Role of domestic shipping in the introduction or secondary spread of nonindigenous species: biological invasions within the Laurentian Great Lakes


Correspondence author. E-mail:,


  1. The most effective way to manage invasive species is to prevent their introduction via vector regulation. While progress has been made in the management of transoceanic ballast water, domestic vessels operating within smaller geographic regions such as the Laurentian Great Lakes, Mediterranean Sea, North Sea or Baltic Sea are often exempt from regulations.
  2. We randomly surveyed unmanaged ballast water moved by domestic vessels within the Laurentian Great Lakes and compared the results with that of exchanged ballast water from transoceanic vessels to assess invasion risk of zooplankton transported by these two types of vessels.
  3. Total abundance and species richness were significantly different between the two vessel types with mean abundance being two magnitudes greater, and species richness being threefold higher in domestic vessels compared with transoceanic vessels. Abundance of restricted taxa – cumulatively the Great Lakes' indigenous and nonindigenous species (NIS) which do not occur in all five lakes – was also significantly higher in domestic vessels (mean densities were 24 170 and 3421 individuals per m3 for domestic and transoceanic vessels, respectively), whereas the abundance of NIS did not differ between vessels (median densities of 2015 and 850 individuals per m3, respectively).
  4. We documented 89 species transported by domestic vessels of which 31 had restricted distribution and eight were NIS. While most NIS were already established in all five lakes, Cercopagis pengoi, a NIS of global concern, and Nitokra hibernica have not been identified from Lake Superior, and both were sampled from ballast water destined for discharge in Lake Superior. Beside the risk of spread of NIS between lakes, domestic shipping can act as a vector for homogenization of indigenous taxa, with at least 21 native species (99 events) being moved outside their historical distribution.
  5. Synthesis and applications. Our study indicates that management of invasive species should consider ecological, not geographical or political boundaries. Domestic vessels operating within a limited geographic region have high potential to introduce or spread species with restricted distribution, demonstrating importance of intraregional ballast water management. Results presented here should interest policy makers and environmental managers who seek to reduce invasion risk.


The world's ecosystems are undergoing rapid changes, with the introduction and spread of nonindigenous species (NIS) being a key stressor on global biodiversity (Lawler et al. 2006; Clavero et al. 2009). Many species fail to establish after arrival to a new environment but those that succeed may have negative consequences on local community composition, ecosystem functioning and/or services to human society (Chapin et al. 2000). Such changes in the environment, coupled with low probabilities for eradication of established NIS populations, highlight the importance of early detection of introduced NIS and management efforts focused on preventing new introductions (Lodge et al. 2006).

The shipping industry has played a major role in the spread of aquatic NIS globally, with ballast water and sediment typically being more important vectors of introduction than hull fouling in the Great Lakes and possibly in other freshwater ecosystems (Sylvester & MacIsaac 2010). To prevent the introduction of new NIS, ballast water management regulations were enacted by the USA, Canada and many other countries (USCG 1993; IMO 2004; Government of Canada 2006; SLSDC 2008). All vessels arriving at the Great Lakes from foreign ports must perform mid-ocean ballast water exchange on filled ballast tanks and/or saltwater flushing on tanks with residual ballast, such that the final salinity of ballast water in tanks upon entering the Great Lakes is at least 30 parts per thousand. Several studies have demonstrated the high efficacy of these regulations for protecting freshwater ports from invasive species (e.g. Briski et al. 2010; Bailey et al. 2011). Nevertheless, 71% of ballast water transfers in the Great Lakes are exempt from regulations, being conducted by domestic vessels which may spread indigenous taxa or established NIS from one lake to another (Rup et al. 2010). Interlake ballast water transfers could be particularly effective at spreading species as domestic voyages are of short duration, with relatively high survivorship of species in ballast (DiBacco et al. 2012).

The Great Lakes–St. Lawrence River basin, however, can be divided into multiple watersheds resulting in at least five ecoregions: Superior, Michigan-Huron, Erie, Ontario and the Lower St. Lawrence (Abell et al. 2000). These ecoregions are characterized by distinct biodiversity, species endemism and community assemblages and consequently should be managed as separate conservation units (Abell et al. 2000). As a biological invasion occurs when any species arrives and establishes somewhere beyond its previous range (Williamson 1996), the translocation of native species with restricted distribution between adjacent ecoregions can also be considered a new introduction, even if at a finer scale than typically considered in invasion biology literature.

