For the past several decades, the North American Great Lakes have suffered from eutrophication. The deteriorating state of the Great Lakes alarmed both the governments of Canada and the United States resulting in the Great Lakes Water Quality Agreement, which has brought about substantial improvements in water quality. While phosphorus abatement resulted in a significant decrease in nutrients, the sudden invasions of exotic species posed a serious threat to Great Lakes food webs. The zebra mussel (Dreissena polymorpha) and the quagga mussel (D. bugensis), followed by other exotic species, infested Lakes Erie and Ontario causing a drastic reduction in phytoplankton biomass and increasing water clarity. In Lake Erie, post-Dreissena declines in phytoplankton size structure and changes in community composition were observed in this study, along with significant declines in primary productivity in the west basin. At the other end of the food web, exotic species such as alewife (Alosa pseudoharengus), rainbow smelt (Osmerus mordax) and white perch (Morone americana) have become important to the Lake Erie commercial fishery, while other native fish species have declined. This paper presents an historical perspective and a general overview of the impact of nonindigenous species in the North American Great Lakes from the base of the food web to the fisheries. Lake Erie has been chosen as a case study to provide a detailed treatment. The expansion and growth of nonindigenous species has been responsible for significant modifications to the structural and functional characteristics of the food webs and fisheries of the Great Lakes. Our experience demonstrates the significance of the impact of exotics and the need to manage this serious problem on a global basis so that the integrity of food webs and fisheries throughout the world can be protected.
This paper is dedicated to Dr. Jack Vallentyne for his contributions to Great Lakes research, especially for the implementation of the ‘ecosystem approach’. These contributions were in evidence in revisions to the Great Lakes Water Quality Agreement and more currently in the management of exotic species.
The North American Great Lakes are an extremely important global natural resource, shared by Canada and the United States. They contain almost 20% of the world's fresh surface water (GLIN, 2005). Since European colonization of North America began in the 1600s, numerous stressors that can be directly attributed to human activities have been shown to affect the integrity of the Great Lakes. An historical perspective of major stressors affecting the Great Lakes during the past century (1930 to the present) is given in (Figure 1). The stressors affecting the Great Lakes can be classified broadly as eutrophication, contamination and establishment of exotic species. Of these, eutrophication of Lakes Ontario and Erie was the first to receive widespread attention. A concerted effort to reverse this trend resulted in the development of the Great Lakes Water Quality Agreement (GLWQA).
The GLWQA of 1972 between Canada and the United States (IJC, 1988) implemented a phosphorus abatement program that proved very successful as evidenced by substantial improvements in water quality (Burns, 1985; Leach, 1999). The GLWQA was revisited in the 1980s to address the issue of contamination of the Great Lakes basin via the discharge of toxic substances. As a result, 42 of the most severely polluted sites were designated as ‘Areas of Concern’ and remedial action plans were established to chart a course for their improvement (Hartig and Zarull, 1992; Munawar et al., 2003). Success in reversing contamination has been less successful than for eutrophication with only two sites (Collingwood Harbour and Severn Sound, both in Georgian Bay, Lake Huron) being successfully delisted as Areas of Concern (Environment Canada, 2003). Nevertheless, the GLWQA was unprecedented in requiring holistic and integrative assessments of ecosystem health. This is one example of the true legacy of Dr. Vallentyne's work.
Since non-indigenous species are one of the major stresses together with nutrient and contaminant enrichment, this paper has been designed to focus on the present status of the exotics in the Great Lakes and their impact on the food web integrity. The paper attempts to integrate the historical perspective with the current situation concerning the impact of non-indigenous species (NIS) on various trophic levels. Lake Erie has been selected as a case study to offer a detailed prognosis of their impacts and to demonstrate how exotic species have affected various components of the Lake Erie food web, namely phytoplankton/primary productivity, zooplankton, benthos and fish. In this sense, the review is confined to those species perceived to have the most significant impacts on each food web component.
It is indeed a great pleasure to present and dedicate this invited paper as part of the J.R. Vallentyne Lecture Series of the Aquatic Ecosystem Health and Management Society (AEHMS). It is befitting to devote this paper to Dr. Vallentyne for his global anti-eutrophication and pro-phosphorus abatement campaigns well documented in The Algal Bowl (1974). Dr. Vallentyne had tremendous influence initiating, developing and implementing the ‘ecosystem approach’ in the management of the Great Lakes and emphasizing the need to think holistically and globally. It is a privilege to deliver this paper to honour Dr. Vallentyne's teachings and views that continue to have far reaching implications.
