important, often
R packages for ENMs, we searched in Google Scholar for the number of novel additional tasks. These included spatially filtering species records
citations for each package or paper presenting that package. (spThin; Aiello-Lammens et al. 2015), generating virtual species
(sdmvspecies, Duan et al. 2015; virtualspecies, Leroy et al. 2016),
3. Results modelling species range dynamics (RangeShiftR, Bocedi et al., 2014, G.
2021), and automatically fine-tuning and validating models to deter­
Our search allowed us to compile a total of 60 R packages for ENMs mine optimal levels of complexity (ENMeval; Muscarella et al., 2014). A
(Table 1) published since 2007 (Fig. 1). We found 39 packages available new version of ENMeval was released in 2021 (Kass et al., 2021). It
through CRAN and 21 currently in development for different purposes provides functions for results visualisation, data partition with spatial
(see Table 1). These packages were cited 9664 times, with eighteen cited criteria, model evaluation with cross-validation and null models, and
more than 100 times each, and two (ENMeval and biomod2) more than niche overlap from model predictions. The package was originally built
1000 times (see Table 2). for Maxent models, but the current version allows possible extensions
for any modelling algorithm. Two other packages released in the same
year provided tools for estimating species’ ecological niches using kernel
interpolators instead of regression-based and machine-learning algo­
Table 1 rithms: ecospat (Di Cola et al., 2017) and hypervolume (Blonder et al.,
Summary of R packages for ecological niche models (last search performed at 2014) ecospat also provides other model-related functions, such as sta­
28/10/2022). tistical comparison of species niches (using the methods proposed by
Package main purpose N CRAN In development Warren et al., 2008, 2010), computing ensembles of small models
Data collection and preparation 18 14 4
(Breiner et al., 2018), analysing non-analogue climates (Elith et al.,
Model calculation 24 14 10 2010, 2011; Mesgaran et al., 2014), and evaluating model performance.
Model evaluation and validation 7 6 1 zoon and SDMTools were released in 2015. zoon (Golding et al., 2018)
Model application 11 5 6 provides functions for comparing and diagnosing models. The results
Total 60 39 21
can be shared from an online repository. SDMTools (VanderWal et al.,

2
N. Sillero et al. Ecological Modelling 476 (2023) 110242

Fig. 1. Increment over time in the number of R packages for ecological niche models.

2015) has tools for post-processing models, including threshold selec­ moving occurrences to the closest non-NA pixel), maps species ranges
tion, calculating validation and landscape metrics, and tracking distri­ and calculates Maxent models with dismo (Hijmans et al., 2021)
bution changes over time. package.
Ten packages were published in 2016. sdm (Naimi and Araújo, 2016) Until 2016, R packages related to ENMs focused mainly on correla­
was the third package, together with dismo (Hijmans et al., 2021) and tive algorithms, although dismo provides some simple functions for
biomod2 (Thuiller et al., 2012, Thuiller et al., 2021), providing a full set mechanistic models. In contrast, NicheMapR was designed specifically
of tools for the four stages of modelling: data preparation, model for mechanistic niche models, including datasets of micro-climate
calculation with several algorithms (GAM, GLM, MARS, SVM, RF, and (Kearney and Porter, 2017, 2020). In addition, SPEDInstabR (Gui­
Maximum Entropy), model evaluation, and model projection. SDMPlay sande et al., 2017) selects environmental variables based on a fluctua­
(Charlene et al., 2021) also offers functions for the four modelling tion index, which identifies the environmental factors that better
stages, but only for maximum entropy and boosted regression trees. discriminate between the conditions in the species’ area and those in the
Another package (maxnet) provided a direct R implementation of the background area.
maximum entropy algorithm, without the need for calling Java to run Six packages were released on CRAN in 2018, providing some more
the Maxent software outside R. This package, written by the main author interesting functionalities. For instance, CENFA (Rinnan and Lawler,
of Maxent, uses the package glmnet to fit an inhomogeneous Poisson 2019) presented a new implementation of ENFA: it includes algorithms
process model, which is equivalent to the maximum entropy algorithm for climate- and ecological-niche factor analyses and for visualisation of
(Phillips et al., 2017). The package rmaxent (Baumgartner and Wilson, spatial variability of species sensitivity, exposure, and vulnerability to
2022), which is also Java-free, includes additional functionalities for climate change. Three other packages provided different tools for sup­
projecting Maxent models, calculating information criteria (AIC, AICc, porting the modelling process. Here, rangeModelMetadata (Merow et al.,
BIC) as implemented in Warren et al. (2010), parsing Maxent .lambdas 2019) catalogues ENMs with hierarchical structures through dictio­
files, and estimating MESS (multivariate environmental similarity sur­ naries. In this way, communicating and sharing the algorithms and pa­
face; Elith et al., 2010, 2011). The package MIAmaxent (Vollering et al., rameters used to compute the models is easier and more accurate. mopa
2019) provides pre-processing (e.g., variable selection) and evaluation (Iturbide et al., 2018) provides functions for transferable models,
functions for maxent and logistic regression models. ENiRG (Cánovas pseudo-absence data generation, and factor contribution quantification.
et al., 2016) is an interface for R and GRASS GIS and implements ENFA. fuzzySim (Barbosa, 2015) includes functions to compute fuzzy versions
modEvA (Barbosa et al., 2016) provides a varied set of tools for evalu­ of species occurrence patterns based on presence-absence data (inverse
ating ENMs, including discrimination and calibration metrics, optimis­ distance interpolation, trend surface analysis, and
ing classification thresholds, and analysing model extrapolation with prevalence-independent favorability), and pairwise fuzzy similarity
MESS on data frames (Elith et al., 2010, 2011). sdmpredictors (Bosch and amongst species’ potential distributions. fuzzySim currently has addi­
Fernandez, 2022) downloads environmental data from different climatic tional functions for data preparation, model consensus and comparison,
datasets, including terrestrial, marine and freshwater environments: such as unique abbreviations of species names, gridding (thinning) point
currently WorldClim (Fick and Hijmans, 2017), ENVIREM (Title and occurrence data onto raster maps, converting species lists to
Bemmels, 2018), Bio-ORACLE (Tyberghein et al., 2012), MARSPEC presence-absence tables, transposing part of a data frame, selecting
(Sbrocco and Barber, 2013) and Freshwater (Domisch et al., 2015). relevant variables for models, assessing the false discovery rate, or
biogeo (Robertson et al., 2016) was the first package developed for analysing and dealing with multicollinearity. Finally, Metrics (Hammer
automatically detecting and correcting errors in species occurrence and Frasco, 2018) provides several metrics for evaluating models
datasets. It includes functions for assessing data quality of occurrence calculated with machine learning algorithms, while bdclean (Gueta
records, identifying potentially erroneous records in geographical and et al., 2018), like biogeo (Guisande et al., 2017), provides functions for
environmental space, and providing possible alternative positions for cleaning and validating species occurrence data.
those records. Another package focused on the analysis of species oc­ kuenm (Cobos et al., 2019b) was released in 2019. It is a package for
currences is red (Cardoso, 2017), which calculates the Extent of Occur­ calibrating Maxent models by selecting optimal parameterisations for
rence, Area of Occupancy, and Red List Index for groups of species. red each study. Therefore, the user can run multiple Maxent models with
also downloads and pre-processes species occurrences (e.g. thinning and different conditions and choose the best one. ESDM (Woodman et al.,