The objective of this study is to quantify the invasion risk posed by domestic ballast water transport considering movement of NIS as well as restricted taxa within the Great lakes (i.e. both indigenous taxa and NIS which do not occur in all five lakes are collectively referred to hereafter as restricted taxa). To parameterize the corresponding risk of introduction and/or spread, we compare the results of a random survey of zooplankton sampled from domestic ballast water with similar data from transoceanic vessels having conducted mid-ocean ballast water exchange (Bailey et al. 2011). Owing to shorter transit and no ballast water management for domestic vs. transoceanic vessels, we tested three a priori hypotheses: (i) zooplankton abundance is higher in ballast water of domestic vessels than transoceanic vessels, (ii) zooplankton species richness is higher in ballast water of domestic vessels than transoceanic vessels and (iii) finally, domestic vessels transport species with restricted distribution to locations from which they have not been reported.

Materials and methods

Between 2007 and 2009, we collected 83 ballast water samples from 72 domestic vessels operating between Great Lakes ports. Vessels were boarded opportunistically at cargo or fuel docks in Corunna, ON, Duluth, MN, Goderich, ON, Sarnia, ON, Superior, WI or Windsor, ON, to collect samples from a variety of source ports throughout the shipping season. In general, a single ballast tank was sampled during each vessel visit when all ballast was loaded at the same location. Three tanks with identical ballast history were sampled on four occasions to confirm similarity of community composition between tanks. In addition, on three occasions, two tanks having different ballast histories were sampled during a single vessel visit. We obtained data about each vessel's ballast history, including date and location of ballast uptake, from vessel personnel.

Tanks were sampled by two different methods. The preferred sample method was to lower a 30-cm-diameter, 53-μm mesh plankton net to the lowest accessible point of the ballast tank through an opened tank access hatch. One to five net hauls were conducted, dependent on haul depth, ensuring that a minimum of 1000 L of water were filtered through the net. When access through a tank hatch could not be obtained, 12·7-cm-outer diameter low-density polyethylene tubing, fitted with a stainless steel check valve, was lowered to the tank bottom through the tank's sounding tube. Fifty litres of water were collected by inertial pumping and filtered through 53-μm mesh. To determine whether differences in sampling methods may influence our results, four tanks were sampled using both methods. After filtration, each sample was preserved in ethanol and sent to taxonomic experts for enumeration and morphological identification. The status of identified taxa (native or NIS; broad or restricted distribution) was determined after extensive literature review.

Statistical Analysis

Statistical comparisons of total abundance and Sorensen's coefficients of similarity calculated for pairs of tanks sampled by different methods and for pairs of tanks within and between vessels indicated no effect of sampling method and confirmed similarity of tanks with identical ballast history (Appendix S1). Following these results, data from multiple tanks sampled during a single vessel transit were averaged and the sampling method was disregarded from further analyses. We compared community composition of unmanaged domestic ballast water with that reported by Bailey et al. (2011) for exchanged ballast water of transoceanic vessels. Bailey et al. (2011) sampled 24 ballast tanks from 16 vessels using net hauls with a minimum of 1000 L of water filtered through the net; methodology was similar to our study, allowing the two studies to be compared. We tested for differences in the cumulative mean abundance of total individuals, restricted individuals and individuals of NIS using the t test or Mann–Whitney U-test (spss 11.5.0; SPSS Inc., 1989–2002; Chicago, IL, USA). A logarithmic transformation was applied to all data sets to meet assumptions of parametric tests. If a Levene's test for homogeneity of variances was significant, or a normal distribution was not achieved, the nonparametric Mann–Whitney U-test was used. A significance level of 95% was used for all statistical analyses.

Species richness of the general vessel population, based on results from our sampled vessels, was estimated by calculating Chao-1, an estimator of species richness based on the number of rare species in a sample (Chao 1984; Chao & Shen 2003). We compared Chao-1 species richness estimates based on all taxa and on restricted taxa recorded from domestic and transoceanic vessels (Bailey et al. 2011) to examine the relative community diversity associated with the two vessel pathways. Sample-based species rarefaction curves were generated for both types of vessels to determine whether results were influenced by sample size. Confidence intervals (95%) were generated to test for significant differences between the two types of vessels (Chao & Shen 2006; Gotelli & Entsminger 2006). Chao-1 estimates were calculated using spade software (Chao & Shen 2006), while rarefaction curves were generated with 5000 random iterations using ecosim (Gotelli & Entsminger 2006).


