Content uploaded by Jennifer G Howeth
Author content
All content in this area was uploaded by Jennifer G Howeth on Nov 06, 2019
Content may be subject to copyright.
BIODIVERSITY
RESEARCH
Predicting invasiveness of species in
trade: climate match, trophic guild and
fecundity influence establishment and
impact of non-native freshwater fishes
Jennifer G. Howeth
1
*, Crysta A. Gantz
2
, Paul L. Angermeier
3,4
,
Emmanuel A. Frimpong
3
, Michael H. Hoff
5
, Reuben P. Keller
6
,
Nicholas E. Mandrak
7
, Michael P. Marchetti
8
, Julian D. Olden
9
,
Christina M. Romagosa
10
and David M. Lodge
2
1
Department of Biological Sciences,
University of Alabama, Tuscaloosa, AL
35487, USA,
2
Department of Biological
Sciences and Notre Dame Environmental
Change Initiative, University of Notre Dame,
Notre Dame, IN 46556, USA,
3
Department
of Fish and Wildlife Conservation, Virginia
Polytechnic Institute and State University,
Blacksburg, VA 24061, USA,
4
U. S.
Geological Survey, Virginia Cooperative Fish
and Wildlife Research Unit, Virginia
Polytechnic Institute and State University,
Blacksburg, VA 24061, USA,
5
Fisheries
Program, United States Fish and Wildlife
Service, Bloomington, MN 55437, USA,
6
Institute of Environmental Sustainability,
Loyola University Chicago, Chicago, IL
60660, USA,
7
Department of Biological
Sciences, University of Toronto Scarborough,
Toronto, ON M1C 1A4, Canada,
8
Department of Biology, St. Mary’s College
of California, Moraga, CA 94556, USA,
9
School of Aquatic and Fishery Sciences,
University of Washington, Seattle, WA
98195, USA,
10
Department of Wildlife
Ecology and Conservation, University of
Florida, Gainesville, FL 32611, USA
*Correspondence: Jennifer G. Howeth,
Department of Biological Sciences, University
of Alabama, Tuscaloosa, AL 35487, USA.
E-mail: jghoweth@ua.edu
ABSTRACT
Aim Impacts of non-native species have motivated development of risk assess-
ment tools for identifying introduced species likely to become invasive. Here,
we develop trait-based models for the establishment and impact stages of fresh-
water fish invasion, and use them to screen non-native species common in
international trade. We also determine which species in the aquarium, biological
supply, live bait, live food and water garden trades are likely to become inva-
sive. Results are compared to historical patterns of non-native fish establishment
to assess the relative importance over time of pathways in causing invasions.
Location Laurentian Great Lakes region.
Methods Trait-based classification trees for the establishment and impact
stages of invasion were developed from data on freshwater fish species that
established or failed to establish in the Great Lakes. Fishes in trade were deter-
mined from import data from Canadian and United States regulatory agencies,
assigned to specific trades and screened through the developed models.
Results Climate match between a species’ native range and the Great Lakes
region predicted establishment success with 75–81% accuracy. Trophic guild
and fecundity predicted potential harmful impacts of established non-native
fishes with 75–83% accuracy. Screening outcomes suggest the water garden
trade poses the greatest risk of introducing new invasive species, followed by
the live food and aquarium trades. Analysis of historical patterns of introduc-
tion pathways demonstrates the increasing importance of these trades relative
to other pathways. Comparisons among trades reveal that model predictions
parallel historical patterns; all fishes previously introduced from the water gar-
den trade have established. The live bait, biological supply, aquarium and live
food trades have also contributed established non-native fishes.
Main conclusions Our models predict invasion risk of potential fish invaders
to the Great Lakes region and could help managers prioritize efforts among
species and pathways to minimize such risk. Similar approaches could be
applied to other taxonomic groups and geographic regions.
Keywords
Aquarium, biological invasions, classification tree, ecological impact, establish-
ment success, exotic species, Laurentian Great Lakes, live food, risk assess-
ment, water garden.
DOI: 10.1111/ddi.12391
ª2015 John Wiley & Sons Ltd http://wileyonlinelibrary.com/journal/ddi 1
Diversity and Distributions, (Diversity Distrib.) (2015) 1–13
A Journal of Conservation Biogeography
Diversity and Distributions
INTRODUCTION
Increased intentional and unintentional transport of live
organisms associated with international commerce has led to
the establishment of many species well beyond their native
geographic ranges (Levine & D’Antonio, 2003; Hulme, 2009).
Biological invasions have subsequently altered patterns of
biodiversity from local to continental scales (Wilcove et al.,
1998; Baiser et al., 2012) and resulted in both ecological and
economic impacts on entire ecosystems (Pimentel et al.,
2005; Strayer et al., 2006). These impacts have motivated the
development of risk assessment tools that identify non-native
species likely to establish and have impact (Leung et al.,
2012; Ib
a~
nez et al., 2014). These tools can be used to priori-
tize both the species and geographic regions for which the
risk of establishment and impact is greatest, thus facilitating
more focused, efficient management actions that target
prevention (Keller & Perrings, 2011; Jenkins, 2013).
Risk assessments for evaluating species imported via inter-
national trade have been under-utilized relative to the mag-
nitude of the invasion threat posed (Keller & Lodge, 2007;
Smith et al., 2009). This is illustrated in freshwater ecosys-
tems where the risk of species invasion is high (Padilla &
Williams, 2004; Keller & Lodge, 2007), especially with respect
to non-native fishes (Olden et al., 2010). Risk assessments
based on species’ ecological traits associated with specific
stages of invasion can offer scientifically rigorous support for
efforts to prevent invasion (Marchetti et al., 2004a; Leung
et al., 2012). Aggregating species assessment results across
specific trade pathways can also help decision-makers deter-
mine which pathways pose the greatest risk of introducing
new invaders, and thus deserve greater management or
policy attention (Pysek et al., 2011).
Fishes constitute approximately 90% of all individual ani-
mals imported into the United States annually (Smith et al.,
2009). Many of these fishes are sold through the aquarium,
biological supply, live bait, live food and water garden trades
(Keller & Lodge, 2007; Smith et al., 2008). Some of these
fishes are subsequently released by humans into the wild. For
example, 2% and 3% of aquarium and water garden owners
in the Great Lakes basin, respectively, release unwanted fishes
(Marson et al., 2009a,b). Further, >6% of aquarists in the
northwestern United States have released live fishes (Strecker
et al., 2011). As different trade pathways select species for
different purposes, it is likely that they pose different overall
risks for the recipient ecosystems. For example, the water
garden trade requires fishes that can survive in outdoor
ambient conditions, while the aquarium trade imposes no
such limitation. Knowledge of which trades introduce the
most high-risk species would facilitate more efficient prioriti-
zation of management actions and policy improvements
(Lodge et al., 2006).
Traits that are highly predictive of non-native fish estab-
lishment differ across studies and geographic regions, but
some traits are consistently correlated with establishment
success (Garc
ıa-Berthou, 2007). For example, temperature
tolerance (Kolar & Lodge, 2002; Bomford et al., 2010),
trophic status (Marchetti et al., 2004b; Ruesink, 2005),
spawning habitat requirements (Mandrak, 1989; Olden et al.,
2006) and history of invasion outside the native range (Kolar
& Lodge, 2002; Ribeiro et al., 2008) have emerged as robust
indicators of establishment success. In contrast, traits associ-
ated with invasion impacts have exhibited greater variation
across studies (Garc
ıa-Berthou, 2007), suggesting that addi-
tional understanding is required to model this stage of
invasion.
The Laurentian Great Lakes of North America serve as an
important region for developing and testing trait-based risk
assessments of potentially invasive species. Hundreds of non-
native species, including at least 65 freshwater fishes, have
been introduced to the Great Lakes basin from multiple
pathways over the last two centuries (Mandrak & Cudmore,
2010). The resulting invasions have altered ecosystem func-
tions in undesirable ways and jeopardized valued services
provided by the Great Lakes (Rothlisberger et al., 2010,
2012). Major trade pathways, including the aquarium, bio-
logical supply, live bait, live food and water garden trades,
have the potential to introduce many more unwanted non-
native freshwater fishes to the Great Lakes region (Ricciardi,
2006; Keller & Lodge, 2007).