3
N. Sillero et al. Ecological Modelling 476 (2023) 110242

2019), also released in 2019, is a package for ensemble modelling of WorldClim), elevation, land use, soil, accessibility, and administrative
species distribution and abundance models. ESDM includes a graphical boundaries. The other three packages (occAssess, sampbias, and bRaca­
user interface (GUI) built with an R Shiny app. The MinBAR package tus) provide functions for detecting errors in species occurrence datasets.
(Rotllan-Puig and Traveset, 2021) estimates the minimum background Specifically, bRacatus (Arlé et al., 2021) presents a method to estimate
area by calculating several maxent models with different parameters the probability of a georeferenced record being true or false and corre­
until determining the best one. SDMtune (Vignali et al., 2020) and sponding to a native or an alien occurrence. It avoids subjective
sdmbench (Angelov, 2018) provide functions for calculating, selecting thresholds of data filtering by implementing a probabilistic framework
and validating models, with several algorithms of machine learning. to propagate uncertainties in posterior analyses. On the other hand,
SDMselect includes functions for the selection of covariates and model occAssess (Boyd et al., 2021) and sampbias (Zizka et al., 2020) detect
specifications. ellipsenm (Cobos et al., 2019b) and humboldt (Brown and geographical biases in species occurrence data.
Carnaval, 2019) provide advanced functions for running analyses in the Finally, five more packages were released in 2022. flexsdm (Velazco
environmental space: ellipsenm (Cobos et al., 2019a) calculates ellipsoid et al., 2022) is another package providing functions for the four
envelopes, calibrates and selects models, replicates models and pro­ modelling stages. It includes algorithms for presence-background
jections, and assesses niche overlaps; and humboldt (Brown and Car­ (Maxent) and presence-absence (GAM, GLM, GBM, SVM, RF, ANN)
naval, 2019) analyses the amount of the species’ observed modelling, together with functions for survey bias correction, sampling
environmental space that is truncated by the available environmental pseudo-absences and background points, data partitioning, and
space, providing also tools for analysing niche overlap and divergence, reducing collinearity in predictors; fitting and evaluating models, en­
and rarefying species occurrence records to reduce survey bias. Coor­ sembles of small models and ensemble models; and prediction, inter­
dinateCleaner (Zizka et al., 2019) cleans occurrence datasets, similar to polation, and overprediction correction. itsdm (Song, 2022) provides
biogeo (Guisande et al., 2017) and bdclean (Gueta et al., 2018): it detects functions for modelling presence-only records with isolation forest
geo-referencing and dating imprecisions in a standardised and repro­ (iForest) and variations (extended iForest and SCiForest). mcera5
ducible way, based on geographical gazetteers and a global database of (Klinges et al., 2022) is the second R package designed for mechanistic
georeferenced biodiversity institutions to identify records from e.g. models: it includes functions for getting microclimate data from ERA5
horticulture or captivity. climate data. specieschrom (Kléparski and Beaugrand, 2022) quantifies
In 2020, the R version of ENMTools (Warren et al., 2010) was finally niche optimum, breadth and overlap, and identifies the most discrimi­
released (Warren et al., 2021). The ENMTools R package not only pro­ nant combination of environmental variables. bdc (Ribeiro et al., 2022)
vides tools and statistical tests for comparing species niches but also includes functions for merging biodiversity datasets; flagging and
offers a full set of tools for the four modelling stages. ENMTools includes removing erroneous occurrence data; cleaning, parsing and harmo­
algorithms for presence-background (Maxent, Poisson point process) nizing scientific names; and flagging and correcting inconsistent
and presence-absence (GAM, GLM) models, and unique evaluation tools, collection dates.
such as Monte Carlo tests and null models. ENMTools is the only pack­
age, to our knowledge, where null models are calculated at the same 3.2. An organised list of packages
time as the empirical models, as a parameter of the algorithm function.
One important difference with other full-set packages, such as biomod2 We present the list of packages in Table 2, organised by the four main
(Thuiller et al., 2021) or sdm (Naimi and Araújo, 2016), is that ENMTools modelling steps: data collection and preparation, model calculation,
has independent functions for each algorithm (similarly to dismo, Hij­ model evaluation and validation, and model application. This list will be
mans et al., 2021), instead of one main model computing function where curated and frequently updated at https://sites.google.com/view/s
all algorithms are included as parameters. blab/r-packages-for-enms.
ENMTools was not the only full-set package released in 2020: Many packages can tackle more than one aim, but they are placed in
ENMTML (de Andrade et al., 2020) provides a single function with all our table in only one of the modelling steps, according to its main
arguments necessary to calculate ENMs, including variable collinearity purpose. We considered the following main purposes in each modelling
control, bias control, accessible area delimitation, pseudo-absence step: (1) data collection and preparation: compilation of biodiversity
allocation, data partition, 13 modelling algorithms (Bioclim, Mahala­ and geographical data, species filtering, preprocessing data for models,
nobis Distance, Domain, ENFA, GLM, GAM, BRT, Random Forests, SVM, creating virtual species; (2) model calculation: mechanistic models,
Maximum Entropy with quadratic and linear and default features, correlative models, ensemble modelling, Bayesian models, functions
Maximum Likelihood, and Gaussian Process), classification thresholds, complementary to other packages, Maxent specific, ENFA specific, niche
model evaluation metrics, over-prediction analysis, ensemble model­ estimation; (3) model evaluation and validation; (4) model application:
ling, and projections over time and space. In this way, ENMTML aims to comparison of models, dispersion, utilities. We indicate in the table the
simplify the entire modelling process by reducing it to a single main purpose of each package, followed by a small description, main
command. contributor(s), main reference, CRAN link, and development website (if
Five more packages were released in 2020. blockCV (Valavi et al., available).
2019) provides functions for generating spatially or environmentally
separated folds for k-fold cross-validation of models. blockCV can be 4. Discussion
used to separate training and test data for other modelling packages.
Similarly, enmSdm (Morelli et al., 2020) is a complement to dismo The number of packages for ENMs has increased over time. Since
(Hijmans et al., 2021), with advanced functions for preparing data, 2010, at least one package has been released on CRAN every year, except
training and evaluating dismo models, and comparing ecological niches. in 2013. The number of packages will likely continue to increase.
embarcadero (Carlson, 2020) calculates ENMs with Bayesian additive Therefore, do we have all the packages we need? Considering our
regression trees (BART), including basic model summary statistics and organised list, some functions for ENMs may be redundant. A varied set
diagnostics, variable importance measures, and plotting. modleR of ENMs can be computed and validated currently through several
(Sánchez-Tapia et al., 2020) provides a modelling workflow based on packages: dismo, biomod2, sdm, ENMeval, ENMTools, ENMTML, flexsdm,
dismo. ssdm (Schimtt et al., 2017) calculates potential species richness SDMtune and sdmbench provide multiple algorithms, for presence-
from a set of stacked individual models. absence, presence-background, or presence-only data. ssdm (Schmitt
Four additional packages were released in 2021. geodata (Hijmans et al., 2017) also provides all the necessary functions for computing and
et al., 2022) downloads geographic and environmental data directly into validating ENMs, but it is designed specifically to calculate potential
R, including species occurrences and variables like climate (from species richness from stacked individual ENMs. Several packages for

4
N. Sillero et al. Ecological Modelling 476 (2023) 110242

Table 2
Curated list of R packages for ecological niche models organised by their main purpose.
Data collection and preparation
N R package Main purpose Description Main contributor(s) Main reference CRAN link Development Citations
site

1 sdmpredictors Geographical Functions for getting Samuel Bosch; Lennert https://cran.r-p https://lifewat 8
data terrestrial and marine Tyberghein; Olivier De roject.org ch.github.
compilation predictors, including Clerck; Salvador /package=sdmp io/sdmpredict
WorldClim, ENVIREM, Fernandez; Lennert redictors ors/
Bio-ORACLE and Schepers
MARSPEC.
2 geodata Geographical Functions for getting Robert J. Hijmans; https://cran.r-pr NA
data geographic data, Aniruddha Ghosh; Alex oject.org/pac
compilation including climate, Mandel kage=geodata
elevation, land use,
soil, species
occurrence,
accessibility, and
administrative
boundaries.
3 mcera5 Geographical Functions for getting James P. Duffy Klinges et al., Not in CRAN https://github. 0
data microclimate data for Dave Klinges 2022 com/dklinge
compilation mechanistic models Lydia Soifer s9/mcera5
from ERA5 climate
data.
4 rgbif Species data Interface to the Global Scott Chamberlain; https://cran.r-p 8
compilation Biodiversity Damiano Oldoni; Vijay roject.org/p
Information Facility Barve; Peter Desmet; ackage=rgbif
(GBIF) API. Laurens Geffert; Dan
Mcglinn; Karthik Ram
rOpenSci; John Waller
5 spThin Model Functions for spatial Matthew E. Aiello- Aiello-Lammens https://cran. 884
preparation thinning of species Lammens; Robert A. et al., 2015 r-project.org/pac
occurrence records. Boria; Aleksandar kage=spThin
Radosavljevic; Bruno
Vilela; Robert P.
Anderson; Robert
Bjornson; Steve Weston
6 biogeo Data Assessing and Robertson, MP; Visser, Robertson et al., https://cran. https://github. 61
preparation improving data quality V; Hui, C 2016 r-project.org/we com/cr
of occurrence record b/packages/bio an/biogeo
datasets geo/index.html

7 CoordinateCleaner Data Standardised cleaning Zizka et al., 2019 https://cran.r-pr 285

preparation of occurrence records oject.org/packa
from biological ge=Coordinate
collection databases Cleaner
8 bdc Data Standardising, Ribeiro, BR; Velazco, Ribeiro et al., https://cran.r-pr https://github. 3
preparation and integrating, and SJE; Guidoni-Martins, K; 2022 oject.or com/brunob
cleaning cleaning biodiversity Tessarolo, G; Jardim, g/web/packages rr/bdc/issues
data Lucas; Bachman, SP; /bdc/index.html
Loyola, R
9 bRacatus Data A method to estimate Arlé, E; Zizka, A; Keil, P; Arlé et al., 2021 https://cran.r-pr https://github. 2
preparation the accuracy and Winter, M; Essl, F; oject.org/web/ com/E
biogeographical status Knight, T; Weigelt, P; packages/b duardoArle
of georeferenced Jiménez-Muñoz, M; Racatus/index. /bRacatus
biological data Meyer, C html