Great Lakes: historical perspective
The history of NIS in the Great Lakes is very well connected with the arrival of the first European settlers. The Europeans facilitated their livelihood by making extensive modifications in the Great Lakes basin permitting access of the exotic species to remote areas upstream of Niagara Falls (Jude et al., 2005). Construction of the Welland canal in the 19th century provided open access to large ships with the huge quantities of ballast water harbouring a diversity of NIS in various stages of growth. The release of ballast water containing organisms foreign to the Great Lakes has been the prime vector of NIS over the past two decades. In addition, there were numerous other vectors such as ship fouling (e.g., sea lamprey), aquarium release (e.g., European ear snail) and intentional release (e.g., chinook salmon, rainbow trout) (Mills et al., 1993; Hall and Mills, 2000). The opening of the St. Lawrence Seaway in 1959 increased the number of exotics brought into the basin by shipping vectors and therefore increased the likelihood that new NIS would become established (Ricciardi, 2001; Grigorovich et al., 2003). While NIS are not a new phenomena in the Great Lakes, numbers have increased exponentially in recent years (Figure 2) with shipping and unintentional release identified as the two most likely vectors of introduction (Figure 3). The increased risk of new species introduction greatly increases the potential for harmful ecosystemic effects, even though the impacts of most invasive species are heretofore unknown (Figure 4). Table 1 provides a list of those species known to be causing food web alterations in the Great Lakes. The roughly 8 million tonnes of freight that enters Great Lakes ports each year from all over the world (Grigorovich et al., 2003) combined with the ability of some organisms to survive in ballast sediment (Bailey et al., 2003) makes exotic species a global threat to the ecological integrity of large lakes.
Non-indigenous species are the most serious threat to the integrity of the Laurentian Great Lakes since they have affected all trophic levels. It is estimated that approximately 170 invasive species have been introduced in the Great Lakes (Mills et al., 1993; Ricciardi, 2001; Holeck et al., 2004; Jude et al., 2005). Little is known however, about bacterial, viral, protozoan and algal invaders (Sheath, 1987; Jude et al., 2005) except where significant impacts have been reported by the disease-causing invaders. Some examples are the protozoan Myxobolus cerebralis that is the cause of whirling disease in salmonids (Ganzhorn et al., 1992), the bacterial pathogen Furunculosis that affects fisheries (Benson and Boydstun, 1999) and the protist parasite Glugea hertwigi imported along with its host, the rainbow smelt (Mills et al., 1993). Microbes in large quantities are transported to the harbours and coastal regions of the Great Lakes. Abundant microbial communities with their high growth rate and ability to adapt easily to a wide range of nutrient and other environmental parameters are ideally suited for establishing in new environments. Scarcity of serious disease outbreaks may simply be due to good fortune rather than reflect an absence of microbial invaders (Jude et al., 2005).
Very little is known about algal invaders in the Great Lakes (Norse, 1991; Titley, 1992). The red alga Bangia atropurpurea, and green algae Enteromorpha and Nitellopsis are known to be common in the Great Lakes (Sheath, 1987; Jude et al., 2005). The increase of ion concentration in the lower Great Lakes since 1850 has facilitated the invasion of marine-based species (Beeton, 1965; Sheath, 1987). Diatoms have contributed significantly to the list of invaders (Mills et al., 1993). Some of the diatom species such as Stephanodiscus binderanus, Actinocyclus normanii fo. Subsalsa and Cyclotella atomus have become established in near-shore waters that are eutrophic and rich in ionic strength.
Macrophyte invaders have become well established in the harbours and wetlands of the Great Lakes basin, representing 10 to 30% of the local flora (Mal et al., 1997; TePas and Charlebois, 2002; Jude et al., 2005). Some of the commonly found species are purple loosestrife (Lythrum salicaria), Eurasian watermilfoil (Myriophyllum spicatum), Phragmites australis, reed canary grass (Phalaris arundinacea), hybrid cattails (Typha x glauca), water clover (Marsilea quadrifolia), fanwort (Cabomba caroliniana), water cress (Rorippa nastrurtium-aquaticum), water chestnut (Trapa natans), yellow floating heart (Nymphoides peltata) and curly pondweed (Potamogeton crispus) (Mills et al., 1993; Jude et al., 2005).