Here, we develop trait-based risk assessment tools for the
establishment and impact stages of non-native fishes in the
Great Lakes. We then use the models to identify traded
non-native fishes that pose a threat of becoming invasive in
the Great Lakes ecosystem and assess the relative risks of
the aquarium, biological supply, live bait, live food and
water garden trades for future invasions. We compare our
analysis of trade pathway risks to historical patterns of non-
native fish establishment in the Great Lakes to evaluate the
relative contribution of trade to invasion over time and to
identify which trades have historically supplied invaders
with the highest rates of establishment success. In addition
to advancing trait-based approaches, our results could
inform current practices and future policies to prevent fish
invasions via trade and thus safeguard the goods and ser-
vices provided by the Great Lakes and other freshwater
ecosystems.
METHODS
Predicting establishment and impact
We identified 65 non-native fish species introduced to the
Great Lakes, 37 of which have established self-sustaining
(reproducing) populations (Mandrak & Cudmore, 2010; see
Table S1). The degree of ecological impacts associated with
established non-native fishes is difficult to determine from
literature because not all species have been studied and
impacts differ across the region. Therefore, we evaluated the
ecological impact of the 37 established species with a ques-
tionnaire (Appendix S1) distributed to university and gov-
ernment agency experts (Table S2) with experience in the
2Diversity and Distributions, 1–13, ª2015 John Wiley & Sons Ltd
J. G. Howeth et al.
Great Lakes region and substantial knowledge of aquatic
invasive species, fisheries and/or ecosystems. Experts were
chosen based upon their specific expertise in particular lake
basins to yield even, complete coverage of the Great Lakes
region. The questionnaire asked experts to assign each spe-
cies to one of four categories ranging from no perceived eco-
logical impact to very high perceived ecological impact in
the Great Lakes (definitions modified from Bradford et al.,
2009 and Mandrak et al., 2012) and to assign either low or
high confidence to their answer. Respondents could decline
to rank a species if they were not familiar with its impact. A
total of 33 experts in Canada and the United States received
the questionnaire via electronic mail on 24 August 2011 and
responses were received from 27 (82%) experts (Table S2).
The 27 respondents cumulatively reported over 557 years of
research experience in the Great Lakes, indicating that the
impact scores come from a scientific community that is well-
informed about fishes in the Great Lakes region.
We calculated the mean and 95% confidence interval (CI)
of responses for each species and then ranked the species
from lowest to highest impact based upon the upper CI
value to incorporate variation in respondent answers. There
were no obvious breaks in the impact rankings across the 37
species (i.e. impact ranks increased approximately linearly
from lowest to highest). Therefore, for the analysis of
impact, we compared the 12 species with the highest and
lowest impacts, respectively (Table 1). This approach maxi-
mized our ability to detect differences in species traits and
invasion risk parameters between low- and high-impact
species.
Species traits and invasion risk parameters
From the literature, we identified 18 species traits and inva-
sion risk parameters (details in Table S3) that have been pre-
viously associated with the establishment and impact stages
of invasion (reviewed in Garc
ıa-Berthou, 2007) and for
which sufficient data typically exist in publicly available data-
bases. These 18 traits and parameters were categorized a pri-
ori as follows. Ecological traits included (1) diet breadth
(sum of diet items, including algae/phytoplankton, vascular
plants, detritus, aquatic/terrestrial invertebrates and larval
fishes, fishes/crayfishes/crabs/frogs, blood, eggs); (2) macro-
habitat association (lentic, lotic); (3) salinity tolerance (nar-
row, wide); (4) temperature tolerance [cold (10–17°C), cold/
cool, cool (18–26°C), cool/warm, warm (>26°C)]; and (5)
trophic guild (herbivore–detritivore, invertivore, invertivore–
piscivore, omnivore, piscivore). Life history traits included
(6) maximum total body length; (7) egg diameter; (8) fecun-
dity (maximum reported per spawning season, per female);
(9) larval size (length at hatching); (10) longevity (maximum
life span); (11) maturation size (proportion, length of female
at maturation as a function of maximum total length); (12)
reproductive guild (non-guarders and open substratum
spawners, non-guarders and brood hiders, guarders and sub-
stratum choosers, guarders and nest spawners, substrate
indifferent); and (13) spawning frequency (batch, serial).
Invasion risk parameters included (14) climate match [(mea-
sured with the online program CLIMATCH (Bureau of Rural
Sciences, 2009) following a US Fish and Wildlife Service pro-
tocol (Appendix S2; Hoff, 2014) based upon Bomford et al.,
2010)]; (15) relatedness [absolute value of the difference in
family-level rank (from Nelson, 2006) between the non-
native species and the most closely related native or non-
native species established in the first invaded Great Lake
(from Mandrak & Cudmore, 2012)]; (16) phylogeny (de-
scription below); (17) prior establishment success (number
of countries where the species has established); and (18) size
of native range (area). Data for all variables were gathered
from online databases, including FishBase (Froese & Pauly,
2014), FishTraits (Frimpong & Angermeier, 2009) and pub-
lished literature (e.g. Kolar & Lodge, 2002; Mims et al.,
2010).
Given the broad taxonomic diversity of the species evalu-
ated (27 families, 48 genera), ecological constraint is more
likely to explain data patterns than phylogenetic constraint
(Westoby et al., 1995). However, to account for any impact
of phylogenetic history (i.e. shared ancestral traits), we fol-
lowed Grafen (1989) and calculated phylogenetic relatedness
by ranking species according to their family membership [i.e.
the degree of their derived characters, ordered from most
ancient to most derived based on Nelson, (2006)].
Statistical analysis
We tested which subsets of the 18 variables were most
strongly associated with the establishment and impact
stages of invasion using a classification tree analysis in the
program CART (v. 6.6, Salford Systems, San Diego, CA,
USA). Classification tree analysis is a machine learning
method that employs binary recursive partitioning to
model categorical response variables (Breiman et al., 1984).
The tree is constructed by repeatedly splitting the data into
two groups (nodes) defined by a threshold value (continu-
ous data) or category (categorical data) of a single inde-
pendent variable that maximizes homogeneity of outcome
(e.g. established vs. not established) within the two groups
created by the split (De’ath & Fabricius, 2000). Classifica-
tion trees are particularly well suited for the analysis of
complex ecological data sets that may support nonlinear
relationships that interact hierarchically, and where data
may be missing for some independent variables (De’ath &
Fabricius, 2000).
We developed classification trees for each stage of invasion
with binary-dependent outcomes: establish or fail for the
establishment stage, and low or high for the impact stage.
Node splitting criteria were based on the Gini homogeneity
index, with no constraints on the minimum node size
(De’ath & Fabricius, 2000). We chose the optimal classifica-
tion tree by performing 10-fold cross-validation and selecting
the smallest tree within one standard error of relative cost
(misclassification rate).
Diversity and Distributions, 1–13, ª2015 John Wiley & Sons Ltd 3
Predicting invaders from trade
Predicting invaders from trade
We compiled a list of species used in five trades that trans-
port live non-native fishes into the Great Lakes region:
aquarium, biological supply, live bait, live food (including
aquaculture) and water garden (Keller & Lodge, 2007). We
obtained Canadian and United States federal data sets that
document live fishes imported through legal, permit-based
processes. We compiled a list of Canadian live fish imports
over the period 1 October 2004 to 30 September 2005 from
data provided by the Canadian Border Service Agency’s
Facility for Information Retrieval Management, which covers
all ports in Canada. From the United States Fish and Wild-
life Service’s Law Enforcement Management Information Sys-
tem, a database that tracks import and export data for all
United States ports authorized to process live animal ship-
ments, we obtained a list of live fish imports over the period
1 October 2004 to 30 November 2005 through a Freedom of
Information Act request following Romagosa et al. (2009).
