Spiny but photogenic: amateur sightings complement herbarium specimens to reveal the bioregions of cacti

Abstract

Premise: Cacti are characteristic elements of the Neotropical flora and of major interest for biogeographic, evolutionary, and ecological studies. Here we test global biogeographic boundaries for Neotropical Cactaceae using specimen-based occurrences coupled with data from visual observations, including citizen science records, as a means to tackle the known collection biases in the family.
Methods: Species richness and record density were assessed separately for preserved specimens and human observations and a bioregional scheme tailored to Cactaceae was produced using the interactive web application Infomap Bioregions based on data from 261,272 point records cleaned through automated and manual steps.
Key Results: We find that areas in Mexico and southwestern USA, Eastern Brazil and along the Andean region have the greatest density of records and the highest species richness. Human observations complement information from preserved specimens substantially, especially along the Andes. We propose 24 cacti bioregions, among which the most species-rich are, in decreasing order: northern Mexico/southwestern USA, central Mexico, southern central Mexico, Central America, Mexican Pacific coast, central and southern Andes, northwestern Mexico/extreme southwestern USA, southwestern Bolivia, northeastern Brazil, Mexico/Baja California.
Conclusions: The bioregionalization proposed shows novel or modified biogeographical boundaries specific to cacti, and can thereby aid further evolutionary, biogeographic, and ecological studies by providing a validated framework for further analyses. This classification builds upon, and is distinctive from, other expert-derived regionalization schemes for other taxa. Our results showcase how observation data, including citizen-science records, can complement traditional specimen-based data for biogeographic research, particularly for taxa with specific specimen collection and preservation challenges and those that are threatened or internationally protected.

Full text

1
Spiny but photogenic: amateur sightings complement herbarium specimens to reveal
the bioregions of cacti
Alice Calvente1,2, Ana Paula Alves da Silva1, Daniel Edler3,7, Fernanda Antunes Carvalho4,
Mariana Ramos Fantinati5, Alexander Zizka6, Alexandre Antonelli7,8,9
1Laboratório de Botânica Sistemática, Departamento de Botânica e Zoologia, Centro de
Biociências, Universidade Federal do Rio Grande do Norte, Av. Senador Salgado Filho,
3000, CEP 59078970, Lagoa Nova, Natal, RN, Brazil; 3Integrated Science Lab, Department
of Physics, Umeå University, Umeå, Sweden; 4Departamento de Genética, Ecologia e
Evolução, Instituto de Ciências Biológicas, Universidade Federal de Minas Gerais, Av.
Antônio Carlos 6627, Pampulha, CEP 31270-901, Belo Horizonte, MG, Brasil;
5Departamento de Ciências Biológicas, Universidade Estadual Paulista - câmpus de Assis,
Av. Dom Antônio, 2100, Parque Universitário, CEP 19806-900, Assis, SP, Brazil;
6Biodiversity of plants, Philipps University Marburg, 35043 Marburg, Germany;
7Gothenburg Global Biodiversity Centre, Department of Biological and Environmental
Sciences, University of Gothenburg, SE-405 30 Gothenburg, Sweden; 8 Royal Botanic
Gardens Kew, TW9 3AE Richmond, United Kingdom; 9 Department of Biology, University
of Oxford, Oxford OX1 3RB, United Kingdom ; 2Author for correspondence
(acalvente@cb.ufrn.br). AC and APAS contributed equally to this work.
Manuscript received _______; revision accepted _______.
Short title: Bioregions of cacti
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

2
Abstract
Premise: Cacti are characteristic elements of the Neotropical flora and of major interest for
biogeographic, evolutionary, and ecological studies. Here we test global biogeographic
boundaries for Neotropical Cactaceae using specimen-based occurrences coupled with data
from visual observations, including citizen science records, as a means to tackle the known
collection biases in the family.
Methods: Species richness and record density were assessed separately for preserved
specimens and human observations and a bioregional scheme tailored to Cactaceae was
produced using the interactive web application Infomap Bioregions based on data from
261,272 point records cleaned through automated and manual steps.
Key Results: We find that areas in Mexico and southwestern USA, Eastern Brazil and
along the Andean region have the greatest density of records and the highest species
richness. Human observations complement information from preserved specimens
substantially, especially along the Andes. We propose 24 cacti bioregions, among which
the most species-rich are, in decreasing order: northern Mexico/southwestern USA, central
Mexico, southern central Mexico, Central America, Mexican Pacific coast, central and
southern Andes, northwestern Mexico/extreme southwestern USA, southwestern Bolivia,
northeastern Brazil, Mexico/Baja California.
Conclusions: The bioregionalization proposed shows novel or modified biogeographical
boundaries specific to cacti, and can thereby aid further evolutionary, biogeographic, and
ecological studies by providing a validated framework for further analyses. This
classification builds upon, and is distinctive from, other expert-derived regionalization
schemes for other taxa. Our results showcase how observation data, including citizen-
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

3
science records, can complement traditional specimen-based data for biogeographic
research, particularly for taxa with specific specimen collection and preservation challenges
and those that are threatened or internationally protected.
Key words: bioregional schemes; Cactaceae; citizen science; iNaturalist; Neotropical
regionalization; succulents; visual observations.
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

4
INTRODUCTION
Biogeographic regions are fundamental to the study of biogeography because they
can inform on dynamic processes of origin, migration, and extinction of evolutionary
lineages through time and space (Antonelli, 2017a; Ferrari, 2017; Morrone, 2018).
Bioregions and related terms, including areas of endemism, biogeographic realms, phyto-
or zoogeographic zones, biomes, ecoregions, or even ecosystems, have been used as
operational units of regionalization (Morrone, 2018). Many of these units closely reflect
continental divisions, as plate tectonics repeatedly isolated and connected continental biotas
through evolutionary time (Antonelli, 2017b; Ficetola et al., 2017). Nevertheless, within
continents, smaller patches of more cohesive biotas are also recognizable, but may be more
difficult to define, depending on criteria (different methods used to delimit operational
units) and context (focus on whole biotas, particular taxonomic groups, or even based on
purely abiotic and geographical aspects).
Biotic and abiotic barriers and influences do not affect all organisms equally.
Consider for instance that the permeability of a barrier can be different for species bearing
animal- or wind-dispersed seeds (e.g, Antonelli, 2009; Nazareno et al., 2021) or even for
species dispersed by non-flying small mammals or migrating bats (which can facilitate the
connection of isolated populations; e.g., Shilton et al., 1999). Hence, while universal
bioregions shared by many taxa delimited by large geographical barriers are important, to
understand the significance of geological, climatic and other earth-history processes on the
evolution of life (Parenti and Ebach, 2009), taxon-specific bioregionalization schemes are
often more valuable for more specific applications, such as ancestral area reconstruction in
historical biogeography (Edler et al., 2017). In this case, it is crucial to define the
operational bioregions prior to analysis, because these bioregions influence the inference of
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

5
ancestral areas and consequently the scenarios of origin and dispersal/vicariance events
through time.
Metrics of dissimilarity and endemism, as well as species richness, abundance, and
rarity, based on occurrence data for species (including the longitude, latitude, date, and
other elements of meta-data) are the primary source of data used to delimit bioregions
(Harold and Mooi, 1994; Olson and Dinerstein, 1998; Olson et al., 2001). In recent
decades, bioregionalization schemes have risen in prominence as a means to analytically
support evolutionary, ecological, biogeographic and conservation studies (Olson et al.,
2001; Morrone, 2018; Montalvo-Mancheno et al., 2020), and have been propelled by the
increase of digitization of biological collections, initiatives to promote public online
occurrence databases, and the organization and publication of taxonomic research and
inventories in electronic format (Soltis, 2017; Heberling et al., 2021). Recent approaches to
bioregionalization also incorporate macroecological principles, ordination, and network
methods (Kreft and Jetz, 2010; Vilhena and Antonelli, 2015; Edler et al., 2017; Droissart et
al., 2018; Colli-Silva et al., 2019), although many biogeographic studies still use bioregions
determined from ‘expert-based’ drawings on maps, a practice that is feasible for many taxa,
but on the other hand it is largely subjective, not reproducible, and prevents estimates of
uncertainty (Edler et al., 2017; Ferrari, 2017).
Several bioregionalization schemes have been proposed for the Neotropical region
(e.g., Olson et al., 2001; De Nova et al., 2012; Hughes et al., 2013; Morrone, 2014; Zizka et
al., 2018; Morrone et al., 2022). As probably the most species-rich region on Earth (Kier et
al., 2005; Antonelli and Sanmartín, 2011; Zizka, 2019) with an exuberant faunal, fungal,
floristic and geo-climatic diversity, the Neotropics comprise a heterogeneous landscape
with a great variety of habitats, including interwoven areas of tropical rainforests,
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

6
seasonally dry tropical forests, savannas, rocky fields, deserts, prairies, swamps, and high
altitude and coastal ecosystems. Furthermore, the Neotropics are home to a diversity of life
history strategies in plants (such as epiphytes, trees, and fire-adapted geophytes) that a
single, general bioregionalization may fail to capture (Antonelli and Sanmartín, 2011;
Hughes et al., 2013).
Cacti as a study system for Neotropical biogeography— Neotropical areas under
arid and semi-arid climates form broad, more or less well-defined units in the schemes
already proposed for Neotropical regions and are noteworthy for their largely unique biota
under high levels of threat (Dinerstein, 2017; Dryflor, 2021; Morrone et al., 2022). Cacti
are among the most characteristic elements of the flora of seasonally dry Neotropical areas
and are one of the groups of angiosperms best adapted to tropical aridity (Anderson, 2001).
Cacti also occur in other Neotropical habitats, such as tropical and subtropical forests,
usually occupying water-stressed niches, as epiphytes or lithophytes (Anderson, 2001;
Barthlott et al., 2015). The cactus family is almost endemic to the Neotropical region; of
around 1,850 known species in 130 genera, only one naturally occurs outside the Americas
–the epiphytic and bird-dispersed Rhipsalis baccifera (J.S.Muell.) Stearn, which also
reaches Africa, Asia, and Sri Lanka (Nyffeler, 2002; Hunt et al., 2006; Nyffeler and Eggli,
2010). The wide distribution, elevated richness and high levels of regional endemism make
Cactaceae a suitable model for investigating the diversity, distribution and evolutionary
history associated with dry habitats (Silva et al., 2018). Studies focusing on such groups
with particular life histories are still poorly documented and may serve as reference to other
groups of organisms with congruent distributions in similar conditions (Miranda et al.,
2018; Colli-Silva and Pirani, 2020).
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

7
The geographic distribution of Cactaceae has been addressed generally, at the
family level, based on taxonomic knowledge and checklists and using different analytical
tools (Barthlott and Hunt, 1993; Anderson, 2001; Barthlott et al., 2015; Amaral et al.,
2022). Traditionally, three centers of diversity are accepted: (1) Mexico and southwestern
USA, which comprises the largest number of species; (2) the central Andean region of
Peru, Bolivia, northern Argentina, and Chile; and (3) Eastern Brazil (Barthlott and Hunt,
1993; Anderson, 2001). These three centers were also identified by Taylor (1997), who
reported a fourth center along central-western and southern Brazil, Paraguay, Uruguay, and
central Argentina. Barthlott et al. (2015) estimated distribution ranges for individual species
using data from the literature and other various sources and expanded the knowledge on
diversity centers, recognizing seven additional subordinated centers: Chihuahua, Puebla-
Oaxaca, Sonora-Sinaloan, Jalisco (all four north/central American); Southern central
Andes, Caatinga and Mata Atlântica (all three south American).
Other approaches have associated distribution data with conservation assessments,
as Cactaceae have historically been affected by illegal trade and habitat loss. For example,
Goettsch et al. (2015, 2018) provided a formal global assessment of the conservation status
of all cactus species based on point occurrences and expert-reviewed range maps. The study
found nearly a third of the species are threatened with extinction (Goettsch et al., 2015,
2018). Amaral et al. (2022) used occurrence and phylogenetic data to investigate spatial
patterns and factors associated with endemism and concluded that legally protected areas
do not guarantee the evolutionary conservation of the family. Furthermore, these authors
suggested that different abiotic factors may contribute to the prediction of endemism in the
group. Pillet et al. (2022) used species distribution models to predict a negative impact of
future climatic changes, with severe extinction risk for most species they assessed.
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