Recent introductions of zooplankton and benthos have threatened the structure and function of the microbial food web, energy flow and eventually the fisheries by altering predator and prey interactions. Commonly known invertebrate introductions include Bythotrephes longimanus (formerly B. cederstroemi; Evans, 1988) and Cercopagis pengoi(MacIsaac et al., 1999; Makarewicz et al., 2001). Amongst the benthos, the Asiatic clam (Corbicula fluminea) is commonly found in the Great Lakes and is limited in its expansion due to its sensitivity to lower temperatures (French and Schloesser, 1991).
The discovery of the zebra mussel, Dreissena polymorpha, in Lake Erie in 1988, occurred at a critical time when the lower Great Lakes were recovering from eutrophication due to the implementation of a phosphorus abatement program. This species is native to the Baltic Sea and transported in the ballast water of transoceanic ships (MacIsaac et al., 1992; Leach, 1993), The sudden appearance of zebra mussels, along with the closely related quagga mussel (Dreissena bugensis) and their subsequent rapid spread throughout the Great Lakes and along the Ohio and Mississippi River corridor southward to Louisiana (New York Sea Grant, 2000), focused the attention of researchers and managers on the potential problem of NIS. Early studies indicated that zebra mussels would dramatically alter ecosystem function by increasing water clarity (Holland, 1993), decreasing algal abundance (Nicholls and Hopkins, 1993) and changing benthic composition (Dermott and Munawar, 1993). More detailed research about exotic species revealed that species invasions, and subsequent habitat alterations, have been occurring in the Great Lakes for over 150 years (Mills et al., 1993) and current species invasions continue to alter the integrity and trophic status of Great Lakes food webs (MacIsaac, 1999; Munawar et al., 1999, 2001; Hall and Mills, 2000).
The ecological integrity of the Great Lakes fisheries has been of concern for a long time because these lakes contain one of the world's largest commercially valuable freshwater fisheries. Table 2 shows the status of commercially harvested species in the Great Lakes. Lake Ontario has the highest number of species that have been extirpated (6) followed by Lake Erie (3). Lake Erie has the greatest number of species under threat (6) followed by Lakes Ontario (4) and Huron (4). Conversely, only 1 commercially harvested species in Lake Ontario and 2 in each of the remaining lakes are considered to have good prospects for survival. The lower lakes, Ontario and Erie, are the most stressed and their watersheds are also the most densely populated by humans.
There has been a progressive depletion of many indigenous fish stocks in the Great Lakes due to various stressors (Figure 1; Table 2). Populations of lake sturgeon (Acipenser fulvescens), deepwater cisco (Coregonus johannae), lake trout (Salvelinus namaycush), lake whitefish (Coregonus clupeaformis) and lake herring (Coregonus artedii) have collapsed and been replaced by other less valuable species. The elimination of indigenous forms such as blue pike (Sander vitreus glaucus), Atlantic salmon (Salmo salar), lake trout and many others adapted to specific environments represents the loss of irreplaceable genetic material. The impacts of exotic species have not gone unnoticed. For example, alewife (Alosa pseudoharangus), rainbow smelt (Osmerux mordax), white perch (Morone americana), eurasian ruffe (Gymnocephalus cernuus) and zebra mussels have been associated with changes in indigenous fish resources, notably yellow perch (Perca flavescens) and bloater (Coregonus hoyi) in Lake Michigan, bloater in Lake Huron and yellow perch in Lake Erie (Munawar et al., 2001). These exotics have now become a permanent component of aquatic communities and help maintain commercial catch at near historic levels.
An attempt was made to control the number of exotic species that could potentially enter the Great Lakes in 1989 with the imposition of (voluntary) ballast water controls for trans-oceanic vessels. The assumption was made that any freshwater species living in the ballast tanks would be killed off by high salinity ocean water (Locke et al., 1991). However, since then, the number of exotic species found in the Great Lakes has continued to grow exponentially (Ricciardi, 2001). One explanation for the continued appearance of invasive species is the ability of certain rotifers and cladocerans to survive in residual ballast sediments after mid-oceanic ballast exchange (Bailey et al., 2005). More than 170 invertebrate species have been identified in vessels entering the Great Lakes, including 22 freshwater species not previously reported (Holeck et al., 2004; Bailey et al., in press).