For the Canadian and United States data sets, each species
was assigned to one or more of the five trades based on
information in the documents and/or from the importer’s
website. Our analysis excluded fishes imported for zoos or
university research, fish species listed on the Convention on
Table 1 Established non-native fishes (n=37 species) in the Great Lakes and their ecological impact based upon expert questionnaire
results. The mean impact score (and 95% CI) and percentage of respondents that had high confidence in their answers are reported
(n=27 respondents). Species are listed in rank order from high to low impact based upon the upper 95% CI. The 12 species perceived
to have the highest (impact class ‘high’) and the 12 species with lowest (impact class ‘low’) impact were used to develop the impact
classification tree reported in Fig. 1. The intermediate impact species (impact class ‘medium’) not used in modelling are also reported.
Species Common name Impact class Impact score (95% CI) % high confidence
Petromyzon marinus Linnaeus 1758 Sea Lamprey High 3.85 (3.68–4.00) 100
Alosa pseudoharengus Wilson 1811 Alewife High 3.74 (3.53–3.95) 100
Neogobius melanostomus Pallas 1814 Round Goby High 3.27 (2.94–3.59) 96
Osmerus mordax Mitchill 1814 Rainbow Smelt High 3.04 (2.65–3.42) 93
Cyprinus carpio Linnaeus 1758 Common Carp High 2.81 (2.24–3.39) 81
Oncorhynchus tschawytscha Walbaum 1792 Chinook Salmon High 2.81 (2.45–3.16) 92
Morone americana Gmelin 1789 White Perch High 2.46 (2.17–2.75) 81
Oncorhynchus mykiss Walbaum 1792 Rainbow Trout High 2.40 (2.05–2.75) 84
Salmo trutta Linnaeus 1758 Brown Trout High 2.25 (1.81–2.69) 75
Ctenopharyngodon idella Valenciennes 1844 Grass Carp High 1.90 (1.31–2.48) 32
Oncorhynchus kisutch Walbaum 1792 Coho Salmon High 2.12 (1.80–2.43) 73
Gymnocephalus cernua Linnaeus 1758 Ruffe High 2.04 (1.75–2.34) 61
Scardinius erythrophthalmus Linnaeus 1758 Rudd Medium 1.71 (1.33–2.08) 53
Gasterosteus aculeatus Linnaeus 1758 Threespine Stickleback Medium 1.50 (0.96–2.04) 44
Esox niger Lesueur 1818 Chain Pickerel Medium 1.62 (1.28–1.95) 46
Proterorhinus marmoratus Pallas 1814 Tubenose Goby Medium 1.43 (0.97–1.89) 48
Aplodinotus grunniens Rafinesque 1819 Freshwater Drum Medium 1.50 (1.14–1.86) 50
Salmo salar Linnaeus 1758 Atlantic Salmon Medium 1.45 (1.07–1.84) 50
Oncorhynchus gorbuscha Walbaum 1792 Pink Salmon Medium 1.50 (1.22–1.78) 64
Lepomis microlophus Gunther 1859 Redear Sunfish Medium 1.33 (0.93–1.74) 50
Gambusia affinis Baird & Girard 1853 Eastern Mosquitofish Medium 1.43 (1.14–1.72) 36
Gambusia holbrooki Girard 1859 Western Mosquitofish Medium 1.43 (1.14–1.72) 36
Carassius auratus Linnaeus 1758 Goldfish Medium 1.38 (1.13–1.62) 42
Noturus insignis Richardson 1836 Margined Madtom Medium 1.18 (0.76–1.60) 55
Cyprinella lutrensis Baird & Girard 1853 Red Shiner Medium 1.38 (1.16–1.59) 25
Morone mississippiensis Jordan & Eigenmann 1887 Yellow Bass Low 1.22 (0.96–1.48) 44
Alosa aestivalis Mitchill 1814 Blueback Herring Low 1.23 (0.99–1.47) 54
Enneacanthus gloriosus Holbrook 1855 Bluespotted Sunfish Low 1.18 (0.91–1.45) 64
Lepomis humilis Girard 1858 Orangespotted Sunfish Low 1.21 (1.00–1.43) 50
Ictiobus cyprinellus*Valenciennes 1844 Bigmouth Buffalo Low 1.17 (0.93–1.41) 58
Apeltes quadracus Mitchill 1815 Fourspine Stickleback Low 1.18 (0.98–1.39) 27
Misgurnus anguillicaudatus Cantor 1842 Oriental Weatherfish Low 1.14 (0.96–1.33) 43
Ictiobus bubalus Rafinesque 1818 Smallmouth Buffalo Low 1.17 (1.10–1.32) 58
Ictiobus niger Rafinesque 1819 Black Buffalo Low 1.17 (1.10–1.32) 58
Carpiodes carpio Rafinesque 1820 River Carpsucker Low 1.10 (0.91–1.29) 50
Phenacobius mirabilis Girard 1856 Suckermouth Minnow Low 1.00 (1.00–1.00) 11
Clinostomus elongatus*Kirtland 1840 Redside Dace Low 1.00 (1.00–1.00) 53
*Introduced beyond the species’ native range in the Great Lakes.
4Diversity and Distributions, 1–13, ª2015 John Wiley & Sons Ltd
J. G. Howeth et al.
International Trade of Endangered Species and not regularly
traded, and marine fishes.
Next, we performed Internet searches for species in the
biological supply, live bait and water garden trades because
they were poorly represented in the federal importation spe-
cies data sets. The search strings ‘trade pathway name supply’
and ‘trade pathway name supply Canada’ were used for each
of the three trades. Vendors listed in up to the top 15 hits
were used to obtain identities of species available from major
suppliers. Four non-native species were documented for the
live bait trade from Internet searches. We obtained addi-
tional bait species from published surveys of bait shops (Lit-
vak & Mandrak, 1999; Drake & Mandrak, 2014). The final
list of species in all five trades excluded fishes already in our
established/failed data set (Table S1) or that were native to
the Great Lakes. For the species in trade, climate match and
trait data were collected using the methods described above.
We used the establishment classification tree to predict
which species would likely establish in the Great Lakes. For
those species predicted to establish, we estimated impact
using the impact classification tree. Trades posing the great-
est threat to the Great Lakes were identified by the propor-
tion and the absolute number of species predicted to
establish and have high impact.
Historical patterns of establishment
We used data from Mandrak & Cudmore (2012) on the
number of fish species both intentionally and unintentionally
introduced to the Great Lakes between 1829 and 2012 to
determine the contribution of different trades to past fish
invasions relative to other introduction pathways (commer-
cial shipping, natural dispersal, stocking, unknown). We
evaluated the introduction pathway of 64 species in Table S1
(all species in Table S1 except Gambusia holbrooki, which has
an unknown introduction date and pathway), in addition to
four other fish species that have been introduced and have
unknown establishment status (Alosa chrysochloris, Ameiurus
catus, Lepisosteus oculatus, Pimephales vigilax), and three
hybrid species bred and stocked for recreation that failed to
establish (Salvelinus fontinalis 9S. namaycush,S. fonti-
nalis 9S. namaycush 9S. namaycush,Salmo trutta 9
S. fontinalis) (N. E. Mandrak, unpublished data). For those
species introduced through trade, assignment of the specific
trade pathway (aquarium, biological supply, live bait, live
food, water garden) was accomplished by cross-referencing
with the species list formulated in the Predicting invaders
from trade subsection [above] in addition to reviewing infor-
mation reported in the United States Geological Survey’s
Nonindigenous Aquatic Species Database (USGS, 2013). Six
species (Table S1) could have originated from multiple trades
and were represented in each possible pathway for the analy-
sis.
To assess the contribution of the five trades collectively to
non-native species introduction over time, we used linear
regression to evaluate trends in the introduction frequency
over 25-year time intervals. Additionally, for each of the five
trades, we calculated the proportion of introduced species
that established.
