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

net/publication/255721977

Species Distribution Modeling and Ecological Niche Modeling: Getting the

Concepts Right

Article in Natureza & Conservação · December 2012

DOI: 10.4322/natcon.2012.019

CITATIONS READS

615 25,688

2 authors:

Andrew Townsend Peterson Jorge Soberón

University of Kansas University of Kansas
1,085 PUBLICATIONS 99,956 CITATIONS 232 PUBLICATIONS 37,271 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Jorge Soberón on 23 April 2014.

The user has requested enhancement of the downloaded file.

 Essays & Perspectives

Natureza & Conservação 10(2):1-6, December 2012

Copyright© 2012 ABECO
Handling Editor: Paulo de Marco Jr.
Brazilian Journal of Nature Conservation

Species Distribution Modeling and Ecological Niche Modeling:

Getting the Concepts Right

A. Townsend Peterson* & Jorge Soberón

Biodiversity Institute, The University of Kansas, Kansas, USA

Abstract
We provide an overview of conceptual considerations in terminology related to ecological niche modeling and species distribution
modeling, two near-synonymous (but not quite), relatively new tools in macroecology and biogeography. We show that a large
majority of published studies taking advantage of these tools use terminology inappropriate to the biogeographic and ecological
basis on which their application is founded. We suggest that only via rigorous and appropriate terminology will these tools
achieve their fullest potential.
Key words: Geographic Range, Niche, Species Distributions, Macroecology, Biogeography.

Introduction
Interest in describing, understanding, and predicting their focus is on a subset of all ecological niche dimensions
geographic and environmental distributions of species is that is defined in coarse-resolution dimensions that are
very old (Wallace 1860; Grinnell 1917). In the last 20 years, relatively unlinked to (i.e., not affected by) the population
methods have been developed to estimate distributional processes of the species in question (Soberón 2007; Soberón
areas on the basis of correlations of known occurrences with & Nakamura 2009; Soberón 2010). A first in-depth synthesis
environmental variables. Usage of these methodologies has of concepts in the field was offered recently (Peterson et al.
literally exploded in recent years, now with hundreds of 2011), but has not as-yet seen broad uptake.
papers being published every year (Lobo et al. 2010). Both the
The debate between ENM and SDM is far from being
fundamental importance of distributional areas as a concept
merely semantic. It is perfectly feasible to model species’
in biogeography and ecology, and myriad applications to
distributions without resorting to a niche definition or even
which these methods can be applied explain the massively
referring to any environmental variables (Jennrich & Turner
increasing interest in this field (Peterson et al. 2011).
1969; Rapoport 1982; Bahn & McGill 2007). On the other
Since the late 1990s, two terminologies have been used to hand, modeling the processes that produce and shape the
refer to correlative summaries of species’ environmental area of distribution, transferring causal factors in time or
associations and the relationships of those associations in space, or interpreting biologically the obtained pattern,
to their geographic distributions: “species distribution obviously requires some hypothesis about the ecology of
models” (SDM; Elith & Leathwick 2009; Franklin 2010) the species, which is clearly a niche-related inquiry.
and “ecological niche models” (ENM; Harrison 1997;
Peterson et al. 1999); other, more neutral terms have also In this paper, we discuss this dichotomy of terminology. To
been used, such as “bioclimatic envelope models” (Araújo characterize how these terminologies are used in the field,
& Peterson 2012). Arguments regarding these terms have we searched Web of Science for journal articles that used
been numerous (Peterson 2006; Elith & Leathwick 2009; “species distribution model” or “ecological niche model”
Franklin 2010; Sillero 2011; Araújo & Peterson 2012; Warren under topic or title. Of the 242 papers that resulted, 13
2012) and rather inconclusive. In general, it appears that were by one or the other of us, and so were removed from
those who use SDM prefer to avoid overinterpreting the consideration. A further three used the term SDM, but in a
ecological significance of the model, or worry that many very different sense more closely related to environmental
dimensions of true ecological niches remain uncharacterized chemistry (Lee et al. 1995; Shen et al. 1995; Shen & Lin
by these methods. The ENM world has emphasized that 2003); these papers were also removed from analysis.
Because these questions involve subtleties of how terms
*
Send correspondence to: A. Townsend Peterson
Biodiversity Institute, The University of Kansas, Lawrence,
are used, we focused on the abstract associated with each
Kansas, 66045, USA publication, and the terminology used therein. Of the 226
e-mail: [email protected] papers analyzed, 127 used SDM, 51 used ENM, and 50 were
2 Peterson & Soberón Natureza & Conservação 10(2):1-6, December 2012