Lake Erie case study
Lake Erie, the smallest of the Great Lakes by volume has been among the most impacted by exotic species including zebra mussels, Bythotrephes and round gobies (Neogobius melanostomus) (Hall and Mills, 2000). Consequently it has been chosen as a case study in this paper for in depth impact assessment. Beginning with the establishment of zebra mussels in the late 1980s, dramatic changes in water clarity (Holland, 1993; Charlton et al., 1999), phytoplankton (Nicholls and Hopkins, 1993; Makarewicz et al., 1999; Munawar and Munawar, 1999), primary productivity (Dahl et al., 1995; Munawar et al., 1999), zooplankton (MacIsaac et al., 1992; Johannsson et al., 1999) and benthos (Dermott and Munawar, 1993; Manny and Schloesser, 1999) have been observed. Zebra mussels have also been shown to alter patterns of nutrient cycling (Johengen et al., 1995; Arnott and Vanni, 1996) and to promote algal blooms (Arnott and Vanni, 1996; Vanderploeg et al., 2001), potentially ushering in a return to pre – phosphorus abatement conditions. These changes have been shown to reverberate through the food web and directly affect higher trophic levels (Dermott et al., 1999; Dermott, 2001; Munawar et al., 2006). At least 32 native fish species have been identified as rare in Lake Erie (Noble, 2002) with introduced species identified as one of the leading causes of this decline (Dextrase and Mandrak, 2005). Massive fall die-offs of fish-eating waterbirds, including common loons, mergansers and grebes, due to Type E botulism have been recurring in Lake Erie since 1999 (Campbell, 2003). The mechanism of infection is not clear, as live fish do not typically develop Type E botulism (Yule et al., 2004), however the presence of zebra mussels, quagga mussels and round gobies in post mortem analyses of gizzard contents has led to speculation that exotic species are a contributing factor (Campbell, 2003). Generally, information about the impact of the exotics on lower trophic levels is scarce or lacking especially on microbes and phytoplankton. Consequently an attempt has been made to include effects of the NIS across trophic levels of the food web.
Phytoplankton and primary productivity samples were collected using an integrated sampler (Schroeder, 1969). Size fractionated primary productivity was measured with a 14carbon tracer following the standard protocol of Munawar and Munawar (1996). In 1988, only > 20 μ m and < 20 μm fractions were determined, whereas > 20 μ m, 2–20 μ m and < 2 μ m fractions were determined in later years. In order to make the data comparable among years, we simply added the 2–20 μ m and < 2 μ m categories together to produce an estimate of < 20 μ m primary productivity. Three stations in the west basin were sampled along a NW-SE transect between 1988 and 1996.
Phytoplankton samples were preserved in Lugol's iodine and identification and enumeration followed the Utermöhl (1958) inverted microscope technique (Munawar et al., 1987). Size composition of the phytoplankton community was assessed using Equivalent Spherical Diameter (ESD). Single index stations from 1988, 1992 and 1998 from each of the three basins were used to make pre- and post-dreissenid mussel comparisons of the phytoplankton community due to the limited availability of pre-mussel data. Chlorophyll a was determined by acetone pigment extraction (Strickland and Parsons, 1968). Values were obtained from Environment Canada's STAR database.
In order to provide an holistic review of changes in the Lake Erie food web mitigated by NIS, new data on phytoplankton and primary productivity, combined with previously published data on zooplankton (Johannsson et al., 2000) and benthos (Dermott et al., 1999; MacDougall et al., 2001) are included in this paper. Sampling locations for primary productivity, phytoplankton, zooplankton and benthos are shown in Figure 5. Furthermore, we present a review of the impacts of exotic species on native fisheries.
Phytoplankton and chlorophyll a
The impact of non-indigenous species in general and dreissenid mussels in particular has been dramatic on the Lake Erie ecosystem. Relatively few studies have explored the impacts of zebra mussels on phytoplankton communities (Nicholls and Hopkins, 1993; Makarewicz et al., 1999; Munawar and Munawar, 1999). Commonly chlorophyll a and secchi depth (e.g. Holland, 1993; Nicholls and Hopkins, 1993; Charlton et al., 1999) have been used to assess the effects of mussel filtration. Impacts on lower trophic levels are therefore difficult to discern. Nicholls and Hopkins (1993) found that chlorophyll a levels declined significantly after the introduction of zebra mussels to Lake Erie, as did phytoplankton abundance, which was attributed to the filtering impacts of zebra mussels. Lakewide and extensive monitoring programs administered by Fisheries & Oceans Canada and Environment Canada revealed significant declines in chlorophyll a concentrations in all 3 basins of Lake Erie (Figure 6) with the west basin being the most heavily impacted. Due to the unavailability of data (lack of monitoring) prior to the advent of zebra mussels, it is not possible to make immediate pre and post mussel comparisons of chlorophyll a. However, it is very obvious from the data that the lowest levels of chlorophyll a were observed after the establishment of zebra mussels. Indeed, other studies attributed this reduction of chlorophyll a to filtering by Dreissena (Nicholls and Hopkins, 1993; Makarawicz et al., 1999). On the other hand, there are others who attribute this reduction of chlorophyll a to phosphorus abatement (Charlton et al., 1999). Regardless, the continued decline in chlorophyll a levels combined with the ability of zebra mussels to selectively graze phytoplankton (Heath et al., 1995) created the circumstances for dramatic changes in the biomass, structure and composition of phytoplankton. Recurring blooms of Microcystis aeruginosa (Cyanophyta), a toxic blue-green alga, have been reported in the west basin with greater frequency in the post-zebra mussel period and has been linked to the selective rejection in zebra mussel pseudofeces of large colonial cyanophyta (Vanderploeg et al., 2001).