8
Although these earlier studies generated a comprehensive general understanding of cacti
distribution, historical biogeography studies on the group have relied on general schemes of
Neotropical bioregionalizations (e.g., Ocampo and Columbus, 2010; Calvente et al., 2011;
Vazquez-Sanchez et al., 2013; Hernandez-Hernandez et al. 2014; Lavor et al., 2018;
Majure et al., 2022) and the validity of such general schemes for cacti remains to be tested.
Most studies of spatial biogeography in plants, including cacti, use preserved
herbarium specimens point records as primary evidence (Funk and Richardson, 2002; Folk
and Siniscalchi, 2021). However, this is problematic for cacti due to a known collection
deficit associated with the difficult handling and preservation of specimens. Cacti are often
ignored by collectors, because they are succulent, spiny and grow in extreme habitats, and
the techniques to adequately preserve cactus specimens require special training (Anderson,
2001; Taylor and Zappi, 2004). The preservation of specimens of rare and endangered cacti
can also be additionally challenging and biased, since many collectors avoid collecting
these taxa, particularly when the whole plant must be collected for adequate preservation. It
is also important to note that all cacti (except species in genera Pereskia, Pereskiopsis and
Quiabentia) are legally protected by the Convention on International Trade in Endangered
Species of Wild Fauna and Flora (CITES, appendices I and II) since 1975.
Despite those challenges, Cactaceae systematics has a wide appeal and even reaches
the general public, since cacti are charismatic, well known due to their popularity as house
plants, and easy to identify at family level. Consequently, synthesizing information about
the distribution of cacti can benefit from networks of amateur enthusiasts and horticultural
societies, and other sources of information such as personal communications, field
observations and pictures. When combined with academic knowledge, these additional
sources of information can provide an integrated understanding about the general
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

9
occurrence of species (e.g., Taylor and Zappi, 2004; Goettsch et al., 2015, 2018). Despite
legitimate concerns about the limits of data based solely on human observations, citizen-
science data nevertheless hold the potential to complement the information provided
exclusively by traditionally preserved specimens, particularly where the sources are well-
referenced images linked to public databases that benefit from being curated to some
degree (Troudet et al., 2018).
Here, we combine data from preserved specimens and human observations gathered
from public databases to analyse diversity patterns for Cactaceae and then produce a
bioregionalization scheme for the family based on network analysis, also integrating
phylogenetic information. We aim to test the application of data from visual observations of
cacti (including citizen-science-based data) in spatial biogeographical analyses and to test
global biogeographic boundaries for the family. In particular, we address the following
questions:
(1) Are occurrence data currently available through public databases, including
human observation records, consistent with scientific knowledge and therefore capable of
providing a comprehensive distribution dataset for Cactaceae?
(2) What are the bioregional boundaries for Cactaceae obtained from a large
dataset?
(3) Is a bioregional scheme tailored for Cactaceae compatible with general
Neotropical bioregionalization schemes?
MATERIALS AND METHODS
We downloaded a total of 261,272 georeferenced records from preserved specimens
(28%) and human observations (72%) for Neotropical Cactaceae (Table 1) using rgbif
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

10
v.1.3.0 (Chamberlain et al., 2019) implemented in R (R Core Team, 2020). To delimit the
Neotropics, we used the polygon defined as: -34.7 32.8, -117.2 32.8, -117.2 -55.8, -34.7 -
55.8, -34.7 32.8 (Fig. 1). We included human observations to explore their potential to
complement preserved specimen data due to the known collection bias in Cactaceae
(Anderson, 2001; Taylor and Zappi, 2004). To access the contribution of each type of
record on the resulting distribution and bioregionalization and to allow a clearer
comparison of record type, we performed analyses separately for: (1) preserved specimens;
(2) human observations (excluding iNaturalist, as these include mostly science-based and
government datasets, but also includes less representative citizen-based datasets; Table S1);
and (3) iNaturalist (the major source of citizen-based occurrences in our dataset, which
included nearly 90% of all human observations; Table 1).
To minimize errors due to coordinate imprecision, data cleaning was performed
through CoordinateCleaner v. 2.0-9 (Zizka et al., 2019) in R. There is no universal solution
to clean and process species occurrence data (Zizka et al., 2020a). To visualize and help us
evaluate the performance of steps of automated data cleaning for cacti, we produced
richness and occurrence density maps of raw and cleaned datasets using the tidyverse
v.1.3.2 (Wickham et al., 2019), speciesgeocodeR v.2.0-10 (Töpel et al., 2017), raster v.3.5-
29 (Hijmans, 2018) and rgdal v.1.5-32 (Bivand et al., 2015) packages. Automated cleaning
excluded records: (1) without geographical coordinates; (2) in a 10 km radius from country
centroids; (3) in the headquarters of biodiversity institutions, such as museums, botanical
gardens and universities; (4) in open sea; (5) with coordinates equal to zero; (6) with equal
latitudes and longitudes; (7) collected prior to 1945 (except for preserved specimens from
Argentina, for which we included all records, because they lacked data for collection
dates); (8) with invalid or imprecise coordinates; (9) not identified at least to species level;
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

11
and (10) duplicated (same coordinates) for the same species (Zizka et al., 2020b). We
further manually standardized names following the classification adopted in Korotkova et
al. (2021). To minimize the registered occurrence of cultivated and naturalized species
(both in preserved specimens and human observation data) we checked all individual
species distribution maps against the geographic distribution cited in the most recent
comprehensive works produced for the family (Hunt et al., 2006; Barthlott et al., 2015,
Hunt, 2016) and excluded non-matching records in QGis 3.6.1 (Qgis, 2021). In this final
manual cleaning we excluded all records of the widely cultivated species Opuntia
cochenillifera (L.) Mill., O. ficus-indica (L.) Mill., Cereus hildmannianus K.Schum.,
Pereskia grandifolia Haw., Schlumbergera truncata (Haw.) Moran and Selenicereus
undatus (Haw.) D.R.Hunt and cleaned individual records of another 109 species (Table S2).
We inferred bioregions using the cleaned datasets of preserved specimens and
human observations separately, and a combined dataset comprising both data sources. We
used the interactive web application Infomap Bioregions v.1.2.0 (Edler et al., 2017) for
bioregionalization, which is an easy-to-use method based on distribution data, bipartite
networks, and network clustering to detect single non-hierarchical solution schemes
(Vilhena and Antonelli, 2015; Kheirkhahzadeh et al., 2016). Infomap Bioregions first bins
the world into grid cells with adaptive resolution based on the density of the data: starting
from a maximum cell size, if there is at least a selected minimum number of records in a
cell, it is included in the analysis. Conversely, if there is more than the pre-selected
maximum number of records, the cell is recursively subdivided into four grid cells until the
maximum capacity is respected or minimum cell size is reached. The software then creates
a network between species and grid cells by connecting each species to all grid cells where
it is found. It uses the Infomap network clustering algorithm to find an optimal partition of
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

12
the network into groups of nodes more tightly inter-connected within than between the
groups. The set of grid cells within each group makes up a bioregion.
Infomap has a resolution parameter called Markov time (called “cluster cost” in
Infomap Bioregions v1) that can be used to zoom in and out for solutions on different
scales (Kheirkhahzadeh et al., 2016). Further analyses were carried out in Infomap
Bioregions v.2.6.1 using newly implemented tools to explore hierarchical solutions
(Rosvall et al., 2011), detect interzones (transition zones) or fuzzy borders between
bioregions where their taxa mix (Bloomfield et al., 2018) and using variable Markov time
to adapt Infomap’s resolution to the network density (Edler et al., 2022). With a constant
Markov time, increasing it to avoid fragmentation of sparse regions tends to collapse dense
regions. Variable Markov time increases the range of scales of bioregions that Infomap
explores by locally increasing Markov time on sparse regions.
Grid cell sizes of 1o x 1o and 2o x 2o and with an adaptive resolution of 1/8o to 4o
were tested to choose the best fit for the data in the Neotropical area; for final figures we
used the 1/2o to 2o adaptive grid cells and a maximum of 200 and a minimum of 3 records
per cell, which showed the best fit as using the resolution of cells smaller than 1/2o and a
minimum of less than 3 records per cell caused the fragmentation in sparse bioregions in
areas with too few records. Using cells larger than 2o caused the loss of definition of
bioregional borders in several areas, such as in Mexico and in the Andean region. We used
a maximum of 200 records per cell as this was close to the number of records in the cell
with the highest number of records. Due to unequal sampling efforts for Cactaceae along
the Neotropics, we tested, but chose not to use, the weight on abundance option for the final
results. Different cluster costs in 10 trial runs were tested (ranging from 0.5–2.0) to allow
the search for larger or smaller bioregions (Kheirkhahzadeh et al., 2016; Edler et al., 2017)
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

13
and the 0.92 cluster cost was selected to fit a wider continental scale inference. All these
subdivisions were supported by the data and merely differ in their value for downstream
applications: a smaller number of bioregions may be most useful for continental-scale
biogeographic analyses, whereas a larger number of fragmented units may hold higher
value for conservation purposes, which could be further explored and ground-truthed. Here
we explore further a larger continental scale scheme highlighting bioregions larger than one
2o x 2o grid cell. Species richness and occurrence density maps were also produced in
adaptive 1/2o to 2o grid cells in Infomap Bioregions (Edler et al., 2017) and were further
edited in QGis 3.6.1 (Qgis, 2021).
Hierarchical solutions were also explored using Markov times varying from 0.7 to
1.2 and under the variable Markov time option. Optimal results with smaller bioregions in
lower levels for South America and North America were found under different Markov
times (1.18 and 1, respectively). Those were nearly identical to the single solution scheme
in lower levels and are shown in the Supplementary Material (Figs. S6 and S7). An
alternative hierarchical solution capturing larger bioregions that consistently emerged in
schemes obtained under varying parameters is shown for North and South America
simultaneously using the variable Markov time option under Markov time=1.
Bioregionalization solutions were also explored using a newly developed method
implemented in Infomap Bioregions 2 that incorporates evolutionary relationships into
species occurrence networks leading to the definition of more historically meaningful
bioregions (Edler et al., 2023). It connects nodes from the phylogenetic tree to the grid cells
where their descendant species occur, weighted by the amount of geographic information
they provide. This makes the network more dense within areas with closely related species
and tends to dissolve bioregional boundaries crossing such areas. Similar to ancestral
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

14
nodes, individual wide-ranging species can also obscure and collapse modular patterns and
transition zones defined by range-restricted species if the links between species and grid
cells are unweighted. As narrowly distributed species are important for unveiling
biogeographic patterns and evolutionary processes (Laffan et al., 2016; Quintero and Jetz
2018), we use range-weighted species by default in Infomap Bioregions 2, treating species
nodes similar to ancestral nodes.
To perform this analysis, we built a phylogenetic backbone for Cactaceae using a
top-down approach departing from a large phylogenetic tree for angiosperms, standardized
according to the botanical nomenclature of The Plant List (GBOTB.extended.TPL.tre)
implemented in the U.PhyloMaker package (Jin and Qian, 2023) and edited to include all
species sampled in our occurrence records dataset. We initially built a tree pruned from the
megatree including species that matched our dataset (616 spp.; Fig. S9). This tree captures
relationships between major clades and genera of Cactaceae consistent with the literature
(e.g. Nyffeler and Eggli, 2010; Guerrero et al., 2018). To avoid making assumptions about
specific placements for species not yet sequenced, while making use of morphologically
informed taxonomic classifications, we then added the remainder of the species binding
them as polytomies to the first diverging node of their respective genus (when the genus
was present in the megatree) or major clades in Cactaceae (when the genus was absent from
the tree). We used the Open Tree of Life (Open Tree of Life, www.tree.opentreeoflife.org,
searched in June 2023) as a reference backbone to bind genera to large clades (tree 2,
available in the supplementary material). As an alternative, we built a second tree with
species from genera absent in the megatree binded to the first diverging node of Cactaceae
(tree 1, available in the supplementary material). Both trees were tested in the
bioregionalization analyses and produced similar results. The final trees match the
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