RESULTS
Predicting establishment and impact
The classification tree for the establishment stage indicates
that a climatic match >71.7% between the native range of
a fish and the Great Lakes region distinguishes species that
established from those that did not (Fig. 1a; relative
cost =0.44, AUROC =0.77). Cross-validation reveals that
81% of established fishes, and 75% of fishes that failed to
establish, are correctly classified with this model. Likewise,
the classification tree for the ecological impact stage accu-
rately distinguishes low- versus high-impact species on the
basis of trophic guild and fecundity (Fig. 1b; relative
cost =0.42, AUROC =0.79). Under cross-validation, 75%
of low impact invaders and 83% of high-impact invaders
were correctly classified with this model. The tree branch-
ing sequence indicates that trophic specialization was the
most important predictor of impact. Species in the top
trophic levels (invertivore–piscivore, piscivore) had a large
ecological impact, while the tree differentiates among
impact levels of lower trophic level species with a split in
fecundity. Species producing more than one million eggs
per spawning season had a high impact, whereas species
with lower fecundities had a low impact (Fig. 1b). An anal-
ysis evaluating the sensitivity of the impact model to a dif-
ferent number of low- and high-impact species indicates
this reported model has the lowest relative cost (error) of
five alternative models (Appendix S3). Further, trophic
guild emerged as the root node in five of the six models
tested, indicating this variable is a robust predictor of inva-
siveness.
Predicting invaders from trade
A total of 787 live freshwater fish species were traded in the
United States and Canada, with 20 species occurring in mul-
tiple trade pathways (Table S4). Screening each of these spe-
cies with the establishment classification tree suggests that 2
of the 13 species in the water garden trade, 2 of the 23 spe-
cies in the live food trade and 4 of the 768 species in the
aquarium trade are capable of establishing in the Great Lakes
(Fig. 2a). No species from the biological supply trade
(n=8) are predicted to establish. The live bait trade con-
tributed no species to the final trade species list, as all bait
species identified were either native to the Great Lakes
(n=3), or already established (n=6). Overall, seven non-
native fish species are predicted capable of establishing in the
Great Lakes from the five trades, and four of these are pre-
dicted to have a large impact (Table 2). All species predicted
to establish from the water garden trade are predicted to
have a large ecological impact, whereas only half of the
Diversity and Distributions, 1–13, ª2015 John Wiley & Sons Ltd 5
Predicting invaders from trade
species predicted to establish from the aquarium and live
food trades are predicted to have a large ecological impact
(Fig. 2b).
Historical patterns of establishment
Since 1829, 45 non-native fish species have been introduced
to the Great Lakes from pathways other than the five
trades, including commercial shipping, natural dispersal,
stocking and unknown sources. Introduction frequency of
these fishes has not increased over time (linear regression;
y=1.15x +0.43, R
2
=0.27, P=0.19, Fig. 3a). Over this
same period, 25 fishes have been introduced through trade
pathways with a significant linear increase in introduction
frequency (y =1.30x 2.71, R
2
=0.78, P=0.004). Of the
species successfully established from trade (n=10), the
majority originated from the live bait trade, followed by
the biological supply and water garden trades (Fig. 3b).
The aquarium and live food trades contributed the fewest
established species. Of the fishes that failed to establish
(n=14), most originated from the aquarium trade with
the remainder originating from the live food trade and bio-
logical supply trade (Table S1). All species introduced
through the live bait and water garden trades successfully
established in the Great Lakes, compared to 75% from bio-
logical supply, 33% from live food and 14% from the
aquarium trade.
DISCUSSION
The present study provides a comprehensive evaluation of
trait-based risks of establishment and ecological impact for
non-native freshwater fishes in the Laurentian Great Lakes
of North America, stratified by five global trade pathways.
We identified climate match, trophic guild and fecundity as
invasion risk factors that are most predictive of past estab-
lishment and impact, and subsequently applied these risk
assessment tools to 787 fishes in international trade to iden-
tify species that present a significant invasion threat in the
future. Application of the tools to these species in trade
that are not yet known to have been released indicates that
the water garden, live food and aquarium trades currently
pose the greatest risk of introducing species that are likely
to establish and generate impacts. These predicted outcomes
parallel the historical importance of the trades in introduc-
ing non-native species to the Great Lakes over a period
Climate match
(%)
≤ 71.7 > 71.7
5 establish, 22 fail
Establish
Fail
32 establish, 6 fail
Trophic guild
Invertivore-piscivore,
piscivore
8 high, 1 low
Fecundity
(number of eggs)
High impact
2 high, 11 low
Herbivore-detritivore,
invertivore, omnivore
≤ 1,013,000 > 1,013,000
Low impact High impact
2 high, 0 low
(a)
(b)
Figure 1 Classification trees predicting
(a) establishment and (b) ecological
impact of non-native fish species in the
Great Lakes. The number of fish species
from the model training data set located
in each response category is indicated at
every terminal node.
6Diversity and Distributions, 1–13, ª2015 John Wiley & Sons Ltd
J. G. Howeth et al.
spanning almost 200 years. The projected importance of the
water garden trade in particular parallels patterns of its his-
torical contribution of introducing successful non-native
fishes. The historical analyses further demonstrate that the
live bait, biological supply, aquarium and live food trades
have also contributed non-native fishes to the Great Lakes
biota.
Accuracy and scientific basis of trait-based risk
assessment tools
Not only did our risk assessment models accurately predict
establishment and impact (75–83% accuracy) in most former
species invasions, but they did so on the basis of only a few,
readily available variables with known ecological relevance.
The match between climate in a species’ native range and
the Great Lakes region proved sufficient to predict the estab-
lishment of introduced fishes, consistent with previous anal-
yses of fishes (Bomford et al., 2010) and other taxa
including terrestrial plants (Petitpierre et al., 2012), birds
(Duncan et al., 2001), mammals (Forsyth et al., 2004), and
amphibians and reptiles (Bomford et al., 2009; Van Wilgen
& Richardson, 2012). This is congruent with the generally
recognized importance of physiological tolerance in shaping
species distributions, often making climate match a robust
predictor of non-native species establishment (Hayes &
Barry, 2008). The predictive ability of climate match is
unique relative to findings from previous invasion risk anal-
yses of the Great Lakes using smaller data sets of introduced
fish species and different predictors of establishment that
exclude aspects of the regional environment (e.g. climate
match) and biota (e.g. relatedness) included in this study
(Kolar & Lodge, 2002; Snyder et al., 2014). These analyses
found positive associations between establishment status and
higher growth rates, previous history of invasiveness, broad
salinity tolerances, broad temperature tolerances and diet
breadth.
The ecological impact model required information about a
species’ trophic guild and average fecundity. These ecological
and life history traits are often important in interspecific
interactions and population growth, respectively. Consistent
with many observations of large top-down population and
Proportion
0.00
0.05
0.10
0.15
0.20
Trade pathway
Aquarium
Biol. supply
Live bait
Live food
Water garden
Number of fish species
0
1
2
3
4
5
High impact
Low impact
(a)
(b)
NA
NA
Figure 2 (a) Proportions of non-native fish species predicted
to establish in the Great Lakes from the aquarium (n=768),
live bait (NA =no contributing species), biological (Biol.)
supply (n=8), live food (n=23) and water garden (n=13)
trades. Identities of species in each trade pathway are in
Table S4. (b) The number of fish species predicted to establish
in each trade pathway and their corresponding predicted
ecological impact. One species (Leuciscus idus) is represented in
two trade pathways (aquarium, water garden). Identities of
species in each trade pathway are in Table 2.
Table 2 Non-native fish species in trade predicted to establish
in the Great Lakes, with corresponding prediction of ecological
impact and associated trade pathway(s). Species were identified
with the classification trees reported in Fig. 1 and are listed in
alphabetical order.