Munawar and Munawar (1999) showed that mean summer biomass of the phytoplankton community of western Lake Erie declined significantly from 5.0 ± 1.7 mg m−3 in 1970 to 1.8 ± 0.3 mg m−3 in 1992, corresponding to a shift in the species composition from Diatomeae to Chlorophyta. Interpretation of both chlorophyll a and phytoplankton data, however, is hindered by an absence of lakewide monitoring data from the 1980s, making it impossible to discern the effects of zebra mussel grazing compared to the impact of phosphorus abatement (Munawar et al., 1999). However, similar results in Lake Ontario have been observed and attributed directly to mussel grazing (Munawar and Munawar, 2003; Munawar et al., 2006) offering some evidence that the changes occurring in Lake Erie may be directly related to the zebra mussel.
Since lake wide data sets are lacking in the 1980s, single station data has been included in each of the three basins for the summer of 1988 and compared with similar data from 1993 and 1998. A sharp drop in phytoplankton biomass, from 10.8 g m−3 in 1988 to 2.6 g m−3 in 1992, was observed in the west basin; however, little change was observed in the central and eastern basins (Figure 7a). A small though highly significant reduction in phytoplankton biomass between pre- and post-Dreissena periods was observed by Makarewicz et al. (1999) in the west basin but not in the central or east basins.
Phytoplankton size composition, measured as Equivalent Spherical Diameter (ESD), showed a decline in all three basins between 1988 and 1992 (Figure 7b). In the west basin, the weighted mean ESD declined from 20.8 to 11.1 μ m between 1988 and 1992. Similar trends were observed in the central basin (from 24.0 to 19.6 μ m) and the eastern basin (from 25.0 to 9.8 μ m). Changes in the size structure of the phytoplankton community were also reflected by structural changes in the taxonomic composition (Figure 8). The west basin showed a shift away from Cyanophyta in 1988 to smaller Chlorophyta in 1992 and a combination of Chlorophyta and Diatomeae in 1998. The central basin also showed an increased prevalence of Chlorophyta in 1992, a combination of Chlorophyta and Diatomeae in 1998 with a decrease in larger dinoflagellates (Dinophyceae). The eastern basin showed a shift away from Dinophyceae in 1988 to Chlorophyta in 1992 followed by Diatomeae in 1998. These findings suggest that larger sized phytoplankton are being grazed by Dreissena, consistent with the findings of Horgan and Mills (1997) who showed that particle sizes up to 150 μ m were readily filtered (including filamentous Cyanophyta).
Changes in the phytoplankton community of Lake Erie are most apparent in the west basin (Makarewicz et al., 1999; Munawar and Munawar, 1999; this study) and the west basin is where zebra mussels were first observed to grow profusely. Our analysis of primary productivity is therefore confined to the west basin and includes immediate pre- (1988) and post- (1993–1996) dreissenid mussel years. Significant reductions in spring primary productivity between 1988 and 1993 were observed for both netplankton (> 20 μ m) and nanoplankton+picoplankton (< 20 μ m), a trend that continued through 1995 (Figure 9). Mean netplankton rates declined from 5.8 ± 1.2 mg C m−3 h−1 during 1988 to 2.4 ± 1.8 mg C m−3 h−1 during 1993. Similarly, mean nanoplankton+picoplankton rates declined from 33.5 ± 2.3 to 8.0 ± 4.1 mg C m−3 h−1 during the same period. While previous research had shown a decline in spring primary productivity in the west basin of Lake Erie coincident with the establishment of the zebra mussel (Munawar and Munawar, 1999), this is the first time that this trend is confirmed with data from several stations. Moreover, this trend towards declining productivity was apparent at all stations in the west basin. The fact that there is no evidence of recovery suggests that dreissenid mussels have irreparably altered the base of the food web, a situation that can only have serious repercussions for higher trophic levels.