15
nomenclature of Korotkova et al. (2021) as our occurrence dataset was standardized
according to it. For this analysis in Infomap Bioregions 2 we used the previously described
settings and the options to integrate the whole tree, 100% tree weight, markov time=1 and
to weight species by range.
RESULTS
Data source— The automated filtering and manual cleaning resulted in the
exclusion of 58% of records and 22% of species names for preserved specimens, 54%
records and 13% of species names for human observations (excluding iNaturalist), and 42%
of records and 21% of species names for iNaturalist data (Table 1). The cleaned complete
dataset with data from all records combined resulted in 137,660 records for 1,248 species
(Table S2), which corresponds to 67% of all accepted species of Cactaceae listed in
Korotkova et al. (2021). Data from preserved specimens included records for 60% of
species, iNaturalist included records for 49% of species, and human observations
(excluding iNaturalist) included records for 20% of species of Cactaceae (Table 1).
Each data category showed similar patterns of record density per species, with most
species with few records and few species with numerous records (Fig. S3). Nevertheless,
for human observations (including iNaturalist), a single species—the saguaro, Carnegiea
gigantea (Engelm.) Britton & Rose—accounted for 16% of records (17,384 records), which
is four times more records than the second most recorded species (Cylindropuntia
leptocaulis (DC.) F.M.Knuth with 3,830 records). For preserved specimens, in contrast, the
most recorded species—Rhipsalis baccifera (J.S.Muell) Stearn—had only a few more
records than the second most recorded species (Opuntia engelmannii Salm-Dyck ex
Engelm. with 853 records), accounting for only 3% of the records (891 records). The vast
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

16
majority of records from iNaturalist were from Mexico and the USA, where 90,772
observations (93%) were recorded for 476 species, and only 7% were recorded in all other
Neotropical countries together (7,391 observations for 442 species).
We found occurrences for species of cacti throughout the Neotropics, although in
some patches of the core Amazonian region few species were documented (Fig. 1). Areas
in Mexico and southwestern USA, Eastern Brazil and along the Andean region had the
greatest density of records considering both preserved specimens and human observations
(Fig. 1a-d, Table 2). Eastern Brazil was better sampled through preserved specimen data,
and iNaturalist showed a greater density of records in Mexico, southeastern USA and
Andean regions (Fig. 1, Table 2).
Species richness maps recovered three main diversity centers for Cactaceae (Fig.
1e-h, Table 2): Mexico/SW USA (589 spp., 81% of records), the Andean region (376 spp.,
6% of records) and Eastern Brazil (175 spp., 4% of records). Although we recorded more
species overall in the Andean region compared with Eastern Brazil, the number of
preserved specimens from the former was lower (2,426 records for 376 species in the
Andean region versus 5,254 for 175 species in Eastern Brazil). The addition of human
observation records (including iNaturalist) increased both spatial coverage and density of
records and increased the species richness estimated for the Andean region by 21% (Fig.1,
Table 2). The addition of human observation records in Central/North America also led to a
noteworthy increase of 6% in the estimation of species richness, while in Eastern Brazil this
increase was only 0.6% (Table 2).
Most of the species (104 of 108 species) for which we obtained data only from
human observations in our dataset occur mainly in the Andean (68%) and Central/North
American (29%) centers of diversity and show local or subregional restricted distributions,
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

17
not exceeding 600km2 in geographic range (Table 3). A significant portion (57%) of these
species have few herbarium records in the Global Biodiversity Information Facility (GBIF,
www.gbif.org, searched in September 2022), have some degree of threat (24%) or are Data
Deficient (18%) in the IUCN Red List (IUCN, 2022), and a 6% were involved in some
degree of taxonomic uncertainty in the last 15 years.
Bioregions— Based on combined human observation and preserved specimen data,
we inferred 24 bioregions for Neotropical Cactaceae using the non-hierarchical solution
method implemented in Infomap Bioregions v.1.2.0 (Fig. 2; Table 4; shape files available
in the supplementary material). A scheme based exclusively on preserved specimens (Fig.
S4) produced very similar results, with all major bioregions coinciding. The scheme based
on combined data shows a clearer definition of bioregional borders in areas with lower
density and spatial coverage of preserved specimen records (Fig. 1), such as in
northwestern and midwestern South America, so we interpreted both results as
complementary (Figs. 2, S4).
Eight of these bioregions include more than 100 species and occupy 10 or more grid
cells: bioregions 19 (Mexico/USA), 17 (central Mexico), 22 (Central America), 14
(Mexican Pacific coast), 1 (mid to southern Andean region), 16 (NW Mexico/extreme SW
USA), 13 (NE Brazil) and 15 (Mexico/Baja California) (Table 4). Bioregions 3 (SW
Bolivia) and 18 (southern central Mexico), although smaller bioregions, also included more
than 100 species. These ten bioregions are also remarkable for the high number (18-99) of
within-bioregion endemics and nine include areas with higher phylogenetic diversity (Table
4; Amaral et al., 2022).
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

18
Six of the most species-rich bioregions are in Mexico and Southwestern USA
(Table 4). Bioregion 19 is continuous, spreading along northern Mexico and southwestern
USA, and it is the most species rich, including 283 species, of which 99 occur exclusively
in its area. Exclusive species of globular to subglobular cacti in the genera Echinocereus,
Coryphantha, Turbinicarpus, Mammillaria, Sclerocactus and Escobaria are indicative of
this bioregion (Table S5). Bioregion 17 is adjacent to the south of Bioregion 19, spreading
continuously in Central Mexico. Exclusive globular to subglobular species in the genera
Stenocactus, Coryphantha, Echinocereus, Mammillaria, Thelocactus and Turbinicarpus are
indicative of this bioregion. Bioregion 14 is adjacent to the west, spreading through the
west coast of Mexico. Various species of columnar, bushy, or globular cacti are indicative
of this bioregion, including exclusive species of Stenocereus, Pereskiopsis, Mammillaria,
Pachycereus, Echinocereus, Selenicereus and Acanthocereus. In the lower levels of an
alternative hierarchical solution for North America, the northern section of Bioregion 14
emerges in a separate bioregion, in the intersection of Bioregion 16 (Fig. S7). Bioregion 16
ranges to the north of Bioregion 14 and to the west of Bioregion 19, and is characterised by
exclusive globular to spiny, cylindrical species of Sclerocactus, Echinocereus, Cochemiea
and Grusonia. Bioregion 15 is mostly in Baja California, and is characterised by species of
globular to spiny, cylindrical, or flattened species of Echinocereus, Ferocactus,
Cylindropuntia, Cochemiea and Opuntia. Finally, Bioregion 18 is in southern Mexico and
Oaxaca, and is characterised by exclusive species of Mammillaria, Stenocereus and
Opuntia with various stem forms.
The other four most species-rich bioregions occur further to the south in the
Americas. Bioregion 22 is a large bioregion that ranges throughout extreme Southern
Mexico and Central America. There, the indicative species are exclusive spiny scandent
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

19
shrubs or epiphytes of Selenicereus, Epiphyllum, Deamia and Acanthocereus and also one
columnar Lemaireocereus, one globular Cochemiea and one flattened Opuntia. Bioregions
1 and 3 are in the Andean region of Argentina, Chile, and Bolivia. In Bioregion 1 the
indicative species are exclusive spiny cylindrical or globular species of Rebutia,
Echinopsis, Maihueniopsis, Tephrocactus, Acanthocalycium and Gymnocalycium and one
columnar Soehrensia. In Bioregion 3 the indicative species are various exclusive globular
species of Lobivia, columnar species of Harrisia, Cleistocactus and Vatricania and one
leafy species of Pereskia. Bioregion 13 is in Eastern Brazil and is characterised by
exclusive species of various columnar genera such as Pilosocereus, Micranthocereus,
Leocereus and Stephanocereus and also globular species of Melocactus.
Other bioregions rich in exclusive species (> 15 spp) are: 2 (Andean region/Chile),
4 (Andean region/Peru), 9 (SE/S Brazil), 21 (Caribbean), 5 (Andean region/Peru and
Ecuador), 11 (extreme S Brazil and Uruguay), 10 (N South America) and 7 (central South
America) (Table 4). Bioregions 7 and 9 also include areas with high phylogenetic diversity
(Table 4; Amaral et al., 2022). Bioregions 23 (Central America), 8 and 12 (both in Central
Brazil), 20 (USA/Florida), 6 (Small area in Peru) and 24 (Galapagos) include less than 50
species and less than 15 exclusive species. However, all of them include a cactus flora with
at least 10 species and/or a high proportion of exclusive species. For example, Bioregion 24
(Galapagos) includes only three species of cacti but all of them are exclusive to this
bioregion. In 19 of the 24 bioregions all indicative cactus species are restricted to their
respective bioregion (Table S5).
The hierarchical solution capturing larger bioregions showed three bioregions in the
first level (Fig. 3b): (1) a northern bioregion comprising North and Central America and
northern South America; (2) one bioregion in the Galapagos Islands; and (3) one bioregion
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

20
including eastern, western and southern South America (Fig. 3). Interzones are highlighted
in central Brazil, Bolivia, Paraguay, Peru, Ecuador and Venezuela, corresponding to areas
in bioregions 5, 6, 7, 8, 10 and 12 (Fig. 3a). Alternative first level hierarchical solutions
(Fig. S8) show a further east/west division in South America. In the second level (Fig. 3c)
13 bioregions emerged in a scheme generally similar to the non-hierarchical solution (Fig.
2), although with a few larger bioregions. The areas corresponding to Bioregions 15 and 16
emerge in a single bioregion, as well as areas corresponding to bioregions 8 and 12. A
single Mexican bioregion includes areas corresponding to bioregions 14, 17 and 18 and a
single Andean bioregion includes areas corresponding to bioregions 2, 4 and 5. A large
bioregion is stretched along Argentina, Uruguay, Bolivia and Paraguay, including areas
corresponding to bioregions 1, 3, 7 and 11.
The two level hierarchical schemes obtained incorporating phylogenetic data (Fig.
4) are very similar to the previous schemes based only on occurrence data (non-hierarchical
and hierarchical; Figs. 2 and 3). The first level includes 15 bioregions with Central and
South American bioregions delimited similarly to the second level scheme and a single
bioregion encompassing the whole Mexican and Southwestern USA center of diversity.
Main differences include a different outline for bioregion 2, including an extreme
Argentinean portion clustering with the southern Chilean region, and a portion of the
Dominican Republic clustering with a large continental bioregion including bioregions 22,
23 and 10. In the second level, five Mexican bioregions emerge with differences only in a
larger bioregion 18, including the southern portion of bioregion 14 and the tip of the
Yucatán peninsula.
DISCUSSION
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

21
The pros and cons of specimen and observation data— Major advantages of
biodiversity distribution data from public repositories or data providers, such as GBIF, are
their free availability and the ease and speed of access. Verification and validation of data
mostly overcomes issues such as taxonomic misidentification and georeferencing errors
that are putatively problematic in such public databases. Data validation steps also help to
identify other common issues of spatial analyses such as incomplete coverage due to
geographically and taxonomically biased sampling (Maldonado et al., 2015; Meyer et al.,
2016). Biased coverage can be particularly problematic for groups such as cacti, for which
collection deficit must be particularly considered when estimating levels of diversity
(Anderson, 2001; Taylor and Zappi, 2004). Nevertheless, the distribution, richness and
density patterns obtained here agree with patterns previously documented and thoroughly
verified and validated in the literature for Cactaceae in the Neotropical region (Barthlott
and Hunt, 1993; Barthlott et al., 2015; Goettsch et al., 2015).
For analysing our large dataset, including both preserved specimens and human
observation point occurrences for all Neotropical cacti, the automated filtering was
particularly useful, and allowed faster standardized, and replicable data validation and
cleaning (Zizka et al., 2019; Zizka et al., 2020b). Comparison to documented knowledge
and estimations based on checklists and species lists validated by specialists (e.g., Hunt et
al., 2006; Barthlott et al., 2015; Hunt, 2016; Korotkova et al., 2021) were also valuable to
validating our dataset and analysing results. Similar validation through comparison has
been performed previously for assessing limitations in inferring biodiversity patterns from
point occurrences in public databases for other plant groups (e.g., Yesson et al., 2007;
Maldonado et al., 2015).
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