nonspecific or vague about the concept used or used other distributional area GO; finally, the presently-inaccessible
terminologies (Table 1). Indeed, two papers used both terms portions of the potential distributional area are termed
(Colacicco-Mayhugh et al. 2010; Roubicek et al. 2010)! As the invadable area GI.
will be analyzed in greater depth below, no clear divisions
Besides G, another space is required: in the multivariate
existed regarding the uses to which these models were put,
space of environmental dimensions, the environments
such that—in effect—we see simply different terms being
associated with G are η(G), or E. Each of the distributional
used for the same thing.
areas discussed above has a corresponding entity in
environmental space: to refer to corresponding elements
Conceptual Framework and Real-world in environmental space of any distributional area, we use
Implications the notation E’ = η(G’); its converse operation is η-1(E’),
All of these questions regarding the complexities of ecological which identifies the geographic location(s) that correspond
niches and geographic distributions require careful and to a given environmental combination E’.
consistent terminology (Peterson et al. 2011). Although The fundamental niche NF, however, is not generally
other ‘vocabularies’ have been offered (Godsoe 2010a; Sillero fully represented in real-world geography—that is, the
2011), they have not successfully captured the complexity fundamental niche is a construct of physiological responses,
of the situation: species are distributed in both geographic and may extend to sets of conditions not represented within
and environmental dimensions (Colwell & Rangel 2009), the study area. The subset of the fundamental niche that is
and both geographic and ecological distributions must be actually represented on relevant landscapes, or the existing
captured and linked in any effective terminology. We use the fundamental niche, can be defined as N F ∩ η ( M ) = N*F
Biotic-Abiotic-Mobility (BAM; Figure 1) framework (Soberón (Jackson & Overpeck 2000), and indeed this estimate may
& Peterson 2005) to contemplate the suites of factors that be reduced still further if sampling is incomplete across the
determine geographic distributions of species, and that landscape: if the area sampled is S, then we estimate a yet-

have been known and discussed for many decades (Grinnell different set of environments N*F = N F ∩ η ( M ∩ S) ⊆ N*F .
1924; Good 1931). This view of species’ distributional This reduction of the fundamental niche is critical in the
ecology captures and frames many of the key phenomena clarifications offered in this paper, as it basically indicates
for which models of niches and distributions have been that any ecological niche that can be estimated on real-world
used: characterizing niches; interpolating distributional landscapes for real-world species will be only partial, and
patterns; and anticipating unknown distributional areas, that these niches can be compared only as conditioned on
geographic potential of invasive species, and responses to the M and S areas corresponding to each of the species in
changing environmental conditions. question(Jiménez-Valverde et al. 2008).
We contemplate two related spaces. In geographic space Note that some crucial points have been made here. If
G, we can consider several different distributional areas. ‘niche modeling’ were genuinely an attempt toestimate
Areas appropriate in terms of abiotic conditions can be fundamental niches, which could then be used to outline
termed GA, or the set of areas that meet the conditions of GA, then they would not be estimating actual distributional
the fundamental niche NF of the species. A similar set of areas, but rather GP, or the potential distribution of the
constraints applies to biotic conditions, wherein a subset species. Hence, niche model outputs must be processed
of G is identified that meets the suite of biotic conditions further (i.e., reduction to the portion of GA that is within M)
necessary for the maintenance of populations of the species. if an actual (occupied) distributional area is to be identified.
Areas that have been accessible to the species over relevant Second, on the environmental side, these approaches
time periods can be termed M, which has been discussed at best can characterize N*F , which is the fundamental
in detail in a previous contribution (Barve et al. 2011). niche conditioned on the associated M and S—these niche
Areas that fulfill both the abiotic and biotic requisites of estimates are not comparable without explicit reference
the species constitute the potential geographic distribution to M and S. These two points have massive implications
of the species GP, and the intersection of the potential for what researchers are ‘doing’ with these techniques at
distribution with the accessible areas is termed the occupied the moment.