Significant declines in mean summer netplankton primary productivity were also observed between pre- and post-zebra mussel invasion periods; however similar trends were not apparent for nanoplankton+picoplankton rates (Figure 9). It is important to note that the same patterns of declining summer netplankton productivity and variable, though not declining, nanoplankton+picoplankton productivity was evident at all stations. The relative stability of nanoplankton+picoplankton rates was consistent with our earlier observations of phytoplankton size compositional changes, showing an increasing emphasis on smaller sized particles. Observed declines in both spring and summer primary productivity have not necessarily resulted in declines in pelagic photosynthesis. Increased light penetration in the water column has increased euphotic depth and maintained areal rates of primary production. Aerial production in the west basin during spring 2001-2002 was ≈ 1.0 g C m−2 d−1 (Fitzpatrick, 2003) similar to that reported for pre-phosphorus abatement 1970 (Glooschenko et al., 1974) and post-zebra mussel 1993 (Dahl et al., 1995).
Zooplankton biomass in Lake Erie, prior to the zebra mussel, was shown to be correlated positively with phytoplankton biomass (Makarewicz, 1993). Zooplankton grazing, especially by Daphnia sp., combined with reduced phosphorus inputs limited the size of the phytoplankton standing crop (Wu and Culver, 1991). In the post-zebra mussel invasion period, declines in zooplankton biomass were observed only in the east basin, where only the nearshore was impacted (Johannsson et al., 1999). Zooplankton biomass remained stable in the west and central basin; however, zooplankton production was found to be 15 to 90% less than potential production lakewide after the zebra mussel invasion (Figure 10), where potential production was estimated from linear regression factoring in primary production and mean zooplankton length but excluded quantification of Dreissena grazing (Johannsson et al., 2000). This loss of zooplankton production was attributed to zebra mussels grazing directly on phytoplankton and therefore diverting energy away from zooplankton (Johannsson et al., 2000). Unlike the phytoplankton community, that was most impacted by zebra mussels in the west basin, the zooplankton community, both in terms of biomass and production, appears to be most impacted in the nearshore areas of the east basin where zebra mussel concentrations are the highest (MacDougall et al., 2001).
The rapid spread of dreissenid mussels across the bottom sediments of Lake Erie has been previously reported (Dermott and Munawar, 1993; Dermott et al., 1998). In 1993, Dreissena were found at all of our sampling locations in the northern portion of the lake and represented more than 90% of the benthic biomass, except for one offshore station in the central basin where dreissenids contributed only 20% (Figure 11). By 1998, however, the situation had changed considerably. In the east basin, non-dreissenid benthic biomass remained stable, while dreissenid biomass increased significantly. The central and west basins, by comparison, experienced significant increases in non-dreissenid benthos at 3 of 4 stations, combined with significant decreases in Dreissena (Figure 11). The long-term impacts of Dreissena on the benthic fauna appeared to be most prevalent in the east basin, because low dissolved oxygen concentrations in the central basin limited dreissenid abundance. While the west basin has continued to show signs of recovery from eutrophication, notably with the re-establishment of mayfly (Hexagenia limbata) colonies (Manny and Schloesser, 1999), the east basin is showing signs of continued deterioration after the Dreissena invasion. The amphipod Diporeia, that previously dominated the benthos, has all but disappeared from the east basin (Dermott et al., 1999). It is generally believed that Diporeia is being out-competed for diatoms by dreissenid mussels (Dermott, 2001; Lozano et al., 2001). Diporeia served as a food source for rainbow smelt and lake whitefish, both mainstays of the commercial fishery. The loss of this food resource is predicted to have direct consequences for the commercial fishery (Dermott et al., 1999; Hoyle et al., 2003).