10 bdclean Data User-friendly Gueta, T; Barve, V; Gueta et al., 2018 https://cran. https://github. 17
preparation biodiversity data Nagarajah, T; Agrawal, r-project.org/we com/bd-
cleaning pipeline A; Carmel, Y b/packages/bdc R/bdclean
lean/index.html
11 SPEDInstabR Data Selection of Cástor Guisande Guisande et al., https://cran. 30
preparation environmental González 2017 r-project.org/p
variables based on a ackage=SP
fluctuation index EDInstabR
12 occAssess Data Assessing potential Boyd et al., 2021 https://github. 5
preparation biases in species com/robbo
occurrence data yd/occAssess/
13 sampbias Data Quantifying Zizka, A; Antonelli A; Zizka et al., 2021 Not in CRAN https://github. 25
preparation geographic sampling Silvestro, D com/azizka/sa
biases in species mpbias
distribution data
14 MinBAR Model Functions for defining Xavier Rotllan-Puig https://cran. https://github. 3
preparation the minimum Anna Traveset r-project.org/pac com/xavi-rp
background area kage=MinBAR /MinBAR
15 SDMselect Not in CRAN 5
(continued on next page)

5
N. Sillero et al. Ecological Modelling 476 (2023) 110242

Table 2 (continued )
Data collection and preparation
N R package Main purpose Description Main contributor(s) Main reference CRAN link Development Citations
site

Model Functions for covariate Sébastien Rochette https://github.

preparation selection and model Travis CI com/statnm
specifications with ap/SDMSelect
cross-validation and
forward model
selection.
16 MIAmaxent Data Functions for variable Julien Vollering; Sabrina Vollering et al., https://cran. https://github. 28
preparation transformation and Mazzoni; Rune 2019 r-project.org/pa com/julien
selection, model Halvorsen Steven ckage=MIAma vollering/
evaluation and Phillips xent MIAmaxent
projection training of
maximum entropy and
standard logistic
regression algorithms.
17 sdmvspecies Virtual species Functions for creating Xiaoquan Kong; Renyan Duan et al., 2015 https://cran. 25
computation virtual species Duan; Minyi Huang r-project.org/p
ackage=sdmvspe
cies
18 virtualspecies Virtual species Functions for creating Boris Leroy; Christine N. Leroy et al., 2016 https://cran.r-pr 165
computation virtual species Meynard; Celine Bellard; oject.org/
Franck Courchamp; package=virtu
Robin Delsol; Willson alspecies
Gaul
Model calculation
N R package Main purpose Description Main contributor(s) Main reference CRAN link Development Citations
site
19 NicheMapR Model Framework for Michael Kearney; Rafael Kearney and Not in CRAN https://mrke. 222
computation microclimate and Schouten; Ilya Maclean Porter, 2017, github.io/
mechanistic niche 2020 https://github.
models. com/mrke/N
icheMapR/
20 dismo Model Algorithms for Robert J. Hijmans; Hijmans et al., https://cran.r-p https://github. 564
computation presence-only Steven Phillips; John 2021 roject.org/p com/rspat
(Bioclim, Domain, Leathwick; Jane Elith ackage=dismo ial/dismo
Mahalanobis distance),
presence-background
(Maxent), and
presence-absence
(BRT).
Functions for
comparing models.
21 enmSDM Data Complement to dismo. Adam B. Smith; Ahmed Not in CRAN https://github. NA
preparation Functions for El-Gabbas com/adamlili
Model preparing data, th/enmSdm
preparation training and evaluating
Model dismo models, and
evaluation comparing ecological
Model niches.
comparison
22 biomod2 Model Algorithms for Wilfried Thuiller; Thuiller et al., https://cran.r-pr 1916
computation presence-only (SRE), Damien Georges; Maya 2003, 2009, W oject.org/pac
presence-background Gueguen; Robin Engler; 2021 kage=biomod2
(Maxent) and Frank Breiner
presence-absence
(GAM, GLM, BRT,
ANN).
Functions for
calibration and
evaluation, ensemble
of models, ensemble
forecasting and
visualisation.
23 sdm Model Algorithms for Babak Naimi; Miguel B. Naimi and https://cran.r-p 473
computation presence-background Araujo Araújo, 2016 roject.org
(Maxent) and /package=sdm
presence-absence
(GAM, GLM, MARS,
SVM, RF).
Functions for ensemble
modelling, calibration
and evaluation, and
projections over time
and space.
24 ENMTools 68
(continued on next page)

6
N. Sillero et al. Ecological Modelling 476 (2023) 110242

Table 2 (continued )
Data collection and preparation
N R package Main purpose Description Main contributor(s) Main reference CRAN link Development Citations
site

Model Algorithms for Dan Warren; Nick Warren et al., https://cran. https://github.
computation presence-background Matzke; Marcel Cardillo; 2021 r-project. com/dan
Model (Maxent) and John Baumgartner; org/packa lwarren/
comparison presence-absence Linda Beaumont; ge=ENMTools ENMTools
(GAM, GLM). Nicholas Huron;
Functions for Monte Marianna Simoes;
Carlo tests and null Teresa Iglesias; Russell
models. Dinnage
25 FLEXSDM Model Algorithms for Santiago José Elías Velazco et al., Not in CRAN 0
preparation presence-background Velazco; Miranda 2022
Model (Maxent) and Brooke Rose; André
computation presence-absence Felipe Alves de Andrade;
(GAM, GLM, GBM, Ignacio Minoli; Janet
SVM, RF, ANN). Franklin
Functions for sampling
bias correction,
sampling pseudo-
absences and
background points,
data partitioning, and
reducing collinearity in
predictors; fitting and
evaluating models,
ensembles of small
models and ensemble
models; models’
predictions,
interpolation and
overprediction
correction.
26 SDMtune Model Functions for training Sergio Vignali https://cran.r-pr 43
computation and testing of Maxent, Arnaud Barras oject.org/pac
Random Forest, neural Veronika Braunisch kage=SDMtune
networks, and BRT.
27 sdmbench Model Functions for Boyan Angelov Not in CRAN https://github. 5
preparation compiling species com/bo
Model occurrence and yanangelov
computation environmental data; /sdmbench
spatial data
partitioning; and
machine learning
algorithms.
28 kuenm Support for Functions for model Marlon E. Cobos M.E. Cobos et al., Not in CRAN https://github. 294
Maxent calibration of Maxent Luis Osorio 2019 com/marlo
models by selection of necobos/
optimal kuenm
parameterizations for
each study.
29 rmaxent Maxent Implementation of John Baums Not in CRAN https://github. 13
implementation tasks related to Maxent com/joh
in models R. nbaums/rm
axent
30 maxnet Maxent Fitting ’Maxent’ Steven Phillips https://cran. NA
implementation models with ’glmnet’. r-project.org/pac
kage=maxnet
31 ENIRG – Ecological Model Interface for R and Fernando Canovas; Cánovas et al., Not in CRAN https://rdrr. 12
Niche in R-GRASS – computation GRASS to overcome Chiara Magliozzi; Jose 2016 io/cran
ENFA issues when working Antonio Palazon- /ENiRG/man
with large data sets. Ferrando; Frederico /ENiRG-pac
Functions from rgrass7 Mestre; Mercedes kage.html
package. Gonzalez-Wanguemert
32 maxlike Model Maxlike Richard Chandler; Andy Chandler and https://cran.r-pr 3
computation implementation, a Royle; Roeland Kindt Royle, 2013; oject.org/pac
likelihood-based Fitzpatrick et al., kage=maxlike
approach for 2013
modelling with
presence-only data.
Maxlike yields
estimates of
occurrence
probability.
33 SDMPlay Model Algorithms for Guillaumot Charlene; Guillaumot et al., https://cran.r-pr 2
computation presence-background Martin Alexis; Eleaume 2017 oject.org/pac
(Maxent) and kage=SDMPlay
(continued on next page)

7
N. Sillero et al. Ecological Modelling 476 (2023) 110242

Table 2 (continued )
Data collection and preparation
N R package Main purpose Description Main contributor(s) Main reference CRAN link Development Citations
site

presence-absence Marc; Danis Bruno;