Table 1. Summary of uses to which different studies have put analyses under the rubrics of ‘ecological niche modeling’ and ‘species
distribution modeling’.

Ecological niche modeling Species distribution modeling Total

Climate change projections 12 35 47
Distributional prediction 19 62 81
Species invasions projections 7 22 29
Niche characterization 2 3 5
Paleodistributional predictions 11 4 15
Total 51 126 177
Species Distribution and Ecological Niche Modeling 3

comparing environments associated with occurrences η (G + )

with those associated with some set of data representing
non-occurrences. Hence, the model is developed in
environmental space, and the target of modeling is some
subset of E, and not an area of G. However, in most modern
studies, the model is then used to classify areas of geographic
space as to whether they present conditions that are in some
sense similar to η (G + ) or not—this assessment of similarity
establishes whether a site is considered as suitable or not
for the species. Hence, in this process, the initial steps
are carried out in geographic space, the model is fitted
in environmental space, and model outputs are generally
visualized in geographic space.

Species Distribution Modeling or

Ecological Niche Modeling?
These seemingly different terms refer to what is in effect
Figure 1. A BAM diagram configuration that may be most the same set of analyses, as can be appreciated from the
relevant to broad geographic questions about species’ geographic Introduction and Table 1. In the strictest sense, one could
distributions. That is, as argued by Soberón (2007), biotic assert that the target of model fitting is inevitably some
considerations may frequently be diffuse and non-limiting, or entity in environmental space, which is the realm of niches,
may be manifested at spatial resolutions so fine as to be nestled
within the coarse-resolution phenomena that are the focus
and not distributions. In that sense, essentially all modern
of most ecological niche modeling and species distribution applications of these techniques should be considered
modeling studies. The circle A represents the parts of the models of ecological niches, which then may be used for
world that contain the abiotic conditions required for a species’ a variety of purposes—understanding actual or potential
survival and growth. The circle M represents the region that has
distributional areas, characterizing ecological niches, etc.
been accessible to the species over a relevant period of time.
The intersection of these two regions is the occupied area GO; Most SDM/ENM studies use something chosen from
GI holds the correct suite of environmental conditions, but
basically the same set of mathematical algorithms, similar
has not been explored by the species; open circles represent
presence data; closed circles indicate absences due to incorrect sorts of occurrence data, and the same sets of environmental
environment; closed squares indicate absences due to lack of variables, and both can produce outputs in the form of
dispersal capacity; and triangles indicate absences owing to both maps; indeed, papers from the two are published in the
incorrect conditions and limited dispersal.Note that absences same journals. As a consequence, many people fail to
can also occur within GO as well. From Saupe et al. (2012).
perceive that the questions posed in modeling distributions
(in a narrow sense) and modeling environments are very
different. That is, in SDM and ENM, researchers attempt to
What is the Target of Modeling?
calculate very different objects, and a distinction between
Some early thinking approaches to understanding these two concepts would focus on how geographic areas
species’ distributions were developed in exclusively of interest are identified and characterized. To model a
spatial dimensions, while others were developed in species’ distribution, in view of the conceptual framework
exclusively environmental dimensions (Hutchinson 1957; offered above, one must estimate NF, which in turn allows
Brown et al. 1996). Indeed, space-only approaches to identification of some hypothesis of GA. Separately, one
mapping species’ distributions continue strong in the must assess M, which is often not a simple enterprise
field of spatial epidemiology, in which, in many studies, (Barve et al. 2011). Then, the species’ actual (occupied)
disease distributions are estimated and characterized only distributional area is GO= GA∩M. So, in sum, genuine
in dimensions of space, and frequently with no reference “species distribution modeling” must include steps of
to environmental conditions (Pfeiffer et al. 2008; Lai et al. niche estimation and steps of assessment of dispersal
2009). The environment-only tradition continues also, but ability or colonization potential (e.g., van Loon et al. 2011).
quietly, hidden as an unstated assumption of many studies The following section (“When are Species’ Distributions
under the rubric of ecological niche modeling or species Modeled Correctly?”) assesses how frequently studies in
distribution modeling (see “When are Species’ Distributions this field take all of these necessary steps.
Modeled Correctly?” below).
Certainly, though, some combinations of these terms and
In essentially all modern applications of these techniques, particular applications are not correct. SDM can focus only
however, the process involves at least two transitions between on the actual distribution—any other use, such as estimating
spaces. That is, known occurrences, G+, are distributed invasive potential or assessing effects of environmental
across geographic space. The modeling process involves change on species’ distributional potential, requires an
4 Peterson & Soberón Natureza & Conservação 10(2):1-6, December 2012