In this section, we explore the relative impacts of non-indigenous species on the Lake Erie fisheries. The non-indigenous species history of Lake Erie has been well documented (e.g., Mills et al., 1993; MacIsaac, 1999; Hall and Mills, 2000) and the reasons for these introductions cover the spectrum from clearly defined management objectives to accidental release. But what all of these introductions share is a lack of foresight regarding the long-term prospects of survival for native fishes. Lake Erie contains one of the world's largest commercial freshwater fisheries. There are 143 fish species, of which 19 are commercially significant (Munawar et al., 2001). The composition and structure of the Lake Erie fishery has been, and continues to be, influenced heavily by commercial fishing, phosphorus abatement and exotic species (Nepszy, 1999). Seventeen introduced fish species have become established in Lake Erie since the mid-1800s, and a further 11 species are known to have been introduced but have failed to become established (Table 3). Undoubtedly, the introduction of additional species, that have not become established, has occurred but has not been documented. These species have been introduced through various vectors including authorized, unauthorized and accidental stocking, colonization through man-made canals, aquarium releases and ballast water releases.
At the same time, 32 native species have become rare enough in Lake Erie to warrant a conservation listing by at least one state or province (Noble, 2002). These include twelve species federally listed by the Committee on the Status of Endangered Wildlife in Canada (COSEWIC) (Table 4). In Canada, the leading causes of imperilment in COSEWIC-listed species are (in order of importance), habitat alteration and destruction, introduced species, and overexploitation (Dextrase and Mandrak, 2005). The impacts of introduced species on native species and ecosystems include competition with, and predation on, native species, habitat alteration or destruction, and introduction of diseases (Hunter, 1996).
The establishment of introduced fishes in Lake Erie has led to the decline, sometimes to the point of imperilment, of several native species. The deliberate stocking of the common carp (Cyprinus carpio; all common and scientific names for fishes according to Nelson et al. (2004); see Table 3 for scientific names of introduced species) and goldfish (Carassius auratis) in the late 1800s, in conjunction with infilling and increased turbidity related to land use changes, likely initiated the decline of coastal wetlands in Lake Erie. It is well documented that the common carp, in particular, uproots large quantities of aquatic vegetation and can decimate vegetated areas of wetlands (Chow-Fraser, 1998). This would impact all native fish species that use wetlands as spawning and nursery areas and, in particular, those species that require wetlands, such as five of the COSEWIC-listed species in Lake Erie (Table 4). However, the decline of these species went largely unnoticed until relatively recently (last 25 years), unlike commercially fished species whose decline was apparent by the late 1800s (Ryan et al., 2003).
The commercial fishery of Lake Erie was originally based on coldwater species such as lake sturgeon, lake trout, lake whitefish, and lake herring, but, as these species declined, the fishery switched to the coolwater yellow perch, sauger (Sander canadensis), blue pike, and walleye (Sander vitreus) (Ryan et al., 2003). The decline of the coldwater species was largely the result of overfishing, but was exacerbated by the introduction of the sea lamprey (Petromyzon marinus), first recorded in Lake Erie in 1921 (Emery, 1985). The lamprey preyed heavily on cold-water species (Ryan et al., 2003) so that they were virtually extirpated from Lake Erie. However, as similar declines have not been documented across the range of these species (e.g., in inland lakes that lack commercial fisheries and sea lamprey), coldwater fish species have not received any conservation protection.
By the mid-1900s, three additional introduced fish species, the alewife (first recorded in 1931), rainbow smelt (1935) and white perch (1953), had become established in Lake Erie as the result of colonization through the Welland Canal from Lake Ontario (alewife, white perch) and following accidental release into Lake Michigan (rainbow smelt) (Table 3). These pelagic species had significant negative impacts on the native coolwater species (e.g., walleye, yellow perch) including predation on, and competition with, their eggs and larvae (Scott and Crossman, 1973; Schaeffer and Margraf, 1987). All of these impacts, but especially over-fishing, led to the extinction of the blue pike by the 1960s (Campbell, 1987).
In an effort to control these very abundant, newly established, pelagic fish species, various Pacific salmonids were intensively stocked from the 1960s (coho and Chinook salmons; Oncorhynchus kisutch and O. tshawytscha) to 1980s (rainbow trout; O. mykiss). Although only limited natural reproduction occurred in the salmons (including the pink salmon (O. gorbuscha), which colonized Lake Erie in the late 1970s following accidental release into Lake Superior (Emery, 1985)), rainbow trout have become established. This followed several unsuccessful attempts to stock Pacific salmon in Lake Erie since the late 1800s (Table 3). These salmonids, as well as the exotic Bythotrephes longimanus and Cercopagis pengoi, have likely contributed to the recent decline of alewife and rainbow smelt in Lake Erie, as they have in the other Great Lakes (Mills et al., 2005; Stetter et al., 2005). Another salmonid species, the brown trout (Salmo trutta), was successfully introduced as early as the late 1800s; however, it has never been very abundant in Lake Erie and its impact on the native ecosystem is likely minimal.