(BRT). Saucede Thomas
Functions for
calibration,
evaluation, and
visualisation.
Designed specifically
for marine
environments.
34 hSDM – Hierarchical Model Functions for Ghislain Vieilledent; Latimer et al., https://cran.r-pr 3
Bayesian Species computation estimating parameters Matthieu Autier; Alan E. 2006 oject.org/
Distribution Models of hierarchical Gelfand; Jérôme Guélat; package=hSDM
Bayesian distribution Marc Kéry
and abundance Andrew M. Latimer;
models. Cory Merow; Frédéric
Mortier; John A.
Silander Jr.; Adam M.
Wilson; Shanshan
Wurinnan
35 CENFA – Climate and Model Algorithms for climate- D. Scott Rinnan Rinnan and https://cran.r-p 33
Ecological Niche computation and ecological-niche Lawler, 2019 roject.org/p
Factor Analysis factor analyses. ackage=CENFA
Functions for
visualisation of spatial
variability of species
sensitivity, exposure,
and vulnerability to
climate change.
It supports processing
of large files and
parallel methods.
36 ENMTML Model ENMTML provides a André F.A. Andrade; de Andrade et al., Not in CRAN https 59
computation single function with all Santiago J.E. Velazco; 2020 ://andrefaa.gi
arguments necessary to Paulo De Marco thub.
calculate ENMs: io/ENMTML/
variable collinearity
control, bias control,
accessible area
delimitation, pseudo-
absence allocation,
data partition, several
algorithms, thresholds,
evaluation metrics,
over-prediction,
ensemble modelling,
and projections over
time and space.
37 embarcadero Model Functions to calculate Colin J. Carlson; Seth Carlson, 2020 Not in CRAN https://github. 34
computation ENMs with Bayesian Musker com/cjc
additive regression arlson/embarc
trees (BARTs), adero/
including basic model
summary statistics and
diagnostics, variable
importance measures,
and plotting.
38 modleR Model A workflow for Andrea Sánchez-Tapia, Sánchez-Tapia https://cran. https://github. 6
computation ecological niche Sara Ribeiro Mortara, et al., 2020 r-project.org/pac com/Model
models based on Diogo Souza Bezerra kage=modelr -R/modleR
’dismo’. Rocha, Felipe Sodré
Mendes Barros,
Guilherme Gall, Marinez
Ferreira de Siqueira
39 hypervolume Niche Functions for Benjamin Blonder; Blonder et al., Not in CRAN 482
calculation estimating shapes and David J. Harris 2014
volumes of high-
dimensional niches
using kernel density
estimation, support
vector machine
delineation, and
convex hulls.
40 ssdm Niche Calculates potential Sylvain Schmitt Schimtt et al., 96
calculation species richness from Robin Pouteau 2017
(continued on next page)

8
N. Sillero et al. Ecological Modelling 476 (2023) 110242

Table 2 (continued )
Data collection and preparation
N R package Main purpose Description Main contributor(s) Main reference CRAN link Development Citations
site

stacked individual Dimitri Justeau https://cran.r-pr

ENMs. Florian de Boissieu oject.org/
Lukas Baumbach package=SSDM
Philippe Birnbaum
41 itsdm Model Functions for presence- Lei Song; Lyndon Estes Song, 2022 https://cran.r-p NA
calculation only species roject.org/p
distribution modelling ackage=itsdm
with isolation forest
(iForest) and its
variations, such as
Extended isolation
forest and SCiForest.
Model evaluation and validation
N R package Main purpose Description Main contributor(s) Main reference CRAN link Development Citations
site
42 PresenceAbsence Model Functions for Elizabeth Freeman Freeman and https://cran.r-pr 465
evaluation evaluating presence- Moisen, 2007 oject.org/web/
absence analysis, packages/Pr
either ENMs or esenceAbsence/i
diagnostic tests. ndex.html
Functions for
calculating thresholds
and AUC.
43 ENMeval Model Functions for Jamie M. Kass; Robert Muscarella et al., https://cran.r-pr 1230
evaluation calculating models Muscarella; Peter J. 2014, 2020 oject.org/pac
with all combinations Galante; Corentin Bohl; kage=ENMeval
of settings, partition Gonzalo E. Buitrago-
data spatially, cross Pinilla; Robert A. Boria;
validation evaluation, Mariano Soley-Guardia;
model selection based Robert P. Anderson
on balance goodness-
of-fit and model
complexity, null
models.
44 SDMTools – Species Model Functions for post Jeremy VanDerWal; Vanderwal et al., Not in CRAN 46
Distribution Modelling evaluation processing and Lorena Falconi; 2015
Tools visualising model Stephanie Januchowski;
outcomes, selecting Luke Shoo; Collin Storlie
thresholds, calculating
accuracy metrics and
landscape
fragmentation
statistics, comparing
models, tracking
changes in
distributions over
time.
45 Metrics Model Functions for Ben Hamner; Michael Hamner and https://cran.r-pr 4
evaluation calculating evaluation Frasco; Erin LeDell Frasco, 2018 oject.org/pac
metrics of machine kage=Metrics
learning models.
46 modEvA – Model Model Functions for A. Márcia Barbosa Barbosa et al., https://cran. https:// 123
evaluation and evaluation performing variation 2016 r-project.org/pac modtools.
analysis partitioning, kage=modEvA wordpress.
discrimination and com/package
calibration metrics, s/modeva/
optimizing prediction
thresholds, performing
multivariate
environmental
similarity surface
(MESS) analysis, and
displaying analytical
plots.
47 usdm – Uncertainty Model Functions for assessing Babak Naimi Naimi et al., 2014 https://cran.r-pr 792
analysis for species evaluation the impact of different oject.org/
distribution models sources of package=usdm
uncertainties on model
performance.
48 blockCV Model Functions for Roozbeh Valavi; Jane Valavi et al., https://cran.r-pr 236
preparation generating spatially or Elith; José Lahoz- 2019 oject.org/pac
Model environmentally Monfort; Gurutzeta kage=blockCV
evaluation separated folds for k- Guillera-Arroita
(continued on next page)

9
N. Sillero et al. Ecological Modelling 476 (2023) 110242

Table 2 (continued )
Data collection and preparation
N R package Main purpose Description Main contributor(s) Main reference CRAN link Development Citations
site

fold cross-validation of
models.
Model application
N R package Main purpose Description Main contributor(s) Main reference CRAN link Development Citations
site
49 ecospat Model Functions for spatial Olivier Broennimann; Di Cola et al., https://cran.r-pr 556
comparison ecology analyses on Valeria Di Cola; Blaise 2017 oject.org/pac
pre, core and post Petitpierre; Frank kage=ecospat
modelling analyses of Breiner; Daniel Scherrer;
species distribution, Manuela D`Amen;
niche quantification Christophe Randin;
and community Robin Engler; Wim
assembly. Hordijk; Heidi Mod;
Julien Pottier; Mirko Di
Febbraro; Loic Pellissier;
Dorothea Pio; Ruben
Garcia Mateo; Anne
Dubuis; Luigi Maiorano;
Achilleas Psomas;
Charlotte Ndiribe;
Nicolas Salamin; Niklaus
Zimmermann; Flavien
Collart; Antoine Guisan
50 ellipsenm Model Functions for ellipsoid Marlon E. Cobos; Luis Not in CRAN https://github. 3
comparison envelopes, calibrating Osorio; Jorge Soberón; com/mar
and selecting models, A. Townsend Peterson; lonecobos/
replicating models and Vijay Barve; Narayani ellipsenm
projections, and Barve
assessing niche
overlap.
51 fuzzySim – Fuzzy Model Functions to compute A. Márcia Barbosa Barbosa, 2015 https://cran. http://fu 102
Similarity in Species combination fuzzy versions of r-project. zzysim.r-forge.
Distributions species occurrence org/packa r-project.org/
patterns based on ge=fuzzySim
presence-absence data
(including inverse
distance interpolation,
trend surface analysis,
and prevalence-
independent
favorability), and
pairwise fuzzy
similarity.
Additional functions
for model consensus
and comparison, and
for data preparation:
unique abbreviations
of species names,
gridding (thinning)
point occurrence data
onto raster maps,
converting species lists
to presence-absence
tables, transposing part
of a data frame,
selecting relevant
variables for models,
assessing the false
discovery rate, or
analysing and dealing
with multicollinearity.
52 zoon Model Functions for Tom August; Tim Lucas; Golding et al., https://cran.r-pr 39
comparison reproducible and Nick Golding; Emiel van 2018 oject.org/
shareable analysis of Loon; Greg McInerny package=zoon
models from an online
repository, comparing
and diagnostic models.
53 Mopa Model ensemble Functions for Maialen Iturbide; Iturbide et al., Not in CRAN 19
transferable models, Joaquin Bedia; 2018
pseudo-absence data SantanderMetGroup
generation, and factor
contribution
quantification.
(continued on next page)

10
N. Sillero et al. Ecological Modelling 476 (2023) 110242

Table 2 (continued )
Data collection and preparation
N R package Main purpose Description Main contributor(s) Main reference CRAN link Development Citations
site