Community Composition of Zooplankton in Ballast Water

Zooplankton abundance in domestic ballast water ranged from 460 to 1 344 200 individuals per m3, with mean abundance (121 369 individuals per m3) significantly higher than that reported previously from transoceanic vessels (5194 individuals per m3; t test = 5·91, d.f. = 94, < 0·05; Fig. 1; Bailey et al. 2011). While Copepoda and Rotifera each represented 37% of total abundance sampled from domestic vessels, followed by Cladocera (15%), Mollusca (10%) and other taxa <1%, Copepoda dominated samples taken from transoceanic vessels (97% abundance; Fig. 1). Rotifera were the most species-rich group sampled from domestic ballast water, with at least 54 species identified, followed by Copepoda (17 species), Cladocera (14 species), Mollusca (two species, only as veliger larvae), Amphipoda (one species) and unidentified Ostracoda (Appendix S2). In contrast, no Rotifera were found on transoceanic vessels; the most species-rich group was Copepoda (20 species), with all remaining taxa represented by <5 species per group (Bailey et al. 2011). The estimated species richness for the general vessel population was significantly higher for domestic vessels (117 species) than for transoceanic vessels (35 species; Fig. 2).

Figure 1.

Mean (±standard error) and median (horizontal line in bar) total abundance (a), restricted taxa abundance (b) and nonindigenous species abundance (c), by taxon in ballast water sampled from 72 unmanaged domestic (grey bars) and 24 exchanged transoceanic (black bars) vessels. Data for transoceanic vessels from Bailey et al. (2011). Note differences in scale of y-axes.

Figure 2.

Sample-based rarefaction curves for 72 domestic (grey lines, ±95% confidence interval) and 24 transoceanic exchanged vessels (black lines, ±95% confidence interval). Also shown are species richness estimates (Chao-1 ± 95% confidence interval) for domestic (grey bar) and transoceanic exchanged vessels (black bar). (a) All taxa and (b) restricted taxa, respectively. Data for transoceanic vessels from Bailey et al. (2011).

Movement of Species with Restricted Distribution in the Great Lakes

The mean abundance of restricted taxa was 24 170 and 3421 individuals per m3 for domestic and transoceanic vessels, respectively (Fig. 1). Although mean abundance of restricted taxa in domestic vessels was significantly higher in comparison with transoceanic vessels (t test = 2·46, d.f. = 94, < 0·05; Fig. 1), NIS abundance was not different (Mann–Whitney U-test, = −1·24, d.f. = 94, > 0·05; Fig. 1). The median abundance of NIS in domestic and transoceanic vessels was 2015 and 850 individuals per m3, respectively (Fig. 1).

The estimated species richness of restricted taxa in the domestic vessel population (48 species) was significantly higher than that for transoceanic vessels (20·5 species; Fig. 2). In addition, 88% of restricted taxa in domestic vessels were Rotifera (27 species; Table 1; Fig. 1), while restricted taxa in transoceanic vessels were dominated (99%) by Copepoda (Bailey et al. 2011). Sixty-eight per cent of the restricted taxa in domestic vessels have not been previously reported from Lake Superior, 40% from Lake Ontario and 26%, 24% and 22% from Lakes Huron, Michigan and Erie, respectively. At least 18 species (59 events) were recorded in ballast water originating from lakes where those species had not been previously reported, while 23 species (101 events) were recorded from ballast water destined to be discharged outside of their historical distribution (Table 1).

Table 1. List of restricted species transported outside their historical distribution in the Great Lakes by domestic ballast water transits
TaxonBallast water sourceBallast water recipient
SHEOMean abundance (No. m−3)SHEMean abundance (No. m−3)
  1. The number of vessel events moving species from or to a location where not previously reported from is denoted in the ballast water source and ballast water recipient columns, respectively, including mean (±standard error) abundance per m3 sampled. Lakes are: S, Superior; H, Huron; E, Erie and O, Ontario. The * indicates uncertain cases. Historical distribution of species based on extensive literature review, including >60 scientific journal publications, taxonomic keys, and unpublished data reports (K. Bowen and O. Johannsson, Fisheries and Oceans Canada) spanning 1894–2011; a list of references consulted is available from the authors on request.