explicit estimate of NF, and as such must fall under the When are Ecological Niches Modeled
rubric of ENM. Indeed, of SDM studies, almost half focus on and Compared Correctly?
distributional prediction (62 of 126 studies, 49.2%); however,
27.8% focus on climate change projections, 17.5% focus on Only five studies focused explicitly on estimating and
projections of distributional potential of invasive species, comparing ecological niches out of the sample that we
3.2% make projections of paleodistributional patterns, considered (Table 1); of these studies, three used SDM
and 2.4% set out explicitly to characterize the species’ terminology and two used ENM terminology, so no
ecological niche (!), as is detailed in Table 1. These 50.8% strong tendency was noted. Considering the need for
of species distribution models thus use inconsistent and niche estimates and their comparisons to be conditioned
illogical terminology, and should rather be termed ENMs. on species-specific estimates of M (Barve et al. 2011) as
discussed above, we assessed each of the five studies as to
When are Species’ Distributions whether some species-specific area was included in the
Modeled Correctly? comparisons.
One paper (Rodder & Engler 2011) discussed metrics of
For those studies that set out to characterize species’
niche overlap and cited key concepts and references, but
geographic distributions (19 ENMs and 62 SDMs), a further
did not enter into detail about means of comparison and
question is whether concepts are managed correctly. As
manipulation that they recommended (Warren et al. 2008).
detailed above (see “Conceptual Framework and Real-world
Of the remaining four papers, three appropriately considered
Implications”), the term ‘distribution’ has many versions,
species-specific areas in their niche comparisons (Godsoe
such that any study purporting to estimate distributions
2010b; Nakazato et al. 2010; Schulte et al. 2012), but one
must either (1) restrict model calibration to accessible
relied on direct comparisons of estimates of niches and
areas (= M), for which assumptions regarding access must
distributions, which is quite perilous and likely misleading
be stated explicitly (Barve et al. 2011), thereby estimating
(Hoisington-Lopez et al. 2012).
GO directly; (2) trim modeled hypotheses of GA to match
hypotheses of accessibility and thereby estimate GO; (3)
Conclusions
incorporate dispersal into analyses explicitly (Robinson et al.
2011; Boulangeat et al. 2012), again estimating GO directly; We reviewed a representative swath of the recent scientific
(4) use true absences as a contrast to presences in model literature, pondering the degree to which these publications
calibration and expect that restrictive factors will be captured use different terminologies, and the degree to which they
indirectly as a consequence (Ward et al. 2009), or (5) state hold logical and conceptual inconsistencies. We found
explicitly that the model outputs are potential distributions considerable cause for concern, such that half of the studies
and not actual distributional area predictions. Hence, we using the terminology “species distribution modeling”
reviewed the methods section of each of the studies that nonetheless take advantage of the models for uses related
had distributional prediction as a main objective to assess to ecological niches. What is more, about two-thirds of
the degree to which they handled the distinction between SDM-termed studies and a non-zero number of ENM-termed
actual and potential distributions appropriately. studies that focus on estimating geographic distributions
Of the 81 studies, we discarded 16 because they were not fail to make appropriate distinctions between actual and
empirical in nature and thereby did not confront these potential distributional areas. Finally, one-quarter of a very
issues directly, and we discarded a further 5 studies because small sample of studies comparing estimates of ecological
access to the full published version was overly difficult, even niches of species used inappropriate methodologies as
well. In sum, an impressive proportion of studies in this
under the University of Kansas’ rather expensive electronic
emerging field of whole-range species-level ecological
journal access expenditures (such difficult-to-access journals
biogeography carry logical and conceptual inconsistencies
will have to survive without our citations!); hence, we
that compromise the rigor of their conclusions.
analyzed a total of 60 papers. Of these 60 papers, 32 made
no indication of any species-specific considerations as to It is certainly tempting to pass these concerns off as ‘just
access to areas or any of the criteria listed above; although words’ (see, e.g., Godsoe 2010a). We argue that this field
some of these analyses were cast on such small spatial is compromised by this lack of rigor, however: not only is
extents that access was probably assumed (but not stated its credibility in the broader scientific community reduced
explicitly), a large proportion of these studies simply did (Hampe 2004; Sinclair et al. 2010), but also many confusions
not state any explicit thought to these considerations. and inappropriate conclusions have been reached for lack
A significant 2 × 2 interaction existed between whether of conceptual rigor (Peterson 2011; Soberón & Peterson
the authors used the SDM versus ENM terminology, and 2011; Araújo & Peterson 2012). “ENM” should be used only
whether accounting was made for the difference between when focus is on estimation of NF or GA or any potential
potential and actual distributions, with those using SDM distribution under changed conditions and circumstances,
tending towards lack of documentation more frequently but care must be taken to distinguish these quantities from
than expected at random (χ2 = 10.01, df = 1, P < 0.002). their ‘existing’ or ‘realized’ manifestations; on the other hand,
Species Distribution and Ecological Niche Modeling 5