54 MigClim Model Functions from Robin Engler; Wim 2012 Not in CRAN 144
dispersion implementing Hordijk; Loic Pellissier Not in CRAN
dispersal into models
over time.
55 RangeShifteR Model Spatially-explicit, Anne-Kathleen Bocedi et al., Not in CRAN https://github. 210
dispersion individual-based Malchow; JetteReeg; 2014, G. 2021 com/RangeSh
simulation platform for Damaris Zurell ifter/RangeSh
modelling species’ iftR-package
range dynamics, such
as expansion and
shifting, and patch
connectivity by linking
complex local
population dynamics
and dispersal
behaviour, while also
taking into account
inter-individual
variability and
evolutionary
processes.
56 ESDM Model Functions for ensemble Sam Woodman Woodman et al., https://cran.r-pr 10
application modelling of species 2019 oject.org/
distribution and package=eSDM
abundance models.
eSDM includes a GUI
(R Shiny app).
57 rangeModelMetadata – Publication of Functions for creating Cory Merow; Brian Merow et al., https://cran. 17
Range modeling metadata data dictionaries MaitnerHannah Owens; 2019 r-project.
Metadata Standards specifying a Jamie Kass; org/package=ra
hierarchical structure Brian Enquist; Rob ngeModelMeta
to catalogue models. Guralnik; Damaris data
Zurrell; Christian Koenig
58 specieschrom Model Quantification of niche Loïck Kléparski Kléparski and Not in CRAN https://github. 1
comparison optimum and breadth, Beaugrand, 2022 com/loic
niche overlapping k-klpr/specie
(index D), and schrom
identification of most
discriminant
combination of
environmental
variables.
59 humboldt Model Functions to analyse Jason L. Brown Brown and Not in CRAN https://jaso 43
comparison the amount of the Carnaval, 2019 nleebrown.git
species’ observed E- hub.io/humbo
space that is truncated ldt/
by the available E-
space.
60 red Model Functions for Pedro Cardoso Cardoso, 2017 https://cran.r-p 42
application performing spatial roject.org
analyses using species /package=red
occurrences or
estimated ranges. It
calculates Maxent
models with dismo
function.

cleaning and filtering occurrence data are also available: spThin, bdclean, computing null models (Raes and ter Steege, 2007; Beale et al., 2008;
CoordinateCleaner, bdc, bRacatus, biogeo. occAssess and sampbias are Bohl et al., 2019). Only ENMTools (Warren et al., 2021) has imple­
specialised in correcting geographical biases in species occurrences. mented the calculation of null models as a parameter of the functions for
Some algorithms are available through specialised and generalistic computing ENM algorithms. ENMeval (Muscarella et al., 2014; Kass
packages: for example, the Maxent algorithm (Phillips et al., 2006, et al., 2021) also provides functions for computing null models. It is
2017) can be computed with maxnet, kuenm, dismo, biomod2, sdm, desirable that validating ENMs with null models becomes a more
ENMeval, ENMTools, flexsdm, SDMPlay, ENMTML, SSDM, MinBAR, red, extended and standard practice (Sillero et al., 2021). There is also room
and SDMtune. However, many of these packages use a function included for improvement in properly evaluating models on true presence-only
in another package (e.g. dismo). On the other hand, are there some data, without requiring unoccupied background localities.
missing functionalities not included in R packages for ENMs? The user If we consider the necessity to gather geographical, environmental
has currently at least 60 available packages providing multiple functions and biodiversity data directly from within R, we will need several
for all modelling steps. However, not all modelling steps are represented packages to run all modelling procedures. However, once the data are
equally: fewer packages are available for evaluating models. In general, compiled, there are several packages offering all necessary functions to
we probably need more functions for validating models, especially for compute, validate and apply the models, namely dismo, biomod2, sdm,

11
N. Sillero et al. Ecological Modelling 476 (2023) 110242

ENMeval, ENMTools, ENMTML, flexsdm, and SDMtune. However, not all published in a scientific journal, even if they were not on CRAN; they
these packages provide the same functionalities. For example, dismo need to be installed following instructions on their development
(Hijmans et al., 2021) does not have ensemble modelling functions, website.
while ENMTools (Warren et al., 2021) does not provide tools to replicate
the models. On the other hand, ENMTools (Warren et al., 2021) validates 5. Conclusions
the models both in the environmental and geographical space, and it
includes the computing of null models as part of the general calculation What may be the future of R packages for ENMs? This question is not
of the model. Recent packages have also reduced the number of steps for easy to answer. As indicated above, packages should implement better
calculating the models. In comparison to biomod2 (Thuiller et al., 2021), evaluation metrics and validating procedures such as null models, and
which requires formatting the data through several functions before improve their communication and integration. With the migration to
computing the models, other packages have reduced the necessary steps modern spatial packages like terra (at least for raster spatial data), we
to a minimum: sdm (Naimi and Araújo, 2016) only needs one function hope that computing speed and stability will increase. Also, packages
for this pre-modelling step, and ENMTML (de Andrade et al., 2020) in­ should facilitate the computation of replicated models for large numbers
tegrates all modelling steps in one single function. of species, without the necessity of using for-loop functions, and provide
The most cited packages are biomod2 (Thuiller et al., 2021) and a more optimised syntax concerning the R language.
ENMeval (Muscarella et al., 2014; Kass et al., 2021). This is likely a
representation of the users’ needs: most users require a package for both Funding
computing and validating models. Some packages such as dismo (Hij­
mans et al., 2021) have fewer citations, probably because they do not This research did not receive any specific grant from funding
have a paper associated. Users frequently forget to cite R packages, or agencies in the public, commercial, or not-for-profit sectors.
they do not know how to cite them, as their citations cannot be seam­
lessly uploaded to a reference manager program.
The communication between packages should also be improved. R is Declaration of Competing Interest
a free open-source software enhanced by a wide community of de­
velopers and users, and therefore it includes a huge diversity of ap­ Joao Carlos Campos reports financial support was provided by FCT -
proaches (Lovelace et al., 2020). For example, several packages provide Fundação para a Ciência e a Tecnologia. Salvador Arenas-Castro reports
tools for reading, importing, and exporting data, either in tabular (ma­ financial support was provided by Spanish Ministry of Universities.
trix, data frame, data table, tibble) or in spatial (vector or raster) formats Salvador Arenas-Castro reports financial support was provided by EU-
(sp, sf, raster, terra, stars, amongst others). Different packages often NextGenerationEU.
require different classes of input objects and also produce output objects
of these classes. This reduces the flexibility for sharing objects between Data availability
functions from different packages. This is important, for example, if we
want to apply spatial statistics (Baddeley et al., 2015) to the results of No data was used for the research described in the article.
widely used packages, such as dismo (Hijmans et al., 2021) and biomod2
(W Thuiller et al., 2021). These two packages currently require input
objects implemented by packages raster (Hijmans, 2022a) and sp Acknowledgements
(Pebesma and Bivand, 2005), which use the rgdal (Bivand et al., 2022)
and rgeos (Bivand and Rundel, 2021) packages for importing and NS is supported by a CEEC2017 contract (CEECIND/02213/2017)
handling spatial data. However, spatstat (Baddeley and Turner, 2005) - from Fundação para a Ciência e a Tecnologia (FCT). SAC is supported by
the most complete package for spatial statistics, currently depends on sp the María Zambrano fellowship program funded by the Spanish Ministry
(Pebesma and Bivand, 2005) and sf (Pebesma, 2018). Therefore, inter­ of Universities and the EU-NextGenerationEU fund. JCC is supported by
connection and communication amongst packages should be improved. a contract from MontObEO project (MTS/BRB/0091/2020), funded by
Most ENM packages currently use sp (Pebesma and Bivand, 2005) FCT.
and raster (Hijmans, 2022a) package formats as spatial data inputs.
However, support for some dependencies or suggestions of these pack­ References
ages (such as rgdal, rgeos and maptools) will end in 2023. Currently, very
few ENM packages (fuzzySim, modEvA, blockCV, dismo in part, and Aiello-Lammens, M.E., Boria, R.A., Radosavljevic, A., Vilela, B., Anderson, R.P., 2015.
biomod2 in the development version on GitHub) have migrated their spThin: an R package for spatial thinning of species occurrence records for use in
ecological niche models. Ecography 38, 1–35. https://doi.org/10.1111/ecog.01132.
code to newer spatial packages like terra (for both raster and vector data; Angelov, B., 2018. sdmbench: R package for benchmarking species distribution models.
Hijmans, 2022b) or sf (for vector data; Pebesma, 2018), which substitute J. Open Source Softw. 3, 847. https://doi.org/10.21105/joss.00847.
the raster and sp packages and are actively maintained, besides being Arlé, E., Zizka, A., Keil, P., Winter, M., Essl, F., Knight, T., Weigelt, P., Jiménez-
Muñoz, M., Meyer, C., 2021. bRacatus : a method to estimate the accuracy and
considerably faster. Future updates of R can introduce problems be­ biogeographical status of georeferenced biological data. Methods Ecol. Evol. 12,
tween versions of the packages as older spatial packages are retired. 1609–1619. https://doi.org/10.1111/2041-210X.13629.
Therefore, code migration to modern spatial packages (e.g. terra or stars Baddeley, A., Rubak, E., Turner, R., 2015. Spatial Point Patterns: Methodology and
Applications with R. Chapman and Hall/CRC, New York, USA.
for raster data, terra or sf for vector data) should be a current priority for
Baddeley, A., Turner, R., 2005. spatstat: an R package for analyzing spatial point
ENM package developers. patterns. J. Stat. Softw. 12, 1–42.
Other packages provide functions (e.g. model computation or vali­ Barbosa, A.M., 2015. fuzzySim: applying fuzzy logic to binary similarity indices in
ecology. Methods Ecol. Evol. 6, 853–858. https://doi.org/10.1111/2041-
dation) for methods rarely used in ENMs. Also, other packages use ENMs
210X.12372.
to perform phylogeographic (e.g. phyr, Li et al. 2020; phylin, Tarroso et Barbosa, A.M., Brown, J.A., Jimenez-Valverde, A., Rea, l R., 2016. R package “modEvA.”
al, 2015) and biogeographical analyses (e.g. rangemap; Cobos et al. R CRAN repos. Rep 45. https://doi.org/10.1111/ddi.12100.
2022). Baumgartner J., Wilson P. (2022). rmaxent: tools for working with Maxent in R. R
package version 0.8.5.9000. https://github.com/johnbaums/rmaxent.
Some packages are in development and not currently on CRAN (e.g. Beale, C.M., Lennon, J.J., Gimona, A., 2008. Opening the climate envelope reveals no
SDMTools, VanderWal et al. 2015; MigClim, Engler et al.,2012; embar­ macroscale associations with climate in European birds. Proc. Natl. Acad. Sci. U. S.
cadero, Carlson 2020). More and more packages (even widely used ones) A. 105, 14908–14912. https://doi.org/10.1073/pnas.0803506105.
Bivand R., Keitt T., Rowlingson B. (2022). rgdal: bindings for the ’Geospatial’ data
are now maintained on GitHub instead of CRAN, as it is often difficult to abstraction library_. R package version 1.5-32. https://CRAN.R-project.org/p
meet CRAN requirements. We included some packages if they were ackage=rgdal.