Species Common name Impact Trade pathway
Carassius carassius
Linnaeus 1758
Crucian Carp Low Live food
Cobitis taenia
Linnaeus 1758
Spined Loach Low Aquarium
Ictalurus furcatus
Valenciennes 1804
Blue Catfish High Water garden
Leuciscus idus
Linnaeus 1758
Ide High Aquarium,
water garden
Misgurnus fossilis
Linnaeus 1758
European
Weatherfish
Low Aquarium
Morone saxatilis
Walbaum 1792
Striped Bass High Live food
Silurus glanis
Linnaeus 1758
European Catfish High Aquarium
Diversity and Distributions, 1–13, ª2015 John Wiley & Sons Ltd 7
Predicting invaders from trade
ecosystem impacts in lake food webs, high-impact non-native
fishes are likely to be top predators (piscivores, invertivore–
piscivores) (Eby et al., 2006; Cucherousset & Olden, 2011).
The impact model further predicted that for non-native spe-
cies in the lower trophic levels (herbivore–detritivore, inverti-
vore, omnivore), high fecundity species have a large
ecological impact. This result aligns with other trait-based
models that find fecundity to be a predictor of invasion suc-
cess in aquatic species including crayfishes (Larson & Olden,
2010) and molluscs (Keller et al., 2007a). The results differ,
however, from previous trait-based impact studies of non-
native freshwater fishes from the same and different ecosys-
tems. For example, a study of invasive Great Lakes fishes
using a different suite of ecological and life history traits
found positive associations between nuisance status (defined
as both ecological and economic impacts from expert opin-
ion) and smaller egg size, wider salinity tolerance and sur-
vivorship in lower water temperatures (Kolar & Lodge,
2002). A related study of invasive Great Lakes fishes found
associations between impact status (method not described)
and egg size, minimum water temperature and history of
invasiveness (Snyder et al., 2014). A study of non-native
freshwater fishes in California, United States, evaluating
impact (defined as average abundance) found prior invasion
success and propagule pressure as reliable indicators (March-
etti et al., 2004b). The lack of consistency in impact predic-
tors for freshwater fishes supports previous findings that
mechanisms of impact often vary across studies (e.g. Garc
ıa-
Berthou, 2007), as do the ways in which impact is perceived
and measured (Cucherousset & Olden, 2011; Lapointe et al.,
2012). They further demonstrate that it is challenging to
identify a single robust trait or variable that predicts impact
of non-native freshwater fishes.
Consistency of the current relative risks of different
trades with historical patterns
Since the early 20th century, the number of fish species
introduced into the Great Lakes region from the live organ-
ism trades has increased relative to other pathways. In the
past few decades, trade has introduced more fishes than all
other pathways combined, reflecting an increase in commer-
cial globalization and the importance of growing human
populations and urban trade hubs (e.g. Chicago, Toronto) in
the Great Lakes region.
The application of our models to species known to be
traded but not known to be established in the Great Lakes
revealed that among the live trades, the water garden trade
has the highest proportion of species likely to establish
(15%); all of which are predicted to have high impact. This
is not surprising given that the water garden trade is
restricted to species that survive well in outdoor ambient
conditions. Accordingly, our historical analysis revealed that
100% of the non-native fish species originating from the
water garden trade have established, paralleling a previous
finding that the water garden trade has supplied the majority
Time period
1829–1830
1831–1856
1857–1882
1883–1908
1909–1934
1935–1960
1961–1986
1987–2012
Number of fish species
0
2
4
6
8
10
12
14
16
18
20
Trade
Other
Number of fish species
0
2
4
6
8
10
12
14
Established
Failed
(a)
(b)
Trade pathway
Aquarium
Biol. supply
Live bait
Live food
Water garden
Figure 3 (a) The number of fish species introduced to the
Great Lakes via different introduction pathways, for the time
period 1829–2012. Data are grouped into 25-year bins, except
for the oldest (2 years) grouping. Species were classified as being
introduced from trade (i.e. commerce in living organisms)
(black bars) or other pathways (grey bars). Analysis includes fish
species that have established, or failed to establish, in the Great
Lakes. (b) The number of fish species that have established
(black bars) or failed to establish (grey bars) in the Great Lakes
from trade, and their associated pathway, including the
aquarium, live bait, biological (Biol.) supply, live food and water
garden trade. Ten species have established, and 14 species have
failed to establish from trade. Six of these species are
represented in multiple trade pathways. Refer to Table S1 for
trade pathway by species.
8Diversity and Distributions, 1–13, ª2015 John Wiley & Sons Ltd
J. G. Howeth et al.
of established aquatic non-native plant species in the Great
Lakes (Keller & Lodge, 2007).
Specifically, our risk assessment models identified Ide
(Leuciscus idus) and Blue Catfish (Ictalurus furcatus) as water
garden species predicted to become established and have an
ecological impact if introduced in the Great Lakes. Blue Cat-
fish has already established outside of its native range within
North America (Bonvechio et al., 2012). Additionally, Blue
Catfish has been genetically crossed with Channel Catfish
(Ictalurus punctatus) in the U.S. aquaculture industry to pro-
duce a high growth hybrid with increased disease resistance
(Arias et al., 2012; Bosworth & Waldbieser, 2014). Blue Cat-
fish and its hybrids thus have the potential to hybridize with
wild native Channel Catfish in the Great Lakes.
The live food trade represents the second-highest risk
pathway in terms of proportion of species predicted to estab-
lish (9%). Similar to the water garden pathway, the live food
trade often selects for species that survive well outdoors
under high density conditions (Naylor et al., 2001). Because
major cities occur in and near the Great Lakes, there is
heightened risk of non-native species sold in urban live fish
markets being released alive into nearby habitats (Rixon
et al., 2005). For example, Crucian Carp (Carassius carassius)
and Striped Bass (Morone saxatilis) are predicted to establish
in the Great Lakes. Striped Bass commonly hybridize with
White Bass (M. chrysops) in live food aquaculture, and both
Striped Bass and its hybrids are frequently found in live fish
markets in Ontario and Qu
ebec, Canada (Rixon et al., 2005).
The aquarium trade ranked third in terms of proportion
of species predicted to establish; but because of the great
number of species imported by this pathway, it was tied for
second in terms of the absolute number of species predicted
to establish and cause high impact. Our analysis identified
four aquarium species that are native to temperate regions of
Europe and/or Asia that would likely establish if introduced
to the Great Lakes: European Weatherfish (Misgurnus fos-
silis), Ide, Spined Loach (Cobitis taenia) and European Cat-
fish (Silurus glanis). The impact model further predicted that
European Catfish, a large-bodied piscivore that has been
introduced outside of its native range in Europe (Copp
et al., 2009a), will become invasive in the Great Lakes if
introduced and established. Eventual introduction of Euro-
pean Weatherfish to the Great Lakes seems likely, given that
10% of aquarium and pet stores in Michigan, USA, and
Ontario, Canada, sell this species (Rixon et al., 2005), and
frequency of occurrence of fish species in stores positively
correlates with regional establishment success (Duggan et al.,
2006).
Although the majority of aquarium species in trade are
tropical in origin and thus predicted not to establish in the
Great Lakes due to the current climate mismatch, we caution
that the currently moderate threat posed by aquarium species
may increase in the near future for two important reasons.
First, increasing demand for unique fishes from hobbyists
(Padilla & Williams, 2004) may bring new temperate fishes
to the region (Keller et al., 2011). Second, elevated regional
temperature associated with climate change will increase the
proportion of warmwater species that can successfully estab-
lish from the aquarium and other trade pathways (Stefan
et al., 2001). The biological supply and live bait trades pose
lower invasion threats to the Great Lakes. No species from
biological supply were predicted to establish; however, the
pathway has historically supplied approximately one-fifth of
successful invaders. Although no species currently in the bait
trade and not already established in the region are predicted
to establish, the bait trade has historically contributed
approximately 40% of successful invaders. This is not sur-
prising, as the bait trade has the highest propagule pressure
stemming from the highest volume and release rates (30%;
Drake & Mandrak, 2014) of all trades. While the use of non-
native baitfishes is illegal in the Great Lakes basin, they do
occur in trade (Drake & Mandrak, 2014), and the bait trade
has the potential to become a high-risk pathway if contami-
nated with non-target invasive species (e.g. bigheaded carps,
Hypopthalmichthys spp.; Cudmore et al., 2012).