Daphnia longiremis 21847·90 ± 1573·09
Cercopagis pengoi 10·30
Nitokra hibernica 1886·77
Skistodiaptomus reighardi 17171·21
Anuraeopsis navicula 1265·00
Anuraeopsis sp.1413·331413·33
Asplanchna herricki124 933·33212 603·33 ± 1233·00
Brachionus bidentatus1135·29 
Brachionus budapestinensis 2525·41 ± 474·31
Brachionus havaenensis1633·451633·45
Brachionus quadradentatus3979·68 ± 302·94 
Brachionus urceolaris21199·39 ± 300·19183·80
Cephalodella sp.11880·00 
Colurella unicata1*146·191*146·19
Conochiloides dossuarius1264·10 
Encentrum sp.1 2·34 
Euchlanis alata1135·29131572·92 ± 1347·87
Euchlanis calpidia1*166·151*166·15
Hexarthra mira 1275·82
Kellicottia bostoniensis6242·36 ± 107·19811265·19 ± 936·77
Keratella cochlearis robusta23340·70 ± 121·75231284·55 ± 114·17
Keratella cochlearis tecta10*4*2968·84 ± 1453·725*7*2*2968·84 ± 1453·72
Keratella taurocephala1*1*12·70 ± 6·401*1*12·70 ± 6·40
Keratella valga tropica41826·00 ± 1169·88141487·86 ± 967·22
Lepadella ovalis1*3*1*135·76 ± 94·942*1*1*135·76 ± 94·94
Lepadella sp.1*1*6854·00 ± 6494·002*6854·00 ± 6494·00
Notholca acuminata 14544·00
Notommata sp. 1224·80
Platyias patulus 1633·45
Polyarthra remata38517·74 ± 4150·481342 603·85 ± 36806·39
Polyarthra vulgaris62448·43 ± 1355·562212 323·39 ± 5855·61
Pompholyx sulcata2173·64 ± 24·753416·83 ± 334·93
Synchaeta kitina31058240·16 ± 4091·362417 181·29 ± 11630·90
Synchaeta pectinata45561·23 ± 2662·21135671·57 ± 3864·70
Trichocerca longiseta150·68311246·71 ± 676·67
Trichocerca rattus 1233·30

Despite similar abundance of NIS between the two vessel types, the taxonomic composition was represented by different taxonomic groups. NIS in domestic vessels were mostly represented by Mollusca (92%, i.e. Dreissena polymorpha and Dreissena rostriformis bugensis, Fig. 1; Table 2), while Copepoda overwhelmingly dominated NIS in transoceanic vessels (99·9%, Fig. 1; Bailey et al. 2011). The cladoceran Cercopagis pengoi and copepod Nitokra hibernica were the only two NIS reported from domestic vessels which have not yet established populations in all five of the Laurentian Great Lakes. Only one of fourteen NIS sampled from transoceanic vessels, the euryhaline copepod Acartia tonsa was considered high risk for establishment in the Great Lakes (Bailey et al. 2011).

Table 2. List of nonindigenous species recorded in domestic ballast water samples
TaxonCommon NameMean Abundance (No. m−3)Occurrence (%)
  1. Mean abundance (±standard error), when present and occurrence in 72 vessels are provided. All taxa are considered established in all five Great Lakes except C. pengoi and N. hibernica, which are not reported from Lake Superior.

Eubosmina coregoniWaterflea2603·73 ± 405·9727 (37·50)
Bythotrephes longimanusSpiny waterflea7·41 ± 2·80 7 (9·72)
Cercopagis pengoiFish-hook waterflea0·30 1 (1·39)
Eurytemora affinisCalanoid copepod1861·91 ± 1394·9412 (16·67)
Nitokra hibernicaHarpacticoid copepod886·77 1 (1·39)
Echinogammus ischnusAmphipod15·18 ± 4·82 1 (1·39)
Dreissena polymorphaZebra mussel9657·29 ± 2584·6656 (77·78)
Dreissena rostriformis bugensisQuagga mussel6986·28 ± 1603·4046 (63·89)


Domestic vessels on the Great Lakes transport a higher abundance of zooplankton in unmanaged ballast water, by two orders of magnitude, compared with transoceanic vessels arriving with exchanged ballast water. Confirming our expectations, we found that the diversity of zooplankton was significantly greater for the domestic vessel population than for transoceanic vessels. We sampled at least 89 taxa in domestic ballast water of which 31 had restricted distribution in the Great Lakes and eight were NIS. As the total volume of ballast moved by domestic vessels is more than three times greater than the volume moved by transoceanic vessels in the Great Lakes (Rup et al. 2010), they represent a tremendously important vector for transfer of species within and between the Great Lakes.