The most recent fishes introduced into Lake Erie are the benthic tubenose goby (Proterorhinus marmoratus) and round goby, which were discovered in the St. Clair River in 1990 (Jude et al., 1992). They quickly spread to lakes Erie, Huron and St. Clair. These fishes, native to the Ponto-Caspian region of Europe, were likely introduced into the Great Lakes basin through ballast water (Jude et al., 1992). Although the round goby feeds extensively on dreissenids, it also feeds on, and competes with, native benthic fish species such as the mottled sculpin (Cottus bairdii) and the logperch (Percina caprodes) (French and Jude, 2001). There are several other benthic fish species at risk (Table 4) that are likely to become further imperilled as the result of similar interactions with gobies.
There has been at least one unanticipated positive effect of introduced species on native fish species at risk in Lake Erie. The pelagic silver chub (Macrhybopsis storeriana) was virtually extirpated in Lake Erie by the late 1950s as a result of the disappearance of burrowing mayflies due to anoxic bottom conditions related to eutrophication (Scott and Crossman, 1973). Silver chub began to increase in abundance by the late 1980s (Holm and Mandrak, 2003). A study of the diets of 110 chub captured in 1999 indicated that they had been feeding primarily on dreissenid mussels, not on the recovering mayflies (N.E. Mandrak, Fisheries & Oceans Canada, Burlington, Ontario, Canada, unpubl. data). Although this may bode well for other native benthic feeding fishes, such as the lake sturgeon, the negative impacts of exotics on fish and other aquatic species at risk (e.g., unionid mussels; Zanatta et al., 2002) will undoubtedly outweigh any positive effects. Furthermore, there is some evidence that increased water clarity in the western basin, likely due to Dreissena filtration, has led to a more benthic food web. Changes in stable isotope ratios in walleye, for example, have indicated a general trend toward feeding on benthic fishes following the Dreissena invasion (Kiriluk et al., 1999). A recent survey of western Lake Erie has shown that planktivore abundance fell from > 80% to less than 50% of the relative abundance (Zhu et al., 2006).
Nonindigenous species pose a significant threat to the North American Great Lakes. Although a number of species were introduced via management practices and by accidental release since the 1800s, the recent outburst of NIS during the past 15 years has caused serious food web alterations. It is not clear, however, whether this intensification of exotics is an artefact of increased awareness or indeed expansion of NIS.
The impact of dreissenid mussels was most apparent in the west basin of Lake Erie. Mussel filtering drastically reduced phytoplankton biomass and productivity, and affected the species and size composition of the community. This resulted in increased water clarity impacting the light regime, which has serious limnological implications. Similarly, after the zebra mussel invasion, zooplankton production was found to be 15 to 90% less than its potential, especially in the nearshore areas of the east basin. The rapid spread of dreissenid mussels across the bottom sediments of Lake Erie resulted in the domination of 90% of the biomass by mussels. It is believed that zebra mussels have played a role in the disappearance of the amphipod Diporeiea that previously dominated the benthos but has since been eliminated from not only Lake Erie, but also from the other Great Lakes. Furthermore, exotic invertebrates and fish have modified food web dynamics resulting in the crash and extirpation of commercially important fisheries. As the number of non-indigenous species in Lake Erie has increased, 32 native fish species have been identified as being at risk. Further introductions may be anticipated at a faster rate due to increased shipping activities and other vectors. It may be necessary for multinational legislation and guidelines to be enacted in order to control all of the vectors responsible for the expansion of exotic species.
Sincere thanks are due to Drs. Marc Babut, Bernard Montuelle (Cemagref) and the organizing committee of the AEHMS 7th Conference held in Lyon, France for the invitation to present the Vallentyne Lecture at the conference. Dr. Hugh MacIsaac, University of Windsor, provided input and advice for the presentation of the plenary lecture. We thank Dr. Ora Johannsson for her advice on zooplankton changes. The constructive criticism, advice and editing of Drs. Ed. Mills, Diane Malley and Sharon Lawrence is greatly appreciated which improved the clarity of the manuscript. Assistance of Jennifer Lorimer, Heather Niblock and Joanne Dzuba is acknowledged with thanks for their assistance in the processing of the manuscript and diagrams.