12
N. Sillero et al. Ecological Modelling 476 (2023) 110242

Bivand R., Rundel C. (2021). rgeos: interface to geometry engine open source (’GEOS’). R Fick, S.E., Hijmans, R.J., 2017. WorldClim 2: new 1-km spatial resolution climate
package version 0.5-9. https://CRAN.R-project.org/package=rgeos. surfaces for global land areas. Int. J. Climatol. 37, 4302–4315. https://doi.org/
Blonder, B., Lamanna, C., Violle, C., Enquist, B.J., 2014. The n-dimensional 10.1002/joc.5086.
hypervolume. Glob. Ecol. Biogeogr. 23, 595–609. https://doi.org/10.1111/ Franklin, J., 2010. Mapping Species Distributions. Cambridge University Press. https://
geb.12146. doi.org/10.1017/CBO9781107415324.004.
Bocedi, G., Palmer, S.C.F., Malchow, A., Zurell, D., Watts, K., Travis, J.M.J., 2021. Freeman, E.A., Moisen, G., 2008. presenceabsence: an R package for presence absence
RangeShifter 2.0: an extended and enhanced platform for modelling spatial eco- analysis. J. Stat. Softw. 23, 238. https://doi.org/10.18637/jss.v023.i11.
evolutionary dynamics and species’ responses to environmental changes. Ecography Golding, N., August, T.A., Lucas, T.C.D., Gavaghan, D.J., Loon, E.E., McInerny, G., 2018.
44, 1453–1462. https://doi.org/10.1111/ecog.05687. The zoon package for reproducible and shareable species distribution modelling.
Bocedi, G., Palmer, S.C.F., Pe’er, G., Heikkinen, R.K., Matsinos, Y.G., Watts, K., Travis, J. Methods Ecol. Evol. 9, 260–268. https://doi.org/10.1111/2041-210X.12858.
M.J., 2014. RangeShifter: a platform for modelling spatial eco-evolutionary Gueta, T., Barve, V., Nagarajah, T., Agrawal, A., Carmel, Y., 2018. Introducing bdclean: a
dynamics and species’ responses to environmental changes. Methods Ecol. Evol. 5, user friendly biodiversity data cleaning pipeline. Biodivers. Inf. Sci. Stand. 2,
388–396. https://doi.org/10.1111/2041-210X.12162. e25564. https://doi.org/10.3897/biss.2.25564.
Bohl, C.L., Kass, J.M., Anderson, R.P., 2019. A new null model approach to quantify Guisan, A., Thuiller, W., Zimmermann, N.E., 2017. Habitat suitability and distribution
performance and significance for ecological niche models of species distributions. models: with applications in R, habitat suitability and distribution models: with
J. Biogeogr. 46, 1101–1111. https://doi.org/10.1111/jbi.13573. applications in R. 10.1017/9781139028271.
Booth, T.H., Nix, H.a., Busby, J.R., Hutchinson, M.F., 2014. BIOCLIM: the first species Guisande, C., García-Roselló, E., Heine, J., González-Dacosta, J., Vilas, L.G., García
distribution modelling package, its early applications and relevance to most current Pérez, B.J., Lobo, J.M., 2017. SPEDInstabR: an algorithm based on a fluctuation
MaxEnt studies. Divers. Distrib. 20, 1–9. https://doi.org/10.1111/ddi.12144. index for selecting predictors in species distribution modeling. Ecol. Inform. 37,
Bosch S., Fernandez S. (2022). sdmpredictors: species distribution modelling predictor 18–23. https://doi.org/10.1016/j.ecoinf.2016.11.004.
datasets. R package version 0.2.14. http://lifewatch.github.io/sdmpredictors/. Guo, Q., Liu, Y., 2010. ModEco: an integrated software package for ecological niche
Boyd, R.J., Powney, G.D., Carvell, C., Pescott, O.L., 2021. occAssess: an R package for modeling. Ecography 4, 637–642.
assessing potential biases in species occurrence data. Ecol. Evol. 11, 16177–16187. Hijmans R. (2022a). raster: geographic data analysis and modeling. R package version
https://doi.org/10.1002/ece3.8299. 3.5-29. https://CRAN.R-project.org/package=raster.
Breiner, F.T., Nobis, M.P., Bergamini, A., Guisan, A., 2018. Optimizing ensembles of Hijmans R. (2022b). terra: spatial data analysis. R package version 1.6-17. https://CRAN.
small models for predicting the distribution of species with few occurrences. R-project.org/package=terra.
Methods Ecol. Evol. 2018, 1–7. https://doi.org/10.1111/2041-210X.12957. Hijmans R.J., Ghosh A., Mandel A. (2022). geodata: download geographic data. R
Brown, J.L., 2014. SDMtoolbox: a python-based GIS toolkit for landscape genetic, package version 0.4-11. https://CRAN.R-project.org/package=geodata.
biogeographic and species distribution model analyses. Methods Ecol. Evol. 5, Hijmans R.J., Phillips S., Leathwick J., Elith J. (2021). dismo: species distribution
694–700. https://doi.org/10.1111/2041-210X.12200. modeling_. R package version 1.3-5. https://CRAN.R-project.org/package=dismo.
Brown, J.L., Bennett, J.R., French, C.M., 2017. SDMtoolbox 2.0: the next generation Hirzel, A.H., Hausser, J., Chessel, D., Perrin, N., 2002. Ecological-niche factor analysis:
python-based GIS toolkit for landscape genetic, biogeographic and species how to compute habitat suitability maps without absence-data? Ecology 7,
distribution model analyses. PeerJ 2017. https://doi.org/10.7717/peerj.4095. 2027–2036.
Brown, J.L., Carnaval, A.C., 2019. A tale of two niches: methods, concepts, and Iturbide, M., Bedia, J., Gutiérrez, J.M., 2018. Tackling uncertainties of species
evolution. Front. Biogeogr. 11, 1–27. https://doi.org/10.21425/F5FBG44158. distribution model projections with package mopa. R. J. 10, 122–139.
Cánovas, F., Magliozzi, C., Mestre, F., Palazón, J.A., González-Wangüemert, M., 2016. Kabacoff, R.I., 2015. R in Action: Data Analysis and Graphics With R. Simon and
ENiRG: R-GRASS interface for efficiently characterizing the ecological niche of Schuster.
species and predicting habitat suitability. Ecography 39, 593–598. https://doi.org/ Kass, J.M., Muscarella, R., Galante, P.J., Bohl, C.L., Pinilla-Buitrago, G.E., Boria, R.A.,
10.1111/ecog.01426. Soley-Guardia, M., Anderson, R.P., 2021. ENMeval 2.0: redesigned for customizable
Cardoso, P., 2017. Red - an R package to facilitate species red list assessments according and reproducible modeling of species’ niches and distributions. Methods Ecol. Evol.
to the IUCN criteria. Biodivers. Data J. 5, e20530. https://doi.org/10.3897/BDJ.5. 12, 1602–1608. https://doi.org/10.1111/2041-210X.13628.