While most NIS reported from domestic ballast water were already present in all five lakes at the time of our study, C. pengoi, a harmful NIS of global concern (ISSG 2011), and N. hibernica are not established in Lake Superior; as both species were sampled from ballast water destined for discharge in Lake Superior, domestic ships are clearly a vector for transfer of established NIS within the region. The high abundance and occurrence of the other NIS, particularly Dreissena spp., provides significant support for this finding, although the already widespread distribution of the remaining NIS prevents definitive conclusions on the role of domestic ballast water in their spread. Nevertheless, the continuing transfer of NIS already broadly distributed is concerning as mixing distant populations may provide opportunities for novel genetic recombination and evolutionary shifts in key life-history or morphological characters, which may result in unexpected surges of NIS population abundance and harmful impacts (Lockwood, Cassey & Blackburn 2005).

In addition to the secondary spread of NIS to new areas, domestic shipping can act as a vector for introduction or spread of indigenous species with restricted distribution to new locations. We recorded at least 23 native species (101 events) being moved to new locations outside their historical distribution, potentially resulting in new introductions to Lakes Superior, Huron and Erie. Although commonly referred to as a single ecosystem, diversity and community composition differ among the five Great Lakes (Abell et al. 2000) and introduction or spread of species from one lake to another could have negative ecological and economical consequences (Chapin et al. 2000; Lawler et al. 2006; Clavero et al. 2009), even if species are not spread far from the native region (Cullingham et al. 2011). Most restricted taxa found in domestic vessels were Rotifera, which until now are reported as NIS only in Italian and Polish inland waters (Gherardi et al. 2008; Ejsmont-Karabin 2011) and are not listed as notorious invaders. However, it is challenging to predict which species may become problematic owing to the context-dependent nature of invasions (Ricciardi, Palmer & Yan 2011).

Considering the long history of ballast movements within the Great Lakes, the continued existence of restricted taxa suggests that intrinsic physical or chemical variability between lakes may prevent the establishment of some species in different parts of system (Grigorovich et al. 2003); however, even after many years of introduction, we would not expect all species which have the potential to establish to have successfully done so owing to long lag-phases and sampling bias, environmental and demographic stochasticity, as well as insufficient genetic diversity to allow for adaptation (Grigorovich et al. 2003; Lockwood, Cassey & Blackburn 2005). Notably, this study reported 18 species (59 events) in ballast water taken from locations where those species have not previously been reported. While it is likely that this finding is confounded by incomplete historical records of species' distributions, our literature review possibly overestimates species' ranges as reports were compiled on a lake-wide basis. Alternatively, it is possible that the individuals in question may have been loaded into tanks at a prior port of call and persisted for multiple voyages; however, the densities of restricted species in this study are at least one order of magnitude greater than the total abundance of invertebrates previously reported in residual ballast water (Duggan et al. 2005). A third and very plausible explanation for species reports from new locations is that interlake spread of some restricted taxa has already occurred.

Management Implications

Low probabilities for the eradication of established NIS populations as well as millions of dollars spent annually on eradication programmes highlight the importance of management efforts focused on preventing new introductions and early detection of introduced NIS (Lodge et al. 2006). Our comparison of zooplankton abundance and species richness in unmanaged domestic ballast water and foreign exchanged ballast water indicates that unmanaged domestic vessels currently pose the greater risk for species invasions. The risk of unintentionally introducing new species could therefore be reduced through ballast water management of all ships, independent of operational area. In the future, ballast water regulations, as well as management of other vectors of invasive species, should consider ecological, not geographical or political boundaries. Recognizing that current requirements for ballast water exchange cannot be met by vessels operating intraregionally, it will be important to develop effective treatment technologies, such as filtration, deoxygenation and/or ultraviolet treatment, which could be utilized onboard vessels. Results presented here should interest policy makers and environmental managers who seek to reduce invasion risk, as well as shipping industry.


We thank participating vessel crews, vessel owners and managers, port facilities and the Canadian Shipowners' and US Lake Carriers' Associations. M. Deneau, L. Quiring, S. Santavy, J. Weakley and the Minnesota Pollution Control Agency provided essential logistical and sampling support. EcoAnalysts, Inc. conducted taxonomic analyses. This research was supported by Transport Canada and Fisheries and Oceans Canada.