22
Data obtained from preserved specimens allowed the identification of three main
centers of diversity in arid and semiarid regions of Mexico/USA, Brazil and around the
Andes, which were already largely acknowledged (Barthlott and Hunt 1993; Taylor, 1997;
Anderson, 2001) and had also emerged as cores of high endemism and phylogenetic
diversity in Amaral et al. (2022). The center of diversity in the Andean region was
particularly underestimated by preserved specimen data and it was enhanced in record
density and species richness with the addition of human observation data. This deficit of
coverage of preserved specimens in the Andean region was particularly highlighted around
the central and southern Andes and may have been caused either due to collection deficit or
missing herbarium data in our dataset. The need for scientific collection in the central
Andean region of Peru and Bolivia to shed light on the taxonomic and conservation status
of several cacti has been highlighted by specialists (e.g., Taylor, 1997) and although there
have been efforts to cover this knowledge deficit (Barthlott et al., 2015; Goettsch et al.,
2015, 2018), preserved specimen data alone might still be insufficient for estimating levels
of diversity in this region.
Our study documented higher species richness for each center of diversity when
human observation data was added, as we obtained occurrence data exclusively from
human observations for 108 species (these had missing data for herbarium specimens in our
dataset). These included many species with particularly restricted distributions, and this
may have hampered their collection and preservation, particularly if they were also located
in remote places and associated with rarity. Forty-seven of these species have been
classified as threatened or Data Deficient (IUCN, 2022), and 31 of these are indeed poorly
represented in scientific collections, for example a few small and globular cacti of
Copiapoa, Parodia and Gymnocalycium from Chile, Bolivia, and Argentina, respectively.
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

23
This is also the case for several species of Mammillaria in Mexico. It is therefore clear that
the data made available from human observations can be particularly valuable for
complementing existing records and knowledge of these species.
On the other hand, all open georeferenced locality data for rare and threatened
species have the potential to facilitate extractive collections and illegal trade; for this
reason, some institutions avoid publicizing occurrence data online for threatened taxa or
generalize the records so that exact locations cannot be determined (Chapman, 2020). The
obscure coordinates option in iNaturalist is one of the tools to avoid the open publication of
precise locality data for threatened species (www.inaturalist.org/, information given under
threatened taxa, searched in September 2022). However, when there are many observations
or records for a narrowly distributed taxon, the mapped occurrence (such as in iNaturalist
or GBIF) can allow the inference of localities and environments where populations occur.
The conservation status tag highlighted for each threatened taxon listed on iNaturalist
nevertheless is a valuable information tool that reaches the community. That together with
further information and recommendations can help to raise awareness and to potentially
expand the network of people involved in conservation actions and to aid in the protection
of taxa.
Another set of species sampled exclusively by human observations in our dataset
(16 spp, 15%) had a relatively good sampling of preserved specimens (more than 10
records per species) in GBIF and a few of them had metadata with geographic coordinates.
Data from these specimens were not included in our dataset as they were eliminated during
automated and manual cleaning due to the quality criteria we followed. Human
observations, such as those obtained from citizen science platforms, are often
georeferenced with acceptable accuracy and migrated directly to the platform and
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

24
subsequently to GBIF, without further editing of the coordinates. This reduces the potential
for generating errors associated with coordinates, making it a potentially powerful source of
data for spatial analysis. In contrast, different sources of errors with the geographical
coordinates can accumulate in scientific collections-based data. Herbarium specimens often
have geographical coordinates data manually transcribed from GPS data collected in the
field or, for older specimens, the coordinates may appear in the labels but may lack
precision or are subject to digitizing errors (Soltis, 2017). Furthermore, the metadata may
undergo subsequent migrations through different databases until it is included in GBIF,
when errors can accumulate.
Lack of data in GBIF may also have affected our assessment of species
underrepresented by preserved specimens. Plant specimens deposited at most of the world’s
about 3,000 herbaria have not yet been fully digitised, including many small herbaria with
important collections of the local flora. Additional material may be available in those
herbaria for several species for which the metadata were not digitized or were not included
in GBIF, and therefore are not included in our dataset. This digitization gap has been
identified previously as a major knowledge impediment for other groups of plants, such as
the Fabaceae (Yesson et al., 2007) and may be the reason for the particularly incomplete
coverage for the Andean region in the preserved specimen dataset. Even though the
representation of herbaria from countries in the Global South in public databases has
increased in the last 20 years (GBIF, www.gbif.org, searched in October 2021),
occurrences from regional Andean herbaria remain underrepresented in our dataset. Of 89
herbaria in Peru, Bolivia, Chile and Argentina with potential collections for cacti registered
in Index Herbariorum (http://sweetgum.nybg.org/science/ih/, searched in October 2021), 25
had datasets for preserved specimens in GBIF, 14 had occurrences for cacti and only seven
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

25
of them included geographical coordinates. In contrast, the collections of herbaria in other
South American countries such as Brazil and Colombia were well represented; we retrieved
data from 73 Brazilian herbaria and from 21 Colombian herbaria. We also retrieved data for
all countries in the Andean region from major and specialized herbarium collections (such
as K, MO, NY, F, DES, E and B), which include duplicates from regional herbaria and may
therefore represent their collections relatively well.
Future initiatives to enhance coverage of records for cacti in public databases should
support not only an increase in sampling for areas and taxa potentially under-represented in
collections, but also the digitization of specimens in local herbaria, and inclusion of the
metadata produced in online databases. Nevertheless, our results support that, for spatial
studies, publicly available human observation records, curated both scientifically and by
citizens, represent a relevant tool to complement the occurrence data obtained only from
preserved specimens. It is important to note, however, that such specimens are the basis of
taxonomic research, providing information and material for many other applications, such
as species description, morphological studies, DNA extraction, studies on herbivory and
other ecological relationships, and biochemical studies (Folk and Siniscalchi, 2021). As
discussed in Troudet et al. (2018), visual observations cannot be used as a basis for these
types of studies and therefore should not be seen as a substitute for preserved specimens.
Although poor data quality and rapid identification error propagation are among the
main concerns for data based on citizen science (Kosmala et al., 2016), initiatives to
promote and support periodic validation and curation of data by specialists in citizen
science databases can help mitigate these issues and make these data more readily reliable
for research (Troudet et al., 2018). The manual cleaning performed in this work based on
mapped point occurrences for individual species allowed a thorough verification of the
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

26
citizen-science-based data, and most of the problematic records (excluded) were related to
species commonly cultivated outside of their native range. Nevertheless, the data from
preserved specimens also included occurrence records of cultivated cacti, which were
excluded during our manual cleaning.
Cacti are particularly well documented in human observation datasets in GBIF, and
it is possible that data from visual observations may be not as informative and
complementary for other plant groups. Many species of cacti are severely threatened and
targeted in conservation action plans, including government-based projects and initiatives
to document the image and occurrence of species observed in natural areas such as in
national parks (Table S1). Also, citizen science observation databases include a large
amount of data for cacti, because they are charismatic and attractive to many people,
including amateurs with a good knowledge of species identification and several of them
work actively to curate observations. Cactaceae has more than ~520,000 research grade
observations in iNaturalist (i.e., independently validated by at least two experts or
“knowledgeable people”, sensu iNaturalist) for 1,719 species in the wild (of a total of
~1,900 in the family). A few charismatic groups show comparable results, such as
Aizoaceae with ~75,000 observations for 1,265 species (~1,722 in the family). For the
highly ornamental Orchidaceae, ~820,000 research grade observations are listed but for
only 8,307 of the ~26,000 species in the family. Numbers are lower and even less
representative for other charismatic and economically important families such as
Myrtaceae, which had ~123,000 research grade observations for 1,890 species (~5,900 in
the family), Bromeliaceae with ~91,000 research grade observations for 1271 species
(~3,728 in the family), Arecaceae with ~86,000 for 777 species (~2,457 in the family), and
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

27
Annonaceae with ~39,000 for 356 species (~2,430 in the family) (www.inaturalist.org/,
searched July 6th, 2023).
Our data also suggest that citizen science, at least for the collection of species
occurrences, is still in its infancy in many Neotropical countries, particularly in Central and
South America, even though there are website translations to several languages including
Spanish. Only 7% of all iNaturalist observations in our dataset were recorded in countries
other than Mexico and USA. A similar pattern, with South American countries showing
fewer contributions to observations in general, is revealed in searches for verifiable
observations of plant families in iNaturalist, even when the size of countries is considered:
184,577 were made in Ecuador, 126,040 in Brazil, 55,954 in Peru and 51,302 in Bolivia,
compared to 1,054,004 in Mexico and 17,289,264 in USA (www.inaturalist.org/, searched
October 16th, 2021). Even though iNaturalist data contributed to enhancing the density of
records in the Andean region in our dataset, the number of observations in general was
lower for Andean countries when compared to northern America countries, to a degree
highly disproportionate to the difference in the number of species. Further efforts from the
scientific community and government initiatives in Central and South America can help to
engage citizens to contribute to databases, promoting the importance of science and basic
science education in general.
The bioregions of cacti— As succulent plants with life history strategies and
adaptations associated with dry environments, cacti are expected to be concentrated in arid
and semi-arid climate regions (Gregory-Wodzicki, 2000; Arakaki et al., 2011). Our data
support this, with higher species richness and major bioregions centered around the
Sonoran (Bioregion 16), Baja Californian (Bioregion 15), Chihuahuan (Bioregion 19) and
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

28
Atacaman (bioregions 1, 2, 3, and 4) deserts and on the Caatinga dry forest (Bioregion 13).
Although a few cactus species inhabit the Amazonian domain (such as the epiphytic and
wide-ranging Rhipsalis baccifera and Epiphyllum phyllanthus), the extra-Amazonian
pattern is pronounced and suggests that the Amazonian tropical forest may also act as a
barrier for the dispersal of most cacti as it does for many taxa from open habitats (Colli-
Silva, 2021). In contrast, the Atlantic Forest, which is the second largest Neotropical forest,
is home to several cactus lineages, such as the epiphytic Rhipsalideae, which can also
inhabit patches of humid forests in the Andean Yungas (Calvente et al., 2011; Barthlott et
al., 2015).
From a Neotropic-wide perspective, topography also seems to influence regional
patterns observed for cacti. Elevation, in combination with many factors such as dynamics
of air masses and humidity, temperature, and exposure to direct sunlight (also associated
with the presence or absence of forested environments) and soil have been associated with
the distribution, phylogenetic diversity and endemism patterns for cacti and other plant
groups (Guerrero et al., 2011; Moeslund et al., 2013; Amaral et al., 2022). Andean and
Mesoamerican higher elevations seem to play a major role in shaping diversity and
distribution patterns for cacti, and consequently may be associated with the delimitation of
bioregions. A major bioregion (Bioregion 1) with high species richness and endemism, is
stretched along higher elevation Andean regions adjacent to five smaller ones (bioregions
2, 3, 4, 5 and 6), which may also have been shaped under elevation dynamics. The three
most species-rich bioregions (bioregions 19, 17 and 18) are also under the influence of
elevation dynamics in Mexico, spreading along the Central Mexican Plateau, interspersed
between Sierra Madre Occidental, Sierra Madre Oriental, the Trans-Mexican Volcanic Belt
and Sierra Madre del Sur.
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