e20530. Kass, J.M., Vilela, B., Aiello-Lammens, M.E., Muscarella, R., Merow, C., Anderson, R.P.,
Carlson, C.J., 2020. Embarcadero: species distribution modelling with Bayesian additive 2018. Wallace: a flexible platform for reproducible modeling of species niches and
regression trees in r. Methods Ecol. Evol. 11, 850–858. https://doi.org/10.1111/ distributions built for community expansion. Methods Ecol. Evol. 9, 1151–1156.
2041-210X.13389. https://doi.org/10.1111/2041-210X.12945.
Chamberlain S., Barve V., Mcglinn D., Oldoni D., Desmet P., Geffert L., Ram K. (2022). Kearney, M.R., Porter, W.P., 2017. NicheMapR – an R package for biophysical modelling:
Rgbif: interface to the global biodiversity information facility API_. R package the microclimate model. Ecography 40, 664–674. https://doi.org/10.1111/
version 3.7.0. https://CRAN.R-project.org/package=rgbif. ecog.02360.
Charlene G., Alexis M., Marc E., Bruno D., Thomas S. (2021). SDMPlay: species Kearney, M.R., Porter, W.P., 2020. NicheMapR – an R package for biophysical modelling:
distribution modelling playground. R package version 2.0. https://CRAN.R-project. the ectotherm and dynamic energy budget models. Ecography 43, 85–96. https://
org/package=SDMPlay. doi.org/10.1111/ecog.04680.
Cobos M.E., Osorio-Olvera L., Soberón J., Peterson A.T., Barve V., Barve N. (2019). Kléparski, L., Beaugrand, G., 2022. The species chromatogram, a new graphical method
ellipsem. R package version 0.3.4. https://github.com/marlonecobos/ellipsenm. to represent, characterize, and compare the ecological niches of different species.
Cobos, M.E., Barve, V., Barve, N., Jiménez-Valverde, A., Nuñez-Penichet, C., 2022. Ecol. Evol. 12 https://doi.org/10.1002/ece3.8830.
rangemap: an R package to explore species’ geographic ranges. Biodivers. Inform. 17 Klinges, D.H., Duffy, J.P., Kearney, M.R., Maclean, I.M.D., 2022. mcera5: driving
https://doi.org/10.17161/bi.v17i.16271. microclimate models with ERA5 global gridded climate data. Methods Ecol. Evol. 13,
Cobos, M.E., Townsend Peterson, A., Barve, N., Osorio-Olvera, L., 2019b. Kuenm: an R 1402–1411. https://doi.org/10.1111/2041-210X.13877.
package for detailed development of ecological niche models using Maxent. PeerJ Latimer, A.M., Wu, S., Gelfand, A.E., Silander Jr., J.A., 2006. Building statistical models
2019, 1–15. https://doi.org/10.7717/peerj.6281. to analyze species distributions. Ecol. Appl. 1, 33–50.
de Andrade, A.F.A., Velazco, S.J.E., De Marco Júnior, P., 2020. ENMTML: an R package Leroy, B., Meynard, C.N., Bellard, C., Courchamp, F., 2016. virtualspecies, an R package
for a straightforward construction of complex ecological niche models. Environ. to generate virtual species distributions. Ecography 39, 599–607. https://doi.org/
Model. Softw. 125, 104615 https://doi.org/10.1016/j.envsoft.2019.104615. 10.1111/ecog.01388.
Di Cola, V., Broennimann, O., Petitpierre, B., Breiner, F.T., D’Amen, M., Randin, C., Li, D., Dinnage, R., Nell, L.A., Helmus, M.R., Ives, A.R., 2020. phyr: an R package for
Engler, R., Pottier, J., Pio, D., Dubuis, A., Pellissier, L., Mateo, R.G., Hordijk, W., phylogenetic species-distribution modelling in ecological communities. Methods
Salamin, N., Guisan, A., 2017. ecospat: an R package to support spatial analyses and Ecol. Evol. 11, 1455–1463. https://doi.org/10.1111/2041-210X.13471.
modeling of species niches and distributions. Ecography 40, 774–787. https://doi. Lovelace, R., Nowosad, J., Muenchow, J., 2020. Geocomputation with R, first issued in
org/10.1111/ecog.02671. paperback. ed, The R Series. CRC Press, Taylor & Francis Group, Boca Raton London,
Domisch, S., Amatulli, G., Jetz, W., 2015. Near-global freshwater-specific environmental New York.
variables for biodiversity analyses in 1km resolution. Sci. Data 2, 150073. 10.1038/ Merow, C., Maitner, B.S., Owens, H.L., Kass, J.M., Enquist, B.J., Jetz, W., Guralnick, R.,
sdata.2015.73. 2019. Species’ range model metadata standards: RMMS. Glob. Ecol. Biogeogr. 1–13.
Duan, R., Kong, X., Huang, M., Wu, G., Wang, Z., 2015. SDMvspecies: a software for https://doi.org/10.1111/geb.12993.
creating virtual species for species distribution modelling. Ecography 108–110. Mesgaran, M.B., Cousens, R.D., Webber, B.L., 2014. Here be dragons: a tool for
https://doi.org/10.1111/ecog.01080. quantifying novelty due to covariate range and correlation change when projecting
Elith, J., Kearney, M., Phillips, S., 2010. The art of modelling range-shifting species. species distribution models. Divers. Distrib. 20 https://doi.org/10.1111/ddi.12209.
Methods Ecol. Evol. 1, 330–342. Morelli, T.L., Smith, A.B., Mancini, A.N., Balko, E.A., Borgerson, C., Dolch, R., Farris, Z.,
Elith, J., Phillips, S.J., Hastie, T., Dudík, M., Chee, Y.E., Yates, C.J., 2011. A statistical Federman, S., Golden, C.D., Holmes, S.M., Irwin, M., Jacobs, R.L., Johnson, S.,
explanation of MaxEnt for ecologists. Divers. Distrib. 17, 43–57. https://doi.org/ King, T., Lehman, S.M., Louis, E.E., Murphy, A., Randriahaingo, H.N.T.,
10.1111/j.1472-4642.2010.00725.x. Randrianarimanana, H.L.L., Ratsimbazafy, J., Razafindratsima, O.H., Baden, A.L.,
Engler, R., Hordijk, W., Guisan, A., 2012. The MIGCLIM R package - seamless integration 2020. The fate of Madagascar’s rainforest habitat. Nat. Clim. Change 10, 89–96.
of dispersal constraints into projections of species distribution models. Ecography https://doi.org/10.1038/s41558-019-0647-x.
35, 872–878. https://doi.org/10.1111/j.1600-0587.2012.07608.x. Muscarella, R., Galante, P.J., Soley-Guardia, M., Boria, R.A., Kass, J.M., Uriarte, M.,
Enrique, M., Muñoz, D.S., Giovanni, R.D., Siqueira, M.F.D., Sutton, T., Brewer, P., 2009. Anderson, R.P., 2014. ENMeval: an R package for conducting spatially independent
openModeller: a generic approach to species ’ potential distribution modelling. evaluations and estimating optimal model complexity for Maxent ecological niche
Geoinformatica 1–25. https://doi.org/10.1007/s10707-009-0090-7. models. Methods Ecol. Evol. 5, 1198–1205. https://doi.org/10.1111/2041-
210X.12261.