“SDM” must include steps to transform areas estimated Good RD, 1931. A theory of plant geography.
from potential to actual, so as to reconstruct distributions New Phytologist, 30:149-171. http://dx.doi.
accurately. If this field is to mature into an important and org/10.1111/j.1469-8137.1931.tb07414.x
unique element in the ecological and biogeographic toolkit, Grinnell J, 1917. Field tests of theories concerning distributional
greater conceptual rigor will be required. control. American Naturalist, 51:115-128. http://dx.doi.
org/10.1086/279591
Acknowledgements Grinnell J, 1924. Geography and evolution. Ecology, 5:225-229.
http://dx.doi.org/10.2307/1929447
Our work was supported by a grant from Microsoft Research
Hampe A, 2004. Bioclimate envelope models: What they detect
(#47780).
and what they hide. Global Ecology and Biogeography, 13:469-
471. http://dx.doi.org/10.1111/j.1466-822X.2004.00090.x
References
Harrison S, 1997. How natural habitat patchiness affects
Araújo MB & Peterson AT, 2012. Uses and misuses of bioclimatic the distribution of diversity in Californian serpentine
envelope modelling. Ecology, 93:1527-1539. http://dx.doi. chaparral. Ecology, 78:1898-1906. http://dx.doi.
org/10.1890/11-1930.1 org/10.1890/0012-9658(1997)078[1898:HNHPAT]2.0.CO;2