29
Fire may also be a relevant factor in our bioregionalization scheme. Closely linked
to the distribution, composition and structure of Neotropical savannas, fire influences the
occurrence of several lineages, such as Melastomataceae (e.g., Microlicieae), Fabaceae
(e.g., Mimosa, Andira), Malvaceae (e.g., Eriotheca), Asteraceae (e.g., Viguiera), and
Poaceae (e.g., Actinocladum) among many others (Soderstrom, 1981; Fritsch et al., 2004;
Simon et al., 2009; Simon and Pennington, 2012). A pattern of seasonal fires in the Cerrado
seems not to favor the occurrence of cacti, which commonly do not have adaptations to
resist these fire cycles as is the case for other succulent taxa (Pennington et al., 2009).
Frequently, cacti in Cerrado occur in rocky outcrops that act as refuges away from the fires
(Taylor and Zappi, 2004; Lavor et al., 2018). These conditions may be the key element
shaping bioregions 8 and 12, which occur in marginal disjunct patches around core Cerrado
regions in Central Brazil.
Although much of the diversity of cacti is confined within continental land masses,
bioregional patterns are also created by marine island systems. In the Pacific, there is the
isolated bioregion 24 placed away from the continent on the Galapagos islands. In the
Caribbean, the bioregional delimitation suggests a complex scenario with disjunct patches
of the same bioregions on different islands and on the continent (bioregions 20 and 21 and
22, 23 and 10 as supported by phylogenetic data) and with a further subdivision of
bioregion 21 in three smaller bioregions in the third level of an alternative hierarchical
solution for this region (Fig. S7). The distance among land masses in the Caribbean is much
shorter, thereby aiding dispersal, but other factors may also have facilitated a dynamic
dispersal of cacti in that region. Pleistocene glacial cycles as well as major meteorological
events such as hurricanes may have increased the opportunity for dispersal among islands,
also influencing the diversification of several Pleistocene aged lineages such as Consolea,
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

30
Harrisia, Melocactus and Pilosocereus (Frank et al., 2013; Lavor et al., 2018; Majure et al.,
2021; Majure et al., 2022).
Major bioregions 9, 10 and 13 and bioregion 11 extend through continuous units
that have emerged in previous Neotropical regional classifications based on patterns
observed for whole vegetations and biotas (Griffith et al., 1998; Omernick and Griffith,
2014; Dinerstein et al., 2017; IBGE, 2019; Morrone et al., 2022; Fig. 5). Among them,
bioregion 11, which extends through the Uruguayan savanna (Dinerstein et al, 2017; Fig.
5b, 5c) in Brazil, Uruguay, and Argentina, has a remarkable cactus flora with several
globose species of Parodia and Frailea inhabiting open grassland habitats. Bioregions 9
and 10 occupy extensive areas of humid forests included in the Tropical and Subtropical
Moist Broadleaf Forests biome (Dinerstein et al., 2017; Fig. 5b). Bioregion 10 also extends
in disjunct patches into the adjacent Tropical & Subtropical Grasslands, Savannas &
Shrublands biome (Dinerstein et al., 2017; Fig. 5b), including forest taxa which can expand
their range into riparian forests along a savanna-like matrix, or lithophyte taxa that occur in
rock outcrops embedded in forests or savannas.
Although Bioregion 13 corresponds mostly to dry and seasonally marked areas of
the Caatinga (an ecoregion in Dinerstein et al., 2017 and Griffith et al., 1998 and a province
in Morrone et al., 2022; Fig. 5), it also expands through adjacent provinces and ecoregions;
it includes ecotonal areas between Caatinga and Cerrado and between Caatinga and
Atlantic Forest already included in the expanded definition of the Caatinga according to the
most recent evaluation of the Instituto Brasileiro de Geografia e Estatística (IBGE, 2019). It
also includes hotspot areas of the campo rupestre provinces Chapada Diamantina and
Southern Espinhaço in the Espinhaço Range (Colli-Silva et al., 2019) with remarkable
endemism for cacti (Amaral et al. 2022).
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

31
Hierarchical schemes and comparison with other classifications— The
North/South pattern shown for Cactaceae in first levels of the hierarchical scheme (Fig. 3)
surpasses the scale of traditional large scale biogeographical divisions for Cactaceae, which
describe three main large diversity centers. The Andean center and the Eastern Brazil center
appear within a continuous large southern bioregion pointing to some degree of
biogeographical connection between them. Alternative first level hierarchical solutions
show a further east/west subdivision of this large southern bioregion with contrasting
relationships for the central area in between them. Those areas in Argentina, Bolivia and
Paraguay can appear either clustered to the East or to the West and occupy large stretches
of transition zones (Fig. S8). Other large transition zones among bioregions occur along the
Andes, in northern South America and in Central America.
Lower level subdivisions along this central South American transition zone were
inconsistent in our results, also when phylogenetic data was integrated, leading to
conflicting solutions with divergent merging and subdivisions of bioregions. A more
conservative approach would be to recognize a single bioregion including bioregions 1, 3
and 7 and 11 and another single bioregion including bioregions 8 and 12 (as in Fig.3). The
same applies to bioregion 6 that may be merged into bioregion 5. Nevertheless, the non-
hierarchical scheme presented here in which these areas are divided into smaller bioregions
(Fig. 2) offers the most consistent outcome for finer-scale analysis based on our dataset,
although the fit of this particular scheme might need to be better evaluated when focusing
on specific analyses, such as ancestral area reconstruction for narrowly distributed lineages.
Overall, phylogenetic data supported solutions based exclusively on occurrence data
for our dataset. Although using different methodological approaches, Holt et al. (2013) and
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

32
Slik et al. (2020) found more pronounced discrepancies when phylogenetic data was
integrated in the classification of zoological regions and tropical forests, respectively.
Amaral et al. (2022) also found similar patterns between taxon richness and phylogenetic
diversity for Cactaceae and nearly half of our non-hierarchical scheme include areas where
they detected high phylogenetic diversity. Nevertheless, incorporating phylogenetic
information here was a powerful tool to highlight that most bioregions outlined are also
meaningful evolutionarily. Differences observed in the outline of a few bioregions
highlights the phylogenetic uniqueness or a narrower definition for a few bioregions. In
some cases, such as for bioregions 14 and 21, some of their bordering portions were
clustered into more widespread bioregions by their shared occurrence of widespread taxa.
With many exclusive lineages such as Eriosyce, Copiapoa and Eulychnia, the phylogenetic
uniqueness of Bioregion 2 was marked, as it emerged distinct from northern Andean
bioregions.
The bioregionalization scheme proposed here for Neotropical cacti is singular in
many aspects. Although overall results are comparable to biogeographic units based on
whole terrestrial biotas (e.g., Olson et al., 2001; Morrone, 2014; Dinerstein et al., 2017;
Morrone et al., 2022; Fig. 5), the exact delineation and regionalization scales are different
in most cases. For example, the second level of the hierarchical scheme (Fig. 3) shows a
single bioregion stretching along the Andean Chile and Peru what is comparable to some
extent of the South American transition zone (Morrone et al., 2022), although north, south
and east borders are much narrower. In Mesoamerica and in the Andean region, bioregions
for cacti are too wide to fit ecoregions and too narrow to fit the biomes in Dinerstein et al.
(2017). Bioregion 3 for example contains the Central Andean Dry Puna and part of the
Central Andean Puna and several other ecoregions of five different biomes (Table 4).
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

33
Nevertheless, Bioregion 2 nearly corresponds to the Chilean Matorral ecoregion within the
Mediterranean Forests, Woodlands & Scrub biome.
In general, the same is observed in Mesoamerican and Andean regions for the
ecoregions of Griffith et al. (1998), with cacti bioregions being larger than third level
ecoregions and narrower than second level ecoregions. Comparison with provinces and
dominions or zones of Morrone et al. (2022) lead to a similar interpretation: overall cacti
bioregions are more generalized than provinces and narrower than dominions. In Brazil on
the other hand, bioregions 9, 13 and 12/8 show a similar outline of operational units defined
in previous schemes (Table 4, Figs. 3 and 4).
These differences in relation to other schemes indicate that the distribution of cacti
seem to exceed the general borders of the whole biota and for others it may be more
restricted. Such singularity in the distribution patterns in a group with unique life history
strategies and adaptations is expected, as these plants may deal differently with extreme
factors that limit the distribution of other groups. For instance, dry regions and habitats are
mostly barriers to the dispersal of forest taxa, but corridors and cradles of diversification for
dry-adapted taxa such as cacti. Richness and endemism analyses conducted with
Neotropical Bromeliaceae also resulted in some degree of singularity when compared to
macroecological schemes based on plant and animal data, likely due to differences in
methodology and evolutionary history of the taxa examined (Zizka et al., 2019).
Differences in source and availability of data, criteria used and scale, which are critical for
the exact delineation of biogeographic units, are other likely reasons for differences among
bioregionalization schemes (Kreft & Jetz, 2010). However, the bioregionalization obtained
here is highly consistent with global biodiversity patterns previously described for cacti.
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

34
Most centers of diversity recognized by Barthlott et al. (2015) emerged inside
corresponding bioregions. The Caatinga center is contained within Bioregion 13. The Mata
Atlântica center is within Bioregion 9, the southern central Andes center is stretched along
bioregions 1 and 3, and the Puebla-Oaxaca center is in Bioregion 18. The very broad
Chihuahua center on the other hand, which has the greatest species and generic richness,
ranges across three bioregions in Mexico (bioregions 16, 17 and 19). Bioregion 16 also
includes the Sonora-Sinaloan center and Bioregion 14 contains the Jalisco center. Although
our results are overall consistent at this continental scale, smaller scale regionalization,
particularly on Andean and Mexican regions with a heterogeneous landscape and
topography associated with high species richness and endemism, may improve the
resolution of regional boundaries, and improve the fit of the scheme, especially for
restricted and narrowly distributed taxa. An improved dataset with the addition of more
high-quality data on species occurrence for still poorly documented areas and taxa and with
more data from herbaria yet to be digitized may bring complementary results on a more
refined smaller scale.
CONCLUSIONS
Citizen-science observations can greatly complement traditional specimen-derived
data for biodiversity information, particularly for taxa such as cacti that are challenging to
collect due to intrinsic or legal factors. We propose a new bioregionalization scheme for
Cactaceae which is overall comparable to other regionalization schemes proposed but
complements them with clearer and better-supported biogeographical boundaries. These
results provide a new comparative backbone to support the investigation of patterns and
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

35
processes underlying the biogeographic history and evolutionary diversification of lineages
of cacti and of the dry Neotropical flora.
ACKNOWLEDGMENTS
The authors thank Pró-reitoria de Pesquisa of UFRN and the Systematics and Evolution
graduate program for providing a visiting professor grant to AA; to the collectors,
observators and institutions that provided and made available the metadata from preserved
specimen and human observations of cacti used in this study; to Rhian Smith for science
and language editing; to Leonardo Versieux for suggestions on the manuscript; and to three
anonymous reviewers for valuable suggestions that helped improve this article. AA
acknowledges financial support from the Swedish Research Council (2019-05191) and the
Royal Botanic Gardens, Kew. APAS acknowledges a PhD fellowship from CAPES. We
acknowledge individual credits for species imagens in Figure 2 and thank: Jan Doležal (E.
aurea), Miguel A. Casado (C. cinerascens), Martin Lowry (L. calorubra), Christian
Bravard (C. brevistylus), Oscar Johnson (B. microsperma), Manuel Roncal (C. substerile),
María Zeta (A. rhodotrichum), Thales Santos (C. pierre-braunianus), Juliana Zuluaga-
Carrero (M. schatzlii), Martin Coronel Varela (P. scopa), Mattheus Mota (C. bicolor),
William Bruno (M. zehntneri), Juan Ramón Manjarrez (S. martinezii), Vince Scheidt,
CNPS (E. maritimus), Marc Faucher (S. erectocentrus), Aaron Balam (S. phyllacanthus),
Ignacio Torres García (M. pectinifera), Ana Luisa Fernández Fuentes (E. longisetus), Eric
M Powell (O. drummondii), Yolanda M. Leon (M. lemairei), Luis Humberto Vicente
Rivera (S. pteranthus), Yamilette Herrera Estévez (E. hookeri), William Stephens (B.
nesioticus).
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