Applications and conclusions
The trait-based risk assessments presented here provide a
science-based, quantitative method for predicting invasion
risk from fish species that may be introduced to the Great
Lakes. Because few parameters are required to achieve high
model accuracy, they could readily be implemented for vol-
untary or regulatory decision-making to achieve the goal,
adopted by United States and Canadian policy-makers, to
reduce the introduction of and harm from invasive species
over the long term (GLWQA, 2012). Currently, most states
and provinces across the Great Lakes basin implement some
form of risk assessment for fishes, but there is little consis-
tency in approach. This leads to inconsistent outcomes in
terms of the number and identity of species considered high
risk. These alternative risk screening approaches include liter-
ature reviews (e.g. Cudmore et al., 2012) and questionnaire-
based assessments (e.g. Fish Invasiveness Screening Kit; Copp
et al., 2009b) which require more data and are thus more
costly to implement (Leung et al., 2012) than our approach.
The speed with which the risk assessment framework devel-
oped in this study can be implemented offers the opportu-
nity to adopt predictive models with minimal cost, and
consequently with greater potential to coordinate preventa-
tive screening efforts among multiple jurisdictions (in this
case, United States and the Canadian province of Ontario)
(Gantz et al., 2015). This result could also encourage the
development of rapid trait-based assessments for application
in other regions and countries.
Although our models are robust, there are at least four
limitations to their management application. First, our estab-
lishment analysis assumes all non-native species have an
equal probability of introduction from trade. There may be
traits predictive of introduction probability in fishes, and
these could differ among trades, but the data to evaluate this
are lacking. Second, we lack direct measures of propagule
Diversity and Distributions, 1–13, ª2015 John Wiley & Sons Ltd 9
Predicting invaders from trade
pressure, which may differ among trades, species and locali-
ties. Propagule pressure is an important determinant of
establishment success (Lockwood et al., 2005; Drake &
Lodge, 2006), yet is commonly missing from quantitative
risk assessments (Leung et al., 2012). The already high accu-
racy of our models might improve if data on propagule pres-
sure become available in the future. For example, Northern
Snakehead (Channa argus) is documented on our list of spe-
cies that have failed to establish in the Great Lakes, but a
previous study indicates the species could establish in the
region based upon suitable climate match (Herborg et al.,
2007). Propagule pressure may explain the failed invasion of
C. argus in the Great Lakes. Third, our data on species in
trade were for only one year, and additional species that we
have not screened have likely entered the marketplace in the
last decade. Fourth, the history of success or failure in past
introductions may become a less reliable guide to the out-
come of future introductions, especially with respect to cli-
mate change. Fortunately, however, the models reported here
could easily be developed under alternative assumptions
about climate match. The climate match variable used in the
establishment model is freely available and should facilitate
application of this screening tool as climate continues to
change.
For short-term application with long-term conservation
implications, our results identify specific species and a subset
of the live trade pathways that pose an imminent high risk
to the Great Lakes. Results indicate that at least seven fish
species currently traded in the water garden, live food and/or
aquarium pathways are likely to establish in the Great Lakes
if introduced. These trade pathways currently experience lit-
tle voluntary or regulatory oversight with respect to risk
assessment of commercial species (Fowler et al., 2007). The
risks of invasion could be lowered by improved management
practices in the private sector and government agencies, and
improved policies that are coordinated across the political
jurisdictions surrounding the Great Lakes. Greater use of risk
assessment tools, including those presented herein, would
also likely deliver net economic benefits (Keller et al.,
2007b).
ACKNOWLEDGMENTS
EPA GLRI funded this work via USFWS award 30181AJ261.
NOAA CSCOR awards NA09NOS4780192, NA10NOS478
0218 provided additional financial support and improved the
project via interactions with managers serving on the Man-
agement Transition Board. The Virginia Cooperative Fish
and Wildlife Research Unit is jointly sponsored by the USGS,
Virginia Polytechnic Institute and State University, Virginia
Department of Game and Inland Fisheries, and Wildlife
Management Institute. Use of trade names or commercial
products does not imply endorsement by the U.S. govern-
ment. This is a publication of the Notre Dame Environmen-
tal Change Initiative.
REFERENCES
Arias, C.R., Cai, W.L., Peatman, E. & Bullard, S.A. (2012)
Catfish hybrid Ictalurus punctatus xI. furcatus exhibits
higher resistance to columnaris disease than the parental
species. Diseases of Aquatic Organisms,100,77–81.
Baiser, B., Olden, J.D., Record, S., Lockwood, J.L. & McKin-
ney, M.L. (2012) Pattern and process of biotic homoge-
nization in the New Pangaea. Proceedings of the Royal
Society B: Biological Sciences,279, 4772–4777.
Bomford, M., Kraus, F., Barry, S.C. & Lawrence, E. (2009)
Predicting establishment success for alien reptiles and
amphibians: a role for climate matching. Biological Inva-
sions,11, 713–724.
Bomford, M., Barry, S.C. & Lawrence, E. (2010) Predicting
establishment success for introduced freshwater fishes: a role
for climate matching. Biological Invasions,12, 2559–2571.
Bonvechio, T.F., Bowen, B.R., Mitchell, J.S. & Bythwood, J.
(2012) Non-indigenous range expansion of the Blue Cat-
fish (Ictalurus furcatus) in the Satilla River, Georgia. South-
eastern Naturalist,11, 355–358.
Bosworth, B. & Waldbieser, G. (2014) General and specific
combining ability of male blue catfish (Ictalurus furcatus)
and female channel catfish (Ictalurus punctatus) for growth
and carcass yield of their F-1 hybrid progeny. Aquaculture,
420, 147–153.
Bradford, M.J., Tovey, C.P. & Herborg, L. (2009) Biological
risk assessment for Yellow Perch (Perca flavescens) in British
Columbia. Fisheries and Oceans Canada, Burlington, ON.
Breiman, L., Friedman, J.H., Olshen, A. & Stone, C.G. (1984)
Classification and regression trees. Wadsworth International
Group, Belmont, CA.
Bureau of Rural Sciences (2009) CLIMATCH free-access
internet-based software. Department of Agriculture, Fish-
eries and Forestry. Bureau of Rural Sciences, Canberra.
Copp, G.H., Britton, J.R., Cucherousset, J., Garc
ıa-Berthou,
E., Kirk, R., Peeler, E. & Stakenas, S. (2009a) Voracious
invader or benign feline? A review of the environmental
biology of European Catfish Silurus glanis in its native and
introduced ranges. Fish and Fisheries,10, 252–282.
Copp, G.H., Vilizzi, L., Mumford, J., Fenwick, G.V., Godard,
M.J. & Gozlan, R.E. (2009b) Calibration of FISK, an inva-
siveness screening tool for nonnative freshwater fishes. Risk
Analysis,29, 457–467.
Cucherousset, J. & Olden, J.D. (2011) Ecological impacts of
nonnative freshwater fishes. Fisheries,36, 215–230.
Cudmore, B.C., Mandrak, N.E., Dettmers, J., Chapman, D.C.
& Kolar, C.S. (2012) Binational ecological risk assessment of
bigheaded carps (Hypophthalmichthys spp.) for the Great
Lakes basin. Canadian Science Advisory Secretariat
Research Document. 2011/114. vi + 57 p. Fisheries and
Oceans Canada, Ottawa, ON.
De’ath, G. & Fabricius, K.E. (2000) Classification and regres-
sion trees: a powerful yet simple technique for ecological
data analysis. Ecology,81, 3178–3192.
10 Diversity and Distributions, 1–13, ª2015 John Wiley & Sons Ltd
J. G. Howeth et al.
Drake, J.M. & Lodge, D.M. (2006) Allee effects, propagule
pressure and the probability of establishment: risk analysis
for biological invasions. Biological Invasions,8, 365–375.