Bahn V & McGill BJ, 2007. Can niche-based distribution Hoisington-Lopez JL, Waits LP & Sullivan J, 2012. Species
models outperform spatial interpolation? Global limits and integrated taxonomy of the Idaho ground
Ecology and Biogeography, 16:733-742. http://dx.doi. squirrel (Urocitellus brunneus): Genetic and ecological
differentiation. Journal of Mammalogy, 93:589-604. http://
org/10.1111/j.1466-8238.2007.00331.x
dx.doi.org/10.1644/11-MAMM-A-021.1
Barve N et al., 2011. The crucial role of the accessible area
Hutchinson GE, 1957. Concluding remarks. Cold Spring Harbor
in ecological niche modeling and species distribution
Symposia on Quantitative Biology, 22:415-427. http://dx.doi.
modeling. Ecological Modelling, 222:1810-1819. http://
org/10.1101/SQB.1957.022.01.039
dx.doi.org/10.1016/j.ecolmodel.2011.02.011
Jackson ST & Overpeck JT, 2000. Responses of plant populations
Boulangeat I, Gravel D & Thuiller W, 2012. Accounting and communities to environmental changes of the Late
for dispersal and biotic interactions to disentangle Quaternary. Paleobiology, 26(Supplement):194-220. http://
the drivers of species distributions and their dx.doi.org/10.1666/0094-8373(2000)26[194:ROPPAC]
abundances. Ecology Letters, 15:584-593. http://dx.doi. 2.0.CO;2
org/10.1111/j.1461-0248.2012.01772.x
Jennrich RI & Turner FB, 1969. Measurment of non-circular
Brown JH, Stevens GC & Kaufman DM, 1996. The geographic home range. Journal of Theoretical Biology, 22:227-237.
range: Size, shape, boundaries, and internal structure. Annual http://dx.doi.org/10.1016/0022-5193(69)90002-2
Review of Ecology and Systematics, 27:597-623. http://dx.doi. Jiménez-Valverde A, Lobo JM & Hortal J, 2008. Not as good as
org/10.1146/annurev.ecolsys.27.1.597 they seem: The importance of concepts in species distribution
modelling. Diversity and Distributions, 14:885-890. http://
Colacicco-Mayhugh MG, Masuoka PM & Grieco JP, 2010.
dx.doi.org/10.1111/j.1472-4642.2008.00496.x
Ecological niche model of Phlebotomus alexandri and
P. papatasi (Diptera: Psychodidae) in the Middle East. Lai P-C, So F-M & Chan K-W, 2009. Spatial Epidemiological
International Journal of Health Geographics, 9. Approaches in Disease Mapping and Analysis. Boca Raton:
CRC Press.
Colwell RK & Rangel TF, 2009. Hutchinson’s duality: The once
and future niche. Proceedings of the National Academy of Lee KW, Cho SH & Park SW, 1995. Studies on the treatment
of waste-water bearing cyanide and heavy-metals by
Sciences USA, 106:19644-19650. http://dx.doi.org/10.1073/
micelle enhanced ultrafiltration technique. Journal of
pnas.0901650106 Environmental Science and Health A, 30:467-484. http://
Elith J & Leathwick J, 2009. Species distribution models: dx.doi.org/10.1080/10934529509376212
Ecological explanation and prediction across space Lobo JM, Jiménez-Valverde A & Hortal J, 2010. The uncertain
and time. Annual Review of Ecology, Evolution, and nature of absences and their importance in species
Systematics, 40:677-697. http://dx.doi.org/10.1146/annurev. distribution modelling. Ecography, 33:103-114. http://
ecolsys.110308.120159 dx.doi.org/10.1111/j.1600-0587.2009.06039.x
Franklin J, 2010. Mapping Species Distributions: Spatial Inference Nakazato T, Warren DL & Moyle LC, 2010. Ecological and
and Prediction. Cambridge: Cambridge University Press. geographic models of species divergence in wild tomatoes.
http://dx.doi.org/10.1017/CBO9780511810602 American Journal of Botany, 97:680-693. http://dx.doi.
org/10.3732/ajb.0900216
Godsoe W, 2010a. I can’t define the niche but I know it
Peterson AT, 2006. Uses and requirements of ecological niche
when I see it: A formal link between statistical theory
models and related distributional models. Biodiversity
and the ecological niche. Oikos, 119:53-60. http://dx.doi. Informatics, 3:59-72.
org/10.1111/j.1600-0706.2009.17630.x
Peterson AT, 2011. Ecological niche conservatism:
Godsoe W, 2010b. Regional variation exaggerates ecological A time-structured review of evidence. Journal
divergence in niche models. Systematic Biology, 59:298-306. of Biogeography, 38:817-827. http://dx.doi.
http://dx.doi.org/10.1093/sysbio/syq005 org/10.1111/j.1365-2699.2010.02456.x
6 Peterson & Soberón Natureza & Conservação 10(2):1-6, December 2012