36
AUTHOR CONTRIBUTIONS
AC, APAS, AA and MRF designed the study. AC, DE and APAS analysed data and results
and AC, APAS and FAC wrote the first draft. All authors reviewed, commented, and
contributed intellectually to improve the analyses and presented results and text.
DATA AVAILABILITY STATEMENT
Datasets used in these studies contain the location of many endangered or legally protected
species and therefore are not publicly archived. Original datasets may be requested through
direct contact with the authors. Phylogenetic trees and shapefiles of the non-hierarchical
scheme are available in the supplementary material and may be used freely.
LITERATURE CITED
Amaral, D. T., I. A. S. Bonatelli, M. Romeiro-Brito, E. M. Moraes and F. F. Franco. 2022.
Spatial patterns of evolutionary diversity in Cactaceae show low ecological
representation within protected areas. Biological Conservation 273: 1–12.
Anderson, E. F. 2001. The cactus family. Timber Press, Portland, Oregon, USA.
Antonelli, A. 2009. Have giant lobelias evolved several times independently? Life form
shifts and historical biogeography of the cosmopolitan and highly diverse subfamily
Lobelioideae (Campanulaceae). BMC Biology 7: 82. 21 pp.
Antonelli, A. 2017a. Comparative biogeography, big data, and common myths. In Friis and
H. Balslev [eds.], Tropical Plant Collections: Legacies from the Past? Essential Tools
for the Future? Scientia Danica. Series B, Biologica.
Antonelli, A. 2017b. Biogeography: Drivers of bioregionalization. Nature ecology &
evolution 1: 0114.
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

37
Antonelli, A. and I. Sanmartín. 2011. Why are there so many plant species in the
Neotropics? Taxon 60: 403–414.
Arakaki, M., P. A. Christin, R. Nyffeler, A. Lendel, U. Eggli, R. M. Ogburn, E. Spriggs, et
al. 2011. Contemporaneous and recent radiations of the world’s major succulent plant
lineages. Proceedings of the National Academy of Science of the United States of
America 108: 8379–8384.
Barthlott, W. and D. R. Hunt. 1993. Cactaceae. In K. Kubitzki, J. G. et al., [eds.], The
families and genera of vascular plants, 161–196. Springer-Verlag, Berlin, Heidelberg,
Alemanha.
Barthlott, W., K. Burstedde, J. L. Geffert, P. L. Ibisch, N. Korotkova, A. Miebach, M. D.
Rafiqpoor, et al. 2015. Schummania 7: Biogeography & biodiversity of cacti.
Germany: Universität Oldenburg.
Bivand, R., T. Keitt, and B. Rowlingson. 2015. Rgdal: Bindings for the ‘Geospatial’ data
abstraction Library. R package version 1.5-32. Website: https://cran.r-
project.org/package=rgdal.
Bloomfield, N. J., N. Knerr, and F. Encinas‐Viso. 2018. A comparison of network and
clustering methods to detect biogeographical regions. Ecography, 41(1), 1-10.
Calvente, A., D. C. Zappi, F. Forest, and L. G. Lohmann. 2011. Molecular Phylogeny,
Evolution, and Biogeography of South American Epiphytic Cacti. International
Journal of Plant Sciences 172: 902–914.
Chamberlain, S., V. Barve, D. Mcglinn, D. Oldoni, P. Desmet, L. Geffert, and K. Ram.
2019. rgbif: Interface to the Global Biodiversity Information Facility API. R package
version 1.3.0. Website: https://cran-r-project.org/package=rgbif.
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

38
Colli-Silva, M., J. R. Pirani, and A. Zizka. 2021. Disjunct plant species in South American
seasonally dry tropical forests responded differently to past climatic fluctuation.
Frontiers Biogeography 13: 1–16.
Colli-Silva, M., T. N. C., Vasconcelos, and J. R. Pirani. 2019. Outstanding plant endemism
levels strongly support the recognition of campo rupestre provinces in mountaintops
of eastern South America. Journal of Biogeography 46: 1723–1733.
Colli-Silva, M., and J. R. Pirani. 2020. Estimating bioregions and undercollected areas in
South America by revisiting Byttnerioideae, Helicteroideae and Sterculioideae
(Malvaceae) occurrence data. Flora 271: 1–15.
De-Nova, J. A., R. Medina, J. C. Montero, A. Weeks, J. A. Rosell, D. M. Olson, L. E.
Eguiarte, et al. 2012. Insights into the historical construction of species‐rich
Mesoamerican seasonally dry tropical forests: The diversification of Bursera
(Burseraceae, Sapindales). New Phytologist 193: 276–287.
Dinerstein, E., D. Olson, A. Joshi, C. Vynne, N. D., Burgess, E. Wikramanayake, N. Hahn,
et al. 2017. An ecoregion-based approach to protecting half the terrestrial realm.
BioScience 67: 534–545.
Droissart, V., G. Dauby, O. J. Hardy, V. Deblauwe, D. J. Harris, S. Janssens, B. A.
Mackinder, et al. 2018. Beyond trees: Biogeographical regionalization of tropical
Africa. Journal of Biogeography 45: 1153–1167.
Edler, D., T. Guedes, A. Zizka, M. Rosvall, and A. Antonelli. 2017. Infomap Bioregions:
Interactive Mapping of Biogeographical Regions from Species Distributions.
Systematic Biology 66: 197–204.
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

39
Edler, D., J. Smiljanić, A. Holmgren, A. Antonelli, and M. Rosvall. 2022. Variable Markov
dynamics as a multi-focal lens to map multi-scale complex networks. arXiv preprint
arXiv:2211.04287
Edler, D., A. Holmgren, A. Rojas, J. Calatayud, M. Rosvall, and A. Antonelli. 2023.
Infomap Bioregions 2 - Exploring the interplay between biogeography and evolution.
arXiv preprint arXiv:2306.17259
EPA (United States Environmental Protection Agency). 2022. Website:
https://www.epa.gov/eco-research/ecoregions-north-america [accessed 4 March 2022]
Ferrari, A. 2017. Biogeographical units matter. Australian Systematic Botany 30: 391–402.
Ficetola, G. S., F. Mazel, and W. Thuiller. 2017. Global determinants of zoogeographical
boundaries. Nature ecology & evolution 1: 0089.
Folk, R. A. and C. M. Siniscalchi. 2021. Biodiversity at the global scale: the synthesis
continues. American Journal of Botany 108: 912–924.
Franck, A. R., B. J. Cochrane, and J. R. Garey. 2013. Phylogeny, biogeography, and
infrageneric classification of Harrisia (Cactaceae). Systematic Botany 38: 210–223.
Fritsch, P. E., F. Almeida, S. S. Renner, A. B. Martins, and B. Cruz. 2004. Phylogeny and
circumscription of the near-endemic Brazilian tribe Microliciae (Melastomataceae).
American Journal of Botany 91: 1105–1114.
Funk, V. A. and K. S. Richardson. 2002. Systematic Data in Biodiversity Studies: Use It or
Lose It. Systematic Biology 51: 303–316.
GBIF.org. 2020. GBIF occurrence [online]. Website: https://doi.org/10.15468/dl.rbvnrw
[accessed 29 June 2020].
Goettsch, B., A. P. Durán, and K. J. Gaston. 2018. Global gap analysis of cactus species
and priority sites for their conservation. Conservation Biology 33: 369–376.
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

40
Goettsch, B., C. Hilton-Taylor, G. Cruz-Piñón, J. P. Duffy, A. Frances, H. M. Hernández,
R. Inger, et al. 2015. High proportion of cactus species threatened with extinction.
Nature Plants 1: 1–7.
Guerrero, P. C., A. P. Durán, and H. E. Walter. 2011. Latitudinal and altitudinal patterns of
the endemic cacti from the Atacama Desert to Mediterranean Chile. Arid of Journal
Environments 75: 991–997.
Guerrero, P. C., L. C. Majure, A. Cornejo-Romero and T. Hernández-Hernández. 2018.
Phylogenetic relationships and evolutionary trends in the cactus family. Journal of
Heredity 2019: 4–21.
Gregory-Wodzicki, K. M. 2000. Uplift history of the Central and Northern Andes: A
review. Geological Society of America Bulletin 112: 1091–1105.
Griffith, G. E., J. M. Omernik, and S. H. Azevedo. 1998. Ecological classification of the
Western Hemisphere. Unpublished report. U.S. Environmental Protection Agency,
Western Ecology Division, Corvallis, OR. 49p. Available at Website:
http://ecologicalregions.info [accessed 4 March 2022]
Harold, A. S. and R. D. Mooi. 1994. Areas of endemism: definition and recognition
criteria. Systematic Biology 43: 261–266.
Heberling, J. M., J. T. Miller, D. Noesgaard, S. B. Weingart, and D. Schigel. 2021. Data
integration enables global biodiversity synthesis. Proceedings of the National
Academy of Sciences 118: 1–7.
Hernandez-Hernandez, T., J. W. Brown, B. O. Schlumpberger, L. E. Eguiarte and S.
Magallón. 2014. Beyond aridification: multiple explanations for the elevated
diversification of cacti in the New World Succulent Biome. New Phytologist 202:
1382–1397.
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

41
Hijmans, R. J. 2019. raster: geographic data analysis and modeling. R package version 3.5-
29. Website: https://cran.r-project.org/packege=raster.
Holt, B. G., J. Lessard, M. K. Borregaard, S. A. Fritz, M. B. Araújo, D. Dimitrov, P. Fabre,
C. H. Graham, et al. 2013. Science 339: 74–78.
Hughes, C. E., R. T. Pennington, and A. Antonelli. 2013. Neotropical plant evolution:
assembling the big picture. Botanical Journal of the Linnean Society 171: 1–18.
Hunt, D. R. 2016. Cites Cactaceae Checklist. England, Royal Botanic Gardens Kew.
Hunt, D. R., N. P. Taylor, and C. Graham. 2006. The New Cactus Lexicon. Text. Milborne
Port, UK: DH Books.
IUCN 2022. The IUCN Red List of Threatened Species [online]. Website:
https://www.iucnredlist.org.
Jin, Y. and H. Qian. 2023. U.PhyloMaker: An R package that can generate large
phylogenetic trees for plants and animals. Plant Diversity 45: 347–352.
Kheirkhahzadeh, M., A. Lancichinetti, and M. Rosvall. 2016. Efficient community
detection of network flows for varying markov times and bipartite networks. Physical
Review E 93: 1–7.
Kier, G., J., Mutke, E. Dinerstein, T. H. Ricketts, W. Kuper, H. Kreft, and W. Barthlott.
2005. Global patterns of plant diversity and floristic knowledge. Journal of
Biogeography 32: 1107–1116.
Korotkova, N., D. Aquino, S. Arias, U. Eggli, A. Franck, C. Gómez-Hinostrosa, P. C.
Guerrero, et al. 2021. Cactaceae at Caryophyllales.org - a dynamic online species-
level taxonomic backbone for the family. Willdenowia 51: 251–270.
Kosmala, M., A. Wiggins, A. Swanson, and B. Simmons. 2016. Assessing data quality in
citizen science. Frontiers in Ecology and the Environment 14: 551–560.
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

42
Kreft, H. and W. Jetz. 2010. A framework for delineating biogeographical regions based on
species distributions. Journal of Biogeography 37: 2029–2053.
Laffan, S. W., D. F. Rosauer, G. Di Virgilio, J. T. Miller, C. E. González-Orozco, N. Knerr,
A. H. Thornhill and B. D. Mishler. 2016. Range-weighted metrics of species and
phylogenetic turnover can better resolve biogeographic transition zones. Methods in
Ecology and Evolution 7: 580–588.
Lavor, P. R., A. Calvente, L. M. Versieux, and I. Sanmartín. 2018. Bayesian spatio
temporal reconstruction reveals rapid diversification and Pleistocene range expansion
in the widespread columnar cactus Pilosocereus. Journal of Biogeography 46: 238–
250.
Majure, L. C., D. Barrios, E. Díaz, B. A. Zumwalde, W. Texto, and N. Negrón-Ortíz. 2021.
Pleistocene aridification underlies the evolutionary history of the Caribbean endemic,
insular, giant Consolea (Opuntioideae). American Journal of Botany 18: 200–215.
Majure, L. C., D. Barrios, E. Díaz, L. F. Bacci, and Y. E. Piñeyro. 2022. Phylogenomics of
the Caribbean melocacti: Cryptic species and multiple invasions. Taxon 71:1–20.
Maldonado, C., C. I. Molina, A. Zizka, C. Persson, C. M. Taylor, J. Albán, E. Chilquillo, et
al. 2015. Estimating species diversity and distribution in the era of Big Data: To what
extent can we trust public databases? Global Ecology and Biogeography 24: 973–
984.
Meyer, C., H. Kreft, R. Guralnick, and W. Jetz. 2015. Global priorities for an effective
information basis of biodiversity distributions. Nature Communications 6: 8221.
Miranda, P. L. S., A. T. Oliveira-Filho, R. T. Pennington, D. M. Neves, T. R. Baker, and K.
G. Dexter. 2018. Using tree species inventories to map biomes and assess their
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

43
climatic overlaps in lowland tropical South America. Global Ecology and
Biogeography 27: 899–912.
Moeslund, J. E., L. Arge, P. K. Bocher, T. Dalgaard, and J. C. Svenning. 2013. Topography
as a driver of local terrestrial vascular plant diversity patterns. Nordic Journal of
Botany 31: 129–144.
Montalvo-Mancheno, C. S., S. Ondei, B. W. Brook, and J. C. Buettel. 2020.
Bioregionalization approaches for conservation: methods, biases, and their
implications for Australian biodiversity. Biodiversity and Conservation 29: 1–17.
Morrone, J. J. 2014. Biogeographical regionalization of the Neotropical region. Zootaxa
3782: 1–110.
Morrone, J. J. 2018. The spectre of biogeographical regionalization. Journal of
Biogeography 45: 1–7.
Morrone, J. J., T. Escalante, G. Rodríguez-Tapia, A. Carmona, M. Arana, and J. D.
Mercado-Gómez. 2022. Biogeographical regionalization of the Neotropical region:
new map and shapefile. Anais da Academia Brasileira de Ciências 94: 1–5.
Nazareno, A. G., L. L. Knowles, C. W. Dick, and L. G. Lohmann. 2021. By Animal, Water,
or Wind: Can Dispersal Mode Predict Genetic Connectivity in Riverine Plant
Species? Frontiers in Plant Science 12: 1–18
Nyffeler, R. 2002. Phylogenetic relationships in the cactus family (Cactaceae) based on
evidence from trnK/matK and trnL-trnF sequences. American Journal of Botany 89:
312–326.
Nyffeler, R. and U. Eggli. 2010. A farewell to dated ideas and concepts: molecular
phylogenetics and a revised suprageneric classification of the family Cactaceae.
Schumannia, 6: 109–149.
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

44
Ocampo, G. and J. T. Columbus. 2010. Molecular phylogenetics of suborder Cactineae
(Caryophyllales), including insights into photosynthetic diversification and historical
biogeography. American Journal of Botany 97: 1827–1847.
Olson, D. M. and E. Dinerstein. 1998. The Global 200: A representation approach to
conserving the Earth’s most biologically valuable ecoregions. Conservation Biology
12: 502–515.
Olson, D. M., E. Dinerstein, E. D. Wikramanayake, N. D. Burgess, G. V. N. Powell, E. C.
Underwood, J. A. D’amico, et al. 2001. Terrestrial Ecoregions of the World: A New
Map of Life on Earth. BioScience 51: 933–938.
Omernik, J. M. and G. E. Griffith. 2014. Ecoregions of the conterminous United States:
evolution of a hierarchical spatial framework. Environmental Management 54: 1249-
1266.
Quintero, I and W. Jetz. 2018. Global elevational diversity and diversification of birds.
Nature 555: 246–250.
Pennington, R. T., M. Lavin, and A. Oliveira-Filho. 2009. Woody Plant Diversity,
Evolution, and Ecology in the Tropics: Perspectives from Seasonally Dry Tropical
Forests. Annual Review of Ecology, Evolution, and Systematics 40: 437–457.
Parenti, L. R., and M. C. Ebach. 2009. Comparative biogeography: discovering and
classifying biogeographical patterns of a dynamic Earth. University of California
Press.
Pillet, M., B. Goettsch, C. Merow, B. Maitner, X. Feng, P. R. Roehrdanz, and B. J. Enquist.
2022. Elevated extinction risk of cacti under climate change. Nature plants 8: 366–
372.
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

45
QGis, D. T. 2021. Quantum GIS geographic information system. Open Source Geospatial
Foundation Project. Website: https://qgis.org/en/site
R Core Team. 2020. R: Language environment for statistical computing. R Foundation for
Statistical computing, Vienna, Austria. Website: https://www.R-project.org.
Rosvall, M., and C. T. Bergstrom. 2011. Multilevel compression of random walks on
networks reveals hierarchical organization in large integrated systems. PloS one, 6(4),
e18209.
Shilton, L. A., J. D. Altringham, S. G. Compton, and R. J. Whittaker. 1999. Old World fruit
bats can be long-distance seed dispersers through extended retention of viable seeds
in the gut. Proceedings of the Royal Society of London, B, Biological Sciences 266:
219–223
Silva, G.A.R., Antonelli, A., Moraes, E.M., Lendel, A., Manfrin, M.H. 2018. The impact of
early Quaternary climate change on the diversification and population dynamics of a
South American cactus species. Journal of Biogeography 45: 76–88.
Simon, F. M., R. Grether, L. P. Queiroz, C. Skema, R. T. Pennington, and C. E. Hughes.
2009. Recent assembly of the Cerrado, a Neotropical plant diversity hotspot, by in
situ evolution of adaptations to fire. Proceedings National Academy of Sciences of the
United States of America 16: 20359–20364.
Simon, F. M. and R. T. Pennington. 2012. Evidence for adaptation to fire regimes in the
Tropical Savannas of the Brazilian Cerrado. International Journal of Plants Sciences
173: 711–723.
Slik, J. W., J. Franklin, V. Arroyo-Rodríguez, R. Field, S. Aguilar, N. Aguirre,
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

46
J. Ahumada, S. Aiba, et al. 2020. Phylogenetic classification of the world’s tropical forests.
Proceedings National Academy of Sciences of the United States of America 115:
1837–1842.
Soderstrom, T. R. 1981. Observations on a Fire-Adapted Bamboo of the Brazilian Cerrado,
Actinocladum verticillatum (Poaceae; Bambusoideae). American Journal of Botany
68: 1200–1211.
Soltis, P. S. 2017. Digitization of herbaria enables novel research. American Journal of
Botany 104: 1281–1284.
Taylor, N. P. and D. C. Zappi. 2004. Cacti of eastern Brazil. Kew Publishing, London, UK.
Taylor, N. P. 1997. Cactaceae. In Oldfield [eds], Cactus and succulent plants: Status Survey
and Conservation Action Plan. 17- 20. Cactus and Succulent Specialist Group, Gland,
Switzerland, and Cambridge.
Töpel, M., A. Zizka, M. F. Calió, R. Scharn, D. Silvestro, and A. Antonelli. 2017.
SpeciesGeoCoder: fast categorization of species occurrences for analysis of
biodiversity, biogeography, ecology, and evolution. Systematic Biology 66: 145–151.
Troudet, J., R. Vignes-Lebbe, P. Grandcolas, and F. Legendre. 2018. The Increasing
Disconnection of Primary Biodiversity Data from Specimens: How Does It Happen
and How to Handle It? Systematic Biology 67: 1110–1119.
Vazquez-Sanchez, M., T. Terrazas, S. Arias, and H. Ochoterena. 2013. Molecular
phylogeny, origin and taxonomic implications of the tribe Cacteae (Cactaceae).
Systematics and Biodiversity 11: 103–116.
Vilhena, D. A. and A. Antonelli. 2015. A network approach for identifying & delimiting
biogeographical regions. Nature Communications 6: 6848.
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

47
Wickham, H. 2018. tidyverse: easily install and load the ‘Tidyverse’. R package version
1.3.2. Website: https://cran.r-project.org/packege=tidyverse
Yesson, C., P. W. Brewer, T. Sutton, N. Caithness, J. S. Pahwa, M. Burgess, W. A. Gray, et
al. 2007. How global is the Global Biodiversity Information Facility? PLOS ONE 2:
1–10.
Zizka, A. 2019. Big data suggest migration and bioregion connectivity as crucial for the
evolution of Neotropical biodiversity. Frontiers of Biogeography 11: 1–7.
Zizka, A., H. ter Steege, M. C. R. Pessoa, and A. Antonelli. 2018. Finding needles in the
haystack: Where to look for rare species in the American tropics. Ecography 41: 321–
330.
Zizka, A., D. Silvestro, T. Andermann, J. Azevedo, C. D. Ritter, D. Edler, H. Farooq, et al.
2019. CoordinateCleaner: Standardized cleaning of occurrence records from
biological collection databases. Methods in Ecology & Evolution 10: 744–751.
Zizka, A., Antunes Carvalho, F., Calvente, A., Rocio Baez-Lizarazo, M., Cabral, A.,
Coelho, J.F.R., Colli-Silva, M., Fantinati, M.R., Fernandes, M.F., Ferreira-Araújo, T.,
Gondim Lambert Moreira, F., Santos, N.M.C., Santos, T.A.B., dos Santos-Costa,
R.C., Serrano, F.C., Alves da Silva, A.P., de Souza Soares, A., Cavalcante de Souza,
P.G., Calisto Tomaz, E., Vale, V.F., Vieira, T.L., Antonelli, A. 2020a. No one-size-
fits-all solution to clean GBIF. PeerJ 8: e9916.
Zizka, A., J. Azevedo, E. Leme, B. Neves, A. F. Costa, D. Caceres, and G. Zizka. 2020b.
Biogeography and conservation status of the pineapple family (Bromeliaceae).
Diversity and Distributions 26: 183–195.
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

48
TABLES
Table 1. Results for raw and cleaned datasets obtained from each record source.
Preserved specimens
Human observations
(excl. iNaturalist)
iNaturalist
All sources
Raw
Cleaned
Raw
Cleaned
Raw
Cleaned
Raw
Cleaned
No. of records
71,957
29,898
21,078
9,599
168,237
98,163
261,272
137,660
No. of species
1,419
1,106
418
362
1,144
902
1,573
1,248
Table 2. Number of records and species in the three main centers of diversity for Neotropical Cactaceae.
Preserved specimens
Human observations
(excl. iNaturalist)
iNaturalist
All sources
Records
Species
Records
Species
Records
Species
Records
Species
Andean region
2,426
298
800
36
4,497
292
8,367
376
Brazil
5,254
174
242
14
378
67
5,920
175
Mexico/USA
17,249
553
7,714
275
86,862
476
111,171
589
Table 3. Source of records, herbarium sampling, conservation status, distribution and taxonomic status for
species recorded only from human observations in our dataset (108 in total) in two main centers of diversity.
All species from the Brazilian center of diversity had preserved specimens in our dataset. The degree of
sampling (<10 records is considered with few records/ > 10 is considered well recorded) is based on the number
of herbarium records documented in GBIF for these species (these records were excluded during automated
filtering steps due to low quality or absent georeferenced data). Threat categories included were CR, EN and
VU according to the IUCN Red List (IUCN, 2022). Taxa included as taxonomic synonyms of other taxa in the
last 15 years were considered as involved in taxonomic uncertainty. Geographic distributions are based on
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