Drake, D.A.R. & Mandrak, N.E. (2014) Ecological risk asso-
ciated with live bait fisheries: a new angle on selective fish-
ing. Fisheries,39, 201–211.
Duggan, I.C., Rixon, C.A.M. & MacIsaac, H.J. (2006) Popu-
larity and propagule pressure: determinants of introduction
and establishment of aquarium fish. Biological Invasions,8,
377–382.
Duncan, R.P., Bomford, M., Forsyth, D.M. & Conibear, L.
(2001) High predictability in introduction outcomes and
the geographical range size of introduced Australian birds:
a role for climate. Journal of Animal Ecology,70, 621–632.
Eby, L., Roach, W., Crowder, L. & Stanford, J. (2006) Effects
of stocking-up freshwater food webs. Trends in Ecology and
Evolution,21, 576–584.
Forsyth, D.M., Duncan, R.P., Bomford, M. & Moore, G.
(2004) Climatic suitability, life-history traits, introduction
effort, and the establishment and spread of introduced
mammals in Australia. Conservation Biology,18, 557–569.
Fowler, A.J., Lodge, D.M. & Hsia, J. (2007) Failure of the
Lacey Act to protect US ecosystems against animal inva-
sions. Frontiers in Ecology and the Environment,5, 353–359.
Frimpong, E.A. & Angermeier, P.L. (2009) FishTraits: a data-
base of ecological and life-history traits of freshwater fishes
of the United States. Fisheries,34, 487–495.
Froese, R. & Pauly, D. (2014) FishBase. World Wide Web
electronic publication, www.fishbase.org.
Gantz, C.A., Gordon, D.R., Jerde, C.L., Keller, R.P., Chadder-
ton, W.L., Champion, P.D. & Lodge, D.M. (2015) Manag-
ing the introduction and spread of non-native aquatic
plants in the Laurentian Great Lakes: a regional risk assess-
ment approach. Management of Biological Invasions,6,45–
55.
Garc
ıa-Berthou, E. (2007) The characteristics of invasive
fishes: what has been learned so far? Journal of Fish Biology,
71,33–55.
Grafen, A. (1989) The phylogenetic regression. Philosophical
Transactions of the Royal Society of London Series B, Biologi-
cal Sciences,326, 119–157.
Great Lakes Water Quality Protocol (2012) Agreement
between the United States government and Canada on Great
Lakes water quality, 75 pp. Great Lakes Water Quality Pro-
tocol, Washington, DC.
Hayes, K.R. & Barry, S.C. (2008) Are there any consistent
predictors of invasion success? Biological Invasions,10,
483–506.
Herborg, L., Mandrak, N.E., Cudmore, B.C. & McIsaac, H.J.
(2007) Comparative distribution and invasion risk of
snakehead (Channidae) and Asian carp (Cyprinidae) spe-
cies in North America. Canadian Journal of Fisheries and
Aquatic Sciences,64, 1723–1735.
Hoff, M.H. (2014) Standard Operating Procedures: Rapid
Screening of Species Risk of Establishment and Impact in
the U.S. United States Fish and Wildlife Service. http://
www.fws.gov/injuriouswildlife/pdf_files/Standard_Operat-
ing_Procedures_01_08_ 14.pdf. (accessed 15 January 2014).
Hulme, P.E. (2009) Trade, transport and trouble: managing
invasive species pathways in an era of globalization. Journal
of Applied Ecology,46,10–18.
Ib
a~
nez, I., Diez, J.M., Miller, L.P., Olden, J.D., Sorte, C.J.B.,
Blumenthal, D.Bradley., B.A., D.M., D’Antonio, C.M.,
Dukes, J.S., Early, R.I., Grosholz, E.D. & Lawler, J.J. (2014)
Integrated assessment of biological invasions. Ecological
Applications,24,25–37.
Jenkins, P.T. (2013) Invasive animals and wildlife pathogens
in the United States: the economic case for more risk
assessments and regulation. Biological Invasions,15, 243–
248.
Keller, R.P. & Lodge, D.M. (2007) Species invasions from
commerce in live aquatic organisms: problems and possible
solutions. BioScience,57, 428–436.
Keller, R.P. & Perrings, C. (2011) International policy
options for reducing the environmental impacts of invasive
species. BioScience,61, 1005–1012.
Keller, R.P., Drake, J.M. & Lodge, D.M. (2007a) Fecundity as
a basis for risk assessment of nonindigenous freshwater
molluscs. Conservation Biology,21, 191–200.
Keller, R.P., Lodge, D.M. & Finnoff, D.C. (2007b) Risk
assessment for invasive species produces net bioeconomic
benefits. Proceedings of the National Academy of Sciences
USA,104, 203–207.
Keller, R.P., Drake, J.M., Drew, M.B. & Lodge, D.M. (2011)
Linking environmental conditions and ship movements to
estimate invasive species transport across the global ship-
ping network. Diversity and Distributions,17,93–102.
Kolar, C.S. & Lodge, D.M. (2002) Ecological predictions and
risk assessment for alien fishes in North America. Science,
298, 1233–1236.
Lapointe, N.W.R., Pendleton, R.M. & Angermeier, P.L.
(2012) A comparison of approaches for estimating relative
impacts of nonnative fishes. Environmental Management,
49,82–95.
Larson, E.R. & Olden, J.D. (2010) Latent extinction and
invasion risk of crayfishes in the southeastern United
States. Conservation Biology,24, 1099–1110.
Leung, B., Roura-Pascual, N., Bacher, S., Heikkila, J., Bro-
tons, L., Burgman, M.A., Dehnen-Schmutz, K., Essl, F.,
Hulme, P.E., Richardson, D.M., Sol, D. & Vila, M. (2012)
TEASIng apart alien species risk assessments: a framework
for best practices. Ecology Letters,15, 1475–1493.
Levine, J.M. & D’Antonio, C.M. (2003) Forecasting biological
invasions with increasing international trade. Conservation
Biology,17, 322–326.
Litvak, M.K. & Mandrak, N.E. (1999) Baitfish trade as a vec-
tor of aquatic introductions. Nonindigenous freshwater
organisms: vectors, biology, and impacts (ed. by R. Claudi
and J.H. Leach), pp. 163–180. Lewis Publishers, New York.
Lockwood, J.L., Cassey, P. & Blackburn, T. (2005) The role
of propagule pressure in explaining species invasions.
Trends in Ecology and Evolution,20, 223–228.
Diversity and Distributions, 1–13, ª2015 John Wiley & Sons Ltd 11
Predicting invaders from trade
Lodge, D.M., Williams, S., MacIsaac, H.J., Hayes, K.R.,
Leung, B., Reichard, S., Mack, R.N., Moyle, P.B., Smith,
M., Andow, D.A., Carlton, J.T. & McMichael, A. (2006)
Biological invasions: recommendations for US policy and
management. Ecological Applications,16, 2035–2054.
Mandrak, N.E. (1989) Potential invasion of the Great Lakes
by fish species associated with climatic warming. Journal of
Great Lakes Research,15, 306–316.
Mandrak, N.E. & Cudmore, B.C. (2010) The fall of native
fishes and the rise of nonnative fishes in the Great Lakes
Basin. Aquatic Ecosystem Health & Management,13, 255–
268.
Mandrak, N.E. & Cudmore, B.C. (2012) Fish species at risk
and nonnative fishes in the Great Lakes Basin: past, present
and future. Great lakes policy and management (ed. by
W.W. Taylor, A.J. Lynch and N.J. Leonard), pp. 167–202.
Michigan State University Press, East Lansing, MI.
Mandrak, N.E., Cudmore, B. & Chapman, P.M. (2012)
National detailed level risk assessment guidelines: assessing
the biological risk of aquatic invasive species in Canada.
Fisheries and Oceans Canada, Burlington, ON.