Peterson AT et al., 2011. Ecological Niches and Geographic Sillero N, 2011. What does ecological modelling model? A
Distributions. Princeton: Princeton University Press. proposed classification of ecological niche models based on
their underlying methods. Ecological Modelling, 222:1343-
Peterson AT, Soberón J & Sánchez-Cordero V, 1999. 1346. http://dx.doi.org/10.1016/j.ecolmodel.2011.01.018
Conservatism of ecological niches in evolutionary time.
Science, 285:1265-1267. http://dx.doi.org/10.1126/ Sinclair SJ, White MD & Newell GR, 2010. How useful are
science.285.5431.1265 species distribution models for managing biodiversity
under future climates? Ecology and Society, 15:8.
Pfeiffer D et al., 2008. Spatial Analysis in Epidemiology. Oxford:
Oxford University Press. http://dx.doi.org/10.1093/acprof Soberón J, 2007. Grinnellian and Eltonian niches and geographic
:oso/9780198509882.001.0001 distributions of species. Ecology Letters, 10:1115-1123. http://
dx.doi.org/10.1111/j.1461-0248.2007.01107.x
Rapoport EH, 1982. Aerography: Geographical Strategies of
Soberón J, 2010. Niche and area of distribution modeling: a
Species. Oxford: Pergamon Press.
population ecology perspective. Ecography, 33:159-167.
Robinson LM et al., 2011. Pushing the limits in marine http://dx.doi.org/10.1111/j.1600-0587.2009.06074.x
species distribution modelling: lessons from the
Soberón J & Nakamura M, 2009. Niches and distributional
land present challenges and opportunities. Global
areas: Concepts, methods, and assumptions. Proceedings of
Ecology and Biogeography, 20:789-802. http://dx.doi. the National Academy of Sciences USA, 106:19644-19650.
org/10.1111/j.1466-8238.2010.00636.x http://dx.doi.org/10.1073/pnas.0901637106
Rodder D & Engler JO, 2011. Quantitative metrics of overlaps Soberón J & Peterson AT, 2005. Interpretation of models of
in Grinnellian niches: advances and possible drawbacks. fundamental ecological niches and species’ distributional
Global Ecology and Biogeography, 20:915-927. http://dx.doi. areas. Biodiversity Informatics, 2:1-10.
org/10.1111/j.1466-8238.2011.00659.x
Soberón J & Peterson AT, 2011. Ecological niche shifts and
Roubicek AJ et al., 2010. Does the choice of climate baseline environmental space anisotropy: A cautionary note. Revista
matter in ecological niche modelling? Ecological Mexicana de Biodiversidad, 82:1348-1353.
Modelling, 221:2280-2286. http://dx.doi.org/10.1016/j.
Van Loon AH et al., 2011. Linking habitat suitability and
ecolmodel.2010.06.021
seed dispersal models in order to analyse the effectiveness
Saupe E et al., 2012. Variation in niche and distribution model of hydrological fen restoration strategies. Biological
performance: The need for a priori assessment of key causal Conservation, 144:1025-1035. http://dx.doi.org/10.1016/j.
factors. Ecological Modelling, 237-238:11-22. http://dx.doi. biocon.2010.12.021
org/10.1016/j.ecolmodel.2012.04.001 Wallace AR, 1860. On the zoological geography of
Schulte U et al., 2012. Cryptic niche conservatism among the Malay Archipelago. Proceedings of the Linnean
evolutionary lineages of an invasive lizard. Global Society of London, 4:172-184. http://dx.doi.
Ecology and Biogeography, 21:198-211. http://dx.doi. org/10.1111/j.1096-3642.1860.tb00090.x
org/10.1111/j.1466-8238.2011.00665.x Ward G et al., 2009. Presence-only data and the EM
Shen YS, Ku Y & Lee KC, 1995. The effect of light absorbency on algorithm. Biometrics, 65:554-563. http://dx.doi.
the decomposition of chlorophenols by ultraviolet-radition org/10.1111/j.1541-0420.2008.01116.x
and UV/H2O2 processes. Water Research, 29:907-914. http:// Warren DL, 2012. In defense of ‘niche modeling’. Trends in
dx.doi.org/10.1016/0043-1354(94)00198-G Ecology & Evolution, 27:497-500. http://dx.doi.org/10.1016/j.
tree.2012.03.010
Shen YS & Lin CC, 2003. The effect of pH on the decomposition
of hydrophenols in aqueous solutions by ultraviolet direct Warren DL, Glor RE & Turelli M, 2008. Environmental niche
photolysis and the ultraviolet-hydrogen peroxide process. equivalency versus conservatism: quantitative approaches
Water Environment Research, 75:54-60. http://dx.doi. to niche evolution. Evolution, 62:2868-2883. http://dx.doi.
org/10.2175/106143003X140827 org/10.1111/j.1558-5646.2008.00482.x

Received: June 2012

First Decision: August 2012
Accepted: August 2012

View publication stats