49
distribution ranges presented for species in Barthlott et al. (2015); species with ranges <200km2 are considered
locally restricted, species with ranges >200km2 and <600km2 are considered sub regionally restricted.
Mexico/ SW USA
Andean region
No. Species
31 (29%)
73 (68%)
Source of records
93% only from iNaturalist, 7% from
iNaturalist and other sources of human
observations
95% only from iNaturalist, 4% from
iNaturalist and other sources of human
observations, 1% only from other sources
of human observations
Herbarium
sampling
68% few records, 32% well recorded
52% few records, 48% well recorded
Conservation status
45% LC, 39% threatened, 16% DD
63% LC or NT, 19% DD or not accessed,
18% threatened
Distribution
81% local, 19% subregional
70% local, 26% subregional, 4 % not
included
Taxonomic status
100% well-defined
90% well-defined, 10% involved in
taxonomic uncertainty
Table 4. Summary data for 24 bioregions of Neotropical Cactaceae: total number of species and number of
species exclusive to the bioregion, estimated size (number of 2ºx2º grid cells), geographical location and
correspondence with biomes of Dinerstein et al. (2017) and biogeographic provinces and dominions of
Morrone et al. (2022). The ten most species-rich bioregions are highlighted in bold. * Indicate bioregions
including areas with higher phylogenetic diversity (0.06-0.241, according to Amaral et al., 2022)
Bioregion
Nº of species/
exclusives
Nº of
grids
Geographic location
Biomes
Biogeographic provinces
(dominions/transition
zones)
Bioregion
01*
134/67
40
Continuous in mid to
southern Andean
region, mostly in
Argentina and Chile
DXS,
MFWS,
MGS,
TBMF,
TGSS,
TSGSS,
TSMBF
Chaco and Pampean
(Chacoan);
Puna, Atacama,
Comechingones, Cuyan
High Andean, and Monte
(South American);
small portions of Yungas
(South Brazilian)
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

50
Bioregion 02
59/49
11
Continuous in Chile
and along the
Argentinian border
DXS,
MFWS,
MGS,
TBMF
Puna, Atacama, and Cuyan
High Andean (South
American)
Bioregion
03*
132/62
9
Continuous mostly in
southwestern Bolivia
DXS,
MGS,
TSDBF,
TSGSS,
TSMBF
Yungas and Rondônia
(South Brazilian); Chaco
(Chacoan); Puna and Monte
(South American)
Bioregion 04
82/42
11
Continuous mostly in
southern Peru and
northern Chile
DXS,
MGS,
TSMBF
Rodônia and Yungas (South
Brazilian); Desert, Puna, and
Atacama (South American)
Bioregion 05
47/24
15
Continuous in Peru
and Ecuador plus a
disjunct small area in
central Colombia
DSX,
FGS,
Ma,
MGS,
TSDBF,
TSGSS,
TSMBF
Cauca, Magdalena, Western
Ecuador and Ecuadorian
(Pacific); Napo (Boreal);
Sabana (Pacific); Ucayali and
Yungas (South Brazilian);
Páramo, Desert and Puna
(South American)
Bioregion 06
10/5
1
Restricted to a small
portion of Central Peru
TSMBF
Rodônia, Ucayali and
Yungas (South Brazilian)
Bioregion 07*
84/15
42
Nearly continuous in
eastern Argentina,
eastern Bolivia,
Paraguay, and smaller
patches in Uruguay
FGS,
MGS,
TGSS,
TSDBF,
TSGSS,
TSMBF
Rodônia, Madeira and
Yungas (South Brazilian);
Cerrado, Chaco and Pampean
(Chacoan); Araucaria,
Esteros del Iberá and Parana
Forest (Paraná); Monte and
Puna (South American)
Bioregion 08
21/7
5
Disjunct patches in
central Brazil (Cerrado
region)
TSGSS,
TSMBF
Cerrado (Chacoan)
Bioregion 09*
91/31
31
Continuous mostly in
coastal northeastern,
southeastern, southern
Brazil and southern
Paraguay
Ma,
TSDBF,
TSGSS,
TSMBF
Caatinga, Cerrado, Chaco
(Chacoan); Araucaria Forest,
Atlantic, Esteros del Iberá,
Parana Forest and Southern
Espinhaço (Paraná)
Bioregion 10
59/16
80
Mostly continuous in
northern South
America and with
disjunct patches
spread in South
America
DXS,
FGS,
Ma,
MGS,
TSCF,
TSDBF,
TSGSS,
TSMBF
All provinces in
Mesoamerican, Pacific,
Boreal Brazilian and South
Brazilian dominions; Xingu-
Tapajós (South-eastern
Amazonian); Cerrado and
Chaco (Chacoan)
Bioregion 11
44/19
7
Continuous in extreme
southern Brazil and
Uruguay
TSGSS,
TSMBF
Pampean (Chacoan); Esteros
del Iberá and Parana Forest
(Paraná)
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

51
Bioregion 12
13/2
9
Continuous in central
Brazil
FGS,
TSDBF,
TSGSS,
TSMBF
Cerrado (Chacoan); Parana
Forest (Paraná); and
Rondônia (South Brazilian)
Bioregion
13*
121/75
35
Continuous in
northeastern Brazil,
including north of
Minas Gerais state
Ma,
TSDBF,
TSGSS,
TSMBF
Caatinga and Cerrado
(Chacoan); Pará (Boreal
Brazilian); Atlantic, Chapada
Diamantina, Parana Forest
and Southern Espinhaço
(Paraná)
Bioregion
14*
156/18
17
Continuous along
Mexican pacific coast
DXS,
Ma,
TSCF,
TSDBF
Sierra Madre Occidental,
Transmexican Volcanic Belt,
and Sierra Madre del Sur
(Mexican); Pacific Lowlands
and Balsas Balsin
(Mesoamerican)
Bioregion 15
102/42
14
Continuous in Mexico
(Baja California) plus
disjunct small patches
along the Mexican
pacific coast
DXS,
Ma,
MFWS,
TSDBF
Pacific Lowlands
(Mesoamerican)
Bioregion
16*
132/23
16
Continuous in
northwestern Mexico
and extreme
southwestern USA
DXS,
Ma,
MFWS,
TCF,
TSCF,
TSDBF
Sierra Madre Occidental
(Mexican); Pacific Lowlands
(Mesoamerican)
Bioregion
17*
223/33
11
Continuous in central
Mexico
DXS,
TSCF,
TSDBF,
TSMBF
Sierra Madre Occidental,
Sierra Madre Oriental, and
Transmexican Volcanic Belt
(Mexican); Balsas Balsin,
Pacific Lowlands and
Veracruzan (Mesoamerican)
Bioregion
18*
165/35
5
Continuous in
southern central
Mexico and Oaxaca
DXS,
TSCF,
TSDBF,
TSMBF
Transmexican Volcanic Belt
and Sierra Madre del Sur
(Mexican); Balsas Basin,
Pacific Lowlands and
Veracruzan (Mesoamerican)
Bioregion
19*
283/99
37
Continuous in
northern Mexico,
southwestern USA
DXS,
Ma,
TBMF,
TCF,
TGSS,
TSCF,
TSGSS,
TSMBF
Sierra Madre Occidental and
Sierra Madre Oriental
(Mexican); Veracruzan
(Mesoamerican)
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

52
Bioregion 20
11/2
12
Continuous in Florida
(USA) plus patches in
Cuba
DXS,
FGS,
Ma,
TBMF,
TGSS
TSCF,
TSDBF,
TSMBF
Cuban (Antillean)
Bioregion 21
39/28
14
Disjunct in the
Caribbean
DXS,
FGS,
Ma,
TSCF,
TSDBF,
TSMBF
Cuban, Jamaica, Hispaniola,
Puerto Rico (Antillean)
Bioregion
22*
161/24
30
Continuous in Central
America
DXS,
Ma,
TSCF,
TSDBF,
TSGSS,
TSMBF
Chiapas Highlands and Sierra
Madre del Sur (Mexican);
Mosquito, Pacific Lowlands,
Veracruzan, Yucatán
Peninsula (Mesoamerican)
Bioregion 23
42/13
6
Continuous in the
portions of Central
America plus a
disjunct patch in
coastal Venezuela
DXS,
TSCF,
TSDBF,
TSMBF
Guatuso- Talamanca,
Puntarenas-Chiriqui and
Venezuelan (Pacific); Pacific
Lowlands (Mesoamerican)
Bioregion 24
3/3
3
Continuous in
Galapagos (Ecuador)
DXS
Galapagos Islands
(Brazilian)
Biomes abbreviations: DXS: Deserts and Xeric Shrublands; FGS: Flooded Grasslands and Savannas; Ma:
Mangroves; MFWS: Mediterranean Forest Woodlands, and Scrub; MGS: Montane Grasslands and
Shrublands; TBMF: Temperate Broadleaf and Mixed Forests; TCF: Temperate Coniferous Forests; TGSS:
Temperate Grasslands, Savannas and Shrublands; TSCF: Tropical and Subtropical Coniferous Forests;
TSDBF: Tropical and Subtropical Dry Broadleaf Forests; TSGSS: Tropical and Subtropical Grasslands,
Savannas and Shrubland; TSMBF: Tropical and Subtropical Moist Broadleaf Forests.
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

53
FIGURE LEGENDS
Figure 1. Density of records (a-d) and species richness (e-h) of Neotropical Cactaceae on
adaptive 0.5o to 2o grid cells, based on: (a, e) Preserved specimens only, (b, f) Human
observations excluding iNaturalist, (c, g) iNaturalist only, (d, h) complete dataset
including preserved specimens and human observations. Darker shades highlight
major diversity centers for the family: North and Central America, Andean Region,
Eastern Brazil. Histograms of records from 1945 to 2021 for preserved specimen (i)
and human observation (j) datasets.
Figure 2. Bioregionalization for Neotropical cacti based on preserved specimens and
human observations resulting in 24 bioregions highlighted in colors and numbered.
The images show the indicative species or the most common (in bold) species for
each bioregion: 1. Echinopsis aurea, 2. Copiapoa cinerascens, 3. Lobivia calorubra,
4. Corryocactus brevistylus, 5. Browningia microsperma, 6. Calymmanthium
substerile, 7. Acanthocalycium rhodotrichum, 8. Cereus pierre-braunianus, 9.
Rhipsalis pulchra, 10. Melocactus schatzlii, 11. Parodia scopa, 12. Cereus bicolor,
13. Melocactus zehntneri, 14. Stenocereus martinezii, 15. Echinocereus maritimus,
16. Sclerocactus erectocentrus, 17. Stenocactus phyllacanthus, 18. Mammillaria
pectinifera, 19. Echinocereus longisetus, 20. Opuntia drummondii, 21. Melocactus
lemairei, 22. Selenicereus pteranthus, 23. Epiphyllum hookeri, 24. Brachycereus
nesioticus (images of species obtained from iNaturalist, individual credits in
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint

54
acknowledgements). Note that some bioregions contain multiple species that are
equally indicative or common (see scores in item S5 below).
Figure 3: Bioregionalization for Neotropical cacti based on a hierarchical solution method
and variable Markov time using preserved specimens and human observations. Two
hierarchical levels show three bioregions in the first level (a, b) and 13 bioregions in
the second level (c). Decreased opacity in cells (a) highlight interzones (fuzzy borders)
between bioregions of the first level.
Figure 4: Bioregional schemes for the Neotropics: maps adapted from (a) Morrone et al.
(2022), (b) Dinerstein et al. (2017), (c) Griffith et al. (1998), and (d) EPA (2022).
Biomes abbreviations: DXS: Deserts and Xeric Shrublands; FGS: Flooded
Grasslands and Savannas; Ma: Mangroves; MFWS: Mediterranean Forest
Woodlands, and Scrub; MGS: Montane Grasslands and Shrublands; TBMF:
Temperate Broadleaf and Mixed Forests; TCF: Temperate Coniferous Forests;
TGSS: Temperate Grasslands, Savannas and Shrublands; TSCF: Tropical and
Subtropical Coniferous Forests; TSDBF: Tropical and Subtropical Dry Broadleaf
Forests; TSGSS: Tropical and Subtropical Grasslands, Savannas and Shrubland;
TSMBF: Tropical and Subtropical Moist Broadleaf Forests.
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (whichthis version posted July 6, 2023. ; https://doi.org/10.1101/2023.03.15.532806doi: bioRxiv preprint