Marchetti, M.P., Moyle, P.B. & Levine, R. (2004a) Invasive
species profiling? Exploring the characteristics of nonnative
fishes across invasion stages in California. Freshwater Biol-
ogy,49, 646–661.
Marchetti, M.P., Moyle, P.B. & Levine, R. (2004b) Alien
fishes in California watersheds: characteristics of successful
and failed invaders. Ecological Applications,14, 587–596.
Marson, D., Cudmore, B., Drake, D.A.R. & Mandrak, N.E.
(2009a) Summary of a survey of aquarium owners in
Canada. Canadian Manuscript Report of Fisheries and
Aquatic Sciences 2905: iv + 20 p. Fisheries and Oceans
Canada, Ottawa, ON.
Marson, D., Cudmore, B., Drake, D.A.R. & Mandrak, N.E.
(2009b) Summary of a survey of water garden owners in
Canada. Canadian Manuscript Report of Fisheries and
Aquatic Sciences 2906: v + 23 p. Fisheries and Oceans
Canada, Ottawa, ON.
Mims, M.C., Olden, J.D., Shattuck, Z.R. & Poff, N.L. (2010)
Life history trait diversity of native freshwater fishes in
North America. Ecology of Freshwater Fish,19, 390–400.
Naylor, R.L., Williams, S.L. & Strong, D.R. (2001) Ecology -
Aquaculture - A gateway for exotic species. Science,294,
1655–1656.
Nelson, J.S. (2006) Fishes of the world. John Wiley and Sons,
Incorporated, Hoboken, NJ.
Olden, J.D., Poff, N.L. & Bestgen, K.R. (2006) Life-history
strategies predict fish invasions and extirpations in the Col-
orado River Basin. Ecological Monographs,76,25–40.
Olden, J.D., Kennard, M.J., Leprieur, F., Tedesco, P.A.,
Winemiller, K.O. & Garcia-Berthou, E. (2010) Conserva-
tion biogeography of freshwater fishes: recent progress and
future challenges. Diversity and Distributions,16, 496–513.
Padilla, D.K. & Williams, S.L. (2004) Beyond ballast water:
aquarium and ornamental trades as sources of invasive spe-
cies in aquatic ecosystems. Frontiers in Ecology and the
Environment,2, 131–138.
Petitpierre, B., Kueffer, C., Broennimann, O., Randin, C., Daeh-
ler, C. & Guisan, A. (2012) Climatic niche shifts are rare
among terrestrial plant invaders. Science,335,1344–1348.
Pimentel, D., Zuniga, R. & Morrison, D. (2005) Update on
the environmental and economic costs associated with
alien-invasive species in the United States. Ecological Eco-
nomics,52, 273–288.
Ribeiro, F., Elvira, B., Collares-Pereira, M.J. & Moyle, P.B.
(2008) Life-history traits of nonnative fishes in Iberian
watersheds across several invasion stages: a first approach.
Biological Invasions,10,89–102.
Ricciardi, A. (2006) Patterns of invasion in the Laurentian
Great Lakes in relation to changes in vector activity. Diver-
sity and Distributions,12, 425–433.
Rixon, C.A.M., Duggan, I.C., Bergeron, N.M.N., Ricciardi, A.
& Macisaac, H.J. (2005) Invasion risks posed by the aquar-
ium trade and live fish markets on the Laurentian Great
Lakes. Biodiversity and Conservation,14, 1365–1381.
Romagosa, C.M., Guyer, C. & Wooten, M.C. (2009) Contri-
bution of the live-vertebrate trade toward taxonomic
homogenization. Conservation Biology,23, 1001–1007.
Rothlisberger, J.D., Lodge, D.M., Cooke, R.M. & Finnoff,
D.C. (2010) Future declines of the binational Laurentian
Great Lakes fisheries: recognizing the importance of envi-
ronmental and cultural change. Frontiers in Ecology and
Environment,8, 239–244.
Rothlisberger, J.D., Finnoff, D.C., Cooke, R.M. & Lodge,
D.M. (2012) Ship-borne nonindigenous species diminish
Great Lakes ecosystem services. Ecosystems,15, 462–476.
Ruesink, J.L. (2005) Global analysis of factors affecting the
outcome of freshwater fish introductions. Conservation
Biology,19, 1883–1893.
Smith, K.F., Behrens, M.D., Max, L.M. & Daszak, P. (2008)
US drowning in unidentified fishes: scope, implications,
and regulation of live fish imports. Conservation Letters,1,
103–109.
Smith, K.F., Behrens, M., Schloegel, L.M., Marano, N., Bur-
giel, S. & Daszak, P. (2009) Reducing the risks of the wild-
life trade. Science,324, 594–595.
Snyder, R.J., Burlakova, L.E., Karatayev, A.Y. & MacNeill,
D.B. (2014) Updated invasion risk assessment for Ponto-
Caspian fishes to the Great Lakes. Journal of Great Lakes
Research,40, 360–369.
Stefan, H.G., Fang, X. & Eaton, J.G. (2001) Simulated fish
habitat changes in North American Lakes in response to
projected climate warming. Transactions of the American
Fisheries Society,130, 459–477.
Strayer, D.L., Eviner, V.T., Jeschke, J.M. & Pace, M.L. (2006)
Understanding the long-term effects of species invasions.
Trends in Ecology & Evolution,21, 645–651.
Strecker, A.L., Campbell, P.M. & Olden, J.D. (2011) The
aquarium trade as an invasion pathway in the Pacific
Northwest. Fisheries,36,74–85.
12 Diversity and Distributions, 1–13, ª2015 John Wiley & Sons Ltd
J. G. Howeth et al.
U.S. Geological Survey (2013) Nonindigenous aquatic species
database. U.S. Geological Survey, Gainesville, FL.
http://nas.er.usgs.gov, accessed 14 July 2013.
Van Wilgen, N.J. & Richardson, D.M. (2012) The roles of
climate, phylogenetic relatedness, introduction effort, and
reproductive traits in the establishment of nonnative rep-
tiles and amphibians. Conservation Biology,26, 267–277.
Westoby, M., Leishman, M.R. & Lord, J.M. (1995) On mis-
interpreting the ‘‘phylogenetic correction’.’. Journal of Ecol-
ogy,83, 531–534.
Wilcove, D.S., Rothstein, D., Dubow, J., Phillips, A. & Losos,
E. (1998) Quantifying threats to imperiled species in the
United States. BioScience,48, 607–615.
SUPPORTING INFORMATION
Additional Supporting Information may be found in the
online version of this article:
Table S1. List of introduced nonnative fish species docu-
mented to have established, or failed to establish, in the
Great Lakes.
Table S2. Experts responding to a questionnaire regarding
the perceived ecological impact of established nonnative fish
species in the Great Lakes.
Table S3. List of 18 selected species traits and invasion risk
parameters used for the establishment and impact classifica-
tion tree analyses.
Table S4. Nonnative fish species (n=787) used in trade and
screened through the developed classification trees.
Appendix S1. Questionnaire regarding risk assessment of
ecological impact by Great Lakes fishes.
Appendix S2. Climate match methods.
Appendix S3. Description and results of the sensitivity analy-
sis for the ecological impact classification tree.
BIOSKETCH
This publication is a product of a North American working
group interested in determinants of invasion success of fresh-
water fishes and the associated development of invasive spe-
cies risk assessment tools. All authors met on three occasions
to design the research. J.G.H and C.A.G collated and anal-
ysed species trait and invasion risk parameter data and con-
ducted the screening analysis. P.L.A, E.A.F, M.H., N.E.M and
J.D.O. contributed trait data. N.E.M contributed historical
data on the Great Lakes and the CBSA import data set.
C.R.M contributed the LEMIS import data set. D.M.L and
R.P.K funded the study. J.G.H. led writing of the manuscript,
with contributions to the conceptual framework and revi-
sions from all authors.
Editor: Mark Robertson
Diversity and Distributions, 1–13, ª2015 John Wiley & Sons Ltd 13
Predicting invaders from trade