13
N. Sillero et al. Ecological Modelling 476 (2023) 110242

Naimi, B., Araújo, M.B., 2016. sdm: a reproducible and extensible R platform for species Song L. (2022). itsdm: isolation forest-based presence-only species distribution modeling.
distribution modelling. Ecography 39, 368–375. https://doi.org/10.1111/ R package version 0.1.3. https://CRAN.R-project.org/package=itsdm.
ecog.01881. Tarroso, P., Velo-Antón, G., Carvalho, S.B., 2015. phylin: an R package for
Naimi, B., Hamm, N.A.S., Groen, T.A., Skidmore, A.K., Toxopeus, A.G., 2014. Where is phylogeographic interpolation. Mol. Ecol. Resour. 15, 349–357. https://doi.org/
positional uncertainty a problem for species distribution modelling? Ecography 37, 10.1111/1755-0998.12312.
191–203. https://doi.org/10.1111/j.1600-0587.2013.00205.x. Thuiller W., Georges D., Gueguen M., Engler R., Breiner F. (2021). biomod2: ensemble
Ondo, I., Thuiller, W., Gueguen, M., Pironon, S., 2019. A new R application for modelling platform for species distribution modeling. R package version 3.5.1. https://CRAN.
species distribution. 10.13140/RG.2.2.23431.83363. R-project.org/package=biomod2.
Osorio-Olvera, L., Soberón, J., Falconi, M., 2019. On population abundance and niche Thuiller, W., 2003. BIOMOD - optimizing predictions of species distributions and
structure. Ecography 42, 1415–1425. https://doi.org/10.1111/ecog.04442. projecting potential future shifts under global change. Glob. Change Biol. 10,
Pebesma, E., 2018. Simple features for R: standardized support for spatial vector data. 1353–1362.
R. J. 10 (1), 439–446. https://doi.org/10.32614/RJ-2018-009. Thuiller, W., Lafourcade, B., Engler, R., Araújo, M.B., 2009. BIOMOD-A platform for
Pebesma, E.J., Bivand, R.S., 2005. Classes and methods for spatial data in R. R News 5 ensemble forecasting of species distributions. Ecography 32, 369–373.
(2). https://cran.r-project.org/doc/Rnews/. Title, P.O., Bemmels, J.B., 2018. ENVIREM: an expanded set of bioclimatic and
Peterson, A.T., Soberón, J., Pearson, R.G., Anderson, R.P., Martínez-Meyer, E., topographic variables increases flexibility and improves performance of ecological
Nakamura, M., Araújo, M.B., 2011. Ecological Niche and Geographical Distributions. niche modeling. Ecography 41, 291–307. https://doi.org/10.1111/ecog.02880.
Princenton University Press, New Jersey. Tyberghein, L., Verbruggen, H., Pauly, K., Troupin, C., Mineur, F., De Clerck, O., 2012.
Phillips, S.J., Anderson, R.P., Dudík, M., Schapire, R.E., Blair, M.E., 2017. Opening the Bio-ORACLE: a global environmental dataset for marine species distribution
black box: an open-source release of Maxent. Ecography 40, 887–893. https://doi. modelling. Glob. Ecol. Biogeogr. 21, 272–281. https://doi.org/10.1111/j.1466-
org/10.1111/ecog.03049. 8238.2011.00656.x.
Phillips, S.J., Anderson, R.P., Schapire, R.E., 2006. Maximum entropy modeling of Valavi, R., Elith, J., Lahoz-Monfort, J.J., Guillera-Arroita, G., 2019. blockCV: an r
species geographic distributions. Ecol. Model. 190, 231–259. package for generating spatially or environmentally separated folds for k-fold cross-
Raes, N., ter Steege, H., 2007. A null-model for significance testing of presence-only validation of species distribution models. Methods Ecol. Evol. 10, 225–232. https://
species distribution models. Ecography 30, 727–736. doi.org/10.1111/2041-210X.13107.
R Core Team, 2022. R: a language and environment for statistical computing. Van Rossum, G., Drake Jr, F.L., 1995. Python Tutorial, 620. Centrum voor Wiskunde en
Ribeiro, B.R., Velazco, S.J.E., Guidoni-Martins, K., Tessarolo, G., Jardim, L., Bachman, S. Informatica, Amsterdam, The Netherlands.
P., Loyola, R., 2022. bdc: a toolkit for standardizing, integrating and cleaning VanDerWal J., Falconi L., Januchowski S., Shoo L., Storlie C. (2015) SDMTools: species
biodiversity data. Methods Ecol. Evol. 13, 1421–1428. https://doi.org/10.1111/ distribution modelling tools. Version: 1.1–221.
2041-210X.13868. Velazco, S.J.E., Rose, M.B., Andrade, A.F.A., Minoli, I., Franklin, J., 2022. flexsdm: an R
Rinnan, D.S., Lawler, J., 2019. Climate-niche factor analysis: a spatial approach to package for supporting a comprehensive and flexible species distribution modelling
quantifying species vulnerability to climate change. Ecography 1494–1503. https:// workflow. Methods Ecol. Evol. 13874. https://doi.org/10.1111/2041-210X.13874,
doi.org/10.1111/ecog.03937. 2041–210X.
Robertson, M.P., Visser, V., Hui, C., 2016. Biogeo: an R package for assessing and Vignali, S., Barras, A.G., Arlettaz, R., Braunisch, V., 2020. SDMtune : an R package to
improving data quality of occurrence record datasets. Ecography 39, 394–401. tune and evaluate species distribution models. Ecol. Evol. 10, 11488–11506. https://
https://doi.org/10.1111/oik.02629. doi.org/10.1002/ece3.6786.
Rotllan-Puig, X., Traveset, A., 2021. Determining the minimal background area for Vollering, J., Halvorsen, R., Mazzoni, S., 2019. The MIAmaxent R package: variable
species distribution models: MinBAR package. Ecol. Model. 439, 109353 https://doi. transformation and model selection for species distribution models. Ecol. Evol. 9,
org/10.1016/j.ecolmodel.2020.109353. 12051–12068. https://doi.org/10.1002/ece3.5654.
Royle, J.A., Chandler, R.B., Yackulic, C., Nichols, J.D., 2012. Likelihood analysis of Warren, D.L., Glor, R.E., Turelli, M., 2008. Environmental niche equivalency versus
species occurrence probability from presence-only data for modelling species conservatism: quantitative approaches to niche evolution. Evolution 62, 2868–2883.
distributions. Methods Ecol. Evol. 3, 545–554. https://doi.org/10.1111/j.2041- https://doi.org/10.1111/j.1558-5646.2008.00482.x (N Y).
210X.2011.00182.x. Warren, D.L., Glor, R.E., Turelli, M., 2010. ENMTools: a toolbox for comparative studies
Sánchez-Tapia, A., Mortara, S.R., Bezerra Rocha, D.S., Mendes Barros, F.S., Gall, G., de of environmental niche models. Ecography 33, 607–611. https://doi.org/10.1111/
Siqueira, M.F., 2020. modleR: a modular workflow to perform ecological niche j.1600-0587.2009.06142.x.
modeling in R (preprint). Ecology. https://doi.org/10.1101/2020.04.01.021105. Warren, D.L., Matzke, N.J., Cardillo, M., Baumgartner, J.B., Beaumont, L.J., Turelli, M.,
Sbrocco, E.J., Barber, P.H., 2013. MARSPEC: ocean climate layers for marine spatial Glor, R.E., Huron, N.A., Simões, M., Iglesias, T.L., Piquet, J.C., Dinnage, R., 2021.
ecology: ecological archives E094-086. Ecology 94. https://doi.org/10.1890/12- ENMTools 1.0: an R package for comparative ecological biogeography. Ecography
1358.1, 979–979. 44, 504–511. https://doi.org/10.1111/ecog.05485.
Schmitt, S., Pouteau, R., Justeau, D., de Boissieu, F., Birnbaum, P., 2017. ssdm: an r Woodman, S.M., Forney, K.A., Becker, E.A., DeAngelis, M.L., Hazen, E.L., Palacios, D.M.,
package to predict distribution of species richness and composition based on stacked Redfern, J.V., 2019. eSDM: a tool for creating and exploring ensembles of predictions
species distribution models. Methods Ecol. Evol. 8, 1795–1803. https://doi.org/ from species distribution and abundance models. Methods Ecol. Evol. 13283.
10.1111/2041-210X.12841. https://doi.org/10.1111/2041-210X.13283, 2041–210X.
Sillero, N., 2011. What does ecological modelling model? A proposed classification of Zizka, A., Antonelli, A., Silvestro, D., 2020. sampbias, a method for quantifying
ecological niche models based on their underlying methods. Ecol. Model. 222, geographic sampling biases in species distribution data. Ecography ecog. 05102.
1343–1346. https://doi.org/10.1016/j.ecolmodel.2011.01.018. https://doi.org/10.1111/ecog.05102.
Sillero, N., Arenas-Castro, S., Enriquez-Urzelai, U., Vale, C.G., Sousa-Guedes, D., Zizka, A., Silvestro, D., Andermann, T., Azevedo, J., Duarte Ritter, C., Edler, D.,
Martínez-Freiría, F., Real, R., Barbosa, A.M., 2021. Want to model a species niche? A Farooq, H., Herdean, A., Ariza, M., Scharn, R., Svantesson, S., Wengström, N.,
step-by-step guideline on correlative ecological niche modelling. Ecol. Model. 456, Zizka, V., Antonelli, A., 2019. CoordinateCleaner: standardized cleaning of
109671 https://doi.org/10.1016/j.ecolmodel.2021.109671. occurrence records from biological collection databases. Methods Ecol. Evol. 10,
744–751. https://doi.org/10.1111/2041-210X.13152.

14

View